The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

Home , DX Cancri, EZ Aquarii, Fomalhaut b, Gliese 229, HD 93250, Henry Draper catalogue

Published in 2010 by Britannica Educational Publishing (a trademark of Encyclopædia Britannica, Inc.) in association with Rosen Educational Services, LLC 29 East 21st Street, New York, NY 10010.

Distributed exclusively by Rosen Educational Services. For a listing of additional Britannica Educational Publishing titles, call toll free (800) 237-9932.

First Edition

Britannica Educational Publishing Michael I. Levy: Executive Editor Marilyn L. Barton: Senior Coordinator, Production Control Steven Bosco: Director, Editorial Technologies Lisa S. Braucher: Senior Producer and Data Editor Yvette Charboneau: Senior Copy Editor Kathy Nakamura: Manager, Media Acquisition Erik Gregersen: Associate Editor, Astronomy and Space Exploration

Rosen Educational Services Jeanne Nagle: Senior Editor Nelson Sá: Art Director Matthew Cauli: Designer Introduction by Greg Roza

Library of Congress Cataloging-in-Publication Data

The Milky Way and beyond / edited by Erik Gregersen.—1st ed. p. cm.—(An explorer’s guide to the universe) “In association with Britannica Educational Publishing, Rosen Educational Services.” Includes index. ISBN 978-1-61530-053-2 (eBook) 1. Milky Way—Popular works. 2. Galaxies—Popular works. I. Gregersen, Erik. QB857.7.M55 2010 523.1'13—dc22 2009037980

On the cover: Thousands of sparkling young stars are nestled within the giant nebula NGC 3603, one of the most massive young star clusters in the Milky Way Galaxy. NASA, ESA, and the Hubble Heritage (STScI/AURA)-ESA/Hubble Collaboration 17 CONTENTS Introduction 12

Chapter 1: The Milky Way Galaxy 21 The Structure and Dynamics of the Milky Way Galaxy 21 Structure of the Spiral System 22 The Nucleus 22 The Central Bulge 22 The Disk 23 The Spiral Arms 24 The Spherical Component 25 The Massive Halo 25 Magnetic Field 25 23 Rotation 26 Mass 26 Star Populations and Movement 28 Stars and Stellar Populations 28 Principal Population Types 28 The Stellar Luminosity Function 31 Density Distribution 35 Density Distribution of Various Types of Stars 36 Variations in the Stellar Density 36 Variation of Star Density with z Distances 37 Stellar Motions 38 Proper Motions 38 Radial Velocities 39 Space Motions 39 High-Velocity Stars 40 Solar Motion 40 Solar Motion Calculations from 33 Radial Velocities 41 Solar Motion Calculations from Proper Motions 41 Solar Motion Calculations from Space Motions 41 Solar Motion Solutions 42

Chapter 2: Stars 45 The Nature of Stars 45 64

Size and Activity 45 Variations in Stellar Size 46 Stellar Activity and Mass Loss 46 The 20 Brightest Stars 47 Distances to the Stars 49 Determining Stellar Distances 49 Nearest Stars 50 The 20 Nearest Stars 51 Stellar Positions 52 Basic Measurements 52 Stellar Motions 52 Light from the Stars 53 Stellar Magnitudes 53 66 Stellar Colours 54 Magnitude Systems 54 Bolometric Magnitudes 55 Stellar Spectra 55 Line Spectrum 56 Spectral Analysis 56 Classification of Spectral Types 57 Bulk Stellar Properties 58 Stellar Temperatures 59 Stellar Masses 59 Visual Binaries 60 Spectroscopic Binaries 61 Eclipsing Binaries 61 Binaries and Extrasolar Planetary Systems 63 Stellar Radii 67 Average Stellar Values 67 69 Stellar Statistics 68 Hertzsprung-Russell Diagram 68 Estimates of Stellar Ages 71 Numbers of Stars Versus Luminosity 72 Mass-Luminosity Correlations 72 Variable Stars 73 Classification 73 95

Pulsating Stars 74 Explosive Variables 76 Peculiar Variables 78 Stellar Atmospheres 80 Stellar Interiors 82 Distribution of Matter 83 Source of Stellar Energy 85

Chapter 3: Star formation and Evolution 89 Birth of Stars and Evolution to the Main Sequence 89 Subsequent Development on the Main 111 Sequence 92 Later Stages of Evolution 93 Evolution of Low-Mass Stars 93 Origin of the Chemical Elements 96 The Most Abundant Chemical Elements 96 Evolution of High-Mass Stars 99 End States of Stars 102 White Dwarfs 103 Neutron Stars 104 Black Holes 108

Chapter 4: Star Clusters 110 Open Clusters 110 Globular Clusters 114 OB and T Associations 117 Dynamics of Star Clusters 119 Clusters in External Galaxies 120 124 Notable Stars and Star Clusters 122 51 Pegasi 122 Alpha Centauri 122 Arcturus 123 Betelgeuse 123 Deneb 123 Fomalhaut 123 Pleiades 124 Polaris 124 125

Sirius 125 Vega 126

Chapter 5: Nebulae 127 Classes of Nebulae 129 Early Observations of Nebulae 131 The Work of the Herschels 131 Advances Brought by Photography and Spectroscopy 132 20th-Century Discoveries 132 Chemical Composition and Physical Processes 133 Interstellar Dust 134 145 Turbulence 136 Galactic Magnetic Field 137 Molecular Clouds 137 Composition 138 Formation of Stars 139 Hydrogen Clouds 140 Reflection Nebulae 141 H II Region 141 Ultracompact H II Regions 144 Supergiant Nebulae 145 Chemical Composition of H II Regions 146 Planetary Nebulae 148 Forms and Structure 148 The Distances of Planetary Nebulae 150 Chemical Composition 150 Positions in the Galaxy 151 Evolution of Planetary Nebulae 151 157 Central Stars 152 The Nature of the Progenitor Stars 153 Supernova Remnants 154 The Crab Nebula 157 The Cygnus Loop 158 Diffuse Ionized Gas 158 Notable Nebulae 159 Cassiopeia A 159 161

Coalsack 160 Great Rift 160 Gum Nebula 160 Horsehead Nebula 160 Lagoon Nebula 161 North American Nebula 161 Orion Nebula 161 R Monocerotis 162 Ring Nebula 162 Trifid Nebula 162

Chapter 6: Galaxies 163 The Evolution of Galaxies 165 164 Historical Survey of the Study of Galaxies 167 The Problem of the Magellanic Clouds 168 Novae in the Andromeda Nebula 169 The Scale of the Milky Way Galaxy 169 The van Maanen Rotation 171 The Shapley-Curtis Debate 172 Hubble’s Discovery of Extragalactic Objects 173 The Distance to the Andromeda Nebula 175 The Golden Age of Extragalactic Astronomy 176 Types of Galaxies 177 Principal Schemes of Classification 177 Elliptical Galaxies 177 Irregular Galaxies 182 Other Classification Schemes and Galaxy Types 182 The External Galaxies 183 181 The Extragalactic Distance Scale 183 Physical Properties of External Galaxies 186 Size and Mass 186 Luminosity 186 Age 187 Composition 187 Structure 188 The Spheroidal Component 188 The Disk Component 189 196

Spiral Arms 189 Gas Distribution 189 Galaxy Clusters 189 Types of Clusters 190 Distribution 190 Interactions Between Cluster Members 191 Galaxies as a Radio Source 191 Radio Galaxies 191 X-ray Galaxies 193 Clusters of Galaxies as Radio and X-ray Sources 193 Quasars 193 Gamma-Ray Bursters 195 203 Notable Galaxies and Galaxy Clusters 195 Andromeda Galaxy 195 Coma Cluster 196 Cygnus A 197 Great Attractor 197 Magellanic Clouds 198 M81 Group 198 Maffei I and II 199 Virgo A 199 Virgo Cluster 200

Appendix: Other Stars and Star Clusters 201

Glossary 212 Further Reading 213 Index 214 207

INTRODuCTION Introduction | 13

or thousands of years, astronomers ergs) of luminosity per second, making it Fhave worked to unlock the mysteries average in size, mass, and brightness. of the heavens. As they observed the stars Scientists use these measurements as and planets that parade majestically benchmarks when discussing other across the sky, one great source of won- stars, which can be bigger or smaller, der was a huge faint trail of light that was more or less bright, depending on their called the Milky Way Galaxy because it type and age. looked like milk, spilled across the dark- While 150 million kilometres may ness of night. When scientists started sound like a very long distance, it is tiny using telescopes in the 1600s, they began in cosmic terms. The next closest star to to understand more about the Milky Way. Earth is Proxima Centauri, which is the But it wasn’t until the 20th century that smallest of three stars making up the triple new technology allowed scientists to take star Alpha Centauri. Proxima Centauri is the full measure of this galaxy and others approximately 4.22 light years from scattered across the universe. What they Earth—or about 3.99 × 1016 metres (2.48 × have uncovered is richly detailed in the 1013 miles). Even the fastest modern pages of this book. spacecraft would take countless lifetimes The primary elements of any galaxy to reach this distant star. are stars. A star is a massive body of gas Scientists have used several diff er- that shines by radiation resulting from ent methods to determine the distances internal energy sources. There are so from Earth to the stars. The earliest many stars in the universe it would be method, which is still used to determine impossible to count them all. Only a very the distances of closer stars, is a trigo- small fraction are actually visible to the nometric parallax, which involves unaided eye. In the Milky Way Galaxy observing a star from two points on alone there are hundreds of billions of opposite sides of Earth’s orbit. This stars. The easiest of these for scientists to technique depends on using the rela- study is the Sun, the star closest to Earth. tively unchanged backdrop of distant It is about 150 million kilometres (93 space for precise measurements. More- million miles) away from us. It has a distant stars require other techniques, radius of about 700,000 km (430,000 most of which rely on comparing miles.). Its mass is about the same as that luminosities. of 330,000 Earth masses. It creates Depending on its mass, a star may approximately 4 x 1023 kilowatts (4 × 1033 “live” 10 million years or 10 billion years.

Milky white stars and interstellar dust cascade across the sky to form the Milky Way Galaxy, a spiral galaxy that is home to the solar system. www.istockphoto.com/Shaun Lowe 14 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

During this time, it goes through many the outer layers of the star to expand. This dramatic changes. Stars begin as clouds is called the red giant stage. It can take of interstellar gases, mostly hydrogen. anywhere from about a million years to Molecules in the clouds slowly begin to hundreds of billions of years for a star collapse and clump together, eventually to reach this stage. forming areas of greater density. The Eventually, due to the lack of hydro- clumps begin to rotate slowly as they gen, the fusion reaction begins to die grow more massive. In time, one or more down. Not all stars die in the same man- of the clumps begins to collapse in on ner. They differ depending on how they itself due to gravity. Often, several stars formed, how dense they are, how big form from the same cloud, resulting in they are, and their age. In less-massive star groups. stars, the outer layers may drift away Mass, gravity, and other forces cause into space, leaving the slowly cooling the cloud to form a disk shape around the core called a white dwarf. More-massive young star core. The temperature within stars can explode as a supernova. The the core continually rises, and when the most-massive stars go supernova, then temperature is high enough, hydrogen collapse in on themselves due to immense fusion begins, which allows stars to radiate gravitational force. The result is a black tremendous amounts of heat and light. hole—a force so powerful nothing, not Soon after, the star has become fully even light, can escape its grasp. formed. With a star the size of the Sun, Stars are not the only bright points this might take tens of millions of years. of light in the night sky. A few of those More massive stars, which burn hotter and twinkling specks are nebulae. A nebula more quickly, may take just a few hundred is an interstellar cloud of gas and dust. thousand years to reach this point. The matter that makes up nebulae is For most of its life, a star continues to called the interstellar medium, which can create helium through hydrogen fusion. be found just about everywhere in the This part of a star’s life is called the main universe, although it is more dense in sequence. Fusion creates the light and nebulae. The composition of nebulae is heat that stars radiate and turns hydro- approximately 90 percent hydrogen and gen into helium in the process. The power nearly 10 percent helium, with a very created by fusion is constantly pushing small amount of other elements mixed outward and fighting against the massive in, namely oxygen, carbon, neon, and gravitational pull of the core. In time, as nitrogen. Many of the characteristics of the hydrogen fuel in the core decreases, nebulae are determined by the physical the core is converted to helium. Hydrogen state of the hydrogen in them. is then burned on the surface of the Based on their components and helium core at a higher rate, which causes behaviour, there are two main Introduction | 15

The interstellar medium (gas and dust) that make up nebulae can take many shapes, including this aptly named Bow Tie Nebula. NASA, ESA, R. Sahai and J. Trauger (Jet Propulsion Laboratory) and the WFPC2 Science Team

categories of nebulae, dark and bright. These are the clouds from which most Dark nebulae, which are also called new stars form through gravitational molecular clouds, are dense and cold. collapse. While molecular clouds can contain Although a very small percentage of approximately 60 diff erent kinds of mol- the interstellar medium, like that found in ecules, the majority of their composition dark nebulae, is in solid grains, that per- is molecular hydrogen (H2). They are centage is very important to the creation opaque because of the relatively high of stars and solar systems. Unlike the concentration of solid grains in them. gases in dark nebulae, solid grains absorb Densities generally are millions of hydro- starlight. In turn, they are able to heat gen molecules per cubic centimetre. and cool the gases. 16 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

Bright nebulae are usually not as Despite their many diff erences, neb- dense as the dark variety. However, as the ulae have basic traits in common. For name suggests, they are visible. There example, all nebulae exhibit chaotic are several diff erent kinds of bright nebu- motions scientists call turbulence. This is lae. Refl ection nebulae are molecular similar to the ripples and whirlpools we clouds just like dark nebulae. However, see when we add a coloured liquid to a they are visible because light from nearby clear liquid. The disorganized fl ow of stars refl ect off of their solid grains. gases creates energy and heat. Scientists H II regions are cosmic clouds that know that turbulence has a great eff ect glow because they have been ionized by on the behaviour of nebulae, but they do the radiation produced by a neighbour- not fully understand why or how. They ing hot star. Ionization occurs when the hope to learn more about turbulence and hydrogen atoms in the cloud separate nebulae by continuing to study known into positive hydrogen ions (H+) and free nebulae and by discovering new ones. electrons, causing the cloud to glow. Galaxies are massive, self-contained Another kind of nebula, called the dif- collections of stars. Scientists believe that fuse nebula, is visible due to the most galaxies formed shortly after the ionization of hydrogen, nitrogen, and birth of the universe, about 13 billion sulfur. Diff use nebulae require the most years ago. They can look very diff erent, energy of all the kinds of nebulae and based on how they formed and evolved. are found in the vicinity of the hottest Some are very small, while others, like and most-massive stars. the Milky Way, have huge spiral arms When a star goes supernova, the result- reaching deep into space. ing explosion can last for several weeks. Scientists have had diffi culty studying After the supernova dies down, a bright, the Milky Way because of a thick layer of colourful nebula, sometimes called a super- interstellar medium that obscures their nova remnant, is left behind. Stars that view of it, even with powerful telescopes. don’t go supernova can create planetary Many think the galaxy’s diameter is about nebulae. This occurs when the envelope 100,000 light-years, and that our sun of gases around the dying star begins to resides in a spiral arm about 30,000 light- expand and spread out. Planetary nebulae years from the galaxy’s center. often have a round, compact shape. The Milky Way is one galaxy in a Scientists believe the Milky Way Galaxy cluster of galaxies called the Local Group. contains about 20,000 planetary nebulae. Galaxy clusters are groups of galaxies

The arms of spiral galaxies, such as the Milky Way and this dusty spiral (pictured), are thought to be produced by density waves that compress and expand galactic material. NASA Headquarters - GRIN Introduction | 17 18 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies that can be hundreds of millions of light- smaller galaxies. In fact, the two galaxies years across. The word “cluster,” however, are moving toward each other and will might be a bit misleading. The Magellanic some day—billions of years from now— Clouds are two satellite galaxies orbiting merge to form a single galaxy. the Milky Way. They are probably Beyond the Andromeda Galaxy, between 160,000 and 190,000 light-years countless other galaxies with countless away. Scientists have learned much about stars and nebulae are scattered to the nebulae and stars by observing these farthest corners of the cosmos. We may neighbouring galaxies. never travel past the distant limits of our The Andromeda Galaxy is the closest own solar system, and those galaxies may spiral galaxy beyond the Milky Way gal- always be just colourful specks visible axy cluster. The Andromeda Galaxy only through our most-powerful tele- cluster is one of the most-distant objects scopes. However, scientists will continue that can be seen from Earth with the to study them and search for new stars, unaided eye. At one time, scientists nebulae, and galaxies in the hope of thought the Andromeda Galaxy was a learning more about our place in the nebula in the Milky Way. However, we vast cosmos. With billions upon billions now know that it is about 2,480,000 light- of galaxies—each of which are home to years from Earth and is twice the size of billions upon billions of stars—there will the Milky Way galaxy. Scientists believe no doubt be plenty for scientists to study this galaxy has a history of “consuming” for many years to come.

CHAPTER 1

The Milky Way Galaxy

n a very dark clear night, if you look upward at the heav- Oens, you will see an irregular luminous band of stars and gas clouds that stretches across the sky. This band is called the Milky Way. The Milky Way is actually a large spiral system, a galaxy, consisting of several billion stars, one of which is the Sun. Although Earth lies well within the Milky Way Galaxy (sometimes simply called the Galaxy), astronomers do not have as complete an understanding of its nature as they do of some external star systems. A thick layer of interstellar dust obscures much of the Galaxy from scrutiny by optical telescopes. Astronomers can determine its large- scale structure only with the aid of radio and infrared telescopes, which can detect the forms of radiation that penetrate the obscuring matter.

THE STRuCTuRE AND DyNAMICS OF THE MILKy WAy GALAxy

The fi rst reliable measurement of the size of the Galaxy was made in 1917 by American astronomer Harlow Shapley. He arrived at his size determination by establishing the spatial distribution of globular clusters. Shapley found that, instead of a relatively small system with the Sun near its centre, as had previously been thought, the Galaxy is immense, with the Sun nearer the edge than the centre. Assuming that the globular clusters outlined the Galaxy, he determined that it 22 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies has a diameter of about 100,000 light- massive black hole surrounded by an years and that the Sun lies about 30,000 accretion disk of high-temperature gas. light-years from the centre. (A light-year is Neither the central object nor any of the the distance traveled by light in one year material immediately around it can be and is roughly 9,460,000,000,000 km, or observed at optical wavelengths because 5,880,000,000,000 miles.) His values have of the thick screen of intervening dust held up remarkably well over the years. in the Milky Way. The object, however, Depending in part on the particular com- is readily detectable at radio wave- ponent being discussed, the stellar disk lengths and has been dubbed Sagittarius of the Milky Way system is just about as A* by radio astronomers. Somewhat large as Shapley’s model predicted, with similar to the centres of active galaxies, neutral hydrogen somewhat more widely though on a lesser scale, the galactic dispersed and dark (i.e., unobservable) nucleus is the site of a wide range of matter perhaps filling an even larger vol- activity apparently powered by the ume than expected. The most distant black hole. stars and gas clouds of the system that Infrared radiation and X-rays are have had their distance reliably deter- emitted from the area, and rapidly mov- mined lie roughly 72,000 light-years from ing gas clouds can be observed there. the galactic centre, while the distance Data strongly indicate that material is of the Sun from the centre has been found being pulled into the black hole from out- to be approximately 25,000 light-years. side the nuclear region, including some gas from the z direction (i.e., perpendic- Structure of the ular to the galactic plane). As the gas Spiral System nears the black hole, its strong gravitational force squeezes the gas into a The Milky Way Galaxy’s structure is fairly rapidly rotating disk, which extends out- typical of a large spiral system. This ward about 5–30 light-years from the structure can be viewed as consisting of central object. Rotation measurements of six separate parts: (1) a nucleus, (2) a the disk and the orbital motions of stars central bulge, (3) thin and thick disks (seen at infrared wavelengths) indicate (4) spiral arms, (5) a spherical component, that the black hole has a mass 4,310,000 and (6) a massive halo. Some of these times that of the Sun. components blend into each other. The Central Bulge The Nucleus Surrounding the nucleus is an extended At the very centre of the Galaxy lies bulge of stars that is nearly spherical in a remarkable object—in all likelihood a shape and that consists primarily of old The Milky Way Galaxy | 23

Image of the centre of the Milky Way Galaxy, produced from the observations made by the Infrared Astronomy Satellite (IRAS). The bulge in the band is the centre of the Galaxy. NASA stars, known as Population II stars, Galaxy resembles other spiral systems, though they are comparatively rich in featuring as it does a bright, fl at arrange- heavy elements. Mixed with the stars are ment of stars and gas clouds that is several globular clusters of similar stars. spread out over its entirety and marked Both the stars and clusters have nearly by a spiral structure. radial orbits around the nucleus. The The disk can be thought of as being bulge stars can be seen optically where the underlying body of stars upon they stick up above the obscuring dust of which the arms are superimposed. This the galactic plane. body has a thickness that is roughly one-fifth its diameter, but different The Disk components have diff erent characteristic thicknesses. The thinnest component, From a distance the most conspicuous often called the “thin disk,” includes the part of the Galaxy would be the disk, dust and gas and the youngest stars, which extends from the nucleus out to while a thicker component, the “thick approximately 75,000 light-years. The disk,” includes somewhat older stars. 24 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

The Spiral Arms the part of the Milky Way Galaxy wherein the solar system is located. Astronomers did not know that the Theoretical understanding of the Galaxy had a spiral structure until 1953, Galaxy’s spiral arms has progressed when the distances to stellar associations greatly since the 1950s, but there is still were first obtained reliably. Because of no complete understanding of the rela- the obscuring interstellar dust and the tive importance of the various effects interior location of the solar system, the thought to determine their structure. The spiral structure is very difficult to detect overall pattern is almost certainly the optically. This structure is easier to dis- result of a general dynamical effect cern from radio maps of either neutral known as a density-wave pattern. The hydrogen or molecular clouds, since both American astronomers Chia-Chiao Lin can be detected through the dust. and Frank H. Shu showed that a spiral Distances to the observed neutral hydro- shape is a natural result of any large-scale gen atoms must be estimated on the basis disturbance of the density distribution of measured velocities used in conjunc- of stars in a galactic disk. When the inter- tion with a rotation curve for the Galaxy, action of the stars with one another is which can be built up from measurements calculated, it is found that the resulting made at different galactic longitudes. density distribution takes on a spiral pat- From studies of other galaxies, it can tern that does not rotate with the stars be shown that spiral arms generally follow but rather moves around the nucleus a logarithmic spiral form such that more slowly as a fixed pattern. Individual stars in their orbits pass in and out of the log r = a − bϕ, spiral arms, slowing down in the arms temporarily and thereby causing the where ϕ is a position angle measured density enhancement. For the Galaxy, from the centre to the outermost part of comparison of neutral hydrogen data the arm, r is the distance from the centre with the calculations of Lin and Shu have of the galaxy, and a and b are constants. shown that the pattern speed is 4 km/sec The range in pitch angles for galaxies is per 1,000 light-years. from about 50° to approximately 85°. The Other effects that can influence a pitch angle is constant for any given gal- galaxy’s spiral shape have been explored. axy if it follows a true logarithmic spiral. It has been demonstrated, for example, The pitch angle for the spiral arms of the that a general spiral pattern will result Galaxy is difficult to determine from the simply from the fact that the galaxy has limited optical data, but most measure- differential rotation; i.e., the rotation ments indicate a value of about 75°. There speed is different at different distances are five optically identified spiral arms in from the galactic centre. Any disturbance, The Milky Way Galaxy | 25 such as a sequence of stellar formation considerably beyond a distance of 100,000 events that are sometimes found drawn light-years from the centre and that its out in a near-linear pattern, will eventu- mass is several times greater than the ally take on a spiral shape simply because mass of the rest of the Galaxy taken of the differential rotation. For example, together. It is not known what its shape is, the outer spiral structure in some galaxies what its constituents are, or how far into may be the result of tidal encounters with intergalactic space it extends. other galaxies or galactic cannibalism. Distortions that also can be included are Magnetic Field the results of massive explosions such as supernova events. These, however, tend It was once thought that the spiral struc- to have only fairly local effects. ture of galaxies might be controlled by a strong magnetic field. However, when the The Spherical Component general magnetic field was detected by radio techniques, it was found to be too The space above and below the disk of weak to have large-scale effects on galac- the Galaxy is occupied by a thinly popu- tic structure. The strength of the galactic lated extension of the central bulge. field is only about 0.000001 times the Nearly spherical in shape, this region is strength of Earth’s field at its surface, a populated by the outer globular clusters, value that is much too low to have dynam- but it also contains many individual field ical effects on the interstellar gas that stars of extreme Population II, such as RR could account for the order represented Lyrae variables and dwarf stars deficient by the spiral-arm structure. This is, how- in the heavy elements. Structurally, the ever, sufficient strength to cause a general spherical component resembles an ellip- alignment of the dust grains in inter- tical galaxy, following the same simple stellar space, a feature that is detected mathematical law of how density varies by measurements of the polarization of with distance from the centre. starlight. In the prevailing model of interstellar The Massive Halo dust grains, the particles are shown to be rapidly spinning and to contain small The least-understood component of the amounts of metal (probably iron), though Galaxy is the giant massive halo that is the primary constituents are ice and car- exterior to the entire visible part. The bon. The magnetic field of the Galaxy can existence of the massive halo is demon- gradually act on the dust particles and strated by its effect on the outer rotation cause their rotational axes to line up in curve of the Galaxy. All that can be said such a way that their short axes are par- with any certainty is that the halo extends allel to the direction of the field. The field 26 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies itself is aligned along the Milky Way band, demonstrated that differential rotation so that the short axes of the particles also leads to a systematic variation of the become aligned along the galactic plane. radial velocities of stars with galactic Polarization measurements of stars at low longitude following the mathematical galactic latitudes confirm this pattern. expression:

Rotation radial velocity = Ar sin 2l ,

The motions of stars in the local stellar where A is called Oort’s constant and is neighbourhood can be understood in approximately 15 km/sec/kiloparsec (1 terms of a general population of stars kiloparsec is 3,260 light-years), r is the that have circular orbits of rotation distance to the star, and l is the galactic around the distant galactic nucleus, with longitude. an admixture of stars that have more A similar expression can be derived highly elliptical orbits and that appear to for measured proper motions of stars. be high-velocity stars to a terrestrial The agreement of observed data with observer as Earth moves with the Sun in Oort’s formulas was a landmark demon- its circular orbit. The general rotation of stration of the correctness of Lindblad’s the disk stars was first detected through ideas about stellar motions. It led to the studies made in the 1920s, notably those of modern understanding of the Galaxy as the Swedish astronomer Bertil Lindblad, consisting of a giant rotating disk with who correctly interpreted the apparent other more spherical, and more slowly asymmetries in stellar motions as the rotating, components superimposed. result of this multiple nature of stellar orbital characteristics. Mass The disk component of the Galaxy rotates around the nucleus in a manner The total mass of the Galaxy, which had similar to the pattern for the planets of seemed reasonably well-established dur- the solar system, which have nearly cir- ing the 1960s, has become a matter of cular orbits around the Sun. Because the considerable uncertainty. Measuring the rotation rate is different at different dis- mass out to the distance of the farthest tances from the centre of the Galaxy, the large hydrogen clouds is a relatively measured velocities of disk stars in dif- straightforward procedure. The measure- ferent directions along the Milky Way ments required are the velocities and exhibit different patterns. The Dutch positions of neutral hydrogen gas, com- astronomer Jan H. Oort first interpreted bined with the approximation that the this effect in terms of galactic rotation gas is rotating in nearly circular orbits motions, employing the radial velocities around the centre of the Galaxy. A rota- and proper motions of stars. He tion curve, which relates the circular The Milky Way Galaxy | 27 velocity of the gas to its distance from the to have been in error. Instead, the curve galactic centre, is constructed. The shape remained almost constant, indicating of this curve and its values are determined that there continue to be substantial by the amount of gravitational pull that amounts of matter exterior to the measured the Galaxy exerts on the gas. Velocities hydrogen gas. This in turn indicates that are low in the central parts of the system there must be some undetected material because not much mass is interior to the out there that is completely unexpected. orbit of the gas; most of the Galaxy is It must extend considerably beyond the exterior to it and does not exert an inward previously accepted positions of the edge gravitational pull. Velocities are high at of the Galaxy, and it must be dark at vir- intermediate distances because most of tually all wavelengths, as it remains the mass in that case is inside the orbit undetected even when searched for with of the gas clouds and the gravitational radio, X-ray, ultraviolet, infrared, and opti- pull inward is at a maximum. At the far- cal telescopes. Until the dark matter is thest distances, the velocities decrease identified and its distribution determined, because nearly all the mass is interior to it will be impossible to measure the total the clouds. mass of the Galaxy, and so all that can be This portion of the Galaxy is said to said is that the mass is several times have Keplerian orbits, since the material larger than thought earlier. should move in the same manner that the The nature of the dark matter in the German astronomer Johannes Kepler Galaxy remains one of the major ques- discovered the planets to move within the tions of galactic astronomy. Many other solar system, where virtually all the mass galaxies also appear to have such unde- is concentrated inside the orbits of the tected matter. The possible kinds of orbiting bodies. The total mass of the GalÂ material that are consistent with the non- axy is then found by constructing detections are few in number. Planets and mathematical models of the system with rocks would be impossible to detect, but different amounts of material distributed it is extremely difficult to understand how in various ways and by comparing the they could materialize in sufficient num- resulting velocity curves with the observed bers, especially in the outer parts of one. As applied in the 1960s, this proce- galaxies where there are no stars or even dure indicated that the total mass of the interstellar gas and dust from which they Galaxy was approximately 200 billion could be formed. Low-luminosity stars, times the mass of the Sun. called brown dwarfs, are so faint that only During the 1980s, however, refinements a few have been detected directly. In the in the determination of the velocity curve 1990s, astronomers carried out exhaus- began to cast doubts on the earlier results. tive lensing experiments involving the The downward trend to lower velocities study of millions of stars in the galactic in the outer parts of the Galaxy was found central areas and in the Magellanic 28 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

Clouds to search for dark objects whose characteristics around the Galaxy. UnderÂ masses would cause lensed brightenings standing of evolutionary differences, of background stars. Some lensing events however, had not yet been achieved, and, were detected, but the number of dark although differences in the chemical objects inferred is not enough to explain abundances in the stars were known, their completely the dark matter in galaxies significance was not comprehended. At and galaxy clusters. It appears likely that this juncture, chemical differences seemed there is more than one form of dark exceptional and erratic and remained matter, with the most important being uncorrelated with other stellar properties. hypothetical types of objects, such as There was still no systematic division of WIMPs (weakly interacting massive stars even into different kinematic fami- particles). lies, in spite of the advances in theoretical work on the dynamics of the Galaxy. Star populations and movement Principal Population Types

The Milky Way Galaxy is made up of In 1944 the German-born astronomer about one hundred billion stars. Stars Walter Baade announced the successful come in many different masses, from a resolution into stars of the centre of the few percent that of the Sun to a hundred Andromeda Galaxy, M31, and its two times greater. Stars also appear in differ- elliptical companions, M32 and NGC 205. ent colours, from a dim, cool red to an He found that the central parts of AndromÂ incandescent blue. Despite their different eda and the accompanying galaxies were properties, stars can be divided into pop- resolved at very much fainter magnitudes ulations. The differences between the than were the outer spiral arm areas of populations can also be seen in how stars M31. Furthermore, by using plates of dif- are distributed and how they move. ferent spectral sensitivity and coloured filters, he discovered that the two ellipti- Stars and Stellar cals and the centre of the spiral had red Populations giants as their brightest stars rather than blue main-sequence stars, as in the case The concept of different populations of of the spiral arms. stars has undergone considerable change This finding led Baade to suggest over the last several decades. Before the that these galaxies, and also the Milky 1940s, astronomers had been aware of Way Galaxy, are made of two populations differences between stars and had largely of stars that are distinct in their physical accounted for most of them in terms of properties as well as their locations. He different masses, luminosities, and orbital applied the term Population I to the stars The Milky Way Galaxy | 29 that constitute the spiral arms of system, contains such objects as open star Andromeda and to most of the stars that clusters, O and B stars, Cepheid variables, are visible in the Milky Way system in emission nebulae, and neutral hydrogen. the neighbourhood of the Sun. He found Its Population II component, spread over that these Population I objects were lim- a more nearly spherical volume of space, ited to the flat disk of the spirals and includes globular clusters, RR Lyrae suggested that they were absent from variables, high-velocity stars, and certain the centres of such galaxies and from the other rarer objects. ellipticals entirely. Baade designated as As time progressed, it was possible Population II the bright red giant stars for astronomers to subdivide the different that he discovered in the ellipticals and in populations in the Galaxy further. These the nucleus of Andromeda. Other objects subdivisions ranged from the nearly that seemed to contain the brightest stars spherical “halo Population II” system to of this class were the globular clusters of the very thin “extreme Population I” sys- the Galaxy. Baade further suggested that tem. Each subdivision was found to the high-velocity stars near the Sun were contain (though not exclusively) charac- Population II objects that happened to be teristic types of stars, and it was even passing through the disk. possible to divide some of the variable- As a result of Baade’s pioneering star types into subgroups according to work on other galaxies in the Local Group their population subdivision. The RR (the cluster of star systems to which the Lyrae variables of type ab, for example, Milky Way Galaxy belongs), astronomers could be separated into different groups immediately applied the notion of two by their spectral classifications and their stellar populations to the Galaxy. It is mean periods. Those with mean periods possible to segregate various compo- longer than 0.4 days were classified as nents of the Galaxy into the two halo Population II, while those with population types by applying both the periods less than 0.4 days were placed idea of kinematics of different popula- in the “disk population.” Similarly, long- tions suggested by their position in the period variables were divided into Andromeda system and the dynamical different subgroups, such that those theories that relate galactic orbital prop- with periods of less than 250 days and of erties with z distances (the distances relatively early spectral type (earlier than above the plane of the Galaxy) for differ- M5e) were considered “intermediate ent stars. For many of these objects, the Population II,” whereas the longer period kinematic data on velocities are the prime variables fell into the “older Population I” source of population classification. The category. As dynamical properties were Population I component of the Galaxy, more thoroughly investigated, many highly limited to the flat plane of the astronomers divided the Galaxy’s stellar 30 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies populations into a “thin disk,” a “thick Sun. Population II stars have considerably disk,” and a “halo.” lower abundances of the heavy elements— An understanding of the physical dif- by amounts ranging from a factor of 5 or ferences in the stellar populations became 10 up to a factor of several hundred. The increasingly clearer during the 1950s total abundance of heavy elements, Z, for with improved calculations of stellar evo- typical Population I stars is 0.04 (given in lution. Evolving-star models showed that terms of the mass percent for all elements giants and supergiants were evolved with atomic weights heavier than helium, objects recently derived from the main a common practice in calculating stellar sequence (a distinctive, primary band of models). The values of Z for halo popula- stars) after the exhaustion of hydrogen in tion globular clusters, on the other hand, the stellar core. As this became better were typically as small as 0.003. understood, it was found that the lumi- A further difference between the two nosity of such giants was not only a populations became clear as the study of function of the masses of the initial main- stellar evolution advanced. It was found sequence stars from which they evolved, that Population II was exclusively made but was also dependent on the chemical up of stars that are very old. Estimates of composition of the stellar atmosphere. the age of Population II stars have varied Therefore, not only was the existence of over the years, depending on the degree giants in the different stellar populations of sophistication of the calculated mod- understood, but differences between the els and the manner in which observations giants with relation to the main sequence for globular clusters are fitted to these of star groups came to be understood in models. They have ranged from 109 years terms of the chemistry of the stars. up to 2 × 1010 years. Recent comparisons At the same time, progress was made of these data suggest that the halo globu- in determining the abundances of stars lar clusters have ages of approximately of the different population types by 1.1–1.3 × 1010 years. The work of American means of high-dispersion spectra obtained astronomer Allan Sandage and his col- with large reflecting telescopes having a laborators proved without a doubt that coudé focus arrangement. A curve of the range in age for globular clusters was growth analysis demonstrated beyond a relatively small and that the detailed doubt that the two population types characteristics of the giant branches of exhibited very different chemistries. In their colour-magnitude diagrams were 1959 H. Lawrence Helfer, George correlated with age and small differences Wallerstein, and Jesse L. Greenstein of in chemical abundances. the United States showed that the giant On the other hand, stars of Population stars in globular clusters have chemical I were found to have a wide range of ages. abundances quite different from those of Stellar associations and galactic clusters Population I stars such as typified by the with bright blue main-sequence stars The Milky Way Galaxy | 31 have ages of a few million years (stars are nearly pure Population II, while irregular still in the process of forming in some of galaxies are dominated by a thick disk of them) to a few hundred million years. Population I, with only a small number of Studies of the stars nearest the Sun indi- Population II stars. Furthermore, the pop- cate a mixture of ages with a considerable ulations vary with galaxy mass; while the number of stars of great age—on the order Milky Way Galaxy, a massive example of of 109 years. Careful searches, however, a spiral galaxy, contains no stars of young have shown that there are no stars in the age and a low heavy-metal abundance, solar neighbourhood and no galactic low-mass galaxies, such as the dwarf clusters whatsoever that are older than irregulars, contain young, low heavy- the globular clusters. This is an indica- element stars, as the buildup of heavy tion that globular clusters, and thus elements in stars has not proceeded far Population II objects, formed first in the in such small galaxies. Galaxy and that Population I stars have been forming since. The Stellar Luminosity Function In short, as the understanding of stellar populations grew, the division into The stellar luminosity function is a Population I and Population II became description of the relative number of understood in terms of three parameters: stars of different absolute luminosities. It age, chemical composition, and kine- is often used to describe the stellar con- matics. (A fourth parameter, spatial tent of various parts of the Galaxy or distribution, appeared to be clearly other groups of stars, but it most com- another manifestation of kinematics.) monly refers to the absolute number of The correlations among these three stars of different absolute magnitudes in parameters were not perfect but seemed the solar neighbourhood. In this form it is to be reasonably good for the Galaxy, usually called the van Rhijn function, even though it was not yet known whether named after the Dutch astronomer Pieter these correlations were applicable to J. van Rhijn. The van Rhijn function is a other galaxies. As various types of galax- basic datum for the local portion of the ies were explored more completely, it Galaxy, but it is not necessarily repre- became clear that the mix of populations sentative for an area larger than the in galaxies was correlated with their immediate solar neighbourhood. InvestÂ Hubble type, which separates galaxies igators have found that elsewhere in the into ellipticals, barred spirals, and spirals Galaxy, and in the external galaxies (as without bars. Spiral galaxies such as the well as in star clusters), the form of the Milky Way Galaxy have Population I con- luminosity function differs in various centrated in the spiral disk and Population respects from the van Rhijn function. II spread out in a thick disk and/or a The detailed determination of the lumi- spherical halo. Elliptical galaxies are nosity function of the solar neighbourhood 32 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies is an extremely complicated process. the best samples to use for determining Difficulties arise because of (1) the incom- the luminosity function of old stars hav- pleteness of existing surveys of stars of ing a low abundance of heavy elements all luminosities in any sample of space (Population II stars). and (2) the uncertainties in the basic data Globular-cluster luminosity func- (distances and magnitudes). In determin- tions show a conspicuous peak at absolute ing the van Rhijn function, it is normally magnitude MV = 0.5, and this is clearly preferable to specify exactly what volume due to the enrichment of stars at that of space is being sampled and to state magnitude from the horizontal branch of explicitly the way in which problems of the cluster. The height of this peak in the incompleteness and data uncertainties data is related to the richness of the hori- are handled. zontal branch, which is in turn related to In general there are four different meth- the age and chemical composition of the ods for determining the local luminosity stars in the cluster. A comparison of function. Most commonly, trigonometric the observed M3 luminosity function parallaxes are employed as the basic with the van Rhijn function shows a sample. Alternative but somewhat less depletion of stars, relative to fainter certain methods include the use of spec- stars, for absolute magnitudes brighter troscopic parallaxes, which can involve than roughly MV = 3.5. This discrepancy much larger volumes of space. A third is important in the discussion of the method entails the use of mean paral- physical significance of the van Rhijn laxes of a star of a given proper motion function and luminosity functions for and apparent magnitude; this yields a clusters of different ages and so will be statistical sample of stars of approxi- dealt with more fully below. mately known and uniform distance. The Many studies of the component stars fourth method involves examining the of open clusters have shown that the distribution of proper motions and tan- luminosity functions of these objects vary gential velocities (the speeds at which widely. The two most conspicuous differ- stellar objects move at right angles to the ences are the overabundance of stars of line of sight) of stars near the Sun. brighter absolute luminosities and the Because the solar neighbourhood is a underabundance or absence of stars of mixture of stars of various ages and dif- faint absolute luminosities. The over- ferent types, it is difficult to interpret the abundance at the bright end is clearly van Rhijn function in physical terms with- related to the age of the cluster (as deter- out recourse to other sources of mined from the main-sequence turnoff information, such as the study of star point) in the sense that younger star clus- clusters of various types, ages, and ters have more of the highly luminous dynamical families. Globular clusters are stars. This is completely understandable The Milky Way Galaxy | 33

Colour-magnitude (Hertzsprung-Russell) diagram for an old globular cluster made up of Population II stars. From Astrophysical Journal, reproduced by permission of the American Astronomical Society in terms of the evolution of the clusters when compared with the van Rhijn func- and can be accounted for in detail by cal- tion, clearly shows a large overabundance culations of the rate of evolution of stars of bright stars due to the extremely young of diff erent absolute magnitudes and age of the cluster, which is on the order of mass. For example, the luminosity func- 10 6 years. Calculations of stellar evolution tion for the young clusters h and χ Persei , indicate that in an additional 10 9 or 10 10 34 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies years all of these stars will have evolved calculations of the rate of evolution of away and disappeared from the bright stars of different masses and luminosi- end of the luminosity function. ties, he showed that it is possible to apply In 1955 the first detailed attempt to a correction to the van Rhijn function in interpret the shape of the general van order to obtain the form of the initial Rhijn luminosity function was made by luminosity function. Comparisons of the Austrian-born American astronomer open clusters of various ages have Edwin E. Salpeter, who pointed out that shown that these clusters agree much the change in slope of this function near more closely with the initial formation

MV = +3.5 is most likely the result of the function than with the van Rhijn func- depletion of the stars brighter than this tion; this is especially true for the very limit. Salpeter noted that this particular young clusters. Consequently, investiga- absolute luminosity is very close to the tors believe that the formation function, turnoff point of the main sequence for as derived by Salpeter, is a reasonable stars of an age equal to the oldest in the representation of the distribution of star solar neighbourhood—approximately 1010 luminosities at the time of formation, years. Thus, all stars of the luminosity even though they are not certain that the function with fainter absolute magni- assumption of a uniform rate of forma- tudes have not suffered depletion of their tion of stars can be precisely true or that numbers because of stellar evolution, as the rate is uniform throughout a galaxy. there has not been enough time for them It has been stated that open-cluster to have evolved from the main sequence. luminosity functions show two discrep- On the other hand, the ranks of stars of ancies when compared with the van Rhijn brighter absolute luminosity have been function. The first is due to the evolution variously depleted by evolution, and so of stars from the bright end of the lumi- the form of the luminosity function in nosity function such that young clusters this range is a composite curve contrib- have too many stars of high luminosity, uted by stars of ages ranging from 0 to as compared with the solar neighbour- 1010 years. hood. The second discrepancy is that Salpeter hypothesized that there very old clusters such as the globular might exist a time-independent function, clusters have too few high-luminosity the so-called formation function, which stars, as compared with the van Rhijn would describe the general initial distri- function, and this is clearly the result of bution of luminosities, taking into stellar evolution away from the main account all stars at the time of formation. sequence. Stars do not, however, disap- Then, by assuming that the rate of star pear completely from the luminosity formation in the solar neighbourhood function; most become white dwarfs and has been uniform since the beginning of reappear at the faint end. In his early this process and by using available comparisons of formation functions with The Milky Way Galaxy | 35 luminosity functions of galactic clusters, the faint end and because of variations Sandage calculated the number of white at the bright end, the local density distri- dwarfs expected in various clusters; pres- bution is not simply derived nor is there ent searches for these objects in a few of agreement between different studies in the clusters (e.g., the Hyades) have sup- the final result. ported his conclusions. In the vicinity of the Sun, stellar den- Open clusters also disagree with the sity can be determined from the various van Rhijn function at the faint end—i.e., surveys of nearby stars and from esti- for absolute magnitudes fainter than mates of their completeness. For example, approximately MV = +6. In all likelihood Wilhelm Gliese’s catalog of nearby stars, a this is mainly due to a depletion of commonly used resource, contains 1,049 another sort, the result of dynamical stars in a volume within a radius of 65 light- effects on the clusters that arise because years. This is a density of about 0.001 stars of internal and external forces. Stars of per cubic light-year. However, even this low mass in such clusters escape from catalog is incomplete, and its incomplete- the system under certain common con- ness is probably attributable to the fact ditions. The formation functions for that it is difficult to detect the faintest stars these clusters may be different from the and faint companions, especially extremely Salpeter function and may exclude faint faint stars such as brown dwarfs. stars. A further effect is the result of the In short, the true density of stars in finite amount of time it takes for stars to the solar neighbourhood is difficult to condense; very young clusters have few establish. The value most commonly faint stars partly because there has not quoted is 0.003 stars per cubic light-year, been sufficient time for them to have a value obtained by integrating the van reached their main-sequence luminosity. Rhijn luminosity function with a cutoff taken M = 14.3. This is, however, distinctly Density Distribution smaller than the true density as calculated for the most complete sampling The density distribution of stars near the volume discussed above and is therefore Sun can be used to calculate the mass an underestimate. Gliese has estimated density of material (in the form of stars) that when incompleteness of the catalogs at the Sun’s distance within the Galaxy. It is taken into account, the true stellar den- is therefore of interest not only from the sity is on the order of 0.004 stars per cubic point of view of stellar statistics but also light-year, which includes the probable in relation to galactic dynamics. In prin- number of unseen companions of multiple ciple, the density distribution can be systems. calculated by integrating the stellar lumi- The density distribution of stars can nosity function. In practice, because of be combined with the luminosity-mass uncertainties in the luminosity function at relationship to obtain the mass density in 36 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies the solar neighbourhood, which includes ations within the plane is dealt with here. only stars and not interstellar material. Density variations are conspicuous This mass density is 4 × 10−24 g/cm3. for early type stars (i.e., stars of higher temperatures) even after allowance has Density Distribution of been made for interstellar absorption. For Various Types of Stars the stars earlier than type B3, for example, large stellar groupings in which the den- To examine what kinds of stars contribute sity is abnormally high are conspicuous to the overall density distribution in the in several galactic longitudes. The Sun, solar neighbourhood, various statistical in fact, appears to be in a somewhat lower sampling arguments can be applied to density region than the immediate sur- catalogs and lists of stars. For rare objects, roundings, where early B stars are such as globular clusters, the volume of relatively scarce. There is a conspicuous the sample must of course be rather large grouping of stars, sometimes called the compared with that required to calculate Cassiopeia-Taurus Group, that has a cen- the density for more common stars. troid at approximately 600 light-years The most common stars and those distance. A deficiency of early type stars that contribute the most to the local stellar is readily noticeable, for instance, in the mass density are the dwarf M (dM) stars, direction of the constellation Perseus at which provide a total of 0.0008 solar distances beyond 600 light-years. Of masses per cubic light-year. It is inter- course, the nearby stellar associations are esting to note that RR Lyrae variables striking density anomalies for early type and planetary nebulae—though many stars in the solar neighbourhood. The are known and thoroughly studied— early type stars within 2,000 light-years contribute almost imperceptibly to the are significantly concentrated at negative local star density. At the same time, white galactic latitudes. This is a manifestation dwarf stars, which are difficult to observe of a phenomenon referred to as the Gould and of which very few are known, are Belt, a tilt of the nearby bright stars in among the more significant contributors. this direction with respect to the galactic plane first noted by the English astrono- Variations in the Stellar Density mer John Herschel in 1847. Such anomalous behaviour is true only for the The star density in the solar neighbour- immediate neighbourhood of the Sun; hood is not perfectly uniform. The most faint B stars are strictly concentrated conspicuous variations occur in the z along the galactic equator. direction, above and below the plane of the Generally speaking, the large varia- Galaxy, where the number density falls off tions in stellar density near the Sun are rapidly. This will be considered separately less conspicuous for the late type dwarf below. The more difficult problem of vari- stars (those of lower temperatures) than The Milky Way Galaxy | 37 for the earlier types. This fact is explained of the ratio of star densities in the centre of as the result of the mixing of stellar spiral arms and in the interarm regions. orbits over long time intervals available The most commonly accepted theoretical for the older stars, which are primarily representation of spiral structure, that of those stars of later spectral types. The the density-wave theory, suggests that this young stars (O, B, and A types) are still ratio is on the order of 0.6, but, for a com- close to the areas of star formation and plicated and distorted spiral structure such show a common motion and common as apparently occurs in the Galaxy, there is concentration due to initial formation no confidence that this figure corresponds distributions. In this connection it is very accurately with reality. On the other interesting to note that the concentration hand, fluctuations in other galaxies can be of A-type stars at galactic longitudes estimated from photometry of the spiral 160° to 210° is coincident with a similar arms and the interarm regions, provided concentration of hydrogen detected by that some indication of the nature of this means of 21-cm (8-inch) line radiation. stellar luminosity function at each posi- Correlations between densities of early tion is available from colours or type stars on the one hand and inter- spectrophotometry. Estimates of the star stellar hydrogen on the other are density measured across the arms of spiral conspicuous but not fixed; there are galaxies and into the interarm regions areas where neutral-hydrogen concentra- show that the large-scale spiral structure tions exist but for which no anomalous of a galaxy of this type is, at least in many star density is found. cases, represented by only a relatively The variations discussed above are small fluctuation in star density. primarily small-scale fluctuations in star It is clear from studies of the external density rather than the large-scale phe- galaxies that the range in star densities nomena so strikingly apparent in the existing in nature is immense. For example, structure of other galaxies. Sampling is too the density of stars at the centre of the difficult and too limited to detect the spiral nearby Andromeda spiral galaxy has structure from the variations in the star been determined to equal 100,000 solar densities for normal stars, although a hint masses per cubic light-year, while the of the spiral structure can be seen in the density at the centre of the Ursa Minor distribution in the earliest type stars and dwarf elliptical galaxy is only 0.00003 stellar associations. In order to determine solar masses per cubic light-year. the true extent in the star-density variations corresponding to these large-scale Variation of Star Density structural features, it is necessary to turn with z Distances either to theoretical representations of the spiral structures or to other galaxies. From For all stars, variation of star density the former it is possible to find estimates above and below the galactic plane 38 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies rapidly decreases with height. Stars of proper motion and radial velocity can be different types, however, exhibit widely measured. Proper motion is the motion of differing behaviour in this respect, and a star across an observer’s line of sight this tendency is one of the important and constitutes the rate at which the clues as to the kinds of stars that occur in direction of the star changes in the celes- different stellar populations. tial sphere. It is usually measured in The luminosity function of stars is dif- seconds of arc per year. Radial velocity is ferent at different galactic latitudes, and the motion of a star along the line of this is still another phenomenon connected sight and as such is the speed with which with the z distribution of stars of different the star approaches or recedes from the types. At a height of z = 3,000 light-years, observer. It is expressed in kilometres per stars of absolute magnitude 13 and fainter second and is given as either a positive or are nearly as abundant as at the galactic negative figure, depending on whether plane, while stars with absolute magnitude the star is moving away from or toward the 0 are depleted by a factor of 100. observer. The values of the scale height for vari- Astronomers are able to measure ous kinds of objects form the basis for the both the proper motions and radial veloc- segregation of these objects into different ities of stars lying near the Sun. They can, population types. Such objects as open however, determine only the radial veloc- clusters and Cepheid variables that have ities of stellar objects in more distant very small values of the scale height are parts of the Galaxy and so must use these the objects most restricted to the plane data, along with the information gleaned of the Galaxy, while globular clusters and from the local sample of nearby stars, to other extreme Population II objects have ascertain the large-scale motions of stars scale heights of thousands of parsecs, in the Milky Way system. indicating little or no concentration at the plane. Such data and the variation of Proper Motions star density with z distance bear on the mixture of stellar orbit types. They show The proper motions of the stars in the the range from those stars having nearly immediate neighbourhood of the Sun are circular orbits that are strictly limited to a usually very large, as compared with very flat volume centred at the galactic those of most other stars. Those of stars plane to stars with highly elliptical orbits within 17 light-years of the Sun, for that are not restricted to the plane. instance, range from 0.49 to 10.31 arc seconds per year. The latter value is that of Stellar Motions Barnard’s star, which is the star with the largest known proper motion. The tan- A complete knowledge of a star’s motion gential velocity of Barnard’s star is 90 in space is possible only when both its km/sec (56 miles/sec), and, from its radial The Milky Way Galaxy | 39 velocity (−108 km/sec [-67 miles/sec]) velocities range from −119 km/sec (-74 and distance (6 light-years), astronomers miles/sec) to +245 km/sec (+152 miles/ have found that its space velocity (total sec). Most values are on the order of ±20 velocity with respect to the Sun) is 140 km/sec (±12 miles/sec), with a mean value km/sec (87 miles/sec). The distance to of −6 km/sec (-4 miles/sec). this star is rapidly decreasing; it will reach a minimum value of 3.5 light-years in Space Motions about the year 11,800. Space motions are made up of a three- Radial Velocities dimensional determination of stellar motion. They may be divided into a set of Radial velocities, measured along the line components related to directions in the of sight spectroscopically using the Galaxy: U, directed away from the galac- Doppler effect, are not known for all of tic centre; V, in the direction of galactic the recognized stars near the Sun. Of the rotation; and W, toward the north galactic 55 systems within 17 light-years, only 40 pole. For the nearby stars the average val- have well-determined radial velocities. ues for these galactic components are as The radial velocities of the rest are not follows: known, either because of faintness or because of problems resulting from the U = km/sec, V = −28 km/sec, and nature of their spectrum. For example, W = −12 km/sec (U = -5 miles/sec, radial velocities of white dwarfs are often V = −17 miles/sec, and W = -7 miles/sec) very difficult to obtain because of the extremely broad and faint spectral lines These values are fairly similar to those for in some of these objects. Moreover, the the galactic circular velocity components, radial velocities that are determined for which give such stars are subject to further compli- cation because a gravitational redshift U = km/sec, V = −12 km/sec, and generally affects the positions of their W = −7 km/sec (U = −5 miles/sec, spectral lines. The average gravitational V = -7 miles/sec, and W = -4 miles/sec) redshift for white dwarfs has been shown to be the equivalent of a velocity of −51 Note that the largest difference between km/sec (-32 miles/sec). To study the true these two sets of values is for the average motions of these objects, it is necessary V, which shows an excess of 16 km/sec to make such a correction to the observed (10 miles/second) for the nearby stars shifts of their spectral lines. as compared with the circular velocity. For nearby stars, radial velocities are Since V is the velocity in the direction of with very few exceptions rather small. For galactic rotation, this can be understood stars closer than 17 light-years, radial as resulting from the presence of stars in 40 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies the local neighbourhood that have sig- The very large value for V indicates that, nificantly elliptical orbits for which the with respect to circular velocity, this star apparent velocity in this direction is has practically no motion in the direction much less than the circular velocity. of galactic rotation at all. As the Sun’s This fact was noted long before the motion in its orbit around the Galaxy is kinematics of the Galaxy was understood estimated to be approximately 250 km/ and is referred to as the asymmetry of sec (155 miles/sec) in this direction, the stellar motion. value V of −288 km/sec (-179 miles/sec) is The average components of the primarily just a reflection of the solar velocities of the local stellar neighbour- orbital motion. hood also can be used to demonstrate the so-called stream motion. Calculations Solar Motion based on the Dutch-born American astronomer Peter van de Kamp’s table of Solar motion is defined as the calculated stars within 17 light-years, excluding the motion of the Sun with respect to a speci- star of greatest anomalous velocity, reveal fied reference frame. In practice, that dispersions in the V direction and calculations of solar motion provide the W direction are approximately half the information not only on the Sun’s motion size of the dispersion in the U direction. with respect to its neighbours in the This is an indication of a commonality of Galaxy but also on the kinematic proper- motion for the nearby stars; i.e., these ties of various kinds of stars within the stars are not moving entirely at random system. These properties in turn can be but show a preferential direction of used to deduce information on the motion—the stream motion—confined dynamical history of the Galaxy and of somewhat to the galactic plane and to the its stellar components. Because accurate direction of galactic rotation. space motions can be obtained only for individual stars in the immediate vicinity High-Velocity Stars of the Sun (within about 100 light-years), solutions for solar motion involving many One of the nearest 55 stars, called Kapteyn’s stars of a given class are the prime source star, is an example of the high-velocity of information on the patterns of motion stars that lie near the Sun. Its observed for that class. Furthermore, astronomers radial velocity is −245 km/sec (152 miles/ obtain information on the large-scale sec), and the components of its space motions of galaxies in the neighbourhood velocity are of the Galaxy from solar motion solutions because it is necessary to know the space U = 19 km/sec, V = −288 km/sec, and motion of the Sun with respect to the W = −52 km/sec (U = 19 km/sec, centre of the Galaxy (its orbital motion) V = −288 km/sec, and W = −52 km/sec) before such velocities can be calculated. The Milky Way Galaxy | 41

The Sun’s motion can be calculated in stellar spectra caused by blends of spec- by reference to any of three stellar motion tral lines. Of course, the K-term that arises elements: (1) the radial velocities of stars, when a solution for solar motions is cal- (2) the proper motions of stars, or (3) the culated for galaxies results from the space motions of stars. expansion of the system of galaxies and is very large if galaxies at great distances Solar Motion Calculations from the Milky Way Galaxy are included. from Radial Velocities Solar Motion Calculations For objects beyond the immediate neigh- From Proper Motions bourhood of the Sun, only radial velocities can be measured. Initially it is necessary Solutions for solar motion based on the to choose a standard of rest (the reference proper motions of the stars in proper frame) from which the solar motion is to be motion catalogs can be carried out even calculated. This is usually done by select- when the distances are not known and the ing a particular kind of star or a portion radial velocities are not given. It is neces- of space. To solve for solar motion, two sary to consider groups of stars of limited assumptions are made. The first is that dispersion in distance so as to have a well- the stars that form the standard of rest are defined and reasonably spatially-uniform symmetrically distributed over the sky, and reference frame. This can be accomplished the second is that the peculiar motions—the by limiting the selection of stars accord- motions of individual stars with respect ing to their apparent magnitudes. The to that standard of rest—are randomly procedure is the same as the above except distributed. Considering the geometry then that the proper motion components are provides a mathematical solution for the used instead of the radial velocities. The motion of the Sun through the average average distance of the stars of the refer- rest frame of the stars being considered. ence frame enters into the solution of In astronomical literature where solar these equations and is related to the term motion solutions are published, there is often referred to as the secular parallax. often employed a “K-term,” a term that is The secular parallax is defined as 0.24h/r, added to the equations to account for where h is the solar motion in astronomi- systematic errors, the stream motions of cal units per year and r is the mean stars, or the expansion or contraction of the distance for the solar motion solution. member stars of the reference frame. Recent determinations of solar motion Solar Motion Calculations from high-dispersion radial velocities have From Space Motions suggested that most previous K-terms (which averaged a few kilometres per sec- For nearby well-observed stars, it is possible ond) were the result of systematic errors to determine complete space motions and 42 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies to use these for calculating the solar motion. interest. If stars within about 80 light- One must have six quantities: α (the right years of the Sun are used exclusively, the ascension of the star); δ (the declination result is often called the standard solar of the star); μα (the proper motion in right motion. This average, taken for all kinds ascension); μδ (the proper motion in decli- of stars, leads to a velocity nation); ρ (the radial velocity as reduced to the Sun); and r (the distance of the star). V = 19.5 km/sec (12 miles/sec) To find the solar motion, one calculates the velocity components of each star of the The apex of this solar motion is in the sample and the averages of all of these. direction of Solar motion solutions give values for the Sun’s motion in terms of velocity com- α = 270°, δ = +30°. ponents, which are normally reduced to a single velocity and a direction. The direc- The exact values depend on the selection tion in which the Sun is apparently moving of data and method of solution. These with respect to the reference frame is values suggest that the Sun’s motion with called the apex of solar motion. In addi- respect to its neighbours is moderate but tion, the calculation of the solar motion certainly not zero. The velocity difference provides dispersion in velocity. Such dis- is larger than the velocity dispersions for persions are as intrinsically interesting as common stars of the earlier spectral types, the solar motions themselves because a but it is very similar in value to the disper- dispersion is an indication of the integrity sion for stars of a spectral type similar to of the selection of stars used as a reference the Sun. The solar velocity for, say, G5 frame and of its uniformity of kinematic stars is 10 km/sec (6 miles/sec), and the properties. It is found, for example, that dispersion is 21 km/sec (13 miles/sec). dispersions are very small for certain kinds Thus, the Sun’s motion can be considered of stars (e.g., A-type stars, all of which fairly typical for its class in its neighbour- apparently have nearly similar, almost hood. The peculiar motion of the Sun is circular orbits in the Galaxy) and are very a result of its relatively large age and a large for some other kinds of objects (e.g., somewhat noncircular orbit. It is generally the RR Lyrae variables, which show a dis- true that stars of later spectral types show persion of almost 100 km/sec [62 miles/ both greater dispersions and greater val- sec] due to the wide variation in the shapes ues for solar motion, and this characteristic and orientations of orbits for these stars). is interpreted to be the result of a mixture of orbital properties for the later spectral Solar Motion Solutions types, with increasingly large numbers of stars having more highly elliptical orbits. The motion of the Sun with respect to The term basic solar motion has been the nearest common stars is of primary used by some astronomers to define the The Milky Way Galaxy | 43 motion of the Sun relative to stars mov- mixture of young and old. Connected ing in its neighbourhood in perfectly with this is the fact that the solar motion circular orbits around the galactic centre. apex shows a trend for the latitude to The basic solar motion differs from the decrease and the longitude to increase standard solar motion because of the non- with later spectral types. circular motion of the Sun and because of The solar motion can be based on ref- the contamination of the local population erence frames defined by various kinds of of stars by the presence of older stars in stars and clusters of astrophysical inter- noncircular orbits within the limits of est. Data of this sort are interesting the reference frame. The most commonly because of the way in which they make it quoted value for the basic solar motion is possible to distinguish between objects a velocity of 16.5 km/sec (10 miles/sec) with different kinematic properties in the toward an apex with a position Galaxy. For example, it is clear that interstellar calcium lines have relatively small α = 265°, δ = 25°. solar motion and extremely small dispersion because they are primarily When the solutions for solar motion connected with the dust that is limited to are determined according to the spectral the galactic plane and with objects that class of the stars, there is a correlation are decidedly of the Population I class. between the result and the spectral class. On the other hand, RR Lyrae variables The apex of the solar motion, the solar and globular clusters have very large val- motion velocity, and its dispersion are all ues of solar motion and very large correlated with spectral type. Generally dispersions, indicating that they are speaking (with the exception of the very extreme Population II objects that do not early type stars), the solar motion veloc- all equally share in the rotational motion ity increases with decreasing temperature of the Galaxy. The solar motion of these of the stars, ranging from 16 km/sec (10 various objects is an important consider- miles/sec) for late B-type and early A-type ation in determining to what population stars to 24 km/sec (15 miles/sec) for late the objects belong and what their kine- K-type and early M-type stars. The dis- matic history has been. persion similarly increases from a value When some of these classes of objects near 10 km/sec (6 miles/sec) to a value of are examined in greater detail, it is possible 22 km/sec (14 miles/sec). The reason for to separate them into subgroups and this is related to the dynamical history of find correlations with other astrophysical the Galaxy and the mean age and mixture properties. Take, for example, globular of ages for stars of the different spectral clusters, for which the solar motion is types. It is quite clear, for example, that correlated with the spectral type of the stars of early spectral type are all young, clusters. The clusters of spectral types whereas stars of late spectral type are a G0–G5 (the more metal-rich clusters) 44 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies have a mean solar motion of 80 ± 82 km/ to be 225 km/sec (140 miles/sec) in the sec (50 ± 51 miles/sec) (corrected for the direction standard solar motion). The earlier type globular clusters of types F2–F9, on the âII = 90°. other hand, have a mean velocity of 162 ± 36 km/sec (101 ± 22 miles/sec), suggesting It is not a firmly established number, but that they partake much less extensively it is used by convention in most studies. in the general rotation of the Galaxy. In order to arrive at a clear idea of the Similarly, the most distant globular clus- Sun’s motion in the Galaxy as well as of ters have a larger solar motion than the the motion of the Galaxy with respect to ones closer to the galactic centre. Studies neighbouring systems, solar motion has of RR Lyrae variables also show correla- been studied with respect to the Local tions of this sort. The period of an RR Group galaxies and those in nearby space. Lyrae variable, for example, is correlated Hubble determined the Sun’s motion with with its motion with respect to the Sun. respect to the galaxies beyond the Local For type ab RR Lyrae variables, periods Group and found the value of 300 km/sec frequently vary from 0.3 to 0.7 days, and (186 miles/sec) in the direction toward the range of solar motion for this range of galactic longitude 120°, latitude +35°. This period extends from 30 to 205 km/sec (18 velocity includes the Sun’s motion in rela- ± 127 miles/sec), respectively. This condition to its proper circular velocity, its tion is believed to be primarily the result circular velocity around the galactic cen- of the effects of the spread in age and tre, the motion of the Galaxy with respect composition for the RR Lyrae variables in to the Local Group, and the latter’s motion the field, which is similar to, but larger with respect to its neighbours. than, the spread in the properties of the One further question can be consid- globular clusters. ered: What is the solar motion with respect Since the direction of the centre of to the universe? In the 1990s the Cosmic the Galaxy is well established by radio Background Explorer first determined a measurements and since the galactic reliable value for the velocity and direction plane is clearly established by both radio of solar motion with respect to the nearby and optical studies, it is possible to universe. The solar system is headed determine the motion of the Sun with toward the constellation Leo with a veloc- respect to a fixed frame of reference ity of 370 km/sec (230 miles/sec). This centred at the Galaxy and not rotating value was confirmed in the 2000s by an (i.e., tied to the external galaxies). The even more sensitive space telescope, the value for this motion is generally accepted Wilkinson Microwave Anisotropy Probe. CHAPTER 2

Stars

he Milky Way Galaxy is made up of one hundred billion Tof those tantalizing points of light called stars, the massive, self-luminous celestial bodies of gas that shine by radiation derived from their internal energy sources. Our Sun is a star. Of the tens of billions of trillions of stars composing the observable universe, only a very small percentage are visible to the naked eye. Many stars occur in pairs, multiple systems, and star clusters. Members of such stellar groups are physically related through common origin and bound by mutual gravitational attraction. Somewhat related to star clusters are stellar associations, which consist of loose groups of physically similar stars insuffi cient mass as a group to remain together as an organization.

THE NATuRE OF STARS

To say that stars are balls of gas that shine through the workings of their internal energy does not do justice to their full nature and complexity. Not all stars are like our Sun. Some stars are massive giants doomed to burn away in merely millions of years. Others are dim brown dwarfs that are in some ways like stars and in others like the even smaller giant planets.

Size and Activity

The Sun seems like an impressive star. It casts aside the gloom of night and bathes the entire planet in its life-giving 46 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies rays. However, when the Sun is consid- Sun. Sirius A and Vega, though much ered among stars, it is merely average in brighter, also are dwarf stars; their higher its size and the activity of its winds. temperatures yield a larger rate of emission per unit area. Aldebaran A, Arcturus, Variations in Stellar Size and Capella A are examples of giant stars, whose dimensions are much larger than With regard to mass, size, and intrinsic those of the Sun. Observations with an brightness, the Sun is a typical star. Its interferometer (an instrument that mea- approximate mass is 2 × 1030 kg (about sures the angle subtended by the diameter 330,000 Earth masses), its approximate of a star at the observer’s position), com- radius 700,000 km (430,000 miles), and bined with parallax measurements, which its approximate luminosity 4 × 1033 ergs yield a star’s distance, give sizes of 12 and per second (or equivalently 4 × 1023 kilo- 22 solar radii for Arcturus and Aldebaran watts of power). Other stars often have A. Betelgeuse and Antares A are examples their respective quantities measured in of supergiant stars. The latter has a radius terms of those of the Sun. some 300 times that of the Sun, whereas The table lists data pertaining to the the variable star Betelgeuse oscillates 20 brightest stars, or, more precisely, stel- between roughly 300 and 600 solar radii. lar systems, since some of them are Several of the stellar class of white double (binary stars) or even triple stars. dwarf stars, which have low luminosities Successive columns give the name of the and high densities, also are listed. Sirius star, its brightness expressed in units B is a prime example, having a radius called visual magnitudes and the spectral one-thousandth that of the Sun, which is type or types to which the star or its com- comparable to the size of Earth. Among ponents belong, the distance in light-years other notable stars, Rigel A is a young (a light-year being the distance that light supergiant in the constellation Orion, waves travel in one Earth year: 9.46 trillion and Canopus is a bright beacon in the km, or 5.88 trillion miles), and the visual Southern Hemisphere often used for luminosity in terms of that of the Sun. All spacecraft navigation. the primary stars (designated as the A component) are intrinsically as bright as Stellar Activity and Mass Loss or brighter than the Sun. Some of the companion stars are fainter. The Sun’s activity is apparently not Many stars vary in the amount of unique. It has been found that stars of light they radiate. Stars such as Altair, many types are active and have stellar Alpha Centauri A and B, and Procyon A winds analogous to the solar wind. The are called dwarf stars. Their dimensions importance and ubiquity of strong stellar are roughly comparable to those of the winds became apparent only through Stars | 47

The 20 brightest stars Visual Luminosity Visual Magnitude1 and Distance in Relative to Name Spectral Type Light-years2 the Sun A3 B3 A3 B3 Sirius –1.47 A1 V 8.44 DA 8.6 20.8 0.00225 Canopus –0.72 F0 Ib 310 13,000 Arcturus –0.04 K1.5 III 36.7 101.6 Alpha 0.01 G2 V 1.34 K0 V 4.4 1.39 0.409 Centauri Vega 0.03 A0 V 25.3 45.2 Capella 0.084 G8 III 0.96 G0 III 42.2 120 53 Rigel 0.12 B8 I 7.5 B9 860 48,000 54 Procyon 0.38 F5 IV-V 10.7 DZ 11.4 6.66 0.0005 Achernar 0.05 B3 V 140 900 Betelgeuse 0.58 (var.) M2 l 500 10,500 Beta B2 0.6 B1 III 4 390 6,400 280 Centauri (uncertain) Altair 0.77 A7 V 16.8 10.1 Alpha 0.814 B0.5 IV 2.09 B1 V 320 3,600 1,100 Crucis Aldebaran 0.85 K5 III 65 141 Spica 1.04 B1 III–IV 250 1,700 Antares 1.09 (var.) M1.5 l 7 B2.5 V 550 8,100 35.2 Pollux 1.15 K0 III 33.7 28.6 Fomalhaut 1.16 A4 V 25.1 15.7 Deneb 1.25 A2 I 1,400 47,000 Beta 1.30 B0.5 IV 280 1,700 Crucis 1 Negative magnitudes are brightest, and one magnitude difference corresponds to a difference in brightness of 2.5 times; e.g., a star of magnitude –1 is 10 times brighter than one of magnitude 1.5. 2 One light-year equals about 9.46 trillion km. 3 A and B are brighter and fainter components, respectively, of the star. A multiple system is ranked by the brightness of its A component. 4 Combined magnitudes of A and B. 48 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies advances in spaceborne ultraviolet and drives the solar wind, is not enough. X-ray astronomy, as well as in radio and Instead, the winds of the hot stars must infrared surface-based astronomy. be driven directly by the pressure of the X-ray observations that were made energetic ultraviolet radiation emitted during the early 1980s yielded some rather by these stars. Aside from the simple unexpected findings. They revealed that realization that copious quantities of nearly all types of stars are surrounded ultraviolet radiation flow from such hot by coronas having temperatures of one stars, the details of the process are not million kelvins (K) or more. Furthermore, well understood. Whatever is going on, it all stars seemingly display active regions, is surely complex, for the ultraviolet including spots, flares, and prominences spectra of the stars tend to vary with much like those of the Sun. Some stars time, implying that the wind is not exhibit starspots so large that an entire steady. In an effort to understand better face of the star is relatively dark, while the variations in the rate of flow, theorists others display flare activity thousands of are investigating possible kinds of insta- times more intense than that on the Sun. bilities that might be peculiar to luminous The highly luminous hot, blue stars hot stars. have by far the strongest stellar winds. Observations made with radio and Observations of their ultraviolet spectra infrared telescopes, as well as with opti- with telescopes on sounding rockets and cal instruments, prove that luminous cool spacecraft have shown that their wind stars also have winds whose total mass- speeds often reach 3,000 km (roughly flow rates are comparable to those of 2,000 miles) per second, while losing the luminous hot stars, though their mass at rates up to a billion times that of velocities are much lower—about 30 km the solar wind. The corresponding mass- (20 miles) per second. Because luminous loss rates approach and sometimes red stars are inherently cool objects exceed one hundred-thousandth of a (having a surface temperature of about solar mass per year, which means that 3,000 K, or half that of the Sun), they emit one entire solar mass (perhaps a tenth of very little detectable ultraviolet or X-ray the total mass of the star) is carried away radiation. Thus, the mechanism driving into space in a relatively short span of the winds must differ from that in lumi- 100,000 years. Accordingly, the most nous hot stars. luminous stars are thought to lose sub- Winds from luminous cool stars, stantial fractions of their mass during unlike those from hot stars, are rich in their lifetimes, which are calculated to be dust grains and molecules. Since nearly only a few million years. all stars more massive than the Sun even- Ultraviolet observations have proved tually evolve into such cool stars, their that to produce such great winds the winds, pouring into space from vast num- pressure of hot gases in a corona, which bers of stars, provide a major source of Stars | 49 new gas and dust in interstellar space, stars. Among the ancient Greeks, the fact thereby furnishing a vital link in the cycle that the stars did not seem to move was of star formation and galactic evolution. evidence that Earth did not move around As in the case of the hot stars, the specific the Sun. The real answer was that the mechanism that drives the winds of the stars were very far away. How far away cool stars is not understood. At this time, was not known until astronomical tech- investigators can only surmise that gas nology had advanced far enough for turbulence, magnetic fields, or both in the parallax techniques to be used in the 19th atmospheres of these stars are somehow century. responsible. Strong winds also are found to be Determining Stellar Distances associated with objects called protostars, which are huge gas balls that have not yet Distances to stars were first determined become full-fledged stars in which energy by the technique of trigonometric paral- is provided by nuclear reactions. Radio lax, a method still used for nearby stars. and infrared observations of deuterium When the position of a nearby star is (heavy hydrogen) and carbon monoxide measured from two points on opposite (CO) molecules in the Orion Nebula sides of Earth’s orbit (i.e., six months have revealed clouds of gas expanding apart), a small angular (artificial) dis- outward at velocities approaching 100 placement is observed relative to a km (60 miles) per second. Furthermore, background of very remote (essentially high-resolution, very-long-baseline inter- fixed) stars. Using the radius of Earth’s ferometry observations have disclosed orbit as the baseline, the distance of the expanding knots of natural maser (coher- star can be found from the parallactic ent microwave) emission of water vapour angle, p. If p = 1" (one second of arc), the near the star-forming regions in Orion, distance of the star is 206,265 times thus linking the strong winds to the pro- Earth’s distance from the Sun—namely, tostars themselves. The specific causes 3.26 light-years. This unit of distance is of these winds remain unknown, but if termed the parsec, defined as the distance they generally accompany star formation, of an object whose parallax equals one astronomers will have to consider the arc second. Therefore, one parsec equals implications for the early solar system. 3.26 light-years. Since parallax is inversely After all, the Sun was presumably once a proportional to distance, a star at 10 par- protostar too. secs would have a parallax of 0.1". The nearest star to Earth, Proxima Centauri (a Distances to the Stars member of the triple system of Alpha Centauri), has a parallax of 0.7723", mean- For thousands of years humanity has ing that its distance is 1/0.7723, or 1.295, wondered about how far it was to the parsecs, which equals 4.22 light-years. 50 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

The parallax of Barnard’s star, the next seen at great distances, whereas the intrin- closest after the Alpha Centauri system, sically faint stars can be observed only if is 0.549", so that its distance is nearly 6 they are relatively close to Earth. light-years. Errors of such parallaxes are The brightest and nearest stars fall now typically 0.005", meaning that there roughly into three categories: (1) giant is a 50 percent probability that a star stars and supergiant stars that are tens or whose parallax is 0.065" lies between 14.3 even hundreds of solar radii and extremely and 16.7 parsecs (corresponding to paral- low average densities—in fact, several laxes of 0.070" and 0.060", respectively) orders of magnitude less than that of water and an equal chance that it lies outside (one gram per cubic centimetre [1 cubic that range. Thus, measurements of trigo- centimetre = .06 cubic inch]); (2) dwarf stars nometric parallaxes are useful for only ranging from 0.1 to 5 solar radii and with the nearby stars within a few hundred masses from 0.1 to about 10 solar masses; light-years. In fact, of the billions of stars and (3) white dwarf stars, with masses com- in the Milky Way Galaxy, only about 700 parable to that of the Sun but dimensions are close enough to have their parallaxes appropriate to planets, meaning that their measured with useful accuracy. For more average densities are hundreds of thou- distant stars indirect methods are used. sands of times greater than that of water. Most of them depend on comparing the These rough groupings of stars cor- intrinsic brightness of a star (found, for respond to stages in their life histories. example, from its spectrum or other The second category is identified with observable property) with its apparent what is called the main sequence and brightness. includes stars that emit energy mainly by converting hydrogen into helium in their Nearest Stars cores. The first category comprises stars that have exhausted the hydrogen in their The table lists information about the 20 cores and are burning hydrogen within a nearest known stars. Only three stars, shell surrounding the core. The white Alpha Centauri, Procyon, and Sirius, are dwarfs represent the final stage in the life among the 20 brightest and the 20 nearest of a typical star, when most available stars. Ironically, most of the relatively sources of energy have been exhausted nearby stars are dimmer than the Sun and and the star has become relatively dim. are invisible without the aid of a telescope. The large number of binary stars and By contrast, some of the well-known bright even multiple systems is notable. These stars outlining the constellations have star systems exhibit scales comparable in parallaxes as small as the limiting value size to that of the solar system. Some, and of 0.001" and are therefore well beyond perhaps many, of the nearby single stars several hundred light-years distance from have invisible (or very dim) companions the Sun. The most luminous stars can be detectable by their gravitational effects Stars | 51

on the primary star; this orbital motion of have been found to have masses on the the unseen member causes the visible order of 0.001 solar mass or less, which star to “wobble” in its motion through is in the range of planetary rather than space. Some of the invisible companions stellar dimensions. Current observations

The 20 nearest stars visual magnitude* and distance in visual luminosity name spectral type light- relative to the Sun A*** B*** years** A*** B*** Proxima 11.09 M5.5 V 4.2 0.00005 Centauri Alpha Centauri 0.01 G2 V 1.34 K0 V 4.4 1.37 0.403 (A and B only) Barnard’s star 9.53 M4 V 6 0.0004 Wolf 359 13.44 M6 7.8 0.00002 Lalande 21185 7.47 M2 V 8.3 0.00513 Sirius –1.43 A1 V 8.44 DA 8.6 20 0.00225 BL Ceti (A), 12.54 M5.5 V 12.99 M6 V 8.7 0.00005 0.00004 UV Ceti (B) Ross 154 10.43 M3.5 9.7 0.00046 Ross 248 12.29 M5.5 V 10.3 0.00009 Epsilon Eridani 3.73 K2 V 10.5 0.26 Lacaille 9352 7.34 M1.5 V 10.7 0.00971 Ross 128 11.13 M4 V 10.9 0.00031 EZ Aquarii 13.33 M5 V 13.27 M6 V 11.3 0.00004 0.00004 Procyon 0.38 F5 IV-V 10.70 DZ 11.4 6.65 0.0005 61 Cygni 5.21 K5 V 6.03 K7 V 11.4 0.0778 0.0366 GJ 725 8.9 M3 V 9.69 M5 V 11.5 0.0027 0.0013 GX Andromedae 8.08 M1.5 V 11.06 M3.5 V 11.6 0.00575 0.00037 Epsilon Indi 4.69 K5 24.47 T1 11.8 0.135 0.000000018 DX Cancri 14.78 M6.5 V 11.8 0.00001 Tau Ceti 3.49 G8 V 11.9 0.412 * Negative magnitudes are brightest, and one magnitude difference corresponds to a difference in brightness of 2.5 times; e.g., a star of magnitude –1 is 10 times brighter than one of magnitude 1.5. ** One light-year equals about 9.46 trillion km. *** A and B are brighter and fainter components, respectively, of star. 52 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies suggest that they are genuine planets, Stellar Motions though some are merely extremely dim stars (sometimes called brown dwarfs). Accurate measurements of position Nonetheless, a reasonable inference make it possible to determine the move- that can be drawn from these data is that ment of a star across the line of sight double stars and planetary systems are (i.e., perpendicular to the observer)—its formed by similar evolutionary processes. proper motion. The amount of proper motion, denoted by μ (in arc seconds Stellar Positions per year), divided by the parallax of the star and multiplied by a factor of 4.74

Even the basic measurement of where a equals the tangential velocity, VT, in star is in the sky can yield much useful kilometres per second in the plane of the information. Such observations can tell if celestial sphere. a star is related to its neighbours and how The motion along the line of sight it moves through the Galaxy. (i.e., toward the observer), called radial velocity, is obtained directly from spectroÂ Basic Measurements scopic observations. If λ is the wavelength of a characteristic spectral line of some

Accurate observations of stellar positions atom or ion present in the star, and λL the are essential to many problems of astron- wavelength of the same line measured in omy. Positions of the brighter stars can the laboratory, then the difference Δλ, or be measured very accurately in the equa- λ − λL, divided by λL equals the radial torial system (the coordinates of which velocity, VR, divided by the velocity of are called right ascension [α, or RA] and light, c—namely, declination [δ, or DEC] and are given for some epoch—for example, 1950.0 or, cur- Δλ/λL = VR/c. rently, 2000.0). Fainter stars are measured by using photographic plates or electronic Shifts of a spectral line toward the imaging devices (e.g., a charge-coupled red end of the electromagnetic spectrum device, or CCD) with respect to the (i.e., positive VR ) indicate recession, and brighter stars, and finally the entire group those toward the blue end (negative VR ) is referred to the positions of known exter- indicate approach. If the parallax is nal galaxies. These distant galaxies are known, measurements of μ and VR enable far enough away to define an essentially a determination of the space motion of fixed, or immovable, system, whereas the star. Normally, radial velocities are positions of both the bright and faint stars corrected for Earth’s rotation and for its are affected over relatively short periods motion around the Sun, so that they refer of time by galactic rotation and by their to the line-of-sight motion of the star with own motions through the Galaxy. respect to the Sun. Stars | 53

Consider a pertinent example. The the brightest star in the sky (save the proper motion of Alpha Centauri is about Sun), is −1.4. Canopus, the second bright- 3.5 arc seconds, which, at a distance of 4.4 est, has a magnitude of −0.7, while the light-years, means that this star moves faintest star normally seen without the aid 0.00007 light-year in one year. It thus has of a telescope is of the sixth magnitude. a projected velocity in the plane of the sky Stars as faint as the 30th magnitude have of 22 km per second (14 miles per second). been measured with modern telescopes, As for motion along the line of sight, Alpha meaning that these instruments can detect Centauri’s spectral lines are slightly blue- stars about four billion times fainter than shifted, implying a velocity of approach of can the human eye alone. about 20 km per second. The true space The scale of magnitudes comprises a motion, equal to (222 + 202)½ or about 30 geometric progression of brightness. km per second (19 miles per second), sug- Magnitudes can be converted to light gests that this star will make its closest ratios by letting ln and lm be the bright- approach to the Sun (at three light-years’ nesses of stars of magnitudes n and m; distance) some 280 centuries from now. the logarithm of the ratio of the two brightnesses then equals 0.4 times the Light from the stars difference between them—i.e.,

The light that stars emit does more than log(lm/ln) = 0.4(n − m). beautify the night sky. It tells us much about the stars themselves. All we know Magnitudes are actually defined in about what the stars are made of, how terms of observed brightness, a quantity massive they are, and the temperatures of that depends on the light-detecting their surfaces all comes from starlight. device employed. Visual magnitudes were originally measured with the eye, Stellar Magnitudes which is most sensitive to yellow-green light, while photographic magnitudes Stellar brightnesses are usually expressed were obtained from images on old photo- by means of their magnitudes, a usage graphic plates, which were most sensitive inherited from classical times. A star of to blue light. the first magnitude is about 2.5 times as Today, magnitudes are measured elec- bright as one of the second magnitude, tronically, using detectors such as CCDs which in turn is some 2.5 times as bright equipped with yellow-green or blue filters as one of the third magnitude, and so on. to create conditions that roughly corre- A star of the first magnitude is therefore spond to those under which the original 2.55 or 100 times as bright as one of the visual and photographic magnitudes were sixth magnitude. The magnitude of Sirius, measured. Yellow-green magnitudes are which appears to an observer on Earth as still often designated V magnitudes, but 54 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies blue magnitudes are now designated B. astronomers suspect that large numbers The scheme has been extended to other of such faint stars exist, but most of these magnitudes, such as ultraviolet (U ), red objects have so far eluded detection. (R), and near-infrared (I ). Other systems vary the details of this scheme. All mag- Stellar Colours nitude systems must have a reference, or zero, point. In practice, this is fixed Stars differ in colour. Most of the stars in arbitrarily by agreed-upon magnitudes the constellation Orion visible to the measured for a variety of standard stars. naked eye are blue-white, most notably The actually measured brightnesses Rigel (Beta Orionis), but Betelgeuse of stars give apparent magnitudes. These (Alpha Orionis) is a deep red. In the tele- cannot be converted to intrinsic bright- scope, Albireo (Beta Cygni) is seen as two nesses until the distances of the objects stars, one blue and the other orange. One concerned are known. The absolute quantitative means of measuring stellar magnitude of a star is defined as the mag- colours involves a comparison of the nitude it would have if it were viewed at yellow (visual) magnitude of the star with a standard distance of 10 parsecs (32.6 its magnitude measured through a blue light-years). Since the apparent visual filter. Hot, blue stars appear brighter magnitude of the Sun is −26.75, its abso- through the blue filter, while the opposite lute magnitude corresponds to a is true for cooler, red stars. diminution in brightness by a factor of In all magnitude scales, one magni- (2,062,650)2 and is, using logarithms, tude step corresponds to a brightness ratio of 2.512. The zero point is chosen so −26.75 + 2.5 × log(2,062,650)2, that white stars with surface tempera- or −26.75 + 31.57 = 4.82. tures of about 10,000 K have the same visual and blue magnitudes. The conven- This is the magnitude that the Sun tional colour index is defined as the blue would have if it were at a distance of 10 magnitude, B, minus the visual magni- parsecs—an object still visible to the tude, V; the colour index, B − V, of the Sun naked eye, though not a very conspicu- is thus ous one and certainly not the brightest in the sky. Very luminous stars, such as +5.47 − 4.82 = 0.65. Deneb, Rigel, and Betelgeuse, have absolute magnitudes of −7 to , while an Magnitude Systems extremely faint star, such as the companion to the star with the catalog name Problems arise when only one colour BD + 4°4048, has an absolute visual mag- index is observed. If, for instance, a star is nitude of +19, which is about a million found to have, say, a B − V colour index of times fainter than the Sun. Many 1.0 (i.e., a reddish colour), it is impossible Stars | 55 without further information to decide surface temperature of 3,000 K has an whether the star is red because it is cool energy maximum on a wavelength scale or whether it is really a hot star whose at 10000 angstroms (Å) in the far-infrared, colour has been reddened by the passage and most of its energy cannot therefore be of light through interstellar dust. AstroÂ measured as visible light. (One angstrom nomers have overcome these difficulties equals 10−10 metre, or 0.1 nanometre.) by measuring the magnitudes of the same Bright, cool stars can be observed at stars through three or more filters, often infrared wavelengths, however, with spe- U (ultraviolet), B, and V. cial instruments that measure the amount Observations of stellar infrared of heat radiated by the star. Corrections for light also have assumed considerable the heavy absorption of the infrared waves importance. In addition, photometric by water and other molecules in Earth’s air observations of individual stars from must be made unless the measurements spacecraft and rockets have made possible are made from above the atmosphere. the measurement of stellar colours over a The hotter stars pose more difficult large range of wavelengths. These data problems, since Earth’s atmosphere extin- are important for hot stars and for assess- guishes all radiation at wavelengths ing the effects of interstellar attenuation. shorter than 2900 Å. A star whose surface temperature is 20,000 K or higher radi- Bolometric Magnitudes ates most of its energy in the inaccessible ultraviolet part of the electromagnetic The measured total of all radiation at all spectrum. Measurements made with wavelengths from a star is called a bolo- detectors flown in rockets or spacecraft metric magnitude. The corrections extend the observable wavelength region required to reduce visual magnitudes to down to 1000 Å or lower, though most bolometric magnitudes are large for very radiation of distant stars is extinguished cool stars and for very hot ones, but they below 912 Å—a region in which absorption are relatively small for stars such as the by neutral hydrogen atoms in intervening Sun. A determination of the true total space becomes effective. luminosity of a star affords a measure of To compare the true luminosities of its actual energy output. When the energy two stars, the appropriate bolometric cor- radiated by a star is observed from Earth’s rections must first be added to each of surface, only that portion to which the their absolute magnitudes. The ratio of the energy detector is sensitive and that can luminosities can then be calculated. be transmitted through the atmosphere is recorded. Most of the energy of stars like Stellar Spectra the Sun is emitted in spectral regions that can be observed from Earth’s surface. On A star’s spectrum contains information the other hand, a cool dwarf star with a about its temperature, chemical 56 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies composition, and intrinsic luminosity. the second level of hydrogen, and thus Spectrograms secured with a slit spectro- the hydrogen lines are dim. By contrast, graph consist of a sequence of images of at very high temperatures—for instance, the slit in the light of the star at successive that of the surface of a blue giant star— wavelengths. Adequate spectral resolu- the hydrogen atoms are nearly all ionized tion (or dispersion) might show the star and therefore cannot absorb or emit any to be a member of a close binary system, line radiation. Consequently, only faint in rapid rotation, or to have an extended dark hydrogen lines are observed. The atmosphere. Quantitative determination characteristic features of ionized metals of its chemical composition then becomes such as iron are often weak in such hotter possible. Inspection of a high-resolution stars because the appropriate electron spectrum of the star may reveal evidence transitions involve higher energy levels of a strong magnetic field. that tend to be more sparsely populated than the lower levels. Another factor is Line Spectrum that the general “fogginess,” or opacity, of the atmospheres of these hotter stars is Spectral lines are produced by transitions greatly increased, resulting in fewer of electrons within atoms or ions. As the atoms in the visible stellar layers capable electrons move closer to or farther from of producing the observed lines. the nucleus of an atom (or of an ion), The continuous (as distinct from the energy in the form of light (or other radia- line) spectrum of the Sun is produced pri- tion) is emitted or absorbed. The yellow marily by the photodissociation of “D” lines of sodium or the “H” and “K” negatively charged hydrogen ions (H−)— lines of ionized calcium (seen as dark i.e., atoms of hydrogen to which an extra absorption lines) are produced by dis- electron is loosely attached. In the Sun’s crete quantum jumps from the lowest atmosphere, when H− is subsequently energy levels (ground states) of these destroyed by photodissociation, it can atoms. The visible hydrogen lines (the absorb energy at any of a whole range of so-called Balmer series), however, are wavelengths and thus produce a continu- produced by electron transitions within ous range of absorption of radiation. The atoms in the second energy level (or first main source of light absorption in the excited state), which lies well above the hotter stars is the photoionization of ground level in energy. Only at high tem- hydrogen atoms, both from ground level peratures are sufficient numbers of atoms and from higher levels. maintained in this state by collisions, radiations, and so forth to permit an Spectral Analysis appreciable number of absorptions to occur. At the low surface temperatures of The physical processes behind the for- a red dwarf star, few electrons populate mation of stellar spectra are well enough Stars | 57 understood to permit determinations of Catalogue and the Bright Star Catalogue temperatures, densities, and chemical list spectral types from the hottest to the compositions of stellar atmospheres. The coolest stars. These types are designated, star studied most extensively is, of course, in order of decreasing temperature, by the the Sun, but many others also have been letters O, B, A, F, G, K, and M. This group investigated in detail. is supplemented by R- and N-type stars The general characteristics of the spec- (today often referred to as carbon, or tra of stars depend more on temperature C-type, stars) and S-type stars. The R-, N-, variations among the stars than on their and S-type stars differ from the others in chemical differences. Spectral features also chemical composition; also, they are depend on the density of the absorbing invariably giant or supergiant stars. With atmospheric matter, and density in turn the discovery of brown dwarfs, objects is related to a star’s surface gravity. Dwarf that form like stars but do not shine stars, with great surface gravities, tend to through thermonuclear fusion, the system have high atmospheric densities; giants of stellar classification has been expanded and supergiants, with low surface gravities, to include spectral types L and T. have relatively low densities. Hydrogen The spectral sequence O through M absorption lines provide a case in point. represents stars of essentially the same Normally, an undisturbed atom radiates chemical composition but of different a very narrow line. If its energy levels are temperatures and atmospheric pressures. perturbed by charged particles passing This simple interpretation, put forward nearby, it radiates at a wavelength near its in the 1920s by the Indian astrophysicist characteristic wavelength. In a hot gas, the Meghnad N. Saha, has provided the physi- range of disturbance of the hydrogen lines cal basis for all subsequent interpretations is very high, so that the spectral line radi- of stellar spectra. The spectral sequence ated by the whole mass of gas is spread out is also a colour sequence: the O- and considerably; the amount of blurring B-type stars are intrinsically the bluest depends on the density of the gas in a and hottest; the M-, R-, N-, and S-type known fashion. Dwarf stars such as Sirius stars are the reddest and coolest. show broad hydrogen features with exten- In the case of cool stars of type M, the sive “wings” where the line fades slowly spectra indicate the presence of familiar out into the background, while supergiant metals, including iron, calcium, magne- stars, with less-dense atmospheres, display sium, and also titanium oxide molecules relatively narrow hydrogen lines. (TiO), particularly in the red and green parts of the spectrum. In the somewhat Classification of Spectral Types hotter K-type stars, the TiO features disappear, and the spectrum exhibits a Most stars are grouped into a small num- wealth of metallic lines. A few especially ber of spectral types. The Henry Draper stable fragments of molecules such as 58 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies cyanogen (CN) and the hydroxyl radical stronger, attaining their maximum inten- (OH) persist in these stars and even in sities in A-type stars, in which the surface G-type stars such as the Sun. The spectra temperature is about 9,000 K. Thereafter, of G-type stars are dominated by the these absorption lines gradually fade as characteristic lines of metals, particularly the hydrogen becomes ionized. those of iron, calcium, sodium, magne- The hot B-type stars, such as Epsilon sium, and titanium. Orionis, are characterized by lines of The behaviour of calcium illustrates helium and of singly ionized oxygen, the phenomenon of thermal ionization. At nitrogen, and neon. In very hot O-type low temperatures a calcium atom retains stars, lines of ionized helium appear. all of its electrons and radiates a spectrum Other prominent features include lines of characteristic of the neutral, or normal, doubly ionized nitrogen, oxygen, and car- atom; at higher temperatures collisions bon and of trebly ionized silicon, all of between atoms and electrons and the which require more energy to produce. absorption of radiation both tend to detach In the more modern system of spectral electrons and to produce singly ionized classification, called the MK system (after calcium atoms. At the same time, these the American astronomers William W. ions can recombine with electrons to pro- Morgan and Philip C. Keenan, who intro- duce neutral calcium atoms. At high duced it), luminosity class is assigned to temperatures or low electron pressures, or the star along with the Draper spectral type. both, most of the atoms are ionized. At For example, the star Alpha Persei is classi- low temperatures and high densities, the fied as F5 Ib, which means that it falls about equilibrium favours the neutral state. The halfway between the beginning of type F concentrations of ions and neutral atoms (i.e., F0) and of type G (i.e., G0). The Ib suf- can be computed from the temperature, fix means that it is a moderately luminous the density, and the ionization potential supergiant. The star Pi Cephei, classified as (namely, the energy required to detach an G2 III, is a giant falling between G0 and K0 electron from the atom). but much closer to G0. The Sun, a dwarf The absorption line of neutral cal- star of type G2, is classified as G2 V. A star cium at 4227 Å is thus strong in cool of luminosity class II falls between giants M-type dwarf stars, in which the pressure and supergiants; one of class IV is called a is high and the temperature is low. In the subgiant. hotter G-type stars, however, the lines of ionized calcium at 3968 and 3933 Å (the Bulk Stellar Properties “H” and “K” lines) become much stronger than any other feature in the spectrum. When a star is considered as a whole, its In stars of spectral type F, the lines of properties reveal much of interest. From neutral atoms are weak relative to those a star’s temperature to how it interacts of ionized atoms. The hydrogen lines are with a companion star, the consideration Stars | 59

2 4 . of bulk stellar properties has been and L = 4πR σT eff will continue to be a major part of astronomical studies. This relation defines the star’s equivalent blackbody, or effective, temperature. Stellar Temperatures Since the total energy radiated by a star cannot be directly observed (except Temperatures of stars can be defined in a in the case of the Sun), the effective tem- number of ways. From the character of the perature is a derived quantity rather than spectrum and the various degrees of an observed one. Yet, theoretically, it is ionization and excitation found from its the fundamental temperature. If the bolo- analysis, an ionization or excitation tem- metric corrections are known, the effective perature can be determined. temperature can be found for any star A comparison of the V and B magni- whose absolute visual magnitude and tudes yields a B − V colour index, which is radius are known. Effective temperatures related to the colour temperature of the are closely related to spectral type and star. The colour temperature is therefore range from about 40,000 K for hot O-type a measure of the relative amounts of stars, through 5,800 K for stars like the radiation in two more or less broad wave- Sun, to about 800 K for brown dwarfs. length regions, while the ionization and excitation temperatures pertain to the Stellar Masses temperatures of strata wherein spectral lines are formed. The mass of most stars lies within the Provided that the angular size of a range of 0.3 to 3 solar masses. One of star can be measured and that the total the most massive stars determined to energy flux received at Earth (corrected date is the O3-type star HD 93250, a giant for atmospheric extinction) is known, that has perhaps 120 solar masses. There the so-called brightness temperature is a theoretical upper limit to the masses can be found. of nuclear-burning stars (the Eddington

The effective temperature, Teff, of a limit), which limits stars to no more than star is defined in terms of its total energy a few hundred solar masses. The physics 4 output and radius. Thus, since σT eff is the of instability and fragmentation probably rate of radiation per unit area for a per- prohibits the formation of stars much fectly radiating sphere and if L is the total more than 100 times the mass of the Sun. radiation (i.e., luminosity) of a star con- On the low mass side, most stars seem to sidered to be a sphere of radius R, such have at least 0.1 solar mass. The theoreti- a sphere (called a blackbody) would cal lower mass limit for an ordinary star emit a total amount of energy equal to its is about 0.075 solar mass, for below this surface area, 4πR2, multiplied by its energy value an object cannot attain a central per unit area. In symbols, temperature high enough to enable it to 60 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies shine by nuclear energy. Instead, it may Visual Binaries produce a much lower level of energy by gravitational shrinkage. If its mass is not Visual binaries can be seen as double much below the critical 0.075 solar mass stars with the telescope. True doubles, as value, it will appear as a very cool, dim distinguished from apparent doubles star known as a brown dwarf. Its evolu- caused by line-of-sight effects, move tion is simply to continue cooling toward through space together and display a eventual extinction. At still somewhat common space motion. Sometimes a com- lower masses, the object would be a giant mon orbital motion can be measured as planet. Jupiter, with a mass roughly 0.001 well. Provided that the distance to the that of the Sun, is just such an object, binary is known, such systems permit a emitting a very low level of energy (apart determination of stellar masses, m1 and from reflected sunlight) that is derived m2, of the two members. The angular from gravitational shrinkage. radius, a", of the orbit (more accurately, Brown dwarfs were late to be discov- its semimajor axis) can be measured ered, the first unambiguous identification directly, and, with the distance known, having been made in 1995. It is estimated, the true dimensions of the semimajor however, that hundreds must exist in the axis, a, can be found. If a is expressed in solar neighbourhood. An extension of astronomical units, which is given by a the spectral sequence for objects cooler (measured in seconds of arc) multiplied than M-type stars has been constructed, by the distance in parsecs, and the period, using L for warmer brown dwarfs and T P, also measured directly, is expressed in for cooler ones. A major observational years, then the sum of the masses of the difference between the two types is the two orbiting stars can be found from an absence (in L-type dwarfs) or presence application of Kepler’s third law. (An (in T-type dwarfs) of methane in their astronomical unit is the average distance spectra. The presence of methane in the from Earth to the Sun, approximately cooler brown dwarfs emphasizes their 149,597,870 km [92,955,808 miles].) In similarity to giant planets. The class Y symbols, has been proposed for objects even cooler 3 2 than those of class T. (m1 + m2) = a /P Masses of stars can be found only from binary systems and only if the scale in units of the Sun’s mass. of the orbits of the stars around each For example, for the binary system 70 other is known. Binary stars are divided Ophiuchi, P is 87.8 years, and the distance into three categories, depending on the is 5.0 parsecs; thus, a is 22.8 astronomical mode of observation employed: visual units, and binaries, spectroscopic binaries, and eclipsing binaries. m1 + m2 = 1.56 Stars | 61 solar masses. From a measurement of the cannot be found for most spectroscopic motions of the two members relative to binaries, since the angle between the the background stars, the orbit of each orbit plane and the plane of the sky can- star has been determined with respect to not be determined. If spectra from both their common centre of gravity. The mass members are observed, mass ratios can ratio, m2/(m1 + m2), is 0.42; the individual be found. If one spectrum alone is masses for m1 and m2, respectively, are observed, only a quantity called the mass then 0.90 and 0.66 solar mass. function can be derived, from which is The star known as 61 Cygni was the calculated a lower limit to the stellar first whose distance was measured (via masses. If a spectroscopic binary is also parallax by the German astronomer observed to be an eclipsing system, the Friedrich W. Bessel in the mid-19th cen- inclination of the orbit and often the val- tury). Visually, 61 Cygni is a double star ues of the individual masses can be separated by 83.2 astronomical units. Its ascertained. members move around one another with a period of 653 years. It was among the Eclipsing Binaries first stellar systems thought to contain a potential planet, although this has not An eclipsing binary consists of two close been confirmed and is now considered stars moving in an orbit so placed in unlikely. Nevertheless, since the 1990s a space in relation to Earth that the light of variety of discovery techniques have con- one can at times be hidden behind the firmed the existence of more than 300 other. Depending on the orientation of planets orbiting other stars. the orbit and sizes of the stars, the eclipses can be total or annular (in the latter, a Spectroscopic Binaries ring of one star shows behind the other at the maximum of the eclipse) or both Spectroscopic binary stars are found from eclipses can be partial. The best known observations of radial velocity. At least example of an eclipsing binary is Algol the brighter member of such a binary can (Beta Persei), which has a period (interval be seen to have a continuously changing between eclipses) of 2.9 days. The brighter periodic velocity that alters the wave- (B8-type) star contributes about 92 per- lengths of its spectral lines in a rhythmic cent of the light of the system, and the way. The velocity curve repeats itself eclipsed star provides less than 8 percent. exactly from one cycle to the next, and The system contains a third star that is the motion can be interpreted as orbital not eclipsed. Some 20 eclipsing binaries motion. In some cases, rhythmic changes are visible to the naked eye. in the lines of both members can be The light curve for an eclipsing measured. Unlike visual binaries, the binary displays magnitude measure- semimajor axes or the individual masses ments for the system over a complete 62 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies light cycle. The light of the variable star temperature star is paired with a larger is usually compared with that of a nearby object of low surface brightness and if (comparison) star thought to be fixed in the distance between the stars is small, the brightness. Often, a deep, or primary, part of the cool star facing the hotter one minimum is produced when the compo- is substantially brightened by it. Just nent having the higher surface brightness before (and just after) secondary eclipse, is eclipsed. It represents the total eclipse this illuminated hemisphere is pointed and is characterized by a flat bottom. A toward the observer, and the total light of shallower secondary eclipse occurs when the system is at a maximum. the brighter component passes in front of The properties of stars derived from the other; it corresponds to an annular eclipsing binary systems are not neces- eclipse (or transit). In a partial eclipse sarily applicable to isolated single stars. neither star is ever completely hidden, Systems in which a smaller, hotter star is and the light changes continuously dur- accompanied by a larger, cooler object ing an eclipse. are easier to detect than are systems that The shape of the light curve during contain, for example, two main-sequence an eclipse gives the ratio of the radii of stars. In such an unequal system, at least the two stars and also one radius in terms the cooler star has certainly been affected of the size of the orbit, the ratio of lumi- by evolutionary changes and probably nosities, and the inclination of the orbital so has the brighter one. The evolutionary plane to the plane of the sky. development of two stars near one If radial-velocity curves are also avail- another does not exactly parallel that of able—i.e., if the binary is spectroscopic as two well-separated or isolated ones. well as eclipsing—additional information Eclipsing binaries include combina- can be obtained. When both velocity tions of a variety of stars ranging from curves are observable, the size of the orbit white dwarfs to huge supergiants (e.g., as well as the sizes, masses, and densities VV Cephei), which would engulf Jupiter of the stars can be calculated. Furthermore, and all the inner planets of the solar sys- if the distance of the system is measur- tem if placed at the position of the Sun. able, the brightness temperatures of the Some members of eclipsing binaries individual stars can be estimated from are intrinsic variables, stars whose energy their luminosities and radii. All of these output fluctuates with time. In many such procedures have been carried out for the systems, large clouds of ionized gas swirl faint binary Castor C (two red-dwarf com- between the stellar members. In others, ponents of the six-member Castor such as Castor C, at least one of the faint multiple star system) and for the bright M-type dwarf components might be a B-type star Mu Scorpii. flare star, one in which the brightness can Close stars may reflect each other’s unpredictably and suddenly increase to light noticeably. If a small, high- many times its normal value. Stars | 63

Binaries and Extrasolar and appear so close to it as to be unde- Planetary Systems tectable from even the nearest star. If candidate stars are treated as possible Near the Sun, most stars are members of spectroscopic binaries, however, then one binaries, and many of the nearest single may look for a periodic change in the stars are suspected of having compan- star’s radial velocity caused by a planet ions. Although some binary members are swinging around it. The effect is very separated by hundreds of astronomical small—even Jupiter would cause a change units and others are contact binaries in the apparent radial velocity of the Sun (stars close enough for material to pass of only about 10 metres (33 feet) per sec- between them), binary systems are most ond spread over Jupiter’s orbital period frequently built on the same scale as that of about 12 years at best. of the solar system—namely, on the order Current techniques using very large of about 10 astronomical units. The divi- telescopes to study fairly bright stars can sion in mass between two components of measure radial velocities with a precision a binary seems to be nearly random. A of a few metres per second, provided that mass ratio as small as about 1:20 could the star has very sharp spectral lines, occur about 5 percent of the time, and such as is observed for Sun-like stars and under these circumstances a planetary stars of types K and M. This means that at system comparable to the solar system is present the radial-velocity method nor- able to form. mally can detect only massive extrasolar The formation of double and multiple planets. Planets like Earth, 300 times less stars on the one hand and that of plane- massive, would cause too small a change tary systems on the other seem to be in radial velocity to be detectable pres- different facets of the same process. ently. Moreover, the closer the planet is to Planets are probably produced as a natu- its parent star, the greater and quicker ral by-product of star formation. Only a the velocity swing, so that detection of small fraction of the original nebula giant planets close to a star is favoured matter is likely to be retained in planets, over planets farther out. since much of the mass and angular Even when a planet is detected, the momentum is swept out of the system. usual spectroscopic binary problem of Conceivably, as many as 100 million stars not knowing the angle between the orbit could have bona fide planets in the Milky plane and that of the sky allows only a Way Galaxy. minimum mass to be assigned to the Because planets are much fainter planet. The first planet discovered with than the stars they orbit, extrasolar plan- this technique was 51 Pegasi in 1995. ets are extremely difficult to detect One exception to this last problem is directly. Jupiter, for example, would be HD 209458, a seventh-magnitude G0 V only one-billionth as bright as the Sun star about 150 light-years away with a 64 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

The extrasolar planet Fomalhaut b in images taken by the Hubble Space Telescope in 2004 and 2006. The black spot at the centre of the image is a coronagraph used to block the light from Fomalhaut, which is located at the white dot. The oval ring is Fomalhaut’s dust belt, and the lines radiating from the centre of the image are scattered starlight. NASA; ESA; P. Kalas; J. Graham, E. Chiang; E. Kite, University of California, Berkeley; M. Clampin, NASA Goddard Space Flight Center; M. Fitzgerald, Lawrence Livermore National Laboratory; and K. Stapelfeldt and J. Krist, NASA/JPL planetary object orbiting it every 3.5 days. even more of a giant than Jupiter itself. Soon after the companion was discovered What was unexpected is its proximity to in 1999 by its eff ect on the star’s radial the parent star—more than 100 times velocity, it also was found to be eclipsing closer than Jupiter is to the Sun, raising the star, meaning that its orbit is oriented the question of how a giant gaseous almost edge-on toward Earth. This fortu- planet that close can survive the star’s nate circumstance allowed determination radiation. The fact that many other extra- of the planet’s mass and radius—0.69 and solar planets have been found to have 1.42 times those of Jupiter, respectively. orbital periods measured in days rather These numbers imply that the planet is than years, and thus to be very close to Stars | 65 their parent stars, suggests that the HD as are the Sun and its two largest planets, 209458 case is not unusual. There are Jupiter and Saturn. Some of these planets also some confirmed cases of planets seem to be distended in size as a result of around supernova remnants called pul- heating by their stars. The lowest mass sars, although whether the planets transiting planets contain larger fractions preceded the supernova explosions that of heavier elements, as do the smaller produced the pulsars or were acquired planets within the solar system. afterward remains to be determined. The majority of extrasolar planets The first extrasolar planets were dis- with orbital periods longer than two covered in 1992. More than 300 extrasolar weeks have quite eccentric (elongated) planets were known by the early years of orbits. Within the solar system, the plan- the 21st century, with more such discover- ets, especially the larger ones, travel on ies being added regularly. Some of those nearly circular paths about the Sun. studied have minimum masses of 40 or Stars that contain a larger fraction of even 60 Jupiters, which means they are heavy elements (i.e., any element aside likely brown dwarfs. from hydrogen and helium) are more Between 5 and 10 percent of stars sur- likely to possess detectable planets. More veyed have planets at least 100 times as massive stars are more likely to host plan- massive as Earth with orbital periods of a ets more massive than Saturn, but this few Earth years or less. Almost 1 percent correlation may not exist for smaller plan- of stars have such giant planets in very ets. Many extrasolar planets orbit stars close orbits, with orbital periods of less that are members of binary star systems, than one week. In contrast, Jupiter, which and it is common for stars with one has the shortest orbital period of any detectable planet to have others. The large planet (i.e., any planet more mas- planets detected so far around stars other sive than Earth) in the solar system, takes than the Sun have masses from nearly nearly 12 years to travel around the Sun. twice to thousands of times that of Earth. Even the closest planet to the Sun, tiny Most, if not all, appear to be too massive Mercury, requires 88 days to complete an to support life, but this too is the result of orbit. Models of planetary formation sug- detection biases and does not indicate gest that giant extrasolar planets detected that planets like Earth are uncommon. very near their stars were formed at Research in the field of extrasolar greater distances and migrated inward as planets is advancing rapidly, as new tech- a result of gravitational interactions with nologies enable the detection of smaller remnants of the circumstellar disks from and more distant planets as well as the which they accumulated. characterization of previously detected The most massive planets that transit planets. Almost all the extrasolar planetary their stars are made primarily of the two systems known appear very different from lightest elements, hydrogen and helium, the solar system, but planets like those 66 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

The planetary system of HR 8799. Christian Marois/Bruce Macintosh/National Research Council Canada(NRC)/Keck Observatory within the solar system would be very dif- have detectable planets, it is still not fi cult to fi nd around other stars with known whether the solar system is nor- current technology. Thus, as more than mal or unusual. The U.S. National 90 percent of those stars surveyed do not Aeronautics and Space Administration’s Stars | 67

Kepler mission was launched in 2009 and other disturbances that change the and—using transit photometry from space, details of their spectra. optimized to achieve unprecedented A more recent method, called speckle sensitivity for small planets with orbital interferometry, has been developed to periods of up to two years—aims to dis- reproduce the true disks of red super- cover whether planets analogous to Earth giant stars and to resolve spectroscopic are common or rare. binaries such as Capella. The speckle In addition to the growing evidence phenomenon is a rapidly changing inter- for existence of extrasolar planets, space- ference-diffraction effect seen in a highly based observatories designed to detect magnified diffraction image of a star infrared radiation have found more than observed with a large telescope. 100 young nearby stars (including Vega, If the absolute magnitude of a star Fomalhaut and Beta Pictoris) to have and its temperature are known, its size disks of warm matter orbiting them. This can be computed. The temperature deter- matter is composed of myriad particles mines the rate at which energy is emitted mostly about the size of sand grains and by each unit of area, and the total lumi- might be taking part in the first stage of nosity gives the total power output. Thus, planetary formation. the surface area of the star and, from it, the radius of the object can be estimated. Stellar Radii This is the only way available for estimating the dimensions of white dwarf Angular sizes of bright red giant and stars. The chief uncertainty lies in choos- supergiant stars were first measured ing the temperature that represents the directly during the 1920s, using the rate of energy emission. principle of interference of light. Only bright stars with large angular size can Average Stellar Values be measured by this method. Provided the distance to the star is known, the Main-sequence stars range from very physical radius can be determined. luminous objects to faint M-type dwarf Eclipsing binaries also provide exten- stars, and they vary considerably in their sive data on stellar dimensions. The surface temperatures, their bolometric timing of eclipses provides the angular (total) luminosities, and their radii. size of any occulting object, and so ana- Moreover, for stars of a given mass, a fair lyzing the light curves of eclipsing spread in radius, luminosity, surface binaries can be a useful means of deter- temperature, and spectral type may exist. mining the dimensions of either dwarf or This spread is produced by stellar evolu- giant stars. Members of close binary tionary effects and tends to broaden the systems, however, are sometimes subject main sequence. Masses are obtained to evolutionary effects, mass exchange, from visual and eclipsing binary systems 68 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies observed spectroscopically. Radii are independently discovered the relations found from eclipsing binary systems, from shown in it. direct measurements in a few favourable As is seen in the diagram, most of the cases, by calculations, and from absolute congregated stars are dwarfs lying closely visual magnitudes and temperatures. around a diagonal line called the main Average values for radius, bolometric sequence. These stars range from hot, O- luminosity, and mass are meaningful only and B-type, blue objects at least 10,000 for dwarf stars. Giant and subgiant stars times brighter than the Sun down through all show large ranges in radius for a given white A-type stars such as Sirius to mass. Conversely, giant stars of very orange K-type stars such as Epsilon nearly the same radius, surface tempera- Eridani and finally to M-type red dwarfs ture, and luminosity can have appreciably thousands of times fainter than the Sun. different masses. The sequence is continuous; the luminosities fall off smoothly with decreasing Stellar Statistics surface temperature; the masses and radii decrease but at a much slower rate; and Some of the most important generaliza- the stellar densities gradually increase. tions concerning the nature and evolution The second group of stars to be recog- of stars can be derived from correlations nized was a group of giants—such objects between observable properties and cer- as Capella, Arcturus, and Aldebaran— tain statistical results. One of the most which are yellow, orange, or red stars important of these correlations concerns about 100 times as bright as the Sun and temperature and luminosity—or, equiva- have radii on the order of 10–30 million lently, colour and magnitude. km (about 6–20 million miles, or 15–40 times as large as the Sun). The giants lie Hertzsprung-Russell Diagram above the main sequence in the upper right portion of the diagram. The cate- When the absolute magnitudes of stars gory of supergiants includes stars of all (or their intrinsic luminosities on a loga- spectral types; these stars show a large rithmic scale) are plotted in a diagram spread in intrinsic brightness, and some against temperature or, equivalently, even approach absolute magnitudes of against the spectral types, the stars do −7 or . A few red supergiants, such as not fall at random on the diagram but the variable star VV Cephei, exceed in tend to congregate in certain restricted size the orbit of Jupiter or even that of domains. Such a plot is usually called a Saturn, although most of them are smaller. Hertzsprung-Russell diagram, named for Supergiants are short-lived and rare the early 20th-century astronomers Ejnar objects, but they can be seen at great Hertzsprung of Denmark and Henry distances because of their tremendous Norris Russell of the United States, who luminosity. Stars | 69

Schematic spectrum–luminosity correlation (Hertzsprung–Russell diagram) of spiral-arm stars in the neighbourhood of the Sun. From Astrophysical Journal, reproduced by permission of the American Astronomical Society

Subgiants are stars that are redder The white dwarf domain lies about 10 and larger than main-sequence stars of magnitudes below the main sequence. the same luminosity. Many of the best These stars are in the last stages of their known examples are found in close binary evolution. systems where conditions favour their The spectrum-luminosity diagram has detection. numerous gaps. Few stars exist above the 70 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies white dwarfs and to the left of the main difficulty. As a consequence, their colours sequence. The giants are separated from rather than their spectra must be mea- the main sequence by a gap named for sured. Since the colours are closely Hertzsprung, who in 1911 became the first related to surface temperature and there- to recognize the difference between fore to spectral types, equivalent spectral main-sequence and giant stars. The types may be used; but it is stellar colours, actual concentration of stars differs not spectral types, that are observed in considerably in different parts of the this instance. diagram. Highly luminous stars are rare, The differences between the two whereas those of low luminosity are very Hertzsprung-Russell diagrams are strik- numerous. ing. In the second figure, there are no The spectrum-luminosity diagram supergiants, and, instead of a domain at applies to the stars in the galactic spiral an absolute magnitude of about 0, the arm in the neighbourhood of the Sun and giant stars form a branch that starts high represents what would be obtained if a and to the right at about −3.5 for very red composite Hertzsprung-Russell diagram stars and flows in a continuous sequence was constructed combining data for a until it reaches an absolute magnitude of large number of the star groups called about 0. At that point the giant branch open (or galactic) star clusters, as, for splits—a main band of stars, all about the example, the double cluster h and χ Persei, same colour, proceeds downward (i.e., to the Pleiades, the Coma cluster, and the fainter stars) to a magnitude of about +3 Hyades. It includes very young stars, a and then connects to the main sequence few million years old, as well as ancient at about +4 by way of a narrow band. The stars, perhaps as old as 10 billion years. main sequence of Population II stars By contrast, the second diagram extends downward to fainter, redder stars exhibits the type of temperature-lumi- in much the same way as in the spiral- nosity (or colour-magnitude) relation arm Population I stars. The main sequence characteristic of stars in globular clus- ends at about spectral type G, however, ters, in the central bulge of the Galaxy, and does not extend up through the A, and in elliptical external galaxies— B, and O spectral types, though occasion- namely, of the so-called stellar Population ally a few such stars are found in the II. (In addition to these oldest objects, region normally occupied by the main Population II includes other very old sequence. stars that occur between the spiral arms The other band of stars formed from of the Galaxy and at some distance above the split of the giant branch is the “horizon- and below the galactic plane.) Because tal branch,” which falls near magnitude these systems are very remote from the +0.6 and fills the aforementioned observer, the stars are faint, and their Hertzsprung gap, extending to increas- spectra can be observed only with ingly blue stars beyond the RR Lyrae Stars | 71 stars, which are indicated by the cross- being their ages. The young cluster h hatched area in the diagram. Among and χ Persei, which is a few million years these blue hot stars are found novae and old, contains stars ranging widely in the nuclei of planetary nebulae, the latter luminosity. Some stars have already so called because their photographic evolved into the supergiant stage (in image resembles that of a distant planet. such a diagram the top of the main Not all globular clusters show identical sequence is bent over). The stars of lumi- colour-magnitude diagrams, which may nosity 10,000 times greater than that of be due to differences in the cluster ages the Sun have already largely depleted the or other factors. hydrogen in their cores and are leaving the main sequence. Estimates of Stellar Ages The brightest stars of the Pleiades cluster, aged about 100 million years, The shapes of the colour-magnitude dia- have begun to leave the main sequence grams permit estimates of globular-cluster and are approaching the critical phase ages. Stars more massive than about 1.3 when they will have exhausted all the solar masses have evolved away from the hydrogen in their cores. There are no main sequence at a point just above giants in the Pleiades. Presumably, the the position occupied by the Sun. The cluster contained no stars as massive as time required for such a star to exhaust some of those found in h and χ Persei. the hydrogen in its core is about 5–6 The cluster known as Praesepe, or the billion years, and the cluster must be at Beehive, at an age of 790 million years, is least as old. More ancient clusters have older than the Pleiades. All stars much been identified. In the Galaxy, globular more luminous than the first magnitude clusters are all very ancient objects, hav- have begun to leave the main sequence; ing ages within a few billion years of there are some giants. The Hyades, about the average of 14 billion years. In the 620 million years old, displays a similar Magellanic Clouds, however, clusters colour-magnitude array. These clusters exist that resemble globular ones, but contain a number of white dwarfs, indi- they contain numerous blue stars and cating that the initially most luminous therefore must be relatively young. stars have already run the gamut of evo- Open clusters in the spiral arms of lution. In a very old cluster such as M67, the Galaxy—extreme Population I—tell a which is 4.5 billion years old, all of the bright somewhat different story. A colour- main-sequence stars have disappeared. magnitude diagram can be plotted for a The colour-magnitude diagrams for number of different open clusters—for globular and open clusters differ quanti- example, the double cluster h and χ Persei, tatively because the latter show a wider the Pleiades, Praesepe, and M67—with the range of ages and differ in chemical com- main feature distinguishing the clusters position. Most globular clusters have 72 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies smaller metal-to-hydrogen ratios than do magnitude +0.6, corresponding to the open clusters or the Sun. The gaps horizontal branch for Population II, and between the red giants and blue main- no stars as bright as absolute magnitude sequence stars of the open clusters −5. The luminosity function for pure (Population I) often contain unstable Population I is evaluated best from open stars such as variables. The Cepheid vari- star clusters, the stars in such a cluster able stars, for instance, fall in these gaps. being at about the same distance. The The giant stars of the Praesepe clus- neighbourhood of the Sun includes ter are comparable to the brightest stars examples of both Populations I and II. in M67. The M67 giants have evolved from the main sequence near an absolute Mass-Luminosity Correlations magnitude of +3.5, whereas the Praesepe giants must have masses about twice as A plot of mass against bolometric lumi- great as those of the M67 giants. Giant nosity for visual binaries for which good stars of the same luminosity may there- parallaxes and masses are available shows fore have appreciably different masses. that for stars with masses comparable to that of the Sun the luminosity, L, varies as Numbers of Stars Versus a power, 3 + β, of the mass M. This relation Luminosity can be expressed as

Of great statistical interest is the relation- L = (M)3+β. ship between the luminosities of the stars and their frequency of occurrence. The The power differs for substantially fainter naked-eye stars are nearly all intrinsically or much brighter stars. brighter than the Sun, but the opposite is This mass-luminosity correlation true for the known stars within 20 light- applies only to unevolved main-sequence years of the Sun. The bright stars are easily stars. It fails for giants and supergiants seen at great distances; the faint ones can and for the subgiant (dimmer) compo- be detected only if they are close. Only if nents of eclipsing binaries, all of which stars of magnitude +11 were a billion times have changed considerably during their more abundant than stars of magnitude −4 lifetimes. It does not apply to any stars in could they be observed to some fixed limit a globular cluster not on the main of apparent brightness. sequence, or to white dwarfs that are The luminosity function depends on abnormally faint for their masses. population type. The luminosity function The mass-luminosity correlation, pre- for pure Population II differs substan- dicted theoretically in the early 20th tially from that for pure Population I. century by the English astronomer Arthur There is a small peak near absolute Eddington, is a general relationship that Stars | 73 holds for all stars having essentially the absolute magnitudes, provided the inter- same internal density and temperature stellar absorption is also known. distributions—i.e., for what are termed the same stellar models. Classification

Variable Stars Variables are often classified as behaving like a prototype star, and the entire class Many stars are variable. Some are geo- is then named for this star—e.g., RR Lyrae metric variables, as in the eclipsing stars are those whose variability follows binaries considered earlier. Others are the pattern of the star RR Lyrae. The most intrinsically variable—i.e., their total important classes of intrinsically variable energy output fluctuates with time. Such stars are intrinsic variable stars are dealt with in this section. (1) Pulsating variables—stars whose A fair number of stars are intrinsi- variations in light and colour are cally variable. Some objects of this type thought to arise primarily from were found by accident, but many were stellar pulsations. These include detected as a result of carefully planned Beta Canis Majoris stars, RR searches. Variable stars are important in Lyrae stars, and Delta Scuti stars, astronomy for several reasons. They usu- all with short regular periods of ally appear to be stars at critical or less than a day; Cepheids, with short-lived phases of their evolution; periods between 1 and 100 days; detailed studies of their light and spec- and long-period variables, semi- tral characteristics, spatial distribution, regular variables, and irregular red and association with other types of stars variables, usually with unstable may provide valuable clues to the life his- periods of hundreds of days. tories of various classes of stars. Certain (2) Explosive, or catastrophic, vari- kinds of variable stars, such as Cepheids ables—stars in which the variations (periodic variables) and novae and super- are produced by the wrenching novae (explosive variables), are extremely away of part of the star, usually important in that they make it possible to the outer layers, in some explo- establish the distances of remote stellar sive process. They include SS systems beyond the Galaxy. If the intrin- Cygni or U Geminorum stars, sic luminosity of a recognizable variable novae, and supernovae (the last is known and this kind of variable star of which are usually regarded as can be found in a distant stellar system, representing an enormous explo- the distance of the latter can be estimated sion involving most of the matter from a measurement of apparent and in a star). 74 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

(3) Miscellaneous and special types Milky Way, in globular clusters. They are of variables—R Coronae Borealis bluer than classic Population I Cepheids stars, T Tauri stars, flare stars, pul- having the same period, and their light sars (neutron stars), spectrum and curves have different shapes. Studies of magnetic variables, X-ray variable the light and velocity curves indicate that stars, and radio variable stars. shells of gas are ejected from the stars as discontinuous layers that later fall back Pulsating Stars toward the surface. These stars exhibit a relation between period and luminosity An impressive body of evidence indicates different from that for Population I that stellar pulsations can account for Cepheids, and thus the distance of a the variability of Cepheids, long-period Cepheid in a remote stellar system can variables, semiregular variables, Beta be determined only if its population type Canis Majoris stars, and even the irregu- is known. lar red variables. Of this group, the Closely associated with Population II Cepheid variables have been studied in Cepheids are the cluster-type, or RR greatest detail, both theoretically and obser- Lyrae, variables. Many of these stars are vationally. These stars are regular in their found in clusters, but some, such as the behaviour; some repeat their light curves prototype RR Lyrae, occur far from any with great faithfulness from one cycle to cluster or the central galactic bulge. Their the next over periods of many years. periods are less than a day, and there is Much confusion existed in the study no correlation between period and lumi- of Cepheids until it was recognized that nosity. Their absolute magnitudes are different types of Cepheids are associ- about 0.6 but somewhat dependent on ated with different groups, or population metal abundance. They are thus about 50 types, of stars. Cepheids belonging to the times as bright as the Sun and so are use- spiral-arm Population I are characterized ful for determining the distance of star by regularity in their behaviour. They clusters and some of the nearer external show continuous velocity curves indica- galaxies, their short periods permitting tive of regular pulsation. They exhibit a them to be detected readily. relation between period and luminosity Long-period variable stars also prob- in the sense that the longer the period of ably owe their variations to pulsations. the star, the greater is its intrinsic bright- Here the situation is complicated by the ness. This period-luminosity relationship vast extent of their atmospheres, so that has been used to establish the distances radiation originating at very different of remote stellar systems. depths in the star is observed at the same Cepheids with different properties time. At certain phases of the variations, are found in Population II, away from the bright hydrogen lines are observed, Stars | 75 overlaid with titanium oxide absorption. with Population II, and those of periods The explanation is an outward-moving of about a year belong to Population I. layer of hot, recombining gas, whose radi- Red semiregular variables such as ation is absorbed by strata of cool gases. the RV Tauri stars show complex light These stars are all cool red giants and and spectral changes. They do not repeat supergiants of spectral types M (normal themselves from one cycle to the next; composition), R and N (carbon-rich), or S their behaviour suggests a simultaneous (heavy-metal-rich). The range in visual operation of two or more modes of oscil- brightness during a pulsation can be 100- lation. Betelgeuse is an example of an fold, but the range in total energy output irregular red variable. In these stars the is much less, because at very low stellar free period of oscillation does not coin- temperatures (1,500–3,000 K) most of the cide with the periodicity of the driving energy is radiated in the infrared as heat mechanism. rather than as light. Finally, among the various types Unlike the light curves of classic of pulsating variable stars, the Beta Cepheids, the light curves of these red Canis Majoris variables are high- variables show considerable variations temperature stars (spectral type B) that from one cycle to another. The visual often show complicated variations in magnitude of the variable star Mira Ceti spectral-line shapes and intensities, (Omicron Ceti) is normally about 9–9.5 at velocity curves, and light. In many cases, minimum light, but at maximum it may they have two periods of variation so lie between 5 and 2. Time intervals similar in duration that complex inter- between maxima often vary considerably. ference or beat phenomena are observed, In such cool objects, a very small change both in radial velocities and in the shapes in temperature can produce a huge of spectral lines. change in the output of visible radiation. A large body of evidence suggests At the low temperatures of the red vari- that all members of this first class of ables, compounds and probably solid variable stars owe their variability to pul- particles are formed copiously, so that the sation. The pulsation theory was first visible light may be profoundly affected proposed as a possible explanation as by a slight change in physical conditions. early as 1879, was applied to Cepheids in Random fluctuations from cycle to cycle, 1914, and was further developed by which would produce negligible effects Arthur Eddington in 1917–18. Eddington in a hotter star, produce marked light found that if stars have roughly the same changes in a long-period variable. kind of internal structure, then the period Long-period variables appear to fall multiplied by the square root of the den- into two groups; those with periods of sity equals a constant that depends on roughly 200 days tend to be associated the internal structure. 76 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

The Eddington theory, though a good have been precisely predicted by the approximation, encountered some severe theory. Stellar pulsation, like other rhyth- difficulties that have been met through mic actions, may give rise to harmonic modifications. If the entire star pulsated phenomena wherein beats reinforce or in synchronism, it should be brightest interfere with one another. Beat and inter- when compressed and smaller while ference phenomena then complicate the faintest when expanded and at its largest. light and velocity changes. The RR Lyrae The radial velocity should be zero at both stars supply some of the best examples, maximum and minimum light. ObservÂ but semiregular variables such as the RV ations contradict these predictions. When Tauri stars or most Delta Scuti stars evi- the star pulsates, all parts of the main dently vibrate simultaneously with two body move in synchronism, but the outer or more periods. observable strata fall out of step or lag behind the pulsation of the inner Explosive Variables regions. Pulsations involve only the outer part of a star; the core, where energy is The evolution of a member of a close generated by thermonuclear reactions, double-star system can be markedly is unaffected. affected by the presence of its compan- Many years ago, careful measure- ion. As the stars age, the more massive ments of the average magnitudes and one swells up more quickly as it moves colours of RR Lyrae stars in the globular away from the main sequence. It becomes cluster M3 showed that all these stars fell so large that its outer envelope falls under within a narrow range of luminosity and the gravitational influence of the smaller colour (or surface temperature) or, equiv- star. Matter is continuously fed from the alently, luminosity and radius. Also, every more rapidly evolving star to the less star falling in this narrow range of bright- massive one, which still remains on the ness and size was an RR Lyrae variable. main sequence. U Cephei is a classic Subsequent work has indicated that simi- example of such a system for which spec- lar considerations apply to most classic troscopic evidence shows streams of gas Cepheids. Variability is thus a character- flowing from the more highly evolved istic of any star whose evolution carries it star to the hotter companion, which is to a certain size and luminosity, although now the more massive of the two. the amplitude of the variability can vary Eventually, the latter will also leave the dramatically. main sequence and become a giant star, In the pulsation theory as now devel- only to lose its outer envelope to the oped, the light and velocity changes of companion, which by that time may have Cepheids can be interpreted not only reached the white dwarf stage. qualitatively but also quantitatively. The Novae appear to be binary stars that light curves of Cepheids, for example, have evolved from contact binaries of the Stars | 77

W Ursae Majoris type, which are pairs of with a surface temperature near 7,000 K, stars apparently similar to the Sun in size the nova brightens rapidly. Then, near but revolving around one another while maximum light, the shell becomes trans- almost touching. One member may have parent, and its total brightness plummets reached the white dwarf stage. Matter fed rapidly, causing the nova to dim. to it from its distended companion The mass of the shell is thought to be appears to produce instabilities that rather small, about 10–100 times the mass result in violent explosions or nova out- of Earth. Only the outer layers of the star bursts. The time interval between seem to be affected; the main mass settles outbursts can range from a few score down after the outburst into a state much years to hundreds of thousands of years. as before until a new outburst occurs. The In ordinary novae the explosion existence of repeating novae, such as the seems to involve only the outer layers, as star T Coronae Borealis, suggests that the star later returns to its former bright- perhaps all novae repeat at intervals rang- ness; in supernovae the explosion is ing up to thousands or perhaps millions catastrophic. Normally, novae are small of years; and probably, the larger the blue stars much fainter than the Sun, explosion, the longer the interval. There though very much hotter. When an out- is strong evidence that novae are compo- burst occurs, the star can brighten very nents of close double stars and, in rapidly, by 10 magnitudes or more in a particular, that they have evolved from few hours. Thereafter it fades; the rate of the most common kind of eclipsing bina- fading is connected with the brightness ries, those of the W Ursae Majoris type. of the nova. The brightest novae, which Stars of the SS Cygni type, also known reach absolute magnitudes of about −10, as dwarf novae, undergo novalike out- fade most rapidly, whereas a typical slow bursts but of a much smaller amplitude. nova, which reaches an absolute magni- The intervals between outbursts are a few tude of −5, can take 10 or 20 times as long months to a year. Such variables are close to decline in brightness. This property, binaries. The development of this partic- when calibrated as the absolute magni- ular type may be possible only in close tude at maximum brightness versus the binary systems. time taken to decline by two magnitudes, There are two major types of super- allows novae to be used as distance indi- novae, designated type I (or SNe I) and cators for nearby galaxies. The changes type II (or SNe II). They can be distin- in light are accompanied by pronounced guished by the fact that type II have spectroscopic changes that can be inter- hydrogen features in their spectra, while preted as arising from alterations in an type I do not. Type II supernovae arise ejected shell that dissipates slowly in from the collapse of a single star more space. In its earliest phases, the expand- massive than about eight solar masses, ing shell is opaque. As its area grows, resulting in either a neutron star or black 78 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies hole. Type I supernovae, on the other Probably all variable stars represent hand, are believed to originate in a binary more or less ephemeral phases in the system containing a white dwarf, rather evolution of a star. Aside from cata- like the case of ordinary novae. Unlike the strophic events of the kind that produce a latter, however, in which only the outer supernova, some phases of stellar vari- layers of the white dwarf seem to be ability might be of such brief duration as affected, in a type I supernova the white to permit recognizable changes during dwarf is probably completely destroyed, an interval of 50–100 years. Other stages although the details are not yet fully may require many thousands of years. For understood. Certainly a supernova’s example, the period of Delta Cephei, the energy output is enormously greater than prototype star of the Cepheid variables, that of an ordinary nova. has barely changed by a detectable Type I supernovae can be further amount since its variability was discov- divided into Ia and Ib, the distinguishing ered in 1784. feature being that the spectra of Ia show a strong silicon line, whereas the spectra of Peculiar Variables Ib show a weak one. Empirical evidence indicates that in a type Ia supernova the R Coronae Borealis variables are giant absolute magnitude at maximum light stars of about the Sun’s temperature can be determined by a combination of whose atmospheres are characterized by data derived from the rate of dimming excessive quantities of carbon and very after maximum, the shape of the light little hydrogen. The brightness of such a curve, and certain colour measurements. star remains constant until the star sud- A comparison of the absolute and appar- denly dims by several magnitudes and ent magnitudes of maximum light, in then slowly recovers its original bright- turn, allows the distance of the supernova ness. (The star’s colour remains the same to be found. This is a matter of great use- during the changes in brightness.) The fulness because type Ia supernovae at dimmings occur in a random fashion maximum light are the most luminous and seem to be due to the huge concen- “standard candles” available for deter- trations of carbon. At times the carbon mining distances to external galaxies vapour literally condenses into soot, and and thus can be observed in more dis- the star is hidden until the smog blan- tant galaxies more than any other kind ket is evaporated. Similar veiling may of standard candle. In 1999, application of sometimes occur in other types of low- this technique led to the totally unexpected temperature stars, particularly in discovery that the expansion of the uni- long-period variables. verse appears to be accelerating rather Flare stars are cool dwarfs (spectral than slowing down due to the presence type M) that display flares apparently of dark energy. very much like, but much more intense Stars | 79 than, those of the Sun. In fact, the flares radio stars also have been recorded on are sometimes so bright that they over- occasion. These probably include flare whelm the normal light of the star. Solar stars, possibly red supergiants such as flares are associated with copious emis- Betelgeuse, the high-temperature dwarf sion of radio waves, and simultaneous companion to the red supergiant Antares, optical and radio-wave events appear to and the shells ejected from Nova Serpentis have been found in the stars UV Ceti, YZ 1970 and Nova Delphini. The radio emis- Canis Minoris, and V371 Orionis. sion from the latter objects is consistent Spectrum and magnetic variables, with that expected from an expanding mostly of spectral type A, show only small shell of ionized gas that fades away as the amplitudes of light variation but often gas becomes attenuated. The central star pronounced spectroscopic changes. of the Crab Nebula has been detected as Their spectra typically show strong lines a radio (and optical) pulsar. of metals such as manganese, titanium, Measurements from rockets, balloons, iron, chromium, and the lanthanides (also and spacecraft have revealed distinct called rare earths), which vary periodi- X-ray sources outside the solar system. cally in intensity. These stars have strong The strongest galactic source, Scorpius magnetic fields, typically from a few hun- X-1, appears to be associated with a hot dred to a few thousand gauss; one star, variable star resembling an old nova. In HD 215441, has a field on the order of all likelihood this is a binary star system 30,000 gauss. Not all magnetic stars are containing a low-mass normal star and a known to be variable in light; these nonluminous companion. objects also seem to have variable mag- A number of globular clusters are netic fields. The best interpretation is sources of cosmic X-rays. Some of this that these stars are rotating about an X-ray emission appears as intense fluc- inclined axis. As with Earth, the magnetic tuations of radiation lasting only a few and rotation axes do not coincide. seconds but changing in strength by as Different ions are concentrated in differ- much as 25 times. These X-ray sources ent areas (e.g., chromium in one area and have become known as bursters, and sev- the lanthanides in another). eral such objects have been discovered The Sun is an emitter of radio waves, outside of globular clusters as well. Some but with present techniques its radio bursters vary on a regular basis, while emission could only just be detected— others seem to turn on and off randomly. even in its most active phases—at the The most popular interpretation holds distance of the nearest star, about four that bursters are the result of binary sys- light-years away. Most discrete radio- tems in which one of the objects—a frequency sources have turned out to be compact neutron star or black hole—pulls objects such as old supernovae, radio gal- matter from the companion, a normal axies, and quasars, though well-recognized star. This matter is violently heated in the 80 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies process, giving rise to X-rays. That the layers (on average) instead of from the emission is often in the form of a burst is limb, and the dependence of temperature probably caused by something interrupt- on depth can be shown to correspond to ing the flow of matter onto (or into) the the transport of energy by radiation, not compact object or by an eclipsing orbit of by convection, at least in the outer layers the binary system. of the Sun’s atmosphere. The amount of limb darkening in any Stellar Atmospheres star depends on the effective temperature of the star and on the variation in tem- To interpret a stellar spectrum quantita- perature with depth. Limb darkening is tively, knowledge of the variation in occasionally an important factor in the temperature and density with depth in the analysis of stellar observations. For star’s atmosphere is needed. Some general example, it must be taken into account theoretical principles are outlined here. to interpret properly the observed light The gradient of temperature in a curves of eclipsing binaries, and here star’s atmosphere depends on the method again the results suggest transport of of energy transport to the surface. One energy via radiation. way to move energy from the interior of a The layers of a normal star are star to its surface is via radiation; photons assumed to be in mechanical, or hydro- produced in the core are repeatedly static, equilibrium. This means that at absorbed and reemitted by stellar atoms, each point in the atmosphere, the pres- gradually propagating to the surface. A sure supports the weight of the overlying second way is via convection, which is layers. In this way, a relation between a nonradiative mechanism involving a pressure and density can be found for any physical upwelling of matter much as in given depth. a pot of boiling water. For the Sun, at In addition to the temperature and least, there are ways of distinguishing the density gradients, the chemical composi- mechanism of energy transport. tion of the atmospheric layers as well as High-speed photographs of the Sun’s the absorptivity, or opacity, of the mate- disk show that the centre of the disk is rial must be known. In the Sun the brighter than the limb. The difference in principal source of opacity is the negative brightness depends on the wavelength of hydrogen ion (H−), a hydrogen atom with the radiation detected; it is large in violet one extra electron loosely bound to it. In light, is small in red light, and nearly van- the atmospheres of many stars, the extra ishes when the Sun is imaged in infrared electrons break loose and recombine with radiation. This limb darkening arises other ions, thereby causing a reemission because the Sun becomes hotter toward of energy in the form of light. At visible its core. At the centre of the disk, radia- wavelengths the main contribution to the tion is received from deeper and hotter opacity comes from the destruction of Stars | 81 this ion by interaction with a photon (the surface, so the point from which the depth above-cited process is termed photodis- or height is measured is arbitrary. The tem- sociation). In hotter stars, such as Sirius perature of the visible layers ranges from A (the temperature of which is about 4,700 to 6,200 K, the density from about 10,000 K), atomic hydrogen is the main 10−7 to 4 × 10−7 gram per cubic cm (1 cubic source of opacity, whereas in cooler stars cm = .06 cubic inch), and the gas pressure much of the outgoing energy is often from 0.002 to 0.14 atmosphere. The visible absorbed by molecular bands of titanium layers of stars such as the Sun have very oxide, water vapour, and carbon monoxide. low densities and pressures compared Additional sources of opacity are absorp- with Earth’s atmosphere, even though the tion by helium atoms and electron temperature is much higher. The strata of scattering in hotter stars, absorption by the solar atmosphere are very opaque hydrogen molecules and molecular ions, compared with the terrestrial atmosphere. absorption by certain abundant metals For stars other than the Sun, the such as magnesium, and Rayleigh scat- dependence of temperature on depth tering (a type of wavelength-dependent cannot be directly determined. CalcuÂ scattering of radiation by particles lations must proceed by a process of named for the British physicist Lord successive approximations, during which Rayleigh) in cool supergiant stars. the flux of energy is taken to be constant At considerable depths in the Sun with depth. Computations have been and similar stars, convection sets in. undertaken for atmospheres of a variety Though most models of stellar atmo- of stars ranging from dwarfs to super- spheres (particularly the outer layers) giants, from cool to hot stars. Their assume plane-parallel stratified layers, validity can be evaluated only by examin- photographs of granulation on the Sun’s ing how well they predict the observed visible surface belie this simple picture. features of a star’s continuous and line Realistic models must allow for rising spectrum, including the detailed shapes columns of heated gases in some areas of spectral-line features. Considering and descent of cooler gases in others. The the known complexities of stellar atmo- motions of the radiating gases are espe- spheres, the results fit the observations cially important when the model is to be remarkably well. used to calculate the anticipated line Severe deviations exist for stars with spectrum of the star. Typical gas veloci- extended and expanding atmospheres. ties are on the order of 2 km (1.4 miles) Matter flowing outward from a star proper second in the Sun; in other stars they duces a stellar wind analogous to the can be much larger. solar wind, but one that is often much Temperature, density, and pressure all more extensive and violent. In the spec- increase steadily inward in the Sun’s trum of certain very hot O-type stars (e.g., atmosphere. The Sun has no distinct solid Zeta Puppis), strong, relatively narrow 82 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies emission lines can be seen; however, in B9-type component shows the regular the ultraviolet, observations from rockets velocity changes expected of a binary star and spacecraft show strong emission but with an absorption (and associated lines with distinct absorption compo- emission) spectrum corresponding to a nents on the shorter wavelength side. higher temperature (near spectral type These absorption features are produced B5) and a blue continuum corresponding by rapidly outflowing atoms that absorb to a very-high-temperature star. The the radiation from the underlying stellar anomalous B5-type spectrum is evidently surface. The observed shifts in frequency excited principally by the hotter source; it correspond to ejection velocities of envelopes the entire system and shows about 100 km (60 miles) per second. few changes in velocity with time. Much gentler stellar winds are found in Supergiant stars have very extended cool M-type supergiants. atmospheres that are probably not even Rapid stellar rotation also can modify approximately in hydrostatic equilibrium. the structure of a star’s atmosphere. Since The atmospheres of M-type supergiant effective gravity is much reduced near stars appear to be slowly expanding out- the equator, the appropriate description ward. Observations of the eclipsing of the atmosphere varies with latitude. binary 31 Cygni show that the K-type Should the star be spinning at speeds supergiant component has an extremely near the breakup point, rings or shells inhomogeneous, extended atmosphere may be shed from the equator. composed of numerous blobs and fila- Some of the most extreme and inter- ments. As the secondary member of this esting cases of rotational effects are system slowly moves behind the larger found in close binary systems. InterpretÂ star, its light shines through larger masses ations of the light and velocity curves of of the K-type star’s atmosphere. If the these objects suggest that the spectro- atmosphere were in orderly layers, the lines scopic observations cannot be reconciled of ionized calcium, for example, produced with simple, orderly rotating stars. by absorption of the light of the B-type Instead, emission and absorption lines star by the K-type star’s atmosphere, sometimes overlap in such a way as to would grow stronger uniformly as the suggest streams of gas moving between eclipse proceeds. They do not, however. the stars. For example, Beta Lyrae, an eclipsing binary system, has a period of Stellar Interiors 12.9 days and displays very large shifts in orbital velocity. The brighter member at Models of the internal structure of stars— visible wavelengths is a B9-type star; the particularly their temperature, density, and other member appears to be a hot, abnor- pressure gradients below the surface— mal object whose spectral lines have not depend on basic principles explained in been observed. The spectrum of the this section. It is especially important Stars | 83 that model calculations take account of perfect gas law. In such neutral gases the the change in the star’s structure with molecular weight is 2 for molecular hydro- time as its hydrogen supply is gradually gen, 4 for helium, 56 for iron, and so on. In converted into helium. Fortunately, given the interior of a typical star, however, the that most stars can be said to be examples high temperatures and densities virtually of an “ideal gas”, the relations between guarantee that nearly all the matter is temperature, density, and pressure have a completely ionized; the gas is said to be a basic simplicity. plasma, the fourth state of matter. Under these conditions not only are the hydro- Distribution of Matter gen molecules dissociated into individual atoms, but also the atoms themselves are Several mathematical relations can be broken apart (ionized) into their constitu- derived from basic physical laws, assum- ent protons and electrons. Hence, the ing that the gas is “ideal” and that a star molecular weight of ionized hydrogen is has spherical symmetry; both these the average mass of a proton and an elec- assumptions are met with a high degree tron—namely, ½ on the atom-mass scale of validity. Another common assumption noted above. By contrast, a completely is that the interior of a star is in hydro- ionized helium atom contributes a mass static equilibrium. This balance is often of 4 with a helium nucleus (alpha parti- expressed as a simple relation between cle) plus two electrons of negligible mass; pressure gradient and density. A second hence, its average molecular weight is 4/3. relation expresses the continuity of As another example, a totally ionized mass—i.e., if M is the mass of matter nickel atom contributes a nucleus of mass within a sphere of radius r, the mass 58.7 plus 28 electrons; its molecular added, ΔM, when encountering an weight is then 58.7/29 = 2.02. Since stars increase in distance Δr through a shell of contain a preponderance of hydrogen volume 4πr2Δr, equals the volume of the and helium that are completely ionized shell multiplied by the density, ρ. In throughout the interior, the average par- symbols, ticle mass, μ, is the (unit) mass of a proton, divided by a factor taking into account ΔM = 4πr2ρΔr. the concentrations by weight of hydrogen, helium, and heavier ions. Accordingly, A third relation, termed the equation the molecular weight depends critically of state, expresses an explicit relation on the star’s chemical composition, par- between the temperature, density, and ticularly on the ratio of helium to pressure of a star’s internal matter. hydrogen as well as on the total content Throughout the star the matter is entirely of heavier matter. gaseous, and, except in certain highly If the temperature is sufficiently high, evolved objects, it obeys closely the the radiation pressure, Pr, must be taken 84 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies into account in addition to the perfect The temperature does not enter at all. At gas pressure, Pg. The total equation of still higher densities the equation of state state then becomes becomes more intricate, but it can be shown that even this complicated equa-

P = Pg + Pr. tion of state is adequate to calculate the internal structure of the white dwarf

Here Pg depends on temperature, stars. As a result, white dwarfs are prob- density, and molecular weight, whereas Pr ably better understood than most other depends on temperature and on the radi- celestial objects. ation density constant, For normal stars such as the Sun, the energy-transport method for the interior a = 7.5 × 10−15 must be known. Except in white dwarfs or in the dense cores of evolved stars, ther- ergs per cubic cm per degree to the fourth mal conduction is unimportant because power. With μ = 2 (as an upper limit) and the heat conductivity is very low. One ρ = 1.4 grams per cubic cm (the mean significant mode of transport is an actual density of the Sun), the temperature at flow of radiation outward through the star. which the radiation pressure would equal Starting as gamma rays near the core, the gas pressure can be calculated. The the radiation is gradually “softened” answer is 28 million K, much hotter than (becomes longer in wavelength) as it the core of the Sun. Consequently, radia- works its way to the surface (typically, in tion pressure may be neglected for the the Sun, over the course of about a mil- Sun, but it cannot be ignored for hotter, lion years) to emerge as ordinary light more massive stars. Radiation pressure and heat. The rate of flow of radiation is may then set an upper limit to stellar proportional to the thermal gradient— luminosity. namely, the rate of change of temperature Certain stars, notably white dwarfs, with interior distance. Providing yet do not obey the perfect gas law. Instead, another relation of stellar structure, this the pressure is almost entirely contrib- equation uses the following important uted by the electrons, which are said to quantities: a, the radiation constant noted be particulate members of a degenerate above; c, the velocity of light; ρ, the den- gas. If μ' is the average mass per free elec- sity; and κ, a measure of the opacity of the tron of the totally ionized gas, the matter. The larger the value of κ, the lower pressure, P, and density, ρ, are such that P the transparency of the material and the is proportional to a 5/3 power of the den- steeper the temperature fall required to sity divided by the average mass per free push the energy outward at the required electron; i.e., rate. The opacity, κ, can be calculated for any temperature, density, and chemical P = 1013(ρ/μ')5/3. composition and is found to depend in a Stars | 85 complex manner largely on the two for- average energy in a star such as the Sun) mer quantities. are capable of producing nuclear events In the Sun’s outermost (though still of this kind. A minimum temperature interior) layers and especially in certain required for fusion is roughly 10 million giant stars, energy transport takes place K. Since the energies of protons are pro- by quite another mechanism: large-scale portional to temperature, the rate of mass motions of gases—namely, convec- energy production rises steeply as tem- tion. Huge volumes of gas deep within perature increases. the star become heated, rise to higher For the Sun and other normal main- layers, and mix with their surroundings, sequence stars, the source of energy lies thus releasing great quantities of energy. in the conversion of hydrogen to helium. The extraordinarily complex flow patterns The nuclear reaction thought to occur in cannot be followed in detail, but when the Sun is called the proton-proton cycle. convection occurs, a relatively simple In this fusion reaction, two protons (1H) mathematical relation connects density collide to form a deuteron (a nucleus of and pressure. Wherever convection does deuterium, 2H), with the liberation of a occur, it moves energy much more effi- positron (the electron’s positively charged ciently than radiative transport. antimatter counterpart, denoted e+). Also emitted is a neutral particle of very small Source of Stellar Energy (or possibly zero) mass called a neutrino, ν. While the helium “ash” remains in the The most basic property of stars is that core where it was produced, the neutrino their radiant energy must derive from escapes from the solar interior within internal sources. Given the great length seconds. The positron encounters an ordi- of time that stars endure (some 10 billion nary negatively charged electron, and the years in the case of the Sun), it can be two annihilate each other, with much shown that neither chemical nor gravita- energy being released. This annihilation tional effects could possibly yield the energy amounts to 1.02 megaelectron required energies. Instead, the cause volts (MeV), which accords well with must be nuclear events wherein lighter Einstein’s equation nuclei are fused to create heavier nuclei, an inevitable by-product being energy. E = mc2 In the interior of a star, the particles move rapidly in every direction because (where m is the mass of the two particles, of the high temperatures present. Every c the velocity of light, and E the liberated so often a proton moves close enough to energy). a nucleus to be captured, and a nuclear Next, a proton collides with the deu- reaction takes place. Only protons of teron to form the nucleus of a light helium extremely high energy (many times the atom of atomic weight 3, 3He. A “hard” 86 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

X-ray (one of higher energy) or gamma- million tons of hydrogen to 670 million ray (γ) photon also is emitted. The most tons of helium every second. According likely event to follow in the chain is a col- to the formula lision of this 3He nucleus with a normal 4He nucleus to form the nucleus of a E = mc2, beryllium atom of weight 7, 7Be, with the emission of another gamma-ray photon. more than four million tons of matter liter- The 7Be nucleus in turn captures a proton ally disappear into radiation each second. to form a boron nucleus of atomic weight This theory provides a good under- 8, 8B, with the liberation of yet another standing of solar-energy generation, gamma ray. although for decades it has suffered The 8B nucleus, however, is very from one potential problem. For the past unstable. It decays almost immediately several decades the neutrino flux from into beryllium of atomic weight 8, 8Be, the Sun has been measured by different with the emission of another positron experimenters, and only one-third of flux and a neutrino. The nucleus itself there- of electron neutrinos predicted by the after decays into two helium nuclei, 4He. theory have been detected. Over that These nuclear events can be represented time, however, the consensus has grown by the following equations: that the problem and its solution lie not with the astrophysical model of the Sun but with the physical nature of neutrinos themselves. In late 1990s and early 21st century, scientists collected evidence that neutrinos oscillate between the state in which they were created in the Sun and a In the course of these reactions, four state that is more difficult to detect when protons are consumed to form one helium they reach Earth. nucleus, while two electrons perish. The main source of energy in hotter The mass of four hydrogen atoms is 4 stars is the carbon cycle (also called the × 1.00797, or 4.03188, atomic mass units; CNO cycle for carbon, nitrogen, and that of a helium atom is 4.0026. Hence, oxygen), in which hydrogen is trans- 0.02928 atomic mass unit, or 0.7 percent formed into helium, with carbon serving of the original mass, has disappeared. as a catalyst. The reactions proceed as Some of this has been carried away by the follows: first, a carbon nucleus, 12C, cap- elusive neutrinos, but most of it has been tures a proton (hydrogen nucleus), 1H, converted to radiant energy. In order to to form a nucleus of nitrogen, 13N, a keep shining at its present rate, a typical gamma-ray photon being emitted in the star (e.g., the Sun) needs to convert 674 process; thus, Stars | 87

12C + 1H → 13N + γ. Thus, the original 12C nucleus reap- pears, and the four protons that have been The light 13N nucleus is unstable, how- added permit the formation of a helium ever. It emits a positron, e+, which nucleus. The same amount of mass has encounters an ordinary electron, e−, and disappeared, though a different fraction the two annihilate one another. A neu- of it may have been carried off by the trino also is released, and the resulting neutrinos. 13C nucleus is stable. Eventually the 13C Only the hottest stars that lie on the nucleus captures another proton, forms main sequence shine with energy pro- 14N, and emits another gamma-ray photon. duced by the carbon cycle. The faint red In symbols the reaction is represented by dwarfs use the proton-proton cycle the equations exclusively, whereas stars such as the Sun shine mostly by the proton-proton 13N → 13C + e+ + ν; reaction but derive some contribution then 13C + 1H → 14N + γ. from the carbon cycle as well. The aforementioned mathematical Ordinary nitrogen, 14N, is stable, but relationships permit the problem of when it captures a proton to form a stellar structure to be addressed notwith- nucleus of light oxygen-15, 15O, the result- standing the complexity of the problem. ing nucleus is unstable against beta An early assumption that stars have a decay. It therefore emits a positron and a uniform chemical composition throughout neutrino, a sequence of events expressed their interiors simplified the calculations by the equations considerably, but it had to be abandoned when studies in stellar evolution proved 14N + 1H → 15O + γ; that the compositions of stars change then 15O → 15N + e+ + ν. with age. Computations need to be carried out by a step-by-step process known Again, the positron meets an elec- as numerical integration. They must take tron, and the two annihilate each other into account that the density and pres- while the neutrino escapes. Eventually the sure of a star vanish at the surface, 15N nucleus encounters a fast-moving pro- whereas these quantities and the temper- ton, 1H, and captures it, but the formation ature remain finite at the core. of an ordinary 16O nucleus by this process Resulting models of a star’s interior, occurs only rarely. The most likely effect of including the relation between mass, this proton capture is a breakdown of 15N luminosity, and radius, are determined and a return to the 12C nucleus—that is, largely by the mode of energy transport. In the Sun and the fainter main-sequence 15N + 1H → 12C + 4He + γ. stars, energy is transported throughout 88 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies the outer layers by convective currents, such depleted cores. Instead, energy is whereas in the deep interior, energy is assumed to be generated in a thin shell transported by radiation. Among the hot- surrounding the inert core where some ter stars of the main sequence, the reverse fuel remains, and it is presumably pro- appears to be true. The deep interiors of duced by the carbon cycle. Such models the stars that derive their energy primar- are called shell-source models. As a star ily from the carbon cycle are in convective uses up increasing amounts of its hydro- equilibrium, whereas in the outer parts gen supply, its core grows in mass, all the the energy is carried by radiation. The while the outer envelope of the star con- observed masses, luminosities, and radii tinues to expand. These shell-source of most main-sequence stars can be models explain the observed luminosi- reproduced with reasonable and uniform ties, masses, and radii of giants and chemical composition. supergiants. Chemically homogeneous models of The depletion of hydrogen fuel is giant and supergiant stars cannot be con- appreciable even for a dwarf, middle-aged structed. If a yellow giant such as Capella star such as the Sun. The Sun seems to is assumed to be built like a main-se- have been shining at its present rate for quence star, its central temperature turns about the last 20 percent of its current out to be so low that no known nuclear age of five billion years. For its observed process can possibly supply the observed luminosity to be maintained, the Sun’s energy output. Progress has been made central temperature must have increased only by assuming that these stars were considerably since the formation of the once main-sequence objects that, in the solar system, largely as a consequence course of their development, exhausted of the depletion of the hydrogen in its the hydrogen in their deep interiors. Inert interior along with an accompanying cores consequently formed, composed increase in molecular weight and temper- mainly of the helium ash left from the ature. During the past five billion years, hydrogen-fusion process. Since no helium the Sun probably brightened by about nuclear reactions are known to occur at half a magnitude; in early Precambrian the few tens of millions of kelvins likely time (about two billion years ago), the to prevail in these interiors, no thermo- solar luminosity must have been some 20 nuclear energy could be released from percent less than it is today. CHAPTER 3

Star Formation and Evolution

hroughout the Milky Way Galaxy (and even near the Sun Titself), astronomers have discovered stars that are well evolved or even approaching extinction, or both, as well as occasional stars that must be very young or still in the process of formation. Evolutionary eff ects on these stars are not negligible, even for a middle-aged star such as the Sun. More massive stars must display more spectacular eff ects because the rate of conversion of mass into energy is higher. While the Sun produces energy at the rate of about two ergs per gram per second, a more luminous main-sequence star can release energy at a rate some 1,000 times greater. Consequently, eff ects that require billions of years to be easily recognized in the Sun might occur within a few million years in highly luminous and massive stars. A supergiant star such as Antares, a bright main-sequence star such as Rigel, or even a more modest star such as Sirius cannot have endured as long as the Sun has endured. These stars must have been formed relatively recently.

BIRTH OF STARS AND EvOLuTION TO THE MAIN SEquENCE

Detailed radio maps of nearby molecular clouds reveal that they are clumpy, with regions containing a wide range of densities—from a few tens of molecules (mostly hydrogen) per cubic centimetre to more than one million. Stars form 90 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies only from the densest regions, termed about 1 to 10 percent of all star births. The cloud cores, though they need not lie at overall efficiency of star formation in the geometric centre of the cloud. Large associations is quite small. Typically less cores (which probably contain subcon- than 1 percent of the mass of a molecular densations) up to a few light-years in size cloud becomes stars in one crossing time seem to give rise to unbound associations of the molecular cloud (about 5 106 years). of very massive stars (called OB associa- Low efficiency of star formation presum- tions after the spectral type of their most ably explains why any interstellar gas prominent members, O and B stars) or remains in the Galaxy after 1010 years of to bound clusters of less massive stars. evolution. Star formation at the present Whether a stellar group materializes as an time must be a mere trickle of the torrent association or a cluster seems to depend that occurred when the Galaxy was young. on the efficiency of star formation. A typical cloud core rotates fairly If only a small fraction of the matter slowly, and its distribution of mass is goes into making stars, the rest being strongly concentrated toward the centre. blown away in winds or expanding H II The slow rotation rate is probably attribut- regions, then the remaining stars end up able to the braking action of magnetic in a gravitationally unbound association, fields that thread through the core and its dispersed in a single crossing time envelope. This magnetic braking forces the (diameter divided by velocity) by the ran- core to rotate at nearly the same angular dom motions of the formed stars. On speed as the envelope as long as the core the other hand, if 30 percent or more of the does not go into dynamic collapse. Such mass of the cloud core goes into making braking is an important process because stars, then the formed stars will remain it assures a source of matter of relatively bound to one another, and the ejection of low angular momentum (by the standards stars by random gravitational encounters of the interstellar medium) for the forma- between cluster members will take many tion of stars and planetary systems. crossing times. It also has been proposed that mag- Low-mass stars also are formed in netic fields play an important role in the associations called T associations after very separation of the cores from their the prototypical stars found in such envelopes. The proposal involves the groups, T Tauri stars. The stars of a T slippage of the neutral component of a association form from loose aggregates lightly ionized gas under the action of of small molecular cloud cores a few the self-gravity of the matter past the tenths of a light-year in size that are ran- charged particles suspended in a back- domly distributed through a larger region ground magnetic field. This slow slippage of lower average density. The formation of would provide the theoretical explana- stars in associations is the most common tion for the observed low overall efficiency outcome; bound clusters account for only of star formation in molecular clouds. Star Formation and Evolution | 91

At some point in the course of the the available raw material is not what evolution of a molecular cloud, one or stops the accretion flow. A rather differ- more of its cores become unstable and ent picture is revealed by observations subject to gravitational collapse. Good at radio, optical, and X-ray wavelengths. arguments exist that the central regions All newly born stars are highly active, should collapse first, producing a con- blowing powerful winds that clear the densed protostar whose contraction is surrounding regions of the infalling gas halted by the large buildup of thermal and dust. It is apparently this wind that pressure when radiation can no longer reverses the accretion flow. escape from the interior to keep the (now The geometric form taken by the opaque) body relatively cool. The proto- outflow is intriguing. Jets of matter seem star, which initially has a mass not much to squirt in opposite directions along the larger than Jupiter, continues to grow by rotational poles of the star (or disk) that accretion as more and more overlying sweep up the ambient matter in two lobes material falls on top of it. The infall shock, of outwardly moving molecular gas—the at the surfaces of the protostar and the so-called bipolar flows. Such jets and swirling nebular disk surrounding it, bipolar flows are doubly interesting arrests the inflow, creating an intense because their counterparts were discov- radiation field that tries to work its way ered some time earlier on a fantastically out of the infalling envelope of gas and larger scale in the double-lobed forms of dust. The photons, having optical wave- extragalactic radio sources, such as lengths, are degraded into longer quasars. wavelengths by dust absorption and The underlying energy source that reemission, so that the protostar is drives the outflow is unknown. Promising apparent to a distant observer only as mechanisms invoke tapping the rotational an infrared object. Provided that proper energy stored in either the newly formed account is taken of the effects of rotation star or the inner parts of its nebular disk. and magnetic field, this theoretical pic- There exist theories suggesting that ture correlates with the radiative spectra strong magnetic fields coupled with rapid emitted by many candidate protostars rotation act as whirling rotary blades to discovered near the centres of molecular fling out the nearby gas. Eventual colli- cloud cores. mation of the outflow toward the rotation An interesting speculation concern- axes appears to be a generic feature of ing the mechanism that ends the infall many proposed models. phase exists: it notes that the inflow pro- Pre-main-sequence stars of low mass cess cannot run to completion. Since first appear as visible objects, T Tauri molecular clouds as a whole contain stars, with sizes that are several times much more mass than what goes into their ultimate main-sequence sizes. They each generation of stars, the depletion of subsequently contract on a timescale of 92 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies tens of millions of years, the main source the star. A star of the Sun’s mass gener- of radiant energy in this phase being the ally requires tens of millions of years to release of gravitational energy. As the inter- reach the main sequence, whereas one of nal temperature rises to a few million much greater mass might take a few hun- kelvins, deuterium (heavy hydrogen) is dred thousand years. first destroyed. Then lithium, beryllium, By the time the star reaches the main and boron are broken down into helium sequence, it is still chemically homoge- as their nuclei are bombarded by protons neous. With additional time, the hydrogen moving at increasingly high speeds. fuel in the core is converted to helium, and When their central temperatures reach the temperature slowly rises. If the star is values comparable to 107 K, hydrogen sufficiently massive to have a convective fusion ignites in their cores, and they core, the matter in this region has a chance settle down to long stable lives on the to be thoroughly mixed, but the outer main sequence. The early evolution of region does not mix with the core. The high-mass stars is similar. The only differ- Sun, by contrast, has no convective core, ence is that their faster overall evolution and the helium-to-hydrogen ratio is maxi- may allow them to reach the main mum at the centre and decreases outward. sequence while they are still enshrouded Throughout the life of the Sun, there has in the cocoon of gas and dust from which been a steady depletion of hydrogen, so they formed. that the concentration of hydrogen at the Detailed calculations show that a centre today is probably only about one- protostar first appears on the Hertzsprung- third of the original amount. The rest has Russell diagram well above the main been transformed into helium. Like the rate sequence because it is too bright for its of formation of a star, the subsequent colour. As it continues to contract, it rate of evolution on the main sequence is moves downward and to the left toward proportional to the mass of the star; the the main sequence. greater the mass, the more rapid the evolution. Whereas the Sun is destined to Subsequent development endure for some 10 billion years, a star of on the main sequence twice the Sun’s mass burns its fuel at such a rate that it lasts about 3 billion years, As the central temperature and density and a star of 10 times the Sun’s mass has a continue to rise, the proton-proton and lifetime measured in tens of millions of carbon cycles become active, and the years. By contrast, stars having a fraction development of the (now genuine) star is of the mass of the Sun seem able to endure stabilized. The star then reaches the main for trillions of years, which is much greater sequence, where it remains for most of its than the current age of the universe. active life. The time required for the con- The spread of luminosities and colours traction phase depends on the mass of of stars within the main sequence can be Star Formation and Evolution | 93 understood as a consequence of evolution. Later stages of evolution At the beginning of their lives as hydrogen-burning objects, stars define a nearly The great spread in luminosities and unique line in the Hertzsprung-Russell colours of giant, supergiant, and subgiant diagram called the zero-age main sequence. stars is also understood to result from Without differences in initial chemical evolutionary events. When a star leaves composition or in rotational velocity, all the main sequence, its future evolution is the stars would start exactly from this precisely determined by its mass, rate of unique line. As the stars evolve, they rotation (or angular momentum), and adjust to the increase in the helium-to- chemical composition and whether it is a hydrogen ratio in their cores and member of a close binary system. Giants gradually move away from the zero-age and supergiants of nearly the same radius main sequence. When the core fuel is and surface temperature may have exhausted, the internal structure of the star evolved from main-sequence stars of dif- changes rapidly; it quickly leaves the main ferent ages and masses. sequence and moves toward the region of giants and supergiants. Evolution of Low-Mass Stars As the composition of its interior changes, the star departs the main Theoretical calculations suggest that, sequence slowly at first and then more as the star evolves from the main rapidly. When about 10 percent of the sequence, the hydrogen-helium core star’s mass has been converted to gradually increases in mass but shrinks helium, the structure of the star changes in size as more and more helium ash is fed drastically. All of the hydrogen in the in through the outer hydrogen-burning core has been burned out, and this shell. Energy is carried outward from central region is composed almost the shell by rapid convection currents. entirely of inert helium, with trace The temperature of the shell rises; the admixtures of heavier elements. The star becomes more luminous; and it energy production now occurs in a thin finally approaches the top of the giant shell where hydrogen is consumed and domain on the Hertzsprung-Russell dia- more helium added to a growing but gram. By contrast, the core shrinks by inert core. The outer parts of the star gravitational contraction, becoming hot- expand outward because of the increased ter and denser until it reaches a central burning there, and as the star swells up, temperature of about 120 million K. At its luminosity gradually increases. The that temperature the previously inert details of the evolutionary process helium is consumed in the production of depend on the metal-to-hydrogen ratio, heavier elements. and the course of evolution differs for When two helium nuclei each of mass stars of different population types. 4 atomic units (4He) are jammed together, 94 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies it might be expected that they would form stage along what is called the asymptotic a nucleus of beryllium of mass 8 atomic giant branch, located slightly above the units (8Be). In symbols, main region of giants in the Hertzsprung- Russell diagram. 4He + 4He → 8Be. In more massive stars, this cycle of events can continue, with the stellar core Actually, however, 8Be is unstable and reaching ever-higher temperatures and breaks down into two helium nuclei. If fusing increasingly heavy nuclei, until the temperature and density are high the star eventually experiences a super- enough, though, the short-lived beryllium nova explosion. In lower-mass stars like nucleus can (before it decays) capture the Sun, however, there is insufficient another helium nucleus in what is essen- mass to squeeze the core to the tempera- tially a three-body collision to form a tures needed for this chain of fusion nucleus of carbon-12—namely, processes to proceed, and eventually the outermost layers extend so far from the 8Be + 4He → 12C. source of nuclear burning that they cool to a few thousand kelvins. The result is an This fusion of helium in the core, object having two distinct parts: a well- called the triple alpha process, can begin defined core of mostly carbon ash (a white gradually in some stars, but in stars with dwarf star) and a swollen spherical shell masses between about half of and three of cooler and thinner matter spread over times the Sun’s mass, it switches on with a volume roughly the size of the solar dramatic suddenness, a process known system. Such shells of matter, called as the helium flash. Outwardly the star planetary nebulae, are actually observed shows no discernible effect, but the course in large numbers in the sky. Of the nearly of its evolution is changed with this new 3,000 examples known in the Milky Way source of energy. Having only recently Galaxy alone, NGC 7027 is the most become a red giant, it now evolves some- intensively studied. what down and then to the left in the Objects called brown dwarfs are Hertzsprung-Russell diagram, becoming intermediate between a planet and a star smaller and hotter. This stage of core and so evolve differently from low-mass helium burning, however, lasts only about stars. Brown dwarfs usually have a mass a hundredth of the time taken for core less than 0.075 that of the Sun, or roughly hydrogen burning. It continues until the 75 times that of Jupiter. (This maximum core helium supply is exhausted, after mass is a little higher for objects with which helium fusion is limited to a shell fewer heavy elements than the Sun.) around the core, just as was the case for Many astronomers draw the line between hydrogen in an earlier stage. This again brown dwarfs and planets at the lower sets the star evolving toward the red giant fusion boundary of about 13 Jupiter Star Formation and Evolution | 95

hydrogen, but they then stabilize, and the fusion stops.) Brown dwarfs are not actually brown but appear from deep red to magenta depending on their temperature. Objects below about 2,200 K, however, do actually have mineral grains in their atmospheres. The surface temperatures of brown dwarfs depend on both their mass and their age. The most massive and youngest brown dwarfs have temperatures as high as 2,800 K, which overlaps with the temperatures of very low-mass stars, or red dwarfs. (By comparison, the Sun has a surface temperature of 5,800 K.) All brown dwarfs eventually cool below the The brown dwarf 2MASSWJ 1207334−393254 minimum main-sequence stellar temper- (centre) as seen in a photo taken by the Very ature of about 1,800 K. The oldest and Large Telescope at the European Southern smallest can be as cool as about 500 K. Observatory, Cerro Paranal, Chile. Orbiting Brown dwarfs were fi rst hypothesized the brown dwarf at a distance of 69 billion km in 1963 by American astronomer Shiv (43 billion miles) is a planet (lower left) that has a mass four times that of Jupiter. ESO Kumar, who called them “black” dwarfs. American astronomer Jill Tarter proposed the name “brown dwarf” in 1975; masses. The diff erence between brown although brown dwarfs are not brown, dwarfs and stars is that, unlike stars, the name stuck because these objects brown dwarfs do not reach stable lumi- were thought to have dust, and the more nosities by thermonuclear fusion of accurate “red dwarf” already described a normal hydrogen. Both stars and brown diff erent type of star. Searches for brown dwarfs produce energy by fusion of deu- dwarfs in the 1980s and 1990s found sev- terium (a rare isotope of hydrogen) in eral candidates; however, none was their fi rst few million years. The cores of confi rmed as a brown dwarf. In order to stars then continue to contract and get distinguish brown dwarfs from stars of hotter until they fuse hydrogen. However, the same temperature, one can search brown dwarfs prevent further contraction their spectra for evidence of lithium because their cores are dense enough to (which stars destroy when hydrogen hold themselves up with electron degen- fusion begins). Alternatively, one can eracy pressure. (Those brown dwarfs look for (fainter) objects below the mini- above 60 Jupiter masses begin to fuse mum stellar temperature. In 1995 both 96 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies methods paid off. Astronomers at the stony meteorites called chondrites and University of California, Berkeley, from the composition of the Sun’s atmo- observed lithium in an object in the sphere, supplemented by data acquired Pleiades, but this result was not immedi- from spectral observations of hot stars ately and widely embraced. This object, and gaseous nebulae. The table lists the however, was later accepted as the first most abundant chemical elements; it binary brown dwarf. Astronomers at represents an average pertaining to all Palomar Observatory and Johns Hopkins cosmic objects in general. University found a companion to a low- mass star called Gliese 229 B. The The most abundant detection of methane in its spectrum chemical elements 9 showed that it has a surface tempera- (by numbers of atoms per 10 ture less than 1,200 K. Its extremely low atoms of hydrogen) luminosity, coupled with the age of its element symbol abundance stellar companion, implies that it is about 50 Jupiter masses. Hence, Gliese 229 B helium He 9.8 × 107 was the first object widely accepted as a carbon C 501,000 brown dwarf. Infrared sky surveys and nitrogen N 100,000 other techniques have now uncovered oxygen O 794,000 hundreds of brown dwarfs. Some of them fluorine F 33 are companions to stars; others are neon Ne 123,000 binary brown dwarfs; and many of them are isolated objects. They seem to form in sodium Na 2,100 much the same way as stars, and there magnesium Mg 38,000 may be 1–10 percent as many brown aluminum Al 3,000 dwarfs as stars. silicon Si 35,000 phosphorus P 320 Origin of the Chemical sulfur S 17,400 Elements chlorine Cl 250 The relative abundances of the chemical argon Ar 3,600 elements provide significant clues regard- potassium K 133 ing their origin. Earth’s crust has been calcium Ca 2,200 affected severely by erosion, fraction- titanium Ti 91 ation, and other geologic events, so that chromium Cr 473 its present varied composition offers few manganese Mn 288 clues as to its early stages. The composi- iron Fe 33,000 tion of the matter from which the solar system formed is deduced from that of nickel Ni 1,800 Star Formation and Evolution | 97

The most obvious feature is that the A large body of evidence now sup- light elements tend to be more abundant ports the idea that only the nuclei of than the heavier ones. That is to say, when hydrogen and helium, with trace amounts abundance is plotted against atomic of other light nuclei such as lithium, mass, the resulting graph shows a decline beryllium, and boron, were produced in with increasing atomic mass up to an the aftermath of the big bang, the hot atomic mass value of about 100. Thereafter explosion from which the universe is the abundance is more nearly constant. thought to have emerged, whereas the Furthermore, the decline is not smooth. heavier nuclei were, and continue to be, Among the lighter elements, those of produced in stars. The majority of them, even atomic number tend to be more however, are fashioned only in the most abundant, and those with an atomic massive stars and some only for a short number divisible by four are especially period of time after supernova favoured. The abundances of lithium, explosions. beryllium, and boron are rare compared The splitting in the spectral sequence with those of carbon, nitrogen, and oxy- among the cooler stars can be understood gen. There is a pronounced abundance in terms of composition differences. The peak for iron and a relatively high peak M-type stars appear to have a normal (i.e., for lead, the most stable of the heavy solar) makeup, with oxygen more abun- elements. dant than carbon and the zirconium The overwhelming preponderance of group of elements much less abundant hydrogen suggests that all the nuclei than the titanium group. The R-type and were built from this simplest element, a N-type stars often contain more carbon hypothesis first proposed many years ago than oxygen, whereas the S-type stars and widely accepted for a time. According appear to have an enhanced content of to this now-defunct idea, all matter was zirconium as compared with titanium. initially compressed into one huge ball of Other abundance anomalies are neutrons. As the universe began to found in a peculiar class of higher tem- expand, its density decreased and the perature stars, called Wolf-Rayet (or W) neutrons decayed into protons and elec- stars, in which objects containing pre- trons. The protons then captured neutrons, dominantly helium, carbon, and oxygen one after another, underwent beta decay are distinguished from those containing (ejection of electrons), and synthesized helium and nitrogen, some carbon, and the heavy elements. A major difficulty little observed oxygen. These stars are with this hypothesis, among various other extremely hot white stars that have problems, is that atomic masses 5 and 8 peculiar spectra thought to indicate are unstable, and there is no known way to either great turbulence within the star or build heavier nuclei by successive neutron a steady, voluminous ejection of material. capture. A typical Wolf-Rayet star is several times 98 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies the diameter of the Sun and thousands of during the late evolution of a star have times more luminous. Only a few hun- been proposed. dred are known, located mostly in the After hydrogen, helium is the most spiral arms of the Milky Way Galaxy. (The abundant element. Most of it was proba- type was first distinguished in 1867 by bly produced in the initial big bang. the French astronomers Charles-Joseph- Helium is the normal ash of hydrogen Étienne Wolf and Georges-Antoine-Pons consumption, and in the dense cores of Rayet.) Significantly, all these abundance highly evolved stars, helium itself is con- anomalies are found in stars thought to sumed to form, successively, carbon-12, be well advanced in their evolutionary oxygen-16, neon-20, and magnesium-24. development. No main-sequence dwarfs By this time in the core of a sufficiently display such effects. massive star, the temperature has reached A most critical observation is the some 700 million K. Under these condi- detection of the unstable element techne- tions, particles such as protons, neutrons, tium in the S-type stars. This element has and helium-4 nuclei also can interact with been produced synthetically in nuclear the newly created nuclei to produce a laboratories on Earth, and its longest- variety of other elements such as fluorine lived isotope, technetium-99, is known to and sodium. Because these “uneven” ele- have a half-life of 200,000 years. The ments are produced in lesser quantities implication is that this element must than those divisible by four, both the have been produced within the past few peaks and troughs in the curve of cosmic hundred thousand years in the stars abundances can be explained. where it has been observed, suggesting As the stellar core continues to shrink furthermore that this nucleosynthetic and the central temperature and density process is at work at least in some stars are forced even higher, a fundamental dif- today. How the star upwells this heavy ficulty is soon reached. A temperature of element from the core (where it is pro- roughly one billion K is sufficient to cre- duced) to the surface (near where it is ate silicon (silicon-28) by the usual observed) in such a short time without method of helium capture. This tempera- the star’s exploding provides an impres- ture, however, is also high enough to sive challenge to theoreticians. begin to break apart silicon as well as Researchers have been able to dem- some of the other newly synthesized onstrate how elements might be created nuclei. A “semi-equilibrium” is set up in in stars by nuclear processes occurring at the star’s core—a balance of sorts between very high temperatures and densities. No the production and destruction (photo- one mechanism can account for all the disintegration) of silicon. Ironically, elements; rather, several distinct pro- though destructive, this situation is suit- cesses occurring at different epochs able for the production of even heavier Star Formation and Evolution | 99 nuclei up to and including iron (iron-56), with supernovae. Supernovae also release again through the successive capture of many of the heavier elements that make helium nuclei. up the components of the solar system, including Earth, into the interstellar Evolution of medium. Spectral analyses show that High-Mass Stars abundances of the heavier elements are greater than normal, indicating that these If the temperature and the density of the elements do indeed form during the core continue to rise, the iron-group course of the explosion. The shell of a nuclei tend to break down into helium supernova remnant continues to expand nuclei, but a large amount of energy is until, at a very advanced stage, it dissolves suddenly consumed in the process. The into the interstellar medium. star then suffers a violent implosion, or Historically, only seven supernovae collapse, after which it soon explodes as a are known to have been recorded before supernova. the early 17th century. The most famous The term supernova is derived from of them occurred in 1054 and was seen in nova (Latin: “new”), the name for another one of the horns of the constellation type of exploding star. Supernovae Taurus. The remnants of this explosion resemble novae in several respects. Both are visible today as the Crab Nebula, are characterized by a tremendous, rapid which is composed of glowing ejecta of brightening lasting for a few weeks, fol- gases flying outward in an irregular lowed by a slow dimming. Spectroscopically, fashion and a rapidly spinning, pulsating they show blue-shifted emission lines, neutron star, called a pulsar, in the centre. which imply that hot gases are blown The supernova of 1054 was recorded by outward. But a supernova explosion, unlike Chinese and Korean observers; it also a nova outburst, is a cataclysmic event for a may have been seen by southwestern star, one that essentially ends its active American Indians, as suggested by cer- (i.e., energy-generating) lifetime. When a tain rock paintings discovered in Arizona star “goes supernova,” considerable and New Mexico. It was bright enough to amounts of its matter, equaling the mate- be seen during the day, and its great lumi- rial of several Suns, may be blasted into nosity lasted for weeks. Other prominent space with such a burst of energy as to supernovae are known to have been enable the exploding star to outshine its observed from Earth in 185, 393, 1006, entire home galaxy. 1181, 1572, and 1604. Supernovae explosions release not The closest and most easily observed only tremendous amounts of radio waves of the hundreds of supernovae that have and X-rays but also cosmic rays. Some been recorded since 1604 was first sighted gamma-ray bursts have been associated on the morning of Feb. 24, 1987, by the 100 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

Canadian astronomer Ian K. Shelton than iron absorbs rather than produces while working at the Las Campanas energy, however, and, since energy is no Observatory in Chile. Designated SN longer available, an iron core is built up 1987A, this formerly extremely faint at the centre of the aging, heavyweight object attained a magnitude of 4.5 within star. When the iron core becomes too just a few hours, thus becoming visible to massive, its ability to support itself by the unaided eye. The newly appearing means of the outward explosive thrust of supernova was located in the Large internal fusion reactions fails to counter- Magellanic Cloud at a distance of about act the tremendous pull of its own gravity. 160,000 light-years. It immediately Consequently, the core collapses. If the became the subject of intense observa- core’s mass is less than about three solar tion by astronomers throughout the masses, the collapse continues until the Southern Hemisphere and was observed core reaches a point at which its constitu- by the Hubble Space Telescope. SN ent nuclei and free electrons are crushed 1987A’s brightness peaked in May 1987, together into a hard, rapidly spinning with a magnitude of about 2.9, and slowly core. This core consists almost entirely of declined in the following months. neutrons, which are compressed in a vol- Supernovae may be divided into two ume only 20 km (12 miles) across but broad classes, Type I and Type II, accord- whose combined weight equals that of ing to the way in which they detonate. several Suns. A teaspoonful of this Type I supernovae may be up to three extraordinarily dense material would times brighter than Type II; they also dif- weigh 50 billion tons on Earth. Such an fer from Type II supernovae in that their object is called a neutron star. spectra contain no hydrogen lines and The supernova detonation occurs they expand about twice as rapidly. when material falls in from the outer lay- The so-called classic explosion, asso- ers of the star and then rebounds off the ciated with Type II supernovae, has as core, which has stopped collapsing and progenitor a very massive star (a PopulaÂ suddenly presents a hard surface to the tion I star) of at least eight solar masses infalling gases. The shock wave generated that is at the end of its active lifetime. by this collision propagates outward and (These are seen only in spiral galaxies, blows off the star’s outer gaseous layers. most often near the arms.) Until this stage The amount of material blasted outward of its evolution, the star has shone by depends on the star’s original mass. means of the nuclear energy released at If the core mass exceeds three solar and near its core in the process of squeez- masses, the core collapse is too great to ing and heating lighter elements such as produce a neutron star; the imploding hydrogen or helium into successively star is compressed into an even smaller heavier elements—i.e., in the process of and denser body—namely, a black hole. nuclear fusion. Forming elements heavier Infalling material disappears into the Star Formation and Evolution | 101 black hole, the gravitational field of which rate of the universe and that rate’s varia- is so intense that not even light can tion with time. Dark energy, a repulsive escape. The entire star is not taken in by force that is the dominant component (73 the black hole, since much of the falling percent) of the universe, was discovered envelope of the star either rebounds from in 1998 with this method. Type Ia super- the temporary formation of a spinning novae that exploded when the universe neutron core or misses passing through was only two-thirds of its present size the very centre of the core and is spun off were fainter and thus farther away than instead. they would be in a universe without dark Type I supernovae can be divided energy. This implies that the expansion into subgroups, Ia, Ib, Ic, on the basis of rate of the universe is faster now than it their spectra. The exact nature of the was in the past, a result of the current explosion mechanism in Type I generally dominance of dark energy. (Dark energy is still uncertain, although Ia supernovae, was negligible in the early universe.) at least, are thought to originate in binary In the catastrophic events leading to systems consisting of a moderately mas- a supernova explosion and for roughly sive star and a white dwarf, with material 1,000 seconds thereafter, a great variety flowing to the white dwarf from its larger of nuclear reactions can take place. These companion. A thermonuclear explosion processes seem to be able to explain the results if the flow of material is sufficient trace abundances of all the known ele- to raise the mass of the white dwarf ments heavier than iron. above the Chandrasekhar limit of 1.44 Two situations have been envisioned, solar masses. Unlike the case of an ordi- and both involve the capture of neutrons. nary nova, for which the mass flow is less When a nucleus captures a neutron, its and only a superficial explosion results, mass increases by one atomic unit and the white dwarf in a Ia supernova explo- its charge remains the same. Such a sion is presumably destroyed completely. nucleus is often too heavy for its charge Radioactive elements, notably nickel-56, and might emit an electron (beta particle) are formed. When nickel-56 decays to to attain a more stable state. It then cobalt-56 and the latter to iron-56, signifi- becomes a nucleus of the next higher ele- cant amounts of energy are released, ment in the periodic table of the elements. providing perhaps most of the light emit- In the first such process, called the slow, ted during the weeks following the or s, process, the flux of neutrons is low. A explosion. nucleus captures a neutron and leisurely Type Ia supernovae are useful probes emits a beta particle; its nuclear charge of the structure of the universe, since they then increases by one. all have the same luminosity. By measur- Beta decay is often very slow, and, if ing the apparent brightness of these the flux of neutrons is high, the nucleus objects, one also measures the expansion might capture another neutron before 102 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies there is time for it to undergo decay. In neighbourhood of starspots and stellar this rapid, or r, process, the evolution of a flares, and they also arise from supernova nucleus can be very different from that in explosions themselves. Some of these a slow process. In supernova explosions, light-element nuclei also might be pro- vast quantities of neutrons can be produced by cosmic rays shattering atoms of duced, and these could result in the rapid carbon, nitrogen, oxygen, and other ele- buildup of massive elements. One inter- ments in the interstellar medium. esting feature of the synthesis of heavy Finally, the peculiar A-type stars com- elements by neutron capture at a high prise a class of cosmic objects with rate in a supernova explosion is that strange elemental abundance anomalies. nuclei much heavier than lead or even These might arise from mechanical uranium can be fashioned. These in turn effects—for example, selective radiation can decay by fission, releasing additional pressure or photospheric diffusion and amounts of energy. element separation—rather than from The superabundant elements in the nuclear effects. Some stars show enhanced S-type stars come from the slow neutron silicon, others enhanced lanthanides. The process. Moreover, the observation of so-called manganese stars show great technetium-99 is ample evidence that overabundances of manganese and gal- these processes are at work in stars today. lium, usually accompanied by an excess Even so, some low-abundance atomic of mercury. The latter stars exhibit weak nuclei are proton-rich (i.e., neutron-defi- helium lines, low rotational velocities, cient) and cannot be produced by either and excess amounts of gallium, stron- the s or the r process. Presumably, they tium, yttrium, mercury, and platinum, as have been created in relatively rare well as absences of such elements as events—e.g., one in which a quantum of aluminum and nickel. When these types hard radiation, a gamma-ray photon, of stars are found in binaries, the two causes a neutron to be ejected. members often display differing chemical In addition, no known nuclear pro- compositions. It is most difficult to envi- cess is capable of producing lithium, sion plausible nuclear events that can beryllium, and boron in stellar interiors. account for the peculiarities of these These lightweight nuclei are probably abundances, particularly the strange iso- produced by the breakdown, or spalla- tope ratios of mercury. tion, of heavier elements, such as iron and magnesium, by high-energy particles End states of stars in stellar atmospheres or in the early stages of star formation. Apparently, The final stages in the evolution of a star these high-energy particles, called cos- depend on its mass and angular momen- mic rays, originate by means of tum and whether it is a member of a electromagnetic disturbances in the close binary. Star Formation and Evolution | 103

White Dwarfs The central region of a typical white dwarf star is composed of a mixture of car- The faint white dwarf stars represent bon and oxygen. Surrounding this core is the endpoint of the evolution of inter- a thin envelope of helium and, in most mediate- and low-mass stars. White dwarf cases, an even thinner layer of hydrogen. stars, so called because of the white A very few white dwarf stars are sur- colour of the first few that were discov- rounded by a thin carbon envelope. Only ered, are characterized by a low the outermost stellar layers are accessible luminosity, a mass on the order of that of to astronomical observations. the Sun, and a radius comparable to that White dwarfs evolve from stars with of Earth. Because of their large mass an initial mass of up to three or four solar and small dimensions, such stars are masses or even possibly higher. After dense and compact objects with average quiescent phases of hydrogen and helium densities approaching 1,000,000 times burning in its core—separated by a first that of water. red-giant phase—the star becomes a red Unlike most other stars that are sup- giant for a second time. Near the end of ported against their own gravitation by this second red-giant phase, the star loses normal gas pressure, white dwarf stars its extended envelope in a catastrophic are supported by the degeneracy pres- event, leaving behind a dense, hot, and sure of the electron gas in their interior. luminous core surrounded by a glowing Degeneracy pressure is the increased spherical shell. This is the planetary- resistance exerted by electrons compos- nebula phase. During the entire course ing the gas, as a result of stellar of its evolution, which typically takes sev- contraction. The application of the so- eral billion years, the star will lose a major called Fermi-Dirac statistics and of fraction of its original mass through stel- special relativity to the study of the equi- lar winds in the giant phases and through librium structure of white dwarf stars its ejected envelope. The hot planetary- leads to the existence of a mass-radius nebula nucleus left behind has a mass of relationship through which a unique 0.5–1.0 solar mass and will eventually cool radius is assigned to a white dwarf of a down to become a white dwarf. given mass; the larger the mass, the White dwarfs have exhausted all their smaller the radius. Furthermore, the exis- nuclear fuel and so have no residual tence of a limiting mass is predicted, nuclear energy sources. Their compact above which no stable white dwarf star structure also prevents further gravita- can exist. This limiting mass, known as tional contraction. The energy radiated the Chandrasekhar limit, is on the order away into the interstellar medium is thus of 1.4 solar masses. Both predictions are provided by the residual thermal energy in excellent agreement with observations of the nondegenerate ions composing its of white dwarf stars. core. That energy slowly diffuses outward 104 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies through the insulating stellar envelope, called a close binary star, are aged: one is and the white dwarf slowly cools down. a red giant and the other a white dwarf. In Following the complete exhaustion of certain cases, the red giant expands into this reservoir of thermal energy, a process the gravitational domain of its compan- that takes several additional billion years, ion. The gravitational field of the white the white dwarf stops radiating and has dwarf is so strong that hydrogen-rich by then reached the final stage of its evo- matter from the outer atmosphere of the lution and becomes a cold and inert red giant is pulled onto the smaller star. stellar remnant. Such an object is some- When a sizable quantity of this material times called a black dwarf. accumulates on the surface of the white White dwarf stars are occasionally dwarf, a nuclear explosion occurs there, found in binary systems, as is the case for causing the ejection of hot surface gases 1 the white dwarf companion to the bright- on the order of ⁄10,000 the amount of mate- est star in the night sky, Sirius. Aside rial in the Sun. According to the prevailing from playing an essential role in Type Ia theory, the white dwarf settles down after supernovae, they are also behind the out- the explosion; however, the flow of hydro- bursts of novae and of other cataclysmic gen-rich material resumes immediately, variable stars. and the whole process that produced the Novae are a class of exploding stars outburst repeats itself, resulting in whose luminosity temporarily increases another explosion about 1,000 to 10,000 from several thousand to as much as years later. Recent research, however, 100,000 times its normal level. A nova suggests that such outbursts may recur reaches maximum luminosity within hours at much longer intervals—every 100,000 after its outburst and may shine intensely years or so. It is explained that a nova for several days or occasionally for a few eruption separates the members of the weeks, after which it slowly returns to its binary system, interrupting the transfer former level of luminosity. Stars that of matter until the two stars move close become novae are nearly always too faint together again after a considerable length before eruption to be seen with the unaided of time. eye. Their sudden increase in luminosity, however, is sometimes great enough to Neutron Stars make them readily visible in the nighttime sky. To observers, such objects may appear When the mass of a star’s remnant core to be new stars; hence the name nova from lies between 1.4 and about 2 solar masses, the Latin word for “new.” it apparently becomes a neutron star with Most novae are thought to occur in a density more than a million times double-star systems in which members greater than even that of a white dwarf. revolve closely around each other. Both These extremely dense, compact stars are members of such a system, commonly thought to be composed primarily of Star Formation and Evolution | 105

this solid layer, where the pressure is lowest, is composed of an extremely dense form of iron. Another important characteristic of neutron stars is the presence of very strong magnetic fi elds , upwards of 1012 Gauss (Earth’s magnetic fi eld is 0.5 Gauss), which causes the surface iron to be polymerized in the form of long chains of iron atoms. The individual atoms become compressed and elongated in the Geminga pulsar, imaged in X-ray wave- direction of the magnetic fi eld and can lengths by the Earth-orbiting XMM-Newton bind together end-to-end. Below the sur- X-ray Observatory. The pair of bright X-ray face, the pressure becomes much too “tails” outline the edges of a cone-shaped high for individual atoms to exist. shock wave produced by the pulsar as it The discovery of pulsars provided moves through space nearly perpendicular the fi rst evidence of the existence of neu- to the line of sight (from lower right to upper tron stars. Pulsars are neutron stars that left in the image). European Space Agency emit pulses of radiation once per rotation. The radiation emitted is usually neutrons. Neutron stars are typically radio waves. However, some objects are about 20 km (12 miles) in diameter. Their known to give off short rhythmic bursts masses range between 1.18 and 1.44 times of visible light, X-rays, and gamma radia- that of the Sun, but most are 1.35 times that tion as well, and others are “radio-quiet” of the Sun. Thus, their mean densities are and emit only at X- or gamma-ray wave- extremely high—about 10 14 times that of lengths. The very short periods of, for water. This approximates the density example, the Crab (NP 0532) and Vela inside the atomic nucleus, and in some pulsars (33 and 83 milliseconds, respec- ways a neutron star can be conceived of tively) rule out the possibility that they as a gigantic nucleus. might be white dwarfs. The pulses result It is not known defi nitively what is at from electrodynamic phenomena gener- the centre of the star, where the pressure ated by their rotation and their strong is greatest; theories include hyperons, magnetic fi elds, as in a dynamo. In the kaons, pions, and strange quark matter. case of radio pulsars, neutrons at the sur- The intermediate layers are mostly neu- face of the star decay into protons and trons and are probably in a “superfl uid” electrons. As these charged particles are state. The outer 1 km (0.6 mile) is solid, in released from the surface, they enter the spite of the high temperatures, which can intense magnetic fi eld (10 12 Gauss; Earth’s be as high as 1,000,000 K. The surface of magnetic fi eld is 0.5 Gauss) that 106 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies surrounds the star and rotates along J1748−2446ad, it has a period of 1.396 mil- with it. Accelerated to speeds approach- liseconds, which corresponds to a spin rate ing that of light, the particles give off of 716 times per second. These spin rates electromagnetic radiation by synchro- are close to the theoretical limit for a tron emission. This radiation is released pulsar because a neutron star rotating as intense radio beams from the pulsar’s only about four times faster would fly magnetic poles. apart as a result of “centrifugal force” at its These magnetic poles do not coin- equator, notwithstanding a gravitational cide with the rotational poles, and so the pull so strong that the star’s escape veloc- rotation of the pulsar swings the radia- ity is about half the speed of light. tion beams around. As the beams sweep These fast pulsars are known as mil- regularly past Earth with each complete lisecond pulsars. They form in supernovae rotation, an evenly spaced series of pulses like slower rotating pulsars; however, is detected by ground-based telescopes. millisecond pulsars often occur in binary Antony Hewish and Jocelyn Bell, star systems. After the supernova, the astronomers working at the University of neutron star accretes matter from its com- Cambridge, first discovered pulsars in panion, causing the pulsar to spin faster. 1967 with the aid of a radio telescope Careful timing of radio pulsars shows specially designed to record very rapid that they are slowing down very gradu- fluctuations in radio sources. Subsequent ally at a rate of typically a millionth of a searches have resulted in the detection of second per year. The ratio of a pulsar’s about 2,000 pulsars. A significant per- present period to the average slow-down centage of these objects are concentrated rate gives some indication of its age. This toward the plane of the Milky Way Galaxy, so-called characteristic, or timing, age the enormous galactic system in which can be in close agreement with the actual Earth is located. age. For example, the Crab Pulsar, which Although all known pulsars exhibit was formed during a supernova explo- similar behaviour, they show consider- sion observed in 1054 CE, has a able variation in the length of their characteristic age of 1,240 years; however, periods—i.e., the intervals between suc- pulsar J0205+6449, which was formed cessive pulses. The period of the slowest during a supernova in 1181 CE, has a pulsar so far observed is about 11.8 seconds characteristic age of 5,390 years. in duration. The pulsar designated PSR Because pulsars slow down so gradu- J1939+2134 was the fastest-known for ally, they are very accurate clocks. Since more than two decades. Discovered in pulsars also have strong gravitational 1982, it has a period of 0.00155 second, or fields, this accuracy can be used to test 1.55 milliseconds, which means it is spin- theories of gravity. American physicists ning 642 times per second. In 2006 an Joseph Taylor and Russell Hulse won the even faster one was reported; known as Nobel Prize for Physics in 1993 for their Star Formation and Evolution | 107 study of timing variations in the pulsar Pulsars also experience much more PSR 1913+16. PSR 1913+16 has a companion drastic period changes, which are called neutron star with which it is locked in a glitches, in which the period suddenly tight orbit. The two stars’ enormous inter- increases and then gradually decreases acting gravitational fields affect the to its pre-glitch value. Some glitches are regularity of the radio pulses, and by tim- caused by “starquakes,” or sudden cracks ing these and analyzing their variations, in the rigid iron crust of the star. Others Taylor and Hulse found that the stars were are caused by an interaction between the rotating ever faster around each other in crust and the more fluid interior. Usually an increasingly tight orbit. This orbital the interior is loosely coupled to the crust, decay is presumed to occur because the so the crust can slow down relative to the system is losing energy in the form of interior. However, sometimes the coupling gravity waves. This was the first experi- between the crust and interior becomes mental evidence for the existence of the stronger, spinning up the pulsar and gravitational waves predicted by Albert causing a glitch. Einstein in his general theory of relativity. Some X-ray pulsars are “accreting” Some pulsars, such as the Crab and pulsars. These pulsars are in binaries; the Vela pulsars, are losing rotational energy neutron star accretes material from its so precipitously that they also emit radia- companion. This material flows to the tion of shorter wavelength. The Crab magnetic polar caps, where it releases Pulsar appears in optical photographs as X-rays. Another class of X-ray pulsars is a moderately bright (magnitude 16) star called “anomalous.” These pulsars have in the centre of the Crab Nebula. Soon periods of more than five seconds, some- after the detection of its radio pulses in times give off bursts of X-rays, and are 1968, astronomers at the Steward ObservÂ often associated with supernova rem- atory in Arizona found that visible light nants. These pulsars arise from highly from the Crab Pulsar flashes at exactly magnetized neutron stars, or magnetars, the same rate. The star also produces reg- which have a magnetic field of between ular pulses of X-rays and gamma rays. 1014 and 1015 Gauss. (The magnetars also The Vela Pulsar is much fainter at optical have been identified with another class of wavelengths (average magnitude 24) and objects, the soft gamma-ray repeaters, was observed in 1977 during a particu- which give off bursts of gamma rays.) larly sensitive search with the large Some pulsars emit only in gamma Anglo-Australian Telescope situated at rays. In 2008 the Fermi Gamma-ray Space Parkes, Australia. It also pulses at X-ray Telescope discovered the first such pul- wavelengths. The Vela Pulsar does, how- sar within the supernova remnant CTA 1; ever, give off gamma rays in regular since then, it has found 11 others. Unlike pulses and is the most intense source of radio pulsars, the gamma-ray emission such radiation in the sky. does not come from the particle beams at 108 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies the poles but arises far from the neutron inward upon itself. The crushing weight star surface. The precise physical process of constituent matter falling in from all that generates the gamma-ray pulses is sides compresses the dying star to a point unknown. of zero volume and infinite density called Many binary X-ray sources, such as the singularity. Details of the structure of Hercules X-1, contain neutron stars. a black hole are calculated from Albert Cosmic objects of this kind emit X-rays by Einstein’s general theory of relativity. The compression of material from companion singularity constitutes the centre of a stars accreted onto their surfaces. black hole and is hidden by the object’s Neutron stars are also seen as objects “surface,” the event horizon. Inside the called rotating radio transients (RRATs) event horizon the escape velocity (i.e., and as magnetars. The RRATs are sources the velocity required for matter to that emit single radio bursts but at irreg- escape from the gravitational field of a ular intervals ranging from four minutes cosmic object) exceeds the speed of light, to three hours. The cause of the RRAT so that not even rays of light can escape phenomenon is unknown. Magnetars are into space. highly magnetized neutron stars that The radius of the event horizon is have a magnetic field of between 1014 and called the Schwarzschild radius, after the 1015 Gauss. German astronomer Karl Schwarzschild, Most investigators believe that neu- who in 1916 predicted the existence of tron stars are formed by supernova collapsed stellar bodies that emit no radi- explosions in which the collapse of the ation. The Schwarzschild radius (Rg) of an central core of the supernova is halted by object of mass M is given by the follow- rising neutron pressure as the core den- ing formula, in which G is the universal sity increases to about 1015 grams per gravitational constant and c is the speed cubic cm. If the collapsing core is more of light: massive than about three solar masses, 2 however, a neutron star cannot be formed, Rg = 2GM/c . and the core would presumably become a black hole. The size of the Schwarzschild radius is proportional to the mass of the collapsing Black Holes star. For a black hole with a mass 10 times as great as that of the Sun, the radius A black hole can be formed by the death would be 30 km (18.6 miles). of a massive star that exceeds about two Black holes cannot be observed solar masses. When such a star has directly on account of both their small exhausted its internal thermonuclear size and the fact that they emit no light. fuels at the end of its life, it becomes They can be “observed,” however, by the unstable and gravitationally collapses effects of their enormous gravitational Star Formation and Evolution | 109 fields on nearby matter. For example, if a Sagittarius A* demonstrated the presence black hole is a member of a binary star of a black hole with a mass equivalent to system, matter flowing into it from its 4,310,000 Suns. companion becomes intensely heated Supermassive black holes have been and then radiates X-rays copiously before seen in other galaxies as well. In 1994 the entering the event horizon of the black Hubble Space Telescope provided con- hole and disappearing forever. One of clusive evidence for the existence of a the component stars of the binary X-ray supermassive black hole at the centre of system Cygnus X-1 is a black hole. the M87 galaxy. It has a mass equal to two Discovered in 1971 in the constellation to three billion Suns but is no larger than Cygnus, this binary consists of a blue the solar system. The black hole’s exis- supergiant and an invisible companion tence can be inferred from its energetic 8.7 times the mass of the Sun that revolve effects on an envelope of gas swirling about one another in a period of 5.6 days. around it at extremely high velocities. Some black holes apparently have The existence of another kind of nonstellar origins. Various astronomers nonstellar black hole has been proposed have speculated that large volumes of by the British astrophysicist Stephen interstellar gas collect and collapse into Hawking. According to Hawking’s theory, supermassive black holes at the centres numerous tiny primordial black holes, of quasars and galaxies. A mass of gas possibly with a mass equal to that of an falling rapidly into a black hole is esti- asteroid or less, might have been created mated to give off more than 100 times as during the big bang, a state of extremely much energy as is released by the identi- high temperatures and density in which cal amount of mass through nuclear the universe is thought to have origi- fusion. Accordingly, the collapse of mil- nated 13.7 billion years ago. These lions or billions of solar masses of so-called mini black holes, like the more interstellar gas under gravitational force massive variety, lose mass over time into a large black hole would account for through Hawking radiation and disap- the enormous energy output of quasars pear. If certain theories of the universe and certain galactic systems. One such that require extra dimensions are correct, supermassive black hole, Sagittarius A*, the Large Hadron Collider (the world’s exists at the centre of the Milky Way most powerful particle accelerator) could Galaxy. In 2005, infrared observations of produce significant numbers of mini stars orbiting around the position of black holes. CHAPTER 4

Star Clusters

ur Sun is not part of a multiple star system. It fl oats Oalone in space, serene and solitary. However, many stars can be found in groups called star clusters. There are two general types of these stellar assemblages. They are held together by the mutual gravitational attraction of their members, which are physically related through common origin. The two types are open (formerly called galactic) clusters and globular clusters.

OPEN CLuSTERS

Open clusters contain from a dozen to many hundreds of stars, usually in an unsymmetrical arrangement. By contrast, globular clusters are old systems containing thousands to hundreds of thousands of stars closely packed in a symmetrical, roughly spherical form. In addition, groups called associations, made up of a few dozen to hundreds of stars of similar type and common origin whose density in space is less than that of the surrounding fi eld, are also recognized. Four open clusters have been known from earliest times: the Pleiades and Hyades in the constellation Taurus, Praesepe (the Beehive) in the constellation Cancer, and Coma Berenices . The Pleiades was so important to some early peoples that its rising at sunset determined the start of their year. The appearance of the Coma Berenices cluster to the Star Clusters | 111

NGC 6705, a rich cluster (left), and NGC 1508, a poor cluster (right). Courtesy of Lick Observatory, University of California naked eye led to the naming of its con- 1,000 open clusters. Probably about half stellation for the hair of Berenice, wife of the known open clusters contain fewer Ptolemy Euergetes of Egypt (3rd century than 100 stars, but the richest have 1,000 BCE ); it is the only constellation named or more. The largest have apparent diam- after a historical fi gure. eters of several degrees, the diameter of Open clusters are strongly concen- the Taurus cluster being 400 arc minutes trated toward the Milky Way. They form a (nearly seven arc degrees) and that of the fl attened disklike system 2,000 light- Perseus cluster being 240 arc minutes. years thick, with a diameter of about The linear diameters range from the 30,000 light-years. The younger clusters largest, 75 light-years, down to 5 light- serve to trace the spiral arms of the years. Increasingly, it has been found that Galaxy, since they are found invariably to a large halo of actual cluster members lie in them. Very distant clusters are hard surrounds the more-noticeable core and to detect against the rich Milky Way extends the diameter severalfold. Cluster background. A classifi cation based on membership is established through com- central concentration and richness is mon motion, common distances, and so used and has been extended to nearly on. Tidal forces and stellar encounters 112 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies lead to the disintegration of open clusters have been detected with radio telescopes. over long periods of time as stars “evapo- Many of the OB clusters mentioned rate” from the cluster. above contain globules—relatively small, Stars of all spectral classes from O to apparently spherical regions of absorb- M (high to low temperatures) are found ing matter. in open clusters, but the frequency of The most-numerous variables con- types varies from one cluster to another, nected with young open clusters are the T as does concentration near the centre. In Tauri type and related stars that occur by some (O or OB clusters), the brightest the hundreds in some nebulous regions stars are blue, very hot spectral types O of the sky. Conspicuously absent from or B. In others, they are whitish yellow, open clusters is the type most common in cooler spectral type F. High-luminosity globular clusters, the RR Lyrae stars. stars are more common than in the solar Other variables include eclipsing binary neighbourhood, and dwarfs are much stars (both Algol type and contact bina- more scarce. The brightest stars in some ries), flare stars, and spectrum variables, open clusters are 150,000 times as bright such as Pleione. The last-named star, one as the Sun. The luminosity of the bright- of the Pleiades, is known to cast off est stars at the upper end of the main shells of matter from time to time, perhaps sequence varies in clusters from about as a result of its high rotational speed (up to −2 visual magnitude. (Visual magni- to 322 km/sec [20 miles/sec]). About two tude is a magnitude measured through a dozen open clusters are known to contain yellow filter, the term arising because the Population I Cepheids, and since the dis- eye is most sensitive to yellow light.) tances of these clusters can be determined Because of the high luminosity of accurately, the absolute magnitudes of their brightest stars, some open clusters those Cepheids are well-determined. This have a total luminosity as bright as that has been of paramount importance in of some globular clusters (absolute mag- calibrating the period-luminosity relation nitude of ), which contain thousands of for Cepheids, and thus in determining times as many stars. In the centre of rich the distance scale of the universe. clusters, the stars may be only one light- The colour- or spectrum-magnitude year apart. The density can be 100 times diagram derived from the individual that of the solar neighbourhood. In some, stars holds vital information. Colour- such as the Pleiades and the Orion clus- magnitude diagrams are available for ters, nebulosity is a prominent feature, about 200 clusters on the UBV photo- while others have none. In clusters metric system, in which colour is younger than 25 million years, masses of measured from the amount of light radi- neutral hydrogen extending over three ated by the stars in the ultraviolet, blue, times the optical diameter of the cluster and visual (yellow) wavelength regions. Star Clusters | 113

In young clusters, stars are found along pronounced. The apparent convergence the luminous bright blue branch, whereas is caused by perspective: the cluster in old clusters, beyond a turnoff only a members are really moving as a swarm in magnitude or two brighter than the Sun, almost parallel directions and with about they are red giants and supergiants. the same speeds. The Hyades is the most- Distances can be determined by prominent example of a moving cluster. many methods—geometric, photometric, (The Hyades stars are converging with a and spectroscopic—with corrections for velocity of 45 km/sec (28 miles/sec) interstellar absorption. For the very toward the point in the sky with position nearest clusters, direct (trigonometric) coordinates right ascension 94 arc parallaxes may be obtained, and these are degrees, declination +7.6 arc degrees.) The inversely proportional to the distance. Ursa Major group, another moving cluster, Distances can be derived from proper occupies a volume of space containing motions, apparent magnitudes of the the Sun, but the Sun is not a member. The brightest stars, and spectroscopically from cluster consists of a compact nucleus of individual bright stars. Colour-magnitude 14 stars and an extended stream. diagrams, fitted to a standard plot of the Stellar groups are composed of stars main sequence, provide a common and presumed to have been formed together reliable tool for determining distance. The in a batch, but the members are now too nearest open cluster is the nucleus of the widely separated to be recognized as a Ursa Major group at a distance of 65 light- cluster. years; the farthest clusters are thousands Of all the open clusters, the Pleiades of light-years away. is the best known and perhaps the most Motions, including radial velocities thoroughly studied. This cluster, with a and proper motion, have been measured diameter of 35 light-years at a distance of for thousands of cluster stars. The radial 380 light-years, is composed of about 500 velocities of open cluster stars are much stars and is 100 million years old. Near smaller than those of globular clusters, the Pleiades in the sky but not so conspic- averaging tens of kilometres per second, uous, the Hyades is the second nearest but their proper motions are larger. Open cluster at 150 light-years. Its stars are clusters share in the galactic rotation. similar to those in the solar neighbour- Used with galactic rotation formulas, the hood, and it is an older cluster (about 615 radial velocities provide another means million years in age). Measurements of of distance determination. the Hyades long formed a basis for astro- A few clusters are known as moving nomical determinations of distance and clusters because the convergence of age because its thoroughly studied main the proper motions of their individual sequence was used as a standard. The stars toward a “convergent point” is higher-than-usual metal abundance in its 114 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies stars, however, complicated matters, and study of the distances and distributions it is no longer favoured in this way. Coma of globular clusters, the American astron- Berenices, located 290 light-years away, omer Harlow Shapley, then of the Mount is an example of a “poor” cluster, contain- Wilson Observatory in California, deter- ing only about 40 stars. There are some mined that its galactic centre lies in the extremely young open clusters. Of these, Sagittarius region. In 1930, from measure- the one associated with the Orion Nebula, ments of angular sizes and distribution which is some 4 million years old, is the of open clusters, Robert J. Trumpler of closest at a distance of 1,400 light-years. Lick Observatory in California showed A still younger cluster is NGC 6611, some that light is absorbed as it travels through of the stars in which formed only a few many parts of space. hundred thousand years ago. At the other More than 150 globular clusters were end of the scale, some open clusters have known in the Milky Way Galaxy by the ages approaching those of the globular early years of the 21st century. Most are clusters. M67 in the constellation Cancer widely scattered in galactic latitude, but is 4.5 billion years old, and NGC 188 in about a third of them are concentrated Cepheus is 6.5 billion years of age. The around the galactic centre, as satellite oldest known open cluster, Collinder 261 systems in the rich Sagittarius-Scorpius in the southern constellation of Musca, is star fields. Individual cluster masses 8.9 billion years old. include up to one million suns, and their linear diameters can be several hundred Globular Clusters light-years; their apparent diameters range from one degree for Omega Though several globular clusters, such Centauri down to knots of one minute of as Omega Centauri and Messier 13 in arc. In a cluster such as M3, 90 percent the constellation Hercules, are visible of the light is contained within a diam- to the unaided eye as hazy patches of eter of 100 light-years, but star counts light, attention was paid to them only and the study of RR Lyrae member stars after the invention of the telescope. The (whose intrinsic brightness varies regu- first record of a globular cluster, in the larly within well-known limits) include a constellation Sagittarius, dates to 1665 (it larger one of 325 light-years. The clusters was later named Messier 22); the next, differ markedly in the degree to which Omega Centauri, was recorded in 1677 stars are concentrated at their centres. by the English astronomer and mathe- Most of them appear circular and are matician Edmund Halley. probably spherical, but a few (e.g., Omega Investigations of globular and open Centauri) are noticeably elliptical. The clusters greatly aided the understanding most elliptical cluster is M19, its major of the Milky Way Galaxy. In 1917, from a axis being about double its minor axis. Star Clusters | 115

Globular clusters are composed of upper limit on the amount of neutral Population II objects (i.e., old stars). The hydrogen in globular clusters. Dark lanes brightest stars are the red giants, bright of nebulous matter are puzzling features red stars with an absolute magnitude of in some of these clusters. Though it is −2, about 600 times the Sun’s brightness difficult to explain the presence of dis- or luminosity. In relatively few globular tinct, separate masses of unformed matter clusters have stars as intrinsically faint as in old systems, the nebulosity cannot be the Sun been measured, and in no such foreground material between the cluster clusters have the faintest stars yet been and the observer. recorded. The luminosity function for M3 About 2,000 variable stars are known shows that 90 percent of the visual light in the 100 or more globular clusters that comes from stars at least twice as bright have been examined. Of these, perhaps as the Sun, but more than 90 percent of 90 percent are members of the class the cluster mass is made up of fainter called RR Lyrae variables. Other variables stars. The density near the centres of that occur in globular clusters are globular clusters is roughly two stars per Population II Cepheids, RV Tauri, and U cubic light-year, compared with one star Geminorum stars, as well as Mira stars, per 300 cubic light-years in the solar neigh- eclipsing binaries, and novae. bourhood. Studies of globular clusters have The colour of a star, as previously shown a difference in spectral properties noted, has been found generally to corre- from stars in the solar neighbourhood—a spond to its surface temperature, and in a difference that proved to be due to a defi- somewhat similar way the type of spec- ciency of metals in the clusters, which trum shown by a star depends on the have been classified on the basis of degree of excitation of the light-radiating increasing metal abundance. Globular atoms in it and therefore also on the tem- cluster stars are between 2 and 300 times perature. All stars in a given globular poorer in metals than stars like the Sun, cluster are, within a very small percentage with the metal abundance being higher of the total distance, at equal distances for clusters near the galactic centre than for from the Earth so that the effect of distance those in the halo (the outermost reaches on brightness is common to all. Colour- of the Galaxy extending far above and magnitude and spectrum-magnitude below its plane). The amounts of other diagrams can thus be plotted for the stars elements, such as helium, may also differ of a cluster, and the position of the stars from cluster to cluster. The hydrogen in in the array, except for a factor that is the cluster stars is thought to amount to 70–75 same for all stars, will be independent of percent by mass, helium 25–30 percent, distance. and the heavier elements 0.01–0.1 percent. In globular clusters all such arrays Radio astronomical studies have set a low show a major grouping of stars along the 116 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies lower main sequence, with a giant branch coalescence of two lower mass stars in a containing more-luminous stars curving “born-again” scenario that turned them from there upward to the red and with a into a single, more-massive, and seemingly horizontal branch starting about halfway younger star farther up the main sequence, up the giant branch and extending toward although this does not fit all cases. the blue. The other enigma is referred to as the This basic picture was explained as “second parameter” problem. Apart from owing to differences in the courses of evo- the obvious effect of age, the shape and lutionary change that stars with similar extent of the various sequences in a glob- compositions but different masses would ular cluster’s colour-magnitude diagram follow after long intervals of time. The are governed by the abundance of metals absolute magnitude at which the brighter in the chemical makeup of the cluster’s main-sequence stars leave the main members. This is the “first parameter.” sequence (the turnoff point, or “knee”) is Nevertheless, there are cases in which a measure of the age of the cluster, assum- two clusters, seemingly almost identical in ing that most of the stars formed at the age and metal abundance, show horizontal same time. Globular clusters in the Milky branches that are quite different—one Way Galaxy prove to be nearly as old as may be short and stubby, and the other may the universe, averaging perhaps 14 billion extend far toward the blue. There is thus years in age and ranging between evidently another, as-yet-unidentified approximately 12 billion and 16 billion parameter involved. Stellar rotation has years, although these figures continue to been mooted as a possible second param- be revised. RR Lyrae variables, when eter, but that now seems unlikely. present, lie in a special region of the Integrated magnitudes (measurements colour-magnitude diagram called the RR of the total brightness of the cluster), clus- Lyrae gap, near the blue end of the hori- ter diameters, and the mean magnitude zontal branch in the diagram. of the 25 brightest stars made possible Two features of globular cluster the first distance determinations on the colour-magnitude diagrams remain enig- basis of the assumption that the apparent matic. The first is the so-called “blue differences were due entirely to distance. straggler” problem. Blue stragglers are The colour-magnitude diagram, or the stars located near the lower main apparent magnitudes of the RR Lyrae sequence, although their temperature variables, however, leads to the best dis- and mass indicate that they already tance estimates. The correction factor for should have evolved off the main interstellar reddening, which is caused by sequence, like the great majority of other the presence of intervening matter that such stars in the cluster. A possible expla- absorbs and reddens stellar light, is sub- nation is that a blue straggler is the stantial for many globular clusters but Star Clusters | 117 small for those in high galactic latitudes, types of RR Lyrae stars were first distin- away from the plane of the Milky Way. guished in 1902. Omega Centauri is Distances range from about 8,000 light- relatively nearby, at a distance of 16,000 years for NGC 6397 to an intergalactic light-years, and it lacks a sharp nucleus. distance of 390,000 light-years for the The cluster designated 47 Tucanae (NGC cluster called AM-1. 104), with an absolute visual magnitude The radial velocities (the speed at of .3 at a similar distance of 13,500 which objects approach or recede from an light-years, has a different appearance observer, taken as positive when the dis- with strong central concentration. It is tance is increasing) measured by the located near the Small Magellanic Cloud Doppler effect have been determined but is not connected with it. For an from integrated spectra for some 138 observer situated at the centre of this globular clusters. The largest negative great cluster, the sky would have the velocity is 384 km/sec (239 miles/sec) for brightness of twilight on the Earth NGC 7006, while the largest positive because of the light of the thousands of velocity is 494 km/sec (307 miles/sec) for stars nearby. In the Northern Hemisphere, NGC 3201. These velocities suggest M13 in the constellation Hercules is the that the globular clusters are moving easiest to see and is the best known. At a around the galactic centre in highly ellip- distance of 22,000 light-years, it has been tical orbits. The globular cluster system thoroughly investigated and is relatively as a whole has a rotational velocity of poor in variables. M3 in Canes Venatici, about 180 km/s relative to the Sun, or 32,000 light-years away, is the cluster 30 km/s on an absolute basis. For one second richest in variables, with well cluster, Omega Centauri, motions of the more than 200 known. Investigation of individual stars around the massive cen- these variables resulted in the placement tre have actually been observed and of the RR Lyrae stars in a special region of measured. Though proper motions of the the colour-magnitude diagram. clusters are very small, those for individual stars provide a useful criterion for OB and T Associations cluster membership. The two globular clusters of highest The discovery of stellar associations absolute luminosity are in the Southern depended on knowledge of the character- Hemisphere in the constellations CentauÂ istics and motions of individual stars rus and Tucana. Omega Centauri, with an scattered over a substantial area. In the (integrated) absolute visual magnitude 1920s it was noticed that young, hot blue of −10.2, is the richest cluster in variables, stars (spectral types O and B) apparently with nearly 300 known in the early 21st congregated together. In 1949 Victor A. century. From this large group, three Ambartsumian, a Soviet astronomer, 118 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies suggested that these stars are members frequently contain a special type of mul- of physical groupings of stars with a com- tiple star, the Trapezium (named for its mon origin and named them O prototype in Orion), as well as super- associations (or OB associations, as they giants, binaries, gaseous nebulae, and are often designated today). He also globules. Associations are relatively applied the term T associations to groups homogeneous in age. The best distance of dwarf, irregular T Tauri variable stars, determinations are from spectroscopic which were first noted at Mount Wilson parallaxes of individual stars—i.e., esti- Observatory by Alfred Joy. mates of their absolute magnitudes made The chief distinguishing feature of from studies of their spectra. Most of the members of a stellar association is those known are closer than 10,000 light- that the large majority of constituent stars years, with the nearest association, have similar physical characteristics. An straddling the boundary between OB association consists of many hot, blue Centaurus and Crux, at 385 light-years. giant stars, spectral classes O and B, and Associations appear to be almost a relatively small number of other objects. spherical, though rapid elongation would A T association consists of cooler dwarf be expected from the shearing effect of stars, many of which exhibit irregular differential galactic rotation. Expansion, variations in brightness. The stars clearly which is on the order of 10 km/sec (6 miles/ must be relatively close to each other in sec), may well mask the tendency to elon- space, though in some cases they might gate, and this is confirmed in some. Tidal be widely dispersed in the sky and are forces break up an association in less than less closely placed than in the open 10 million years through differences in the clusters. attraction by an outside body on members The existence of an OB association is in different parts of the association. usually established through a study of A good example of an OB association the space distribution of early O- and is Per OB 1 at a distance of some 7,500 B-type stars. It appears as a concentra- light-years, which spreads out from the tion of points in a three-dimensional plot double cluster h and χ Persei. A large of galactic longitude and latitude and group of 20 supergiant stars of spectral distance. More than 70 have been cata- type M belongs to Per OB 1. Associations loged and are designated by constellation with red supergiants may be in a relatively abbreviation and number (e.g., Per OB 1 advanced evolutionary stage, almost ready in the constellation Perseus). In terms of to disintegrate. dimensions, they are larger than open The T associations (short for T Tauri clusters, ranging from 100 to 700 light- associations) are formed by groups of T years in diameter, and usually contain Tauri stars associated with the clouds of one or more open clusters as nuclei. They interstellar matter (nebulae) in which Star Clusters | 119 they occur. About three dozen are recog- globular clusters dotted around it. The nized. A T Tauri star is characterized by richest parts of the spiral arms of the pin- irregular variations of light, low luminos- wheel would be marked by dozens of ity, and hydrogen line (H-alpha) emission. open clusters. If this panorama could be It is a newly formed star of intermediate seen as a time-lapse movie, the great mass that is still in the process of contrac- globular clusters would wheel around the tion from diffuse matter. The small galactic centre in elliptical orbits with motions of T Tauri stars relative to a periods of hundreds of millions of years. given nebula indicate that they are not The open clusters and stellar associa- field stars passing through the nebula. tions would be seen to form out of knots They are found in greatest numbers in of diffuse matter in the spiral arms, grad- regions with bright O- and B-type stars. ually disperse, run through their life cycle, T associations occur only in or near and fade away, while the Sun pursued its regions of galactic nebulosity, either course around the galactic centre for bright or dark, and only in obscured billions of years. regions showing the presence of dust. Young open clusters and associations, Besides T Tauri stars, they include related occupying the same region of space as variables, nonvariable stars, and Herbig- clouds of ionized hydrogen (gaseous Haro objects—small nebulosities 10,000 nebulae), help to define the spiral arms. A astronomical units in diameter, each con- concentration of clusters in the bright taining several starlike condensations in inner portion of the Milky Way between configurations similar to the Trapezium, galactic longitudes 283° and 28° indicates Theta Orionis, in the sword of Orion. an inner arm in Sagittarius. Similarly, the These objects are considered to be star two spiral arms of Orion and Perseus are groups at the very beginning of life. defined between 103° and 213°, with a The constellation of Cygnus has five bifurcation of the Orion arm. Associations T associations, and Orion and Taurus show the existence of spiral structure in have four each. The richest is Ori T2, with the Sun’s vicinity. Older clusters, whose more than 400 members; it has a diameter main sequence does not reach to the blue of 50 by 90 light-years and lies at a dis- stars, show no correlation with spiral tance of 1,300 light-years around the arms because in the intervening years variable star T Ori. their motions have carried them far from their place of birth. Dynamics of star clusters All the O- and B-type stars in the Galaxy might have originated in OB Seen from intergalactic space, the Milky associations. The great majority, if not all, Way Galaxy would appear as a giant of the O-type stars were formed and still luminous pinwheel, with more than 150 exist in clusters and associations. Though 120 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies only 10 percent of the total number of merge with larger ones that do not neces- B-type stars are now in OB associations sarily have the same properties. This has or clusters, it is likely that all formed in complicated the picture of chemical evo- them. At the other (fainter) end of the lution. The case of the globular cluster range of stellar luminosities, the number Omega Centauri suggests this merging of dwarf variable stars in the nearby T also may happen on smaller scales. Its associations is estimated at 12,000. These stars are unusual, perhaps unique, in associations are apparently the main having a variety of chemical composi- source of low-luminosity stars in the tions, as though they came from more neighbourhood of the Sun. than one earlier cluster. While large numbers of associations In a study of star clusters, a time have formed and dispersed and provided a panorama unfolds—from the oldest population of stars for the spiral arms, the objects existing in the Galaxy, the globu- globular clusters have survived relatively lar clusters, through clusters in existence unchanged except for the evolutionary only half as long, to extremely young differences that time brings. They are too open clusters and associations that have massive to be disrupted by the tidal forces come into being since humans first trod of the Galaxy, though their limiting dimen- the Earth. sions are set by these forces when they most closely approach the galactic centre. Clusters in external Impressive as they are individually, their galaxies total mass of 10 million suns is small compared with the mass of the Galaxy as a The study of clusters in external galaxies whole—only about 1/10,000. Their sub- began in 1847, when Sir John Herschel at stance is that of the Galaxy in a very early the Cape Observatory (in what is now stage. The Galaxy probably collapsed from South Africa) published lists of such a gaseous cloud composed almost entirely objects in the nearest galaxies, the of hydrogen and helium. About 14 billion Magellanic Clouds. During the 20th cen- years ago, before the last stages of the tury the identification of clusters was collapse, matter forming the globular extended to more remote galaxies by the clusters may have separated from the use of large reflectors and other more rest. The fact that metal-rich clusters are specialized instruments, including near the galactic nucleus while metal- Schmidt telescopes. poor clusters are in the halo or outer Clusters have been discovered and fringes may indicate a nonuniform dis- studied in many external galaxies, par- tribution of elements throughout the ticularly members of the Local Group (a primordial mass. However, there is evi- group of about 40 stellar systems to which dence that galaxies are given to the Galaxy belongs). At their great dis- cannibalism, in which smaller galaxies tances classification is difficult, but it has Star Clusters | 121 been accomplished from studies of the 122 associations with a mean diameter of colours of the light from an entire cluster 250 light-years, somewhat richer and (integrated colours) or, for relatively few, larger than in the Galaxy. Sixteen of the from colour-magnitude diagrams. associations contain coexistent clusters. Clusters have been found by the hun- Also, 15 star clouds (aggregations of dreds in some of the nearest galaxies. At many thousands of stars dispersed over the distance of the Magellanic Clouds, a hundreds or even thousands of light- cluster like the Pleiades would appear as years) are recognized. a faint 15th magnitude object, subtend- In the great Andromeda spiral galaxy ing 15 seconds of arc instead of several (M31) some 2.2 million light-years away, degrees. Nevertheless, it is estimated that about 500 globular clusters are known. the Small Magellanic Cloud, at a distance Colour studies of some of these clusters of 200,000 light-years, contains about reveal that they have a higher metal con- 2,000 open clusters. In the Large tent than globular clusters of the Galaxy. Magellanic Cloud, at a distance of 163,000 Nearly 200 OB associations are known, light-years, over 1,200 of an estimated with distances up to 80,000 light-years 4,200 have been cataloged. Most of them from the nucleus. The diameters of their are young blue-giant open clusters such dense cores are comparable to those of as NGC 330 and NGC 1866. The open galactic associations. NGC 206 (OB 78) is clusters contain some Cepheid variables the richest star cloud in M31, having a total and in chemical composition are similar mass of 200,000 suns and bearing a strong to, but not exactly the same as, those of resemblance to the double cluster in the Galaxy. The globular clusters fall into Perseus. Some globular clusters have been two distinct groups. Those of the first found around the dwarf elliptical compan- group, the red, have a large metal defi- ions to M31, NGC 185, and NGC 205. ciency similar to the globular clusters in M33 in the constellation Triangulum—a the Galaxy, and some are known to con- spiral galaxy with thick, loose arms (an tain RR Lyrae variables. The globular Sc system in the Hubble classification clusters of the second group are large and scheme)—has about 300 known clusters, circular in outline, with colours much not many of which have globular charac- bluer than normal galactic globular clus- teristics. Of the six dwarf spheroidal ters and with ages of about one million to galaxies in the Local Group, only the one one billion years. They are similar to the in the constellation Fornax has clusters. open clusters of the Magellanic Clouds Its five globular clusters are similar to the but are very populous. The observed dif- bluest globular clusters of the GalÂaxy. No ferences between clusters in the Galaxy clusters have been discovered in the irreg- and the Magellanic Clouds result from ular galaxies NGC 6822 and IC 1613. small differences in helium or heavy- Beyond the Local Group, at a distance element abundances. There are at least of 45 million light-years, the giant 122 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies elliptical galaxy M87 in the Virgo cluster astronomers announced the detection of a of galaxies is surrounded by an estimated planet orbiting it. The extrasolar planet is 13,000 globular star clusters. Inspection not visible from Earth, but its presence of other elliptical galaxies in Virgo shows was deduced from the wobble that its grav- that they too have globular clusters ity induces in the parent star’s motion in a whose apparent magnitudes are similar 4.23-day cycle. It has a mass 47 percent to those in M87, though their stellar that of Jupiter and orbits surprisingly population is substantially smaller. It close (7.8 million km [4.8 million miles]) to appears that the mean absolute magni- the star—much closer than Mercury, which tudes of globular clusters are constant orbits the Sun at a distance of 57.9 million and independent of the absolute lumi- km (35.9 million miles). nosity of the parent galaxy. The total number of clusters now Alpha Centauri known in external galaxies far exceeds the number known in the Milky Way system. Alpha Centauri is a triple star, the faintest component of which, Proxima Notable stars and Centauri, is the closest star to the Sun, at star clusters about 4.2 light-years’ distance. The two brighter components, about 1/5 light-year Of the billions of stars and star clusters farther from the Sun, revolve around each in the universe, some have stood out other with a period of about 80 years, among the rest for a variety of reasons.In while Proxima may be circling them with the section that follows, greater detail is a period probably of 500,000 years. The presented on many of those that have brightest component star resembles the distinguished themselves as distinctive Sun in spectral type, diameter, and abso- objects in the sky or in the history of lute magnitude. Its apparent visual astronomy itself. magnitude is 0.0. The second brightest component, of visual magnitude 1.4, is a 51 Pegasi redder star. The third component, of 11th magnitude, is a red dwarf star. The fifth-magnitude star 51 Pegasi is As seen from Earth, the system is located 48 light-years away from Earth in the fourth brightest star (after Sirius, the constellation Pegasus and was the first Canopus, and Arcturus); the red dwarf sunlike star confirmed to possess a planet. Proxima is invisible to the unaided eye. 51 Pegasi, which has physical properties Alpha Centauri lies in the southern (luminosity and temperature, for example) constellation Centaurus and can be very similar to those of the Sun, became seen only from south of about 40° north the focus of attention in 1995 when latitude. Star Clusters | 123

Arcturus episodes of mass loss over the past 100,000 years. The largest of these shells has a Arcturus (also called Alpha Boötis, ) is radius of nearly 7.5 light-years. one of the five brightest stars in the night sky and the brightest star in the northern Deneb constellation Boötes, with an apparent visual magnitude of −0.05. It is an orange- Deneb, which is also called Alpha Cygni, coloured giant star 36.7 light-years from is one of the brightest stars, with an Earth. It lies in an almost direct line with apparent magnitude of 1.25. Its name is the tail of Ursa Major (the Great Bear); Arabic for “tail” (of the Swan, Cygnus). hence its name, derived from the Greek This star, at about 1,500 light-years’ dis- words for “bear guard.” tance, is the most remote (and brightest intrinsically) of the 20 apparently bright- Betelgeuse est stars. It lies in the northern constellation Cygnus and, with Vega and Betelgeuse, or Alpha Orionis, is the Altair, forms the prominent “Summer brightest star in the constellation Orion, Triangle.” marking the eastern shoulder of the hunter. Its name is derived from the Arabic word Fomalhaut bat al-dshauzâ, which means “the giant’s shoulder.” Betelgeuse has a variable Fomalhaut, or Alpha Piscis Austrini, is apparent magnitude of about 0.6 and is the 17th star (in order of apparent bright- one of the most luminous stars in the ness). It is used in navigation because of night sky. It is easily discernible to even its conspicuous place in a sky region oth- the casual observer, not only because of erwise lacking in bright stars. It lies in its brightness and position in the bril- the southern constellation Piscis liant Orion but also because of its Austrinus, 25 light-years from Earth. A deep-reddish colour. The star is approxi- white star, it has an apparent magnitude mately 640 light-years from Earth. of 1.16. A sixth-magnitude companion Betelgeuse is a red supergiant star star, HR 8721, is yellow and orbits at a dis- roughly 950 times as large as the Sun, tance of about 0.9 light-year. A belt of making it one of the largest stars known. dust orbits between 19.9 and 23.6 billion For comparison, the diameter of Mars’s km (12.4 and 14.7 billion miles) from the orbit around the Sun is 328 times the star. Images taken with the Hubble Space Sun’s diameter. Infrared studies from Telescope in 2004 and 2006 showed a spacecraft have revealed that Betelgeuse planet, Fomalhaut b, orbiting inside the is surrounded by immense shells of mate- dust belt at a distance of 17.8 billion km rial evidently shed by the star during (11.1 billion miles) from the star. These 124 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

material and more than 1,000 stars, of which six or seven can be seen by the unaided eye and have fi gured promi- nently in the myths and literature of many cultures. In Greek mythology the Seven Sisters ( Alcyone , Maia , Electra , Merope , Taygete , Celaeno , and Sterope , names now assigned to individual stars), daugh- ters of Atlas and Pleione, were changed into the stars. The heliacal (near dawn) rising of the Pleiades in spring of the Northern Hemisphere has marked from ancient times the opening of seafaring and farming seasons, as the morning setting of the group in autumn signifi ed Bright nebulosity in the Pleiades (M45, NGC the seasons’ ends. Some South American 1432), distance 430 light-years. Cluster stars Indians use the same word for “Pleiades” provide the light, and surrounding clouds of and “year.” dust reﬂ ect and scatter the rays from the The cluster was fi rst examined tele- stars. Hale Observatories © 1961 scopically by Galileo , who found more than 40 members; it was fi rst photographed by Paul and Prosper Henry in 1885. were the fi rst confi rmed images of an extrasolar planet . The planet has a mass Polaris three times that of Jupiter and an orbital period of 872 years. Polaris, which is also called Alpha Ursae Fomalhaut was associated with the Minoris, is Earth’s present northern Roman goddess Ceres (associated with polestar, or North Star. It can be found the analogous Sicilian and Greek god- at the end of the “handle” of the so-called dess Demeter) and was worshipped; in Little Dipper in the constellation Ursa astrology it is one of four royal stars. Minor . Polaris is actually a triple star, the brighter of two visual components Pleiades being a spectroscopic binary with a period of about 30 years and a Cepheid The Pleiades (catalog number M45) is variable with a period of about 4 days. Its an open cluster of young stars in the changes in brightness are too slight to zodiacal constellation Taurus, about 430 be detected with the unaided eye. light-years from the solar system. It con- Apparent visual magnitude of the Polaris tains a large amount of bright nebulous system is 2.00. Star Clusters | 125

Sirius than the Sun. Its distance from the solar system is about 8.6 light-years, only twice Sirius—which is also called Alpha Canis the distance of the nearest known star Majoris, or the Dog Star—is the brightest beyond the Sun. Its name probably comes star in the night sky, with apparent visual from a Greek word meaning “sparkling,” magnitude of −1.44. It is a binary star in or “scorching.” the constellation Canis Major. The bright Sirius was known as Sothis to the component of the binary is a blue-white ancient Egyptians, who were aware that it star 23 times as luminous as the Sun and made its fi rst heliacal rising (i.e., rose just somewhat larger and considerably hotter before sunrise) of the year at about the

Sirius A and B (lower left) photographed by the Hubble Space Telescope. NASA, ESA, H. Bond (STScI), and M. Barstow (University of Leicester) 126 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies time the annual floods were beginning in eccentricity and with average separation the Nile River delta. They long believed of the stars of about 20 times the Earth’s that Sothis caused the Nile floods; and distance from the Sun. Despite the glare they discovered that the heliacal rising of of the bright star, the seventh-magnitude the star occurred at intervals of 365.25 companion is readily seen with a large days rather than the 365 days of their telescope. This companion star, known calendar year, a correction in the length as Sirius B, is about as massive as the of the year that was later incorporated in Sun, though much more condensed, and the Julian calendar. Among the ancient was the first white dwarf star to be Romans, the hottest part of the year was discovered. associated with the heliacal rising of the Dog Star, a connection that survives in Vega the expression “dog days.” That Sirius is a binary star was first Vega, or Alpha Lyrae, is the brightest star reported by the German astronomer in the northern constellation Lyra and Friedrich Wilhelm Bessel in 1844. He had the fifth brightest in the night sky, with a observed that the bright star was pursu- visual magnitude of 0.03. It is also one of ing a slightly wavy course among its the Sun’s closer neighbours, at a distance neighbours in the sky and concluded that of about 25 light-years. Vega’s spectral it had a companion star, with which it type is A (white) and its luminosity class revolved in a period of about 50 years. V (main sequence). It will become the The companion was first seen in 1862 by northern polestar by about 14,000 CE Alvan Clark, an American astronomer because of the precession of the equinoxes. and telescope maker. Vega is surrounded by a disk of circums- Sirius and its companion revolve tellar dust that may be similar to the solar together in orbits of considerable system’s Kuiper Belt. CHAPTER 5

Nebulae

cattered between the stars in interstellar space are various Stenuous clouds of gas and dust. These clouds are called nebulae (Latin: “mist” or “cloud”). The term was formerly applied to any object outside the solar system that had a diffuse appearance rather than a pointlike image, as in the case of a star. This defi nition, adopted at a time when very distant objects could not be resolved into great detail, unfortunately includes two unrelated classes of objects: the extragalactic nebulae, the enormous collections of stars and gas now called galaxies, and the galactic nebulae, which are composed of the interstellar medium (the gas between the stars, with its accompanying small solid particles) within a single galaxy. Today, the term nebula generally refers exclusively to the interstellar medium . In a spiral galaxy the interstellar medium makes up 3 percent to 5 percent of the galaxy’s mass, but within a spiral arm its mass fraction increases to about 20 percent. About 1 percent of the mass of the interstellar medium is in the form of “dust”—small solid particles that are effi cient in absorbing and scattering radiation. Much of the rest of the mass within a galaxy is concentrated in visible stars, but there is also some form of dark matter that accounts for a substantial fraction of the mass in the outer regions. The most conspicuous property of interstellar gas is its clumpy distribution on all size scales observed, from the 128 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

The Cat’s Eye nebula. NASA, ESA, HEIC, and The Hubble Heritage Team (STScI/AURA) size of the entire Milky Way Galaxy large-scale variations are seen by direct (about 1020 metres, or hundreds of thou- observation; the small-scale variations sands of light-years) down to the are observed by fl uctuations in the inten- distance from Earth to the Sun (about sity of radio waves, similar to the 1011 metres, or a few light-minutes). The “twinkling” of starlight caused by Nebulae | 129 unsteadiness in the Earth’s atmosphere. together making up about two atoms per Various regions exhibit an enormous thousand. range of densities and temperatures. On the basis of appearance, nebulae Within the Galaxy’s spiral arms about can be divided into two broad classes: half the mass of the interstellar medium dark nebulae and bright nebulae. Dark is concentrated in molecular clouds, in nebulae appear as irregularly shaped which hydrogen occurs in molecular black patches in the sky and blot out the form (H2) and temperatures are as low light of the stars that lie beyond them. as 10 kelvins (K). These clouds are incon- They are very dense and cold molecular spicuous optically and are detected clouds; they contain about half of all inter- principally by their carbon monoxide stellar material. Typical densities range (CO) emissions in the millimetre wave- from hundreds to millions (or more) of length range. Their densities in the hydrogen molecules per cubic centime- regions studied by CO emissions are typ- tre. These clouds are the sites where new ically 1,000 H2 molecules per cubic cm stars are formed through the gravitational (1 cubic cm = .06 cubic in). At the other collapse of some of their parts. Most of extreme is the gas between the clouds, the remaining gas is in the diffuse inter- with a temperature of 10 million K and a stellar medium, relatively inconspicuous density of only 0.001 H+ ion per cubic cm. because of its very low density (about 0.1 Such gas is produced by supernovae, the hydrogen atom per cubic cm) but detect- violent explosions of unstable stars. able by its radio emission of the 21-cm line of neutral hydrogen. Classes of nebulae Bright nebulae appear as faintly luminous glowing surfaces; they either emit All nebulae observed in the Milky Way their own light or reflect the light of Galaxy are forms of interstellar matter. nearby stars. They are comparatively Their appearance differs widely, depend- dense clouds of gas within the diffuse ing not only on the temperature and interstellar medium. They have several density of the material observed but also subclasses: (1) reflection nebulae, (2) H II on how the material is spatially situated regions, (3) diffuse ionized gas, (4) with respect to the observer. Their planetary nebulae, and (5) supernova chemical composition, however, is fairly remnants. uniform. It corresponds to the composi- Reflection nebulae reflect the light of tion of the universe in general in that a nearby star from their constituent dust approximately 90 percent of the constitu- grains. The gas of reflection nebulae is ent atoms are hydrogen and nearly all the cold, and such objects would be seen as rest are helium, with oxygen, carbon, dark nebulae if it were not for the nearby neon, nitrogen, and the other elements light source. 130 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

H II regions are clouds of hydrogen than the much more spectacular H II ionized (separated into positive H + ions regions, planetary nebulae, or supernova and free electrons) by a neighbouring hot remnants that occupy a tiny fraction of star. The star must be of stellar type O or the volume. B, the most massive and hottest of nor- Planetary nebulae are ejected from mal stars in the Galaxy, in order to stars that are dying but are not massive produce enough of the radiation required enough to become supernovae—namely, to ionize the hydrogen. red giant stars . That is to say, a red giant Diff use ionized gas , so pervasive has shed its outer envelope in a less- among the nebular clouds, is a major violent event than a supernova explosion component of the Galaxy. It is observed and has become an intensely hot star by faint emissions of positive hydrogen, surrounded by a shell of material that nitrogen, and sulfur ions (H + , N + , and S+ ) is expanding at a speed of tens of detectable in all directions. These emissions collectively require far more power

Planetary Nebula Hen 1357, as photo- A star-forming region in the Orion graphed by the Hubble Space Telescope. Nebula (M42, NGC 1976). This compos- Located about 18,000 light-years from ite image shows an area one light-year Earth in the constellation Ara the Altar, square near the edge of a cavity of ion- this expanding cloud of gas was ized hydrogen heated by ultraviolet expelled from an aging star in the neb- radiation from a star cluster at the ula’s centre. National Aeronautics and nebula’s centre. National Aeronautics Space Administration and Space Administration Nebulae | 131 kilometres per second. Planetary nebulae seeking. A by-product of their search was typically appear as rather round objects the discovery of many bright nebulae. of relatively high surface brightness. Several catalogs of special objects were Their name is derived from their super- compiled by comet researchers; by far the ficial resemblance to planets—i.e., their best known is that of the Frenchman regular appearance when viewed tele- Charles Messier, who in 1781 compiled a scopically as compared with the chaotic catalog of 103 nebulous, or extended, forms of other types of nebula. objects in order to prevent their confu- Supernova remnants are the clouds sion with comets. Most are clusters of of gas expanding at speeds of hundreds stars, 35 are galaxies, and 11 are nebulae. or even thousands of kilometres per sec- Even today many of these objects are ond from comparatively recent explosions commonly referred to by their Messier of massive stars. If a supernova remnant catalog number; M20, for instance, is the is younger than a few thousand years, it great Trifid Nebula, in the constellation may be assumed that the gas in the neb- Sagittarius. ula was mostly ejected by the exploded star. Otherwise, the nebula would consist The Work of the Herschels chiefly of interstellar gas that has been swept up by the expanding remnant of By far the greatest observers of the early older objects. and middle 19th century were the English astronomers William Herschel and his Early observations son John. Between 1786 and 1802 William of nebulae Herschel, aided by his sister Caroline, compiled three catalogs totaling about In 1610, two years after the invention of 2,500 clusters, nebulae, and galaxies. the telescope, the Orion Nebula, which John Herschel later added to the catalogs looks like a star to the naked eye, was 1,700 other nebulous objects in the south- discovered by the French scholar and ern sky visible from the Cape Observatory naturalist Nicolas-Claude Fabri de in South Africa but not from London and Peiresc. In 1656 Christiaan Huygens, the 500 more objects in the northern sky vis- Dutch scholar and scientist, using his ible from England. own greatly superior instruments, was The catalogs of the Herschels formed the first to describe the bright inner the basis for the great New General region of the nebula and to determine Catalogue (NGC) of J.L. Dreyer, pub- that its inner star is not single but a com- lished in 1888. It contains the location pact quadruple system. and a brief description of 7,840 nebulae, Early 18th-century observational galaxies, and clusters. In 1895 and 1908 astronomers gave high priority to comet it was supplemented by two Index 132 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

Catalogues (IC) of 5,386 additional the spectrum of the Orion Nebula showed objects. The list still included galaxies as bright emission lines of glowing gases, well as true nebulae, for they were often with conspicuous hydrogen lines and at this time still indistinguishable. Most some green lines even brighter. By con- of the brighter galaxies are still identi- trast, the spectrum of galaxies was found fied by their NGC or IC numbers to be stellar, so a distinction between gal- according to their listing in the New axies and nebulae—that nebulae are General Catalogue or Index Catalogues. gaseous and galaxies are stellar—was appreciated at this time, although the Advances Brought by true sizes and distances of galaxies were Photography and not demonstrated until the 20th century. Spectroscopy 20th-century discoveries The advent of photography, which allows the recording of faint details invisible to The 20th century witnessed enormous the naked eye and provides a permanent advances in observational techniques as record of the observation for study of fine well as in the scientific understanding of details at leisure, caused a revolution in the physical processes that operate in the understanding of nebulae. In 1880 the interstellar matter. In 1930 a German opti- first photograph of the Orion Nebula was cal worker, Bernhard Schmidt, invented made, but really good ones were not an extremely fast wide-angled camera obtained until 1883. These early photo- ideal for photographing faint extended graphs showed a wealth of detail nebulae. Photographic plates became pro- extending out to distances unsuspected gressively more sensitive to an by visual observers. ever-widening range of colours, but pho- Much can be learned about the physi- tography has been completely replaced cal nature of an astronomical object by by photoelectric devices. Most images studying its spectrum—i.e., the resolution are now recorded with so-called charge- of its light into different wavelengths (or coupled devices (CCDs) that act as arrays colours). Study of the spectrum of an of tiny photoelectric cells, each recording object provides a decisive test as to the light from a small patch of sky. Modern whether it is composed of unresolved CCDs consist of square arrays of up to stars (as are galaxies) or glowing gas. 4,000 cells on each side, or 16 million inde- Stars radiate at all wavelengths, almost pendent photocells, capable of observing always with dark absorption lines super- the sky simultaneously. Electronic detec- imposed, while hot, transparent gas tors are up to 100 times more sensitive clouds radiate only emission lines at cer- than photography, can record a much wider tain wavelengths characteristic of their range of light levels, and are sensitive to constituent gases. In 1864 observation of a much wider range of wavelengths, from Nebulae | 133

0.1 micrometre (3.93700787 x 10-6 inch) in astronomers to record simultaneously a the ultraviolet (accessible only from sat- wide range of wavelengths with very high ellites orbiting above Earth’s atmosphere) spectral resolution (i.e., to distinguish to more than 1.2 micrometres (4.72440945 slightly differing wavelengths). For even x 10-5 inch) in the infrared. higher spectral resolution astronomers Spacecraft allow the observation of employ Fabry-Pérot interferometers. radiation normally absorbed by Earth’s Spectra provide powerful diagnostics of atmosphere: gamma and X-rays (which the physical conditions within nebulae. have very short wavelengths), far-ultravio- Images and spectra provided by Earth- let radiation (with wavelengths shorter than orbiting satellites, especially the Hubble about 0.3 micrometre [1.18110236 x 10-5 Space Telescope, have yielded data of inch], below which atmospheric ozone is unprecedented quality. strongly absorbing), and infrared (from Ground-based observations also have about 3 micrometres to 1 mm [.00012 inches played a major role in recent advances in to .039 inch]), strongly absorbed by atmo- scientific understanding of nebulae. The spheric water vapour and carbon dioxide. emission of gas in the radio and sub- Gamma rays, X-rays, and ultraviolet radia- millimetre wavelength ranges provides tion reveal the physical conditions in the crucial information regarding physical hottest regions in space (extending to some conditions and molecular composition. 100 million kelvins in shocked supernova Large radio telescope arrays, in which gas). Infrared radiation reveals the condi- several individual telescopes function tions within dark cold molecular clouds, collectively as a single enormous instru- into which starlight cannot penetrate ment, give spatial resolutions in the radio because of absorbing dust layers. regime far superior to any yet achieved The primary means of studying nebu- by optical means. lae is not by images but by spectra, which show the relative distribution of the radia- Chemical composition tion among various wavelengths (or colours and physical processes for optical radiation). Spectra can be obtained by means of prisms (as in the ear- Many characteristics of nebulae are lier part of the 20th century), diffraction determined by the physical state of their gratings, or crystals, in the case of X-rays. A constituent hydrogen, by far the most particularly useful instrument is the echelle abundant element. For historical reasons, spectrograph, in which one coarsely ruled nebulae in which hydrogen is mainly ion- grating spreads the electromagnetic radia- ized (H+) are called H II regions, or diffuse tion in one direction, while another finely nebulae. Those in which hydrogen is ruled grating disperses it in the perpendic- mainly neutral are designated H I regions, ular direction. This device, often used both and those in which the gas is in molecu- in spacecraft and on the ground, allows lar form (H2) are referred to as molecular 134 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies clouds. The distinction is important Ultraviolet photons with energies of because of major differences in the radia- more than 11.2 electron volts can dissoci- tion that is present in the various regions ate molecular hydrogen (H2) into two H and consequently in the physical condi- atoms. In H I regions there are enough tions and processes that are important. of these photons to prevent the amount of

Radiation is a wave but is carried by H2 from becoming large, but the destruc- packets called photons. Each photon has tion of H2 as fast as it forms takes its toll a specified wavelength and precise on the number of photons of suitable energy that it carries, with gamma rays energies. Furthermore, interstellar dust is (short wavelengths) carrying the most a fairly efficient absorber of photons and X-rays, ultraviolet, optical, infrared, throughout the optical and ultraviolet microwave, and radio waves following in range. In some regions of space the order of decreasing energies (or increas- number of photons with energies higher ing wavelengths). Neutral hydrogen than 11.2 volts is reduced to the level atoms are extremely efficient at absorb- where H2 cannot be destroyed as fast as ing ionizing radiation—that is, an energy it is produced on grain surfaces. In this per photon of at least 13.6 electron volts case, H2 becomes the dominant form of (or, equivalently, a wavelength of less hydrogen present. The gas is then part than 0.0912 micrometre [3.59055118 x 10-6 of a molecular cloud. The role of inter- in]). If the hydrogen is mainly neutral, no stellar dust in this process is crucial radiation with energy above this thresh- because H2 cannot be formed efficiently old can penetrate except for photons with in the gas phase. energies in the X-ray range and above (thousands of electron volts or more), in Interstellar Dust which case the hydrogen becomes somewhat transparent. The absorption by Only about 0.7 percent of the mass of the neutral hydrogen abruptly reduces the interstellar medium is in the form of solid radiation field to almost zero for energies grains, but these grains have a profound above 13.6 electron volts. This dearth of effect on the physical conditions within hydrogen-ionizing radiation implies that the gas. Their main effect is to absorb no ions requiring more ionizing energy stellar radiation; for photons unable to than hydrogen can be produced, and the ionize hydrogen and for wavelengths out- ionic species of all elements are limited side absorption lines or bands, the dust to the lower stages of ionization. Within grains are much more opaque than the H II regions, with almost all the hydrogen gas. The dust absorption increases with ionized and thereby rendered nonabsorb- photon energy, so long-wavelength radia- ing, photons of all energies propagate, tion (radio and far-infrared) can penetrate and ions requiring energetic radiation for dust freely, near-infrared rather well, and their production (e.g., O++) occur. ultraviolet relatively poorly. Nebulae | 135

Dark, cold molecular clouds, within (1) grains, either free-flying or as compo- which all star formation takes place, owe nents of composite grains that also their existence to dust. Besides absorbing contain silicates, and (2) individual, freely starlight, the dust acts to heat the gas floating aromatic hydrocarbon molecules, under some conditions (by ejecting elec- with a range varying from 70 to several trons produced by the photoelectric hundred carbon atoms and some hydro- effect, following the absorption of a stel- gen atoms that dangle from the outer lar photon) and to cool the gas under edges of the molecule or are trapped in other conditions (because the dust can the middle of it. radiate energy more efficiently than the It is merely convention that these gas and so in general is colder). The larg- molecules are referred to as dust, since est chemical effect of dust is to provide the smallest may be only somewhat larger the only site of molecular hydrogen for- than the largest molecules observed mation on grain surfaces. It also removes with a radio telescope. Both of the dust some heavy elements (especially iron components are needed to explain spec- and silicon) that would act as coolants to troscopic features arising from the dust. the gas. The optical appearance of most In addition, there are probably mantles of nebulae is significantly modified by the hydrocarbon on the surfaces of the grains. obscuring effects of the dust. The size of the grains ranges from per- The chemical composition of the gas haps as small as 0.0003 micrometre phase of the interstellar medium alone, (1.18110236 x 10-8 in) for the tiniest hydro- without regard to the solid dust, can be carbon molecules to a substantial fraction determined from the strength of narrow of a micrometre; there are many more absorption lines that are produced by the small grains than large ones. gas in the spectra of background stars. The dust cannot be formed directly Comparison of the composition of the from purely gaseous material at the low gas with cosmic (solar) abundances densities found even in comparatively shows that almost all the iron, magne- dense interstellar clouds, which would be sium, and silicon, much of the carbon, considered an excellent laboratory vac- and only some of the oxygen and nitro- uum. For a solid to condense, the gas gen are contained in the dust. The density must be high enough to allow a absorption and scattering properties of few atoms to collide and stick together the dust reveal that the solid grains are long enough to radiate away their energy composed partially of silicaceous mate- to cool and form a solid. Grains are known rial similar to terrestrial rocks, though of to form in the outer atmospheres of cool an amorphous rather than crystalline supergiant stars, where the gas density is variety. The grains also have a carbona- comparatively high (perhaps 109 times ceous component. The carbon dust what it is in typical nebulae). The grains probably occurs in at least two forms: are then blown out of the stellar 136 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies atmosphere by radiation pressure (the the outer edge of a cold, quiescent dark mechanical force of the light they absorb molecular cloud and ionizes an H II and scatter). Calculations indicate that region in its vicinity. The pressure refracting materials, such as the constitu- strongly increases in the newly ionized ents of the grains proposed above, should zone, so the ionized gas flows out through condense in this way. the surrounding material. There are also There is clear indication that the dust expanding structures resembling bubbles is heavily modified within the interstellar surrounding stars that are ejecting their medium by interactions with itself and outer atmospheres into stellar winds. with the interstellar gas. The absorption and scattering properties of dust show Turbulence that there are many more smaller grains in the diffuse interstellar medium than in Besides these organized flows, nebulae of dense clouds. Apparently in the dense all types always show chaotic motions medium the small grains have coagulated called turbulence. This is a well-known into larger ones, thereby lowering the phenomenon in gas dynamics that results ability of the dust to absorb radiation when there is low viscosity in flowing with short wavelengths (namely, ultra- fluids, so the motions become chaotic violet, near 0.1 micrometre). The eddies that transfer kinetic and magnetic gas-phase abundances of some elements, energy and momentum from large scales such as iron, magnesium, and nickel, also down to small sizes. On small-enough are much lower in the dense regions scales viscosity always becomes impor- than in the diffuse gas, although even in tant, and the energy is converted into the diffuse gas most of these elements heat, which is kinetic energy on a molec- are missing from the gas and are there- ular scale. Turbulence in nebulae has fore condensed into dust. These profound, but poorly understood, effects systematic interactions of gas and dust on their energy balance and pressure show that dust grains collide with gas support. atoms much more rapidly than one would Turbulence is observed by means of expect if the dust and gas simply drifted the widths of the emission or absorption together. There must be disturbances, lines in a nebular spectrum. No line can probably magnetic in nature, that keep be precisely sharp in wavelength, because the dust and gas moving with respect to the energy levels of the atom or ion from each other. which it arises are not precisely sharp. The motions of gas within nebulae of Actual lines are usually much broader all types are clearly chaotic and compli- than this intrinsic width because of the cated. There are sometimes large-scale Doppler effect arising from motions of flows, such as when a hot star forms on the atoms along the line of sight. The Nebulae | 137 emission line of an atom is shifted to large scales that can undergo turbulent longer wavelengths if it is receding cascading to heat. from the observer and to shorter wavelengths if it is approaching. Part of the Galactic Magnetic Field observed broadening is easily explained by thermal motions, since v2, the aver- There is a pervasive magnetic field that aged squared speed, is proportional to threads the spiral arms of the Milky Way T/m, where T is the temperature and m is Galaxy and extends to thousands of light- the mass of the atom. Thus, hydrogen years above the galactic plane. The atoms move the fastest at any given evidence for the existence of this field temperature. comes from radio synchrotron emission Observations show that, in fact, produced by very energetic electrons hydrogen lines are broader than those of moving through it and from the polariza- other elements but not as much as tion of starlight that is produced by expected from thermal motions alone. elongated dust grains that tend to be Turbulence represents bulk motions, aligned with the magnetic field. The mag- independent of the mass of the atoms. netic field is very strongly coupled to the This chaotic motion of gas atoms of all gas because it acts upon the embedded masses would explain the observations. electrons, even the few in H I regions, and The physical question, though, is what the electrons impart some motion of the maintains the turbulence. Why do the other constituents by means of collisions. turbulent cascades not carry kinetic The gas and field are effectively confined energy from large-size scales into ever- to moving together, even though the gas shorter-size scales and finally into heat? can slip along the field freely. The field The answer is that energy is continu- has an important influence upon the tur- ously injected into the gases by a variety bulence because it exerts a pressure of processes. One involves strong stellar similar to gas pressure, thereby influenc- winds from hot stars, which are blown off ing the motions of the gas. The resulting at speeds of thousands of kilometres per complex interactions and wave motions second. Another arises from the violently have been studied in extensive numerical expanding remnants of supernova explo- calculations. sions, which sometimes start at 20,000 km (12,000 miles) per second and gradu- Molecular Clouds ally slow to typical cloud speeds (10 km [6 miles] per second). A third process is A molecular cloud, or a dark nebula, is an the occasional collision of clouds moving interstellar clump or cloud that is opaque in the overall galactic gravitational poten- because of its internal dust grains. The tial. All these processes inject energy on form of such dark clouds is very irregular: 138 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies they have no clearly defined outer bound- and the internal magnetic field provide aries and sometimes take on convoluted support against the clouds’ own gravity. serpentine shapes because of turbulence. The chemistry and physical condi- The largest molecular clouds are visible tions of the interior of a molecular cloud to the naked eye, appearing as dark are quite different from those of the patches against the brighter background surrounding low-density interstellar of the Milky Way Galaxy. An example is medium. In the outer parts of the dark the Coalsack in the southern sky. Stars cloud, hydrogen is neutral. Deeper within are born within molecular clouds. it, as dust blocks out an increasing amount of stellar ultraviolet radiation, Composition the cloud becomes darker and colder. Approaching the centre, the predominant The hydrogen of these opaque dark clouds form of gaseous carbon changes succes- + exists in the form of H2 molecules. The sively from C on the outside to neutral C largest nebulae of this type, the so-called (C0) and finally to the molecule carbon giant molecular clouds, are a million times monoxide (CO), which is so stable that it more massive than the Sun. They contain remains the major form of carbon in the much of the mass of the interstellar gas phase in the darkest regions. At great medium, are some 150 light-years across, depths within the cloud, other molecules and have an average density of 100 to 300 can be seen from their microwave transi- molecules per cubic centimetre (1 cubic tions, and more than 150 chemical species cm = .06 cubic in) and an internal temper- have been identified within the constitu- ature of only 7 to 15 K. Molecular clouds ent gas. consist mainly of gas and dust but contain Because of the comparatively low many stars as well. The central regions of densities and temperatures, the chemistry these clouds are completely hidden from is very exotic, as judged by terrestrial view by dust and would be undetectable experiments; some rather unstable species except for the far-infrared thermal emis- can exist in space because there is not sion from dust grains and the microwave enough energy to convert them to more- emissions from the constituent molecules. stable forms. An example is the near This radiation is not absorbed by dust and equality of the abundances of the inter- readily escapes the cloud. The material stellar molecule HNC (hydroisocyanic within the clouds is clumped together on acid) and its isomer HCN (hydrocyanic all size scales, with some clouds ranging acid); in ordinary terrestrial conditions down to the masses of individual stars. there is plenty of energy to allow the The density within the clumps may reach nitrogen and carbon atoms in HNC to 5 up to 10 H2 molecules per cubic centime- exchange positions and produce HCN, tre or more. Small clumps may extend by far the preferred species for equilib- about one light-year across. Turbulence rium chemistry. In the cold clouds, Nebulae | 139 however, not enough energy exists for the the protostar is a rotating disk larger exchange to occur. There is less than one- than the solar system that collapses into thousandth as much starlight within a “protoplanets” and comets. cloud as in the interstellar space outside These ideas are given encouraging the cloud, and the heating of the material confirmation by observations of molecu- in the cloud is provided primarily by lar clouds in very long wavelength infrared cosmic rays. Cooling within the cloud radiation. Some of the brightest infra- occurs chiefly by transitions between red sources are associated with such dark low-lying levels of the carbon monoxide dust clouds; a good example is the class molecule. of T Tauri variables, named for their pro- The emission lines from C+, C0, and totype star in the constellation Taurus. CO show that the edges of the molecular The T Tauri stars are known for a variety clouds are very convoluted spatially, with of reasons to be extremely young. The stellar ultraviolet radiation able to pene- variables are always found in or near trate surprisingly far throughout the cloud molecular clouds; they often are also despite the absorption of dust. Stellar powerful sources of infrared radiation, radiation can apparently enter the cloud corresponding to warm clouds of dust through channels where the dust (and gas) heated by the T Tauri star to a few hun- density is lower than average. The clump- dred kelvins. There are some strong iness of the interstellar material has infrared sources (especially in the con- profound effects on its properties. stellation of Orion) that have no visible stars with them; these are presumably Formation of Stars “cocoon stars” completely hidden by their veils of dust. In the inner regions of molecular clouds One of the remarkable features of an important event takes place: the for- molecular clouds is their concentration mation of stars from the gravitational in the spiral arms in the plane of the collapse of dense clumps within the neb- Milky Way Galaxy. While there is no defi- ula. Initially the cloud consists of a chaotic nite boundary to the arms, which have jumble of smaller clouds, each of which is irregularities and bifurcations, the nebu- destined to be an individual stellar sys- lae in other spiral galaxies are strung out tem. Each system has a rotary motion along these narrow lanes and form a arising from the original motions of the beautifully symmetric system when material that is falling into it. Because of viewed from another galaxy. The nebulae this spin, the collapsing cloud flattens as are remarkably close to the galactic plane; it shrinks. Eventually most of its mass is most are within 300 light-years, only 1 in a rotating condensation near its centre, percent of the Sun’s distance from the a “protostar” destined to become one or centre. The details of the explanation of more closely spaced stars. Surrounding why the gas is largely confined to the 140 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies spiral arms is beyond the scope of this their very existence to processes that book. Briefly, the higher density of the occur within stars. stars in the arms produces sufficient gravity to hold the gas to them. Hydrogen Clouds Why doesn’t the gas simply condense into stars and disappear? The present A different type of nebula is the hydro- rate of star formation is about one solar gen cloud, or the H I region; this region mass per year in the entire Galaxy, which is interstellar matter in which hydrogen is contains something like 2 × 109 solar mostly neutral, rather than ionized or masses of gas. Clearly, if the gas received molecular. Most of the matter between no return of material from stars, it would the stars in the Milky Way Galaxy, as well be depleted in roughly 2 × 109 years, about as in other spiral galaxies, occurs in the one-sixth the present age of the Galaxy. form of relatively cold neutral hydrogen There are several processes by which gas gas. Neutral hydrogen clouds are easily is returned to the interstellar medium. detectable at radio wavelengths because Possibly the most important is the ejec- they emit a characteristic energy at a tion of planetary nebula shells; other wavelength of 21 cm (8 in). processes are ejection of material from Neutral hydrogen is dominant in massive O- and B-type normal stars or clouds that have enough starlight to dis- from cool M giants and supergiants. The sociate molecular hydrogen into atoms rate of gas ejection is roughly equal to but lack hydrogen-ionizing photons from the rate of star formation, so that the mass hot stars. These clouds can be seen as of free gas is declining very slowly. (Some separate structures within the lower- gas is also falling into the Galaxy that has density interstellar medium or else on never been associated with any galaxy.) the outer edges of the molecular clouds. This cycling of gas through stars has Because a neutral cloud moves through had one major effect: the chemical com- space as a single entity, it often can be position of the gas has been changed by distinguished by the absorption line that the nuclear reactions inside the stars. its atoms or ions produce at their com- There is excellent evidence that the mon radial velocity in the spectrum of a Galaxy originally consisted of 77 percent background star. hydrogen by mass and that almost all of If neutral clouds at a typical pressure the rest of the constituent matter was were left alone until they could reach an helium. All heavy elements have been equilibrium state, they could exist at produced inside stars by being subjected either of two temperatures: “cold” (about to the exceedingly high temperatures 80 K) or “warm” (about 8,000 K), both and densities in the central regions. Thus, determined by the balance of heating and most of the atoms and molecules on cooling rates. There should be little mate- Earth, as well as in human bodies, owe rial in between. Observations show that Nebulae | 141 these cold and warm clouds do exist, but cluster is of this type. It was discovered in roughly half the material is in clouds at 1912 that the spectrum of this nebula intermediate temperatures, which implies mimics the absorption lines of the nearby that turbulence and collisions between stars, whereas bright nebulae that emit clouds can prevent the equilibrium states their own light show their own charac- from being reached. Cold H I regions are teristic emission lines. The brightest heated by electrons ejected from the dust reflection nebulae are illuminated by grains by interstellar ultraviolet radiation B-type stars that are very luminous but incident upon such a cloud from outside. have temperatures lower than about Cooling is mainly by C+ because passing 25,000 K, cooler than the O-type stars electrons or hydrogen atoms can excite it that would ionize the hydrogen in the gas from its normal energy state, the lowest, and produce an H II region. to one slightly higher, which is then fol- The extent and brightness of reflec- lowed by emission of radiation at 158 tion nebulae show conclusively that dust micrometres. This line is observed to be grains are excellent reflectors in the broad very strong in the spectrum of the Milky range of wavelengths extending from the Way Galaxy as a whole, which indicates ultraviolet (as determined from observa- that a great deal of energy is removed tions from space) through the visible. from interstellar gas by this process. Cold Optical observations suggest that about H I regions have densities of 10 to 100 60–70 percent of the light is reflected hydrogen atoms per cubic cm. Warm H I rather than absorbed, while the corre- regions are cooled by excitation of the sponding fraction for Earth is only 35 n = 2 level of hydrogen, which is at a much percent and for the Moon a mere 5 per- higher energy than the lowest level of C+ cent. Grains reflect light almost as well as and therefore requires a higher tempera- fresh snow, more because of their favour- ture for its excitation. The density of 0.5 able size (which promotes scattering atom per cubic cm (1 cubic cm = .06 in) is rather than absorption) than their chemi- much lower than in the colder regions. At cal composition. Calculations show that any particular density there is far more neu- even graphite, which is black in bulk, tral hydrogen available for cooling than C+. reflects visible light well when dispersed into small particles. Reflection Nebulae H II Region The reflection nebulae are interstellar clouds that would normally be dark nebu- Nebulae that are full of ionized hydrogen lae but whose dust reflects the light from atoms are H II regions. (These regions a nearby bright star that is not hot enough are also called diffuse nebulae or emission to ionize the cloud’s hydrogen. The nebulae.) The energy that is responsible famous nebulosity in the Pleiades star for ionizing and heating the hydrogen in 142 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies an emission nebula comes from a central powered by clusters of massive hot stars star that has a surface temperature in rather than by any single stellar body. A excess of 20,000 K. The density of these typical H II region within the Galaxy clouds normally ranges from 10 to 100,000 measures about 30 light-years in diame- particles per cubic cm (1 cubic cm = .06 ter and has an average density of about in); their temperature is about 8,000 K. 10 atoms per cubic cm. The mass of such Like molecular clouds, H II regions a cloud amounts to several hundred solar typically have little regular structure or masses. The only H II region visible to sharp boundaries. Their sizes and masses the naked eye is the beautiful Orion vary widely. There is even a faint region Nebula . It is located in the constellation of ionized gas around the Sun and other named for the Greek mythological comparatively cool stars, but it cannot be hunter and is seen as the central “star” in observed from nearby stars with existing Orion’s sword. The entire constellation is instruments. enveloped in faint emission nebulosity, The largest H II regions (none of powered by several stars in Orion’s belt which occur in the Milky Way Galaxy) rather than by the star exciting the much are 500 light-years across and contain at smaller Orion Nebula. The largest H II least 100,000 solar masses of ionized region in terms of angular size is the Gum gas. These enormous H II regions are Nebula , discovered by Australian astronomer Colin S. Gum. It measures 40° in angular diameter and is mainly ionized by two very hot stars (Zeta Puppis and Gamma Velorum). High-resolution studies of H II regions reveal one of the surprises that make the study of astrophysics delight- ful. Instead of the smooth structure that might be expected of a gas, a delicate tracery of luminous fi laments can be detected down to the smallest scale that can be resolved. In the Orion Nebula this is about 6 billion km (4 billion miles), or A plume of gas (lower right) in the Orion about the radius of the orbit of Pluto Nebula. A highly supersonic shock wave— around the Sun. Even fi ner details almost moving at a speed of more than 238,000 km surely exist, and there is evidence from (148,000 miles) per hour—was produced by a spectra that much of the matter may be beam of material emanating from a newly gathered into dense condensations, or formed star. National Aeronautics and Space Administration knots, the rest of the space being comparatively empty. Unrestrained gas would Nebulae | 143 fill a vacuum between the visible fila- is merely a conspicuous ionized region ments in about 200 years, an astronomical on the nearby face of a much larger dark instant. The nebular gas must be cloud; the H II region is almost entirely restrained from expansion by the pres- produced by the ionization provided by a sure of million-degree tenuous material single hot star, one of the four bright between the filaments. Its pressure, how- central stars (the Trapezium) identified ever, is comparable to that in the visible by Dutch astronomer Christiaan Huygens “warm” (8,000 K) gas of the H II region. in 1656. The shape of the Orion Nebula Hence, the density of the hot material is appears at visible wavelengths as irregu- several hundred times lower, which effec- lar. However, much of this seeming chaos tively prevents it from being observable is spurious, caused by obscuration of dust except in X-rays. The space throughout in dark foreground neutral material rather the plane of the Milky Way Galaxy is than by the actual distribution of ionized largely filled with this hot component, material. Radio waves can penetrate the which is mainly produced and heated by dust unhindered, and the radio emission supernovae. from the ionized gas reveals it to be quite In H II regions, hot gas also arises circular in shape and surprisingly sym- from the stellar winds of the exciting metrical as seen in projection on the sky. stars. These winds create a large cavity or The foreground dark material obscures bubble in the denser, cooler gas originally about half the ionized nebula. surrounding such a star. In the interior of An H II region on the outer edge of a the bubble, the radially flowing stellar large molecular cloud can induce star for- wind passes through a transition in which mation. For instance, behind the bright its radial motion is converted into heat. Orion Nebula, deeper within the dark The hot gas then fills most of the cavity cold Orion molecular cloud, new stars are (perhaps 90 percent or more) and serves being formed today. At present, none of to separate the filaments of the warm, the new stars is massive and hot enough comparatively dense H II region. Within to produce its own H II region, but pre- the condensations of visible plasma, there sumably some of them eventually will be. are neutral globules in which the gas is When an H II region is produced from quite cold (about 100 K) but is dense cold molecular gas by the formation of a enough (typically, 10,000 atoms per cubic hot star, the temperature is raised from cm) to have about the same pressure as roughly 25 to 8,000 K, and the number of the hot and warm materials. In short, an particles per cubic centimetre is almost

H II region is much more complicated quadrupled because each H2 molecule is than its visual radiation would suggest. split into two ions and two electrons. Gas H II regions are almost always accom- pressure is proportional to the product of panied by molecular clouds on their the temperature and number of particles borders. The Orion Nebula, for example, per cubic centimetre (regardless of their 144 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies mass, so electrons are as important as the The very dense, very young H II much heavier ions). Thus, the pressure in regions within molecular clouds are an H II region is some 800 times the pres- called “ultracompact” because of their sure of the cold gas from which it formed. small sizes and high densities. These The excess pressure causes a violent nebulae are observed only at the wave- expansion of the gas into the dense cloud. lengths of radio and far-infrared radiation, Rapid star formation may occur in the both of which are able to penetrate the compressed region, producing an expand- thick dust in the clouds. They are ing group of young stars. Such groups, extremely bright at wavelengths of 50 the so-called O Associations (with O micrometres (.002 in). There are about stars) or T Associations (with T Tauri 2,500 in the Milky Way Galaxy, represent- stars), have been observed. The compo- ing 10 to 20 percent of the total O-type nent stars simultaneously generate star population. Usually only a light- extremely fast outflows from their atmo- month in size, 100 times smaller than a spheres. These winds create regions of typical H II region, they show densities in hot, tenuous gas surrounding the associa- the ionized region of 105 hydrogen atoms tion. Eventually the massive stars in the per cubic cm. They cannot be at rest with association explode as supernovae, which respect to the surrounding gas; if they further disturb the surrounding gas. were, the immense pressure exerted by their dense hot gas would cause a violent Ultracompact H II Regions expansion. (Their lifetimes would be only about 3,000 years—exceedingly short on This picture of the evolution of H II an astronomical timescale—and not regions and molecular clouds is one of nearly as many could be seen as the num- constant turmoil, a few transient O stars ber observed by astronomers.) Rather, serving to keep the material stirred, in their gas is kept confined because they constant motion, continually producing are moving through the surrounding new stars and churning clouds of gas and cloud at speeds of about 10 km (6 miles) dust. In this way some of the stellar per second, and what is observed is the thermonuclear energy is converted into cloud of freshly ionized gas ahead of the kinetic energy of interstellar gas. This them that has not yet had time to expand. process begins just after the formation of The ultracompact H II regions leave the massive star that will power the behind a trail of ionized material that is mature H II region. The star begins pro- not as bright as the confined gas ahead of ducing copious amounts of ultraviolet them. This trail gradually fades as it radiation, converting the surrounding H2 recombines after the ionizing star has to atomic hydrogen, and then ionizing it to passed. The radio radiation is produced very dense high-pressure H+. by the ionized gas, but the far-infrared Nebulae | 145 radiation is emitted by the 5,000 solar provided by a single star. Indeed, there masses of surrounding dust warmed by are clumps of ionized gas ionized by tight the luminosity of the embedded star. groupings of single stars that are embedded in rather diff use material. These Supergiant Nebulae objects are more than 10 times as luminous as any in the Milky Way Galaxy and The most energetic H II regions within are about 200 light-years in diameter. If nearby galaxies have over 1,000 times they were located at the Orion Nebula, more ionizations per second than does they would cover the entire constellation the Orion Nebula, too many to be of Orion with brightly glowing gas. These

The inner part of the 30 Doradus Nebula, the most luminous nebula in the entire Local Group of galaxies, located in the Large Magellanic Cloud. National Optical Astronomy Observatories 146 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies supergiant nebulae are more than 10 carbon, have important ions that do not times as luminous as any in the Galaxy. show easily observable lines. Elaborate The entire Local Group—the cluster computer calculations predict the ioniza- of galaxies consisting of the Milky Way tion structure of gas ionized by a hot star Galaxy, the great spiral galaxy in AndroÂ whose temperature is determined by its meda, the smaller spiral in Triangulum, and spectrum. The calculations then provide more than 50 other stellar assemblages— predictions of the abundances of the contains but one supergiant nebula, the invisible ions, relative to the observable, object called 30 Doradus, in the Large and the total elemental abundance Magellanic Cloud. It contains a stellar follows. cluster called R136, the source of most of The main difficulty with this straight- the energy radiated by the nebula. This forward procedure is that there are two grouping consists of dozens of the most methods for determining the observed massive known stars of the Milky Way ionic abundances; each should be reli- Galaxy, all packed into a volume only a able, but they give quite different answers. thousandth of a typical stellar spacing in By far the easier method of determining size. How such a cluster could form is a ionic abundances is to observe the bright fascinating puzzle. There are other super- lines produced by collisions between the giant nebulae outside the Local Group, ion and energetic electrons. The bright- some of which radiate 10 times the energy est lines from this process in the entire of 30 Doradus. spectrum of H II regions arise from either O+ or O+2. These bright lines (and others Chemical Composition of such as N+) are the basis of the abundance H II Regions determinations in other galaxies. Alternatively, the ionic abundances The chemical composition of H II regions can be determined from the very faint (the numbers of atoms of each chemical emission lines that follow recombination, element, relative to hydrogen) can be the process by which the higher stage of estimated from nebular spectra. Each ele- ionization captures an electron (usually ment is found in more than one stage of at low energies) into a high level of the ionization, so the first step is to use the ion. Following recombination, there is a emission-line strengths of each stage of cascade from the high energy levels to ionization, relative to those of the hydro- the ground state, with photons in the gen lines, to obtain the abundance of that observed emission line being emitted at particular stage of ionization. All abun- each downward transition. These emis- dant elements have some stages of sion lines are fainter than the hydrogen ionization that produce observable emissions by roughly the ratio of abun- emission lines. On the other hand, some dances, which is more than 1,000, so only elements, such as argon, sulfur, and in bright nebulae can this method be Nebulae | 147 used. However, modern spectrographs on However, no such variations have been large telescopes have provided strengths convincingly detected. Some astronomers of these faint recombination lines for have proposed that there are chemical many objects, with good agreement inhomogeneities within nebulae that give among observers. Checks are provided rise to the differences between the abun- by comparison of several lines of the dances derived from recombination and same stage of ionization. The relative collisionally excited lines, but how such strengths of the observed lines agree well variations could be maintained is with the expectations of the cascading unexplained. process in the excited ions. Nevertheless, abundances are esti- The results are interesting and con- mated on the basis of a simple troversial. For H II regions that are bright interpretation of the mysterious postu- enough for the faint heavy-element lated temperature excursions, but no recombination lines to be measured, the better procedure has been suggested to carbon, nitrogen, neon, and oxygen abun- deal with the startling discrepancies in dances from recombination lines are the derived abundances. The other class uniformly about 1.8 times those from the of ionized nebulae, the planetary nebulae, collisionally excited lines. A common show the same effect. The local estimated interpretation is that there are strong abundances (say, for the Orion Nebula) temperature fluctuations within the nebu- are roughly solar. The abundances of lae. In the warmer regions the collisionally heavy elements per million hydrogen excited lines are strongly overproduced atoms are 500 for carbon, 80 for nitrogen, per heavy ion, so fewer heavy ions are 600 for oxygen, and 100 for neon. There needed to account for the observed line is a gradient of these abundances within strengths. The hydrogen lines are hardly the Milky Way Galaxy. At a distance half- affected by the postulated temperature way to the centre, 12,000 light-years inward, fluctuations. The temperature fluctuations, they are 50 percent larger than locally. which must be large (about 20 percent of Beyond Earth the gradient seems to per- the average), are unexplained. Turbulence sist, but there are very few observations. and magnetic fields are prime suspects. Helium is about 0.1 times as abundant as The nebular temperature can be esti- hydrogen, by number of atoms, through- mated directly from collisional lines out the Milky Way Galaxy. alone by comparing emission lines from Except for a few cases, compositions high-energy levels, populated by collisions, of nebulae in galaxies outside the Milky with lines from lower levels. This process Way Galaxy are measured by collision- can be carried out at various places on ally excited lines. The Large Magellanic the sky within each nebula, and large- Cloud has compositions that are uni- scale temperature fluctuations would formly about one-half those of the local appear as variations from place to place. Milky Way for oxygen, neon, argon, and 148 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies sulfur and are one-quarter for carbon and nitrogen. It appears that the fi rst group of elements must be manufactured together, presumably in massive stars, and ejected together into the interstellar gas. Stars of a diff erent (probably lower) mass must produce carbon and nitrogen. Planetary nebulae suggest the same scenario. The abundance of helium in nebulae has received considerable attention because the helium content of the oldest objects provides clues to the origin of the universe. The value cited above for the Orion Nebula is in agreement with the pre- Ring Nebula (M57, NGC 6720) in the constel- dictions of the big-bang model . lation Lyra, a planetary nebula consisting mainly of gases thrown o by the star in the Planetary Nebulae centre. Hale Observatories © 1959

Some nebulae, called planetary nebulae, are expanding shells of luminous gas Forms and Structure expelled by dying stars. Observed tele- scopically, they have a relatively round Compared with diff use nebulae, planetary compact appearance rather than the nebulae are small objects, having a radius chaotic patchy shapes of other nebulae— typically of 1 light-year and containing a hence their name, which was given mass of gas of about 0.3 solar mass. One because of their resemblance to planetary of the largest-known planetary nebulae, disks when viewed with the instruments the Helix Nebula (NGC 7293) in the con- of the late 1700s, when the fi rst planetary stellation Aquarius, subtends an angle of nebulae were discovered. about 20 minutes of arc—two-thirds the There are believed to be about 20,000 angular size of the Moon. Planetary nebu- objects called planetary nebulae in the lae are considerably denser than most H Milky Way Galaxy, each representing gas II regions, typically containing 1,000– expelled relatively recently from a central 10,000 atoms per cubic cm (1 cubic cm = star very late in its evolution. Because of .06 in) within their dense regions, and the obscuration of dust in the Galaxy, have a surface brightness 1,000 times only about 1,800 planetary nebulae have larger. Many are so far away that they been cataloged. Planetary nebulae are appear stellar when photographed important sources of the gas in the inter- directly, but the conspicuous examples stellar medium. have an angular size up to 20 minutes of Nebulae | 149 arc across, with 10–30 seconds of arc being relatively cool (25,000 K) to some of the usual. Those that show a bright disk have hottest known (200,000 K). much more-regular forms than the chaotic In the nebulae with hot stars, most H II regions, but there are still usually of the helium is doubly ionized, and appre- some brightness fluctuations over the disk. ciable amounts of five-times-ionized The planetaries generally have regular, oxygen and argon and four-times-ionized sharp outer boundaries; often they have a neon exist. In H II regions helium is relatively regular inner boundary as well, mainly once ionized and neon and argon giving them the appearance of a ring. only once or twice. This difference in Many have two lobes of bright material, the states of the atoms results from the resembling arcs of a circle, connected by a temperature of the planetary nucleus (up bridge, somewhat resembling the letter Z. to about 150,000 K), which is much higher Most planetaries show a central star, than that of the exciting star of the H II called the nucleus, which provides the regions (less than 60,000 K for an O star, ultraviolet radiation required for ionizing the hottest). High stages of ionization are the gas in the ring or shell surrounding it. found close to the central star. The rare Those stars are among the hottest known heavy ions, rather than hydrogen, absorb and are in a state of comparatively rapid the photons of several hundred electron evolution. volt energies. Beyond a certain distance As with H II regions, the overall from the central star, all the photons of structural regularity conceals large-scale energy sufficient to ionize a given species fluctuations in density, temperature, of ion have been absorbed, and that spe- and chemical composition. High- cies therefore cannot exist farther out. resolution images of a planetary nebula Detailed theoretical calculations have usually reveal tiny knots and filaments rather successfully predicted the spectra down to the resolution limit. The spec- of the best-observed nebulae. trum of the planetary nebula is basically The spectra of planetary nebulae reveal the same as that of the H II region; it another interesting fact: they are expand- contains bright lines from hydrogen ing from the central star at 24–56 km and helium recombinations and the (15–35 miles) per second. The gravitational bright, collisionally excited forbidden pull of the star is quite small at the dis- lines and faint recombination lines of tance of the shell from the star, so the other ions. (Recombination is the pro- shell will continue its expansion until it cess by which the higher stage of finally merges with the interstellar gas ionization captures an electron [usually around it. The expansion is proportional at low energies] into a high level of the to the distance from the central star, con- ion.) The central stars show a much sistent with the entire mass of gas having greater range of temperatures than been ejected at one brief period from the those in H II regions, ranging from star in some sort of instability. 150 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

The Distances of 30,000 years, after which the nebula is so Planetary Nebulae tenuous that it cannot be distinguished from the surrounding interstellar gas. Estimating the distance to any particular This lifetime is much shorter than the planetary nebula is challenging because lifetimes of the parent stars, so the nebu- of the variety of shapes and masses of the lar phase is relatively brief. ionized gas. There is uncertainty about the amount of ionizing radiation from the Chemical Composition central star that escapes from the nebula and the amount of hot low-density mate- Planetary nebulae are chemically rial that fills part of the volume but does enriched in elements produced by nuclear not emit appreciable radiation. Thus, processing within the central star. Some planetary nebulae are not a homogeneous are carbon-rich, with twice as much car- class of objects. bon as oxygen, while there is more oxygen Distances are estimated by obtaining than carbon in the Sun. Others are over- measurements for about 40 objects that abundant in nitrogen; the most luminous happen to have especially favourable ones, observed in external galaxies, are properties. The favourable properties conspicuous examples. Helium is mod- involve association with other objects estly enhanced in many. There are objects whose distance can be estimated inde- that contain almost no hydrogen; it is as pendently, such as membership in a if the gas had been ejected from these stellar cluster or association with a star of object at the very end of the nuclear- known properties. Statistical methods, burning process. Planetary nebulae also calibrated by these objects, provide rough show a clear indication of the general estimates (about 30 percent errors) of heavy-element abundance gradient in distances for all others. The statistical the Galaxy, presumably a reflection of the method involves assuming that all shells original composition of the stars that have similar masses when all of the shell gave rise to the present nebulae. is ionized and correcting for the fraction As in the case of H II regions, plane- that is neutral for the rest. tary nebulae show discrepancies between From the best available distance the determinations of abundances of determination, the true size of any nebula heavy elements from faint recombination can be found from its angular size. lines as opposed to those determined Typically, planetary nebulae are a few from collisionally excited lines, but in a tenths of a light-year in radius. If this dis- much more severe form. There are some tance is divided by the expansion speed, nebulae for which the two methods give the age of the nebula since ejection is the same abundances. However, the most obtained. Values range up to roughly extreme discrepancies are factors of 30 or Nebulae | 151 more in the oxygen abundances. Perhaps Positions in the Galaxy this wild variation in planetary nebulae is not surprising, since they surely have One of the best indicators of the average regions of material that are strongly age of astronomical objects is their position enriched in heavy elements and deficient and motion in the Galaxy. The youngest in hydrogen. These regions originate in are in the spiral arms, near the gas from the complicated nuclear processing of the which they have formed; the oldest are expelled material ejected from the evolved not concentrated in the plane of the central star. They would have strong cool- Galaxy, nor are they found within the spi- ing from the heavy-element emissions ral arms. By these criteria, the planetaries and thus much lower temperatures than reveal themselves to be rather middle- the regions of normal composition. These aged; they are moderately but not strongly regions would contribute very little to the concentrated in the plane; rather, they are hydrogen emission lines because they concentrated toward the galactic centre, are hydrogen-poor. as the older objects are. Their motions in Some, but not all, planetary nebulae the Galaxy follow elliptical paths, whereas contain internal dust. In general, this dust circular orbits are characteristic of cannot be seen directly but can be younger stars. They belong to the type of detected from the infrared radiation it distribution often called a “disk popula- emits after being heated by nebular and tion,” to distinguish them from the stellar radiation. The presence of dust Population II (very old) and Population I implies that planetary nebulae are even (young) objects proposed by the German richer in heavy elements than gas-phase American astronomer Walter Baade. abundance studies suggest. There is a wide variation in the ages of Among nebulae so far discovered, planetaries, and some are very young two are particularly deviant in chemical objects. composition: one is in the globular cluster M15 and the other in the halo (tenuous Evolution of outer regions) of the Galaxy. Both have Planetary Nebulae very low heavy-element content (down from normal by factors of about 50) but A description of the evolution of a plane- normal helium. Both objects are very old, tary nebula begins before the ejection of suggesting that the primeval gas in the the nebula itself. As will be discussed Galaxy had a low heavy-element content below, the central star is a red giant before but an almost normal amount of helium. the ejection. In such a phase it experi- The origin of most helium in the Galaxy ences a rapid loss of mass, up to 0.01 was the big bang, the initial explosion of Earth mass per day, in the form of a com- the universe itself. paratively slowly expanding stellar wind. 152 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

At this stage the red giant might be heav- the gas is mingled with the general inter- ily obscured by dust that forms from the stellar gas. A curious feature of several heavy elements in the wind. Eventually planetaries is that faint rings surround- the nature of both the star and its wind ing the bright inner nebula can be changes. The star becomes hotter because observed; they are the remnants of a pre- its hot core is exposed by the loss of the vious shell ejected earlier by the star. overlying atmosphere. The inner gas is ionized by radiation from the hot star. Central Stars The ionization zone moves steadily outward through the slowly moving material Many central stars are known from their of what was formerly the stellar wind. The spectra to be very hot. A common type of expansion speed of the gas is typically 30 spectrum has very broad emission lines km (19 miles) per second. Nebulae in this of carbon or nitrogen, as well as of ion- stage are bright but have starlike images ized helium, superimposed upon a bluish as seen from Earth, because they are too continuum. These spectra are indistin- small to show a disk. The gas is at a rela- guishable from those from the very bright tively high density—about one million rare stars known as Wolf-Rayet stars, but atoms per cubic centimetre (1 cubic cm = the planetary nuclei are about 100 times .06 cubic in)—but becomes more dilute as fainter than true Wolf-Rayet objects. The the gas expands. During this stage the stars appear to be losing some mass at the nebula is surrounded by neutral hydro- present time, though evidently not enough gen. It appears to expand faster than the to contribute appreciably to the shell. individual atoms of gas in it are moving; The presence of the nebula allows a the ionized shell is “eating into” the neu- fairly precise determination of the central tral material as the density falls. star’s evolution. The temperature of the The middle stage of evolution occurs star can be estimated from the nebula when the density has dropped to the from the amounts of emission of ionized point at which the entire mass of gas is helium and hydrogen by a method ionized. After this stage is reached, some devised by the Dutch astronomer H. of the ultraviolet radiation escapes into Zanstra. The amount of ionized-helium space, and the expansion of the nebula is radiation is determined by the number of caused entirely by the motion of the gas. photons with energy of more than 54 Most planetaries are now in this middle electron volts, while hydrogen is ionized stage. Finally the central star becomes by photons in excess of 13.6 electron less luminous and can no longer provide volts. The relative numbers of photons enough ultraviolet radiation to keep even in the two groups depend strongly on the dilute nebula ionized. Once again the temperature, since the spectrum shifts outer regions of the nebula become neu- dramatically to higher energies as the tral and therefore invisible. Eventually temperature of the star increases. Hence, Nebulae | 153 the temperature can be found from the stars. It then cools and after about 10,000 observed strengths of the hydrogen and years becomes a very dense white dwarf helium lines. The rate of evolution of the star, scarcely larger than Earth but with a stars can be determined from the sizes of density of thousands of kilograms per their nebulae, as the time since ejection cubic centimetre (1 cubic cm = .06 cubic of the shell is the radius of the nebula in). From this point it cools very slowly, divided by the expansion rate. The energy becoming redder and fainter indefinitely. output, or luminosity, of the central star While there is not yet a very detailed can be estimated from the brightness of theoretical picture of this contraction, a the nebula, because the nebula is con- few results have emerged rather clearly: verting the star’s invisible ultraviolet (1) white dwarf stars must obtain nearly radiation (which contains the greater part all their energy from the contraction of the star’s luminous energy) into visible noted above, not from nuclear sources; radiation. therefore, (2) they must contain practically The resulting theoretical description no hydrogen or helium, except perhaps in of the star’s evolution is quite interesting. a very thin shell on their surfaces. These While there seem to be real differences in conditions would have to be met for the stars at a given stage, the trends are quite evolution to take place so quickly. clear. The central stars in young planetary The absence of hydrogen in the star’s nebulae are about as hot as the massive interior is quite surprising; the planetary O and B stars—35,000–40,000 K—but nebulae are all found to have a normal roughly 10 times fainter. They have half hydrogen abundance of about 1,000 times the diameter of the Sun but are 1,000 as many hydrogen atoms as heavy ele- times as luminous. As the nebula expands, ments, such as oxygen. Thus, it can be the star increases its brightness and tem- concluded that the mechanism of expul- perature, but its radius decreases steadily. sion of the envelope must be very efficient It reaches a maximum energy output at ejecting the hydrogen-rich outer layers when it is roughly 10,000 times as lumi- of a star while leaving heavy-element- nous as the Sun, about 5,000 years after rich material behind. the initial expansion. This is a very small fraction of the star’s age of several billion The Nature of the years; it represents a period equivalent to Progenitor Stars about half an hour in a human life. From this point on, the star becomes fainter, The progenitor must have mass not much but for some time the temperature con- in excess of a solar mass because of the tinues to increase while the shrinkage of distribution of the planetaries in the GalÂ the star continues. At its hottest the star axy. Very massive stars are young and is perhaps 200,000 K, almost five times more closely confined than are nebulae hotter than the hottest of most of the to the galactic plane. Also, the mass of the 154 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies nebula is roughly 0.3 solar mass, and the later phase occurs when convection car- mass of a typical white dwarf (the final ries carbon, the product of helium state of the central star) is roughly 0.7 burning, to the surface. solar mass. Next, the expansion velocity of the nebula is probably comparable to Supernova Remnants the velocity of escape from its progenitor, which implies that the progenitor was a Like the planetary nebulae, supernova red giant star, large and cool, completely remnants are clouds left behind by a unlike the small, hot, blue, nuclear star dying star. In this case, a star goes super- remaining after the ejection. Likely can- nova, a spectacular explosion in which a didates are members of the class of star ejects most of its mass in a violently long-period variable stars, which have expanding cloud of debris. At the bright- about the right size and mass and are est phase of the explosion, the expanding known to be unstable. Symbiotic stars cloud radiates as much energy in a single (i.e., stars with characteristics of both day as the Sun has done in the past three cool giants and very hot stars) also are million years. Such explosions occur candidates. Novae, stars that brighten roughly every 50 years within a large temporarily while ejecting a shell explo- galaxy. They have been observed less sively, are definitely not candidates; the frequently in the Milky Way Galaxy nova shell is expanding at hundreds of because most of them have been hidden kilometres per second. by the obscuring clouds of dust. Galactic The cause of the ejection is the out- supernovae were observed in 1006 in ward force of radiation on the outer layers Lupus, in 1054 in Taurus, in 1572 in CasÂ of red giant stars. The ejection is trig- siopeia (Tycho’s nova, named after Tycho gered by a rapid variation in the nuclear Brahe, its observer), and finally in 1604 in luminosity in the interior of the giant, Serpens, called Kepler’s nova. The stars caused by instability in the helium-burn- became bright enough to be visible in the ing shell. The ejection takes place during daytime. more than one phase of the giant’s evolu- The only naked-eye supernova to tion. Nitrogen-rich nebulae develop occur since 1604 was Supernova 1987A in during an early episode when convection the Large Magellanic Cloud (the galaxy inside the star carries nitrogen, produced nearest to the Milky Way system), visible from carbon in a series of nuclear reac- only from the Southern Hemisphere. On tions (i.e., the carbon-nitrogen cycle of Feb. 23, 1987, a blue supergiant star bright- hydrogen burning), to the surface. A later ened to gradually become third magnitude, ejection takes place with an enrichment easily visible at night, and it has subse- of both nitrogen and helium, which also quently been followed in every wavelength is produced by hydrogen burning. A still band available to scientists. The spectrum Nebulae | 155

within the molecular cloud in which it formed, the expanding remnant might compress the surrounding interstellar gas and trigger subsequent star formation. The remnants contain strong shock waves that create fi laments of material emitting gamma-ray photons with energies up to 10 14 electron volts and accelerating electrons and atomic nuclei up to cosmic-ray energies, from 10 9 up to 10 15 electron volts per particle. In the solar neighbourhood, A small part of the Cygnus Loop supernova remnant, these cosmic rays carry about which marks the edge of an expanding blast wave from an enormous stellar explosion that occurred about as much energy per cubic 10,000 years ago. The remnant is located in the con- metre as starlight in the plane stellation Cygnus, the Swan. National Aeronautics and of the galaxy, and they carry it Space Administration to thousands of light-years above the plane. Much of the radiation from showed hydrogen lines expanding at supernova remnants is synchrotron radi- 12,000 km per second (7,456 miles per ation, which is produced by electrons second), followed by a long period of slow spiraling in a magnetic fi eld at almost the decline. There are 270 known supernova speed of light. This radiation is dramati- remnants, almost all observed by their cally diff erent from the emission from strong radio emission, which can pene- electrons moving at low speeds: it is (1) trate the obscuring dust in the galaxy. strongly concentrated in the forward Supernova remnants are very impor- direction, (2) spread out over a broad tant to the structure of galaxies. They range of frequencies, with the average are a major source of heating of inter- frequency increasing with the electron’s stellar gas by means of the magnetic energy, and (3) highly polarized. Electrons turbulence and violent shocks that they of many diff erent energies produce produce. They are the main source of radiation at essentially all wavelengths, most heavy elements, from oxygen on from radio through infrared, optical, and up. If the exploding massive star is still ultraviolet up to X- and gamma rays. 156 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

About 50 supernova remnants con- that is comparable to or greater than its tain pulsars, the spinning neutron star own; the expansion has by then slowed remnants of the former massive star. The substantially. The dense material, mostly name comes from the exceedingly regu- interstellar at its outer edge, radiates larly pulsed radiation that propagates away its remaining energy for hundreds into space in a narrow beam that sweeps of thousands of years. The final phase is past the observer similarly to the beam reached when the pressure within the from a lighthouse. There are several rea- supernova remnant becomes comparable sons why most supernova remnants do to the pressure of the interstellar medium not contain visible pulsars. Perhaps the outside the remnant, so the remnant loses original pulsar was ejected because its distinct identity. In the later stages of there was a recoil from an asymmetrical expansion, the magnetic field of the gal- explosion, or the supernova formed a axy is important in determining the black hole instead of a pulsar, or the beam motions of the weakly expanding gas. of the rotating pulsar does not sweep past Even after the bulk of the material has the solar system. merged with the local interstellar medium, Supernova remnants evolve through there might be remaining regions of very four stages as they expand. At first, they hot gas that produce soft X-rays (i.e., expand so violently that they simply those of a few hundred electron volts) sweep all older interstellar material observable locally. before them, acting as if they were The recent galactic supernovae expanding into a vacuum. The shocked observed are in the first phases of the gas, heated to millions of kelvins by the evolution suggested above. At the sites of explosion, does not radiate its energy Kepler’s and Tycho’s novae, there exist very well and is readily visible only in heavy obscuring clouds, and the optical X-rays. This stage typically lasts several objects remaining are now inconspicu- hundred years, after which time the shell ous knots of glowing gas. Near Tycho’s has a radius of about 10 light-years. As nova, in Cassiopeia, there are similar the expansion occurs, little energy is lost, optically insignificant wisps that appear but the temperature falls because the to be remnants of yet another supernova same energy is spread into an ever-larger explosion. To a radio telescope, however, volume. The lower temperature favours the situation is spectacularly different: the more emission, and during the second Cassiopeia remnant is the strongest radio phase the supernova remnant radiates its source in the entire sky. Study of this energy at the outermost, coolest layers. remnant, called Cassiopeia A, reveals that This phase can last thousands of years. a supernova explosion occurred there in The third stage occurs after the shell has approximately 1680, missed by observers swept up a mass of interstellar material because of the obscuring dust. Nebulae | 157

The Crab Nebula bluish amorphous inner region of the Crab Nebula is radiating synchrotron At the site of the 1054 supernova is one of radiation , and the spectrum extends up to the most remarkable objects in the sky, gamma-ray energies. The Crab is the the Crab Nebula, now about 10 light-years second brightest X-ray source in the sky, across. Photographed in colour, it is after Scorpius X-1 (an X-ray binary star). revealed as a beautiful red lacy network After almost 1,000 years, the nebula is of long and sinuous glowing hydrogen still losing 100,000 times as much energy fi laments surrounding a bluish structure- per second as the Sun. less region whose light is strongly On the basis of this huge outpouring polarized. The fi laments emit the spec- of energy, it is easy to calculate how long trum characteristic of a diff use nebula. the nebula can shine without a new sup- The gas is expanding at 1,100 km (700 ply of energy. The electrons emitting the miles) per second—slower than the X-rays should decay, or drop to lower 10,000–20,000 km per second in the shells energies, in about 30 years—far less than of new supernovae in other galaxies. The the age of the nebula. The source of energy of the electrons that emit the X-rays was discovered in 1969 to be a pulsar, which has been found to fl ash optically, as well as at radio wavelengths, blinking on and off with a period of 0.033 second. This period is slowly increasing (at the rate of 0.0012 second per century), which implies that the pulsar is slowing down and thereby losing its energy to the nebula. The corresponding rate of energy loss is about equal to the nebula’s rate of energy loss, convincing evidence that a The Crab Nebula (M1, NGC 1952) in the constellation tiny, extremely dense pulsar Taurus is a gaseous remnant of the galactic supernova can supply the energy to the of 1054 CE. The nebula, 6,500 light-years away, is nebula. The Crab Nebula is expanding at 1,100 km (700 miles) per second. Hale Observatories © 1959 unique in being a young supernova remnant and relatively 158 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies free from obscuration, while Tycho’s and small clouds of denser material; many Kepler’s supernovae are conspicuous lines of reasoning from other evidence radio sources, radiating by synchrotron lead to the same result. The present speed emission; in neither case has a detectable of the fi laments is about 100 km per second pulsar been found. (62 miles per second); the approximate age of the Cygnus Loop is 10,000 years. The Cygnus Loop Diffuse ionized gas The best-observed old supernova remnant is the Cygnus Loop (or the Veil A major component of the interstellar Nebula ), a beautiful fi lamentary object medium, or the warm ionized medium roughly in the form of a circular arc in (WIM), is the diff use ionized gas, dilute Cygnus. Its patchiness is striking: the interstellar material that makes up about loop consists of a series of wisps rather 90 percent of the ionized gas in the Milky than a continuous cloud of gas. The most Way Galaxy. It produces a faint emission- likely interpretation of this patchiness is line spectrum that is seen in every that the interstellar medium into which direction. It was fi rst detected from a thin the shock wave is propagating contains haze of electrons that aff ect radio radiation passing through the Milky Way Galaxy. Similar layers are now seen in many other galaxies. The American astronomer Ronald Reynolds and his col- laborators have mapped ionized hydrogen and a few other ions (N+ , S+ , and O++ ). The total power required for the ionization is amazingly large: about 15 percent of the luminosity of all O and B stars. This energy output is about equal to the total power provided by supernovae, but the latter radiate most of their energy either in non- ionizing radiation or in providing kinetic energies to their expanding shells. Other potential energy sources fall far short. Unlike H II regions, the diff use ion- Veil Nebula (NGC 6992) in the constella- ized gas is found far from the galactic tion Cygnus, which glows as it collides plane as well as close to it. Pulsars (spin- with dust and gas in interstellar space. ning neutron stars emitting pulsed radio Palomar Observatory; photograph © Cal- ifornia Institute of Technology 1959 waves) occasionally reside at large distances from the plane and emit radio Nebulae | 159 waves. The electrons in the diff use ion- NOTABLE NEBuLAE ized gas slow these waves slightly in a manner that depends on the frequency, Nearly all nebulae are beautiful objects allowing observers to determine the when seen up close, through a tele- number of electrons per square metre (1 scope. Some of the most notable square metre = 10.8 square feet) on the nebulae also are fascinating subjects path to the pulsar. These observations when considered in depth, with attention show that the diff use ionized gas extends to detail. more than 3,000 light-years above and below the galactic plane, which is much Cassiopeia A farther than the 300-light-year thickness of distributions of molecular clouds, H II Cassiopeia A is the strongest source of regions, and O and B stars. radio emission in the sky beyond the On average, the densities of the elec- solar system, located in the direction of trons are only about 0.05 per cubic cm (a the constellation Cassiopeia about 9,000 fi fth of the average density in the galactic light-years from Earth. Cassiopeia A, plane), and only 10 to 20 percent of the abbreviated Cas A, is the remnant of a volume is occupied by gas even at this supernova explosion caused by the col- low density. The rest of the volume can be lapse of a massive star. The light from the fi lled by very hot, even lower density gas or by magnetic pressure. In the diff use ionized gas, the comparatively low stages of ionization of the common elements (O + , N+ , and S+ ) are much more abundant relative to higher stages (O ++ , N ++ , and S ++ ) than in typical diff use nebulae. Such an eff ect is caused by the extremely low density of the diff use ionized gas; in this case, even hot stars fail to produce high stages of ionization. Thus, it seems possible to explain the peculiar ionization of the diffuse ionized gas with ionization powered by O and B stars, which are mostly found in the plane of the Milky Way Galaxy. Apparently the stars are able to ionize Cassiopeia A supernova remnant, in a composite image synthesized from observations gathered in passages through the clouds enveloping di erent spectral regions by three space-based them so that a substantial part of the ion- observatories. NASA/JPL/California Institute of izing radiation can escape into the regions Technology far from the galactic plane. 160 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies event is estimated to have reached Earth through about one-third of their extent. about 1667, which makes Cas A the From the constellation Cygnus, the rift youngest known supernova remnant in reaches through Aquila and Sagittarius, the Milky Way Galaxy. Although the where the centre of the Galaxy lies hid- explosion must have been very powerful, den behind it, to Centaurus. The clouds no contemporary record exists of its hav- of dark material making up the Great Rift ing been observed, so the explosion may are several thousand light-years from have happened behind an interstellar the Earth. dust cloud. Today the remnant is also weakly observable at visible, infrared, Gum Nebula and X-ray wavelengths, and it appears as an expanding ring of material approxi- The largest known emission nebula in mately five arc minutes in diameter. The terms of angular diameter as seen from expansion rate of the remnant has been Earth is the Gum Nebula, which extends used to estimate how long ago the explo- about 35° in the southern constellations sion occurred. Puppis and Vela. A complex of diffuse, glowing gas too faint to be seen with Coalsack the unaided eye, it was discovered by the Australian-born astrophysicist Colin S. The Coalsack is a dark nebula in the Crux Gum, who published his findings in 1955. constellation (Southern Cross). Easily The Gum Nebula lies roughly 1,000 light- visible against a starry background, it is years from Earth and is about 1,000 perhaps the most conspicuous dark neb- light-years in diameter. It may be the ula. Starlight coming to Earth through it remnant of an ancient supernova—i.e., is reduced by 1 to 1.5 magnitudes. The violently exploding star. Coalsack is about 500 light-years from Earth and 50 light-years in diameter. It fig- Horsehead Nebula ures in legends of peoples of the Southern Hemisphere and has been known to EuroÂ The Horsehead Nebula (catalog number peans since about 1500. The Northern IC 434) is an H II region in the constella- Coalsack, in the constellation Cygnus, is tion Orion. The nebula consists of a cloud similar in nature and appearance but of ionized gas lit from within by young, hot somewhat less prominent. stars; a dark cloud containing interstellar dust lies immediately in front. The dust Great Rift absorbs the light from part of the ionized cloud. A portion of this dark cloud has a The Great Rift is a complex of dark nebu- shape somewhat resembling a horse’s lae that seems to divide the bright clouds head. The nebula is located 400 parsecs of the Milky Way Galaxy lengthwise (1,300 light-years) from the Sun. It has a Nebulae | 161 diameter of approximately 4 parsecs (13 light-years) and a total mass of about 250 solar masses.

Lagoon Nebula

The Lagoon Nebula (catalog numbers NGC 6523 and M8) is an H II region located in the constellation Sagittarius at 1,250 parsecs (4,080 light-years) from the solar system. The nebula is a cloud of interstellar gas and dust approximately 10 parsecs (33 light-years) in diameter. A group of young, hot stars in the cloud ionize the nearby gas. As the atoms in the gas recom- Lagoon Nebula (M8, NGC 6523) in the constellation bine, they produce the light Sagittarius. This bright di use nebula is so large that light emitted by the nebula. from the stars involved does not penetrate its boundaries, Interstellar dust within the and the bright nebula appears to be seen against a larger, darker one. Palomar Observatory; photograph © California nebula absorbs some of Institute of Technology 1961 this light and appears almost to divide the nebula, thus producing a lagoonlike shape. America, with the dusty region being shaped like the Gulf of Mexico. The North North American Nebula American Nebula is approximately 520 parsecs (1,700 light-years) from the Sun. The North American Nebula (NGC 7000) It has a diameter of about 30 parsecs (100 is an ionized-hydrogen region in the con- light-years) and a total mass equal to stellation Cygnus . The nebula is a cloud about 4,000 solar masses. of interstellar gas ionized from within by young, hot stars. Interstellar dust particles Orion Nebula in part of this cloud absorb the light emitted by recombining atoms. The shape of The bright diff use Orion Nebula is faintly the nebula roughly resembles that of North visible to the unaided eye in the sword of 162 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies the hunter’s figure in the constellation erratically, reflecting or re-radiating energy Orion. The nebula lies about 1,350 light- from the star. years from Earth and contains hundreds of very hot (O-type) young stars clustered Ring Nebula about a nexus of four massive stars known as the Trapezium. Radiation from these The Ring Nebula (catalog numbers NGC stars excites the nebula to glow. It was 6720 and M57) is a bright planetary neb- discovered in 1610 by the French scholar ula in the constellation Lyra, about 2,300 Nicolas-Claude Fabri de Peiresc and light-years from the Earth. It was discov- independently in 1618 by the Swiss ered in 1779 by the French astronomer astronomer Johann Cysat. It was the first Augustin Darquier. Like other nebulae of nebula to be photographed (1880), by its type, it is a sphere of glowing gas Henry Draper in the United States. thrown off by a central star. Seen from a Images of the nebula continued to great distance, such a sphere appears improve, and technological advances in brighter at the edge than at the centre the late 1980s enabled scientists to pho- and thus takes on the appearance of a tograph infrared-emitting objects in the luminous ring. It is a popular object for Orion Nebula that had never before been amateur astronomers. observed optically. The Hubble Space Telescope in 1991 revealed the sharpest Trifid Nebula details yet available of known features of the nebula, including what appeared to be The Trifid Nebula (catalog numbers a jet (an energetic outflow) related to the NGC 6514 and M 20) is a bright, diffuse birth of a young star. nebula in the constellation Sagittarius, lying several thousand light-years from R Monocerotis the Earth. It was discovered by the French astronomer Legentil de La R Monocerotis (NGC 2261) is a stellar Galaisière before 1750 and named by the infrared source and nebula in the constel- English astronomer Sir John Herschel lation Monoceros (Greek: Unicorn). The for the three dark rifts that seem to divide star, one of the class of dwarf stars called the nebula and join at its centre. Of T Tauri variables, is immersed in a cloud about the ninth magnitude optically, of matter that changes in brightness the Trifid is also a radio source. CHAPTER 6

Galaxies

he Milky Way Galaxy is just one of many galaxies, the Tsystems of stars and interstellar matter that make up the universe. Many galaxies are so enormous that they contain hundreds of billions of stars. Nature has provided an immensely varied array of galaxies, ranging from faint, diff use dwarf objects to brilliant spiral-shaped giants. Virtually all galaxies appear to have been formed soon after the universe began, and they pervade space, even into the depths of the farthest reaches penetrated by powerful modern telescopes. Galaxies usually exist in clusters, some of which in turn are grouped into larger clusters that measure hundreds of millions of light-years across. These so-called superclusters are separated by nearly empty voids, and this causes the gross structure of the universe to look somewhat like a network of sheets and chains of galaxies. Galaxies diff er from one another in shape, with variations resulting from the way in which the systems were formed and subsequently evolved. Galaxies are extremely varied not only in structure but also in the amount of activity observed. Some are the sites of vigorous star formation, with its attendant glowing gas and clouds of dust and molecular complexes. Others, by contrast, are quiescent, having long ago ceased to form new stars. Perhaps the most conspicuous activity in galaxies occurs in their nuclei, where evidence suggests that 164 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

The Whirlpool Galaxy (left), also known as M51, an Sc galaxy accompanied by a small, irregular companion galaxy, NGC 5195 (right). NASA, ESA, S. Beckwith (STScI), and The Hubble Heritage Team (STScI/AURA) in many cases supermassive objects— then, however, galaxies have become one probably black holes—lurk. These central of the focal points of astronomical inves- black holes apparently formed several tigation. The notable developments and billion years ago; they are now observed achievements in the study of galaxies are forming in galaxies at large distances surveyed here. Included in the discussion (and, therefore, because of the time it are the external galaxies (i.e., those lying takes light to travel to Earth, at times in outside the Milky Way Galaxy, the local the far distant past) as brilliant objects galaxy to which the Sun and Earth called quasars. belong), their distribution in clusters and The existence of galaxies was not rec- superclusters, and the evolution of galax- ognized until the early 20th century. Since ies and quasars. Galaxies | 165

The evolution of galaxies ionized by radiation, structure apparently had already been established in the form The study of the origin and evolution of of density fluctuations. At a crucial point galaxies and the quasar phenomenon has in time, there condensed from the expand- only just begun. Many models of galaxy ing matter small clouds (protogalaxies) formation and evolution have been con- that could collapse under their own gravi- structed on the basis of what we know tational field eventually to form galaxies. about conditions in the early universe, For the latter half of the 20th century, which is in turn based on models of the there were two competing models of gal- expansion of the universe after the big axy formation: “top-down” and “bottom-up.” bang (the primordial explosion from In the top-down model, galaxies formed which the universe is thought to have out of the collapse of much larger gas originated) and on the characteristics of clouds. In the bottom-up model, galaxies the cosmic microwave background (the formed from the merger of smaller enti- observed photons that show us the light- ties that were the size of globular clusters. filled universe as it was when it was a few In both models the angular momentum hundred thousand years old). of the original clouds determined the According to the big-bang model, the form of the galaxy that eventually evolved. universe expanded rapidly from a highly It is thought that a protogalaxy with a compressed primordial state, which large amount of angular momentum resulted in a significant decrease in den- tended to form a flat, rapidly rotating sys- sity and temperature. Soon afterward, the tem (a spiral galaxy), whereas one with dominance of matter over antimatter (as very little angular momentum developed observed today) may have been estab- into a more nearly spherical system (an lished by processes that also predict elliptical galaxy). proton decay. During this stage many The transition from the 20th to the types of elementary particles may have 21st century coincided with a dramatic been present. After a few seconds, the transition in our understanding of the universe cooled enough to allow the for- evolution of galaxies. It is no longer mation of certain nuclei. The theory believed that galaxies have evolved predicts that definite amounts of hydro- smoothly and alone. Indeed, it has become gen, helium, and lithium were produced. clear that collisions between galaxies Their abundances agree with what is have occurred all during their evolution— observed today. About one million years and these collisions, far from being rare later the universe was sufficiently cool for events, were the mechanism by which atoms to form. galaxies developed in the distant past When the universe had expanded to and are the means by which they are be cool enough for matter to remain in changing their structure and appearance neutral atoms without being instantly even now. Evidence for this new 166 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies understanding of galactic evolution “eaten” by the giant central galaxy, as comes primarily from two sources: more well as clouds of M31 stars ejected by the detailed studies of nearby galaxies with strong tidal forces of the collision. new, more sensitive instruments and deep More spectacular are galaxies pres- surveys of extremely distant galaxies, ently in the process of collision and seen when the universe was young. accretion in the more distant, but still Recent surveys of nearby galaxies, nearby, universe. The symptoms of the including the Milky Way Galaxy, have collision are the distortion of the galax- shown evidence of past collisions and ies’ shape (especially that of the spiral capture of galaxies. For the Milky Way arms), the formation of giant arcs of stars the most conspicuous example is the by tidal action, and the enhanced rate Sagittarius Galaxy, which has been of star and star cluster formation. Some of absorbed by our Galaxy. Now its stars lie the most massive and luminous young spread out across the sky, its seven star clusters observed anywhere lie in the globular clusters intermingling with the regions where two galaxies have come globular clusters of the Milky Way Galaxy. together, with their gas and dust clouds Long tails of stars around the Milky Way colliding and merging in a spectacular were formed by the encounter and act as cosmic fireworks display. clues to the geometry of the event. A sec- A second type of evidence for the fact ond remnant galaxy, known as the Canis that galaxies grow by merging comes Major Dwarf Galaxy, can also be traced from very deep surveys of the very dis- by the detection of star streams in the tant universe, especially those carried out outer parts of our Galaxy. These galaxies with the Hubble Space Telescope (HST). support the idea that the Milky Way These surveys, especially the Hubble Galaxy is a mix of pieces, formed by the Deep Field and the Hubble Ultra Deep amalgamation of many smaller galaxies. Field, found galaxies so far away that the The Andromeda Galaxy (M31) also light observed by the HST left them when has a past involving collisions and accre- they were very young, only a few hundred tion. Its peculiar close companion, M32, million years old. This enables the direct shows a structure that indicates that it detection and measurement of young was formerly a normal, more massive gal- galaxies as they were when the universe axy that lost much of its outer parts and was young. The result is a view of a very possibly all of its globular clusters to M31 different universe of galaxies. Instead of in a past encounter. Deep surveys of the giant elliptical galaxies and grand spirals, outer parts of the Andromeda Galaxy the universe in its early years was popu- have revealed huge coherent structures lated with small, irregular objects that of star streams and clouds, with proper- looked like mere fragments. These were ties indicating that these include the the building blocks that eventually outer remnants of smaller galaxies formed bigger galaxies such as the Milky Galaxies | 167

Way. Many show active formation of stars especially successful in mimicking gal- that are deficient in heavy elements axy collisions and in helping to explain because many of the heavy elements had the presence of various tidal arms and not yet been created when these stars jets observed by astronomers. were formed. In summary, the current view of The rate of star formation in these galactic history is that present-day galax- early times was significant, but it did not ies are a mix of giant objects that accreted reach a peak until about one billion years lesser galaxies in their vicinities, espe- later. Galaxies from this time show a max- cially early in the formation of the imum in the amount of excited hydrogen, universe, together with some remnant which indicates a high rate of star forma- lesser, or dwarf, galaxies that have not yet tion, as young, very hot stars are necessary come close enough to a more massive for exciting interstellar hydrogen so that galaxy to be captured. The expansion of it can be detected. Since that time, so the universe gradually decreases the like- much matter has been locked up in stars lihood of such captures, so some of the (especially white dwarfs) that not enough dwarfs may survive to old age—eventu- interstellar dust and gas are available ally dying, like their giant cousins, when to achieve such high rates of star all of their stars become dim white dwarfs formation. or black holes and slowly disappear. An important development that has helped our understanding of the way gal- Historical survey of the axies form is the great success of computer study of galaxies simulations. High-speed calculations of the gravitational history of assemblages Most nebulae look as amorphous and of stars, interstellar matter, and dark ephemeral as clouds in the sky. However, matter suggest that after the big bang the in 1845, the Irish astronomer William universe developed as a networklike Parsons discovered that some nebulae arrangement of material, with gradual had a spiral shape. Why did these objects condensation of masses where the strands have such a well-ordered appearance? of the network intersected. In simulations The dispute over the nature of what of this process, massive galaxies form, were once termed spiral nebulae stands but each is surrounded by a hundred or as one of the most significant in the devel- so smaller objects. The small objects may opment of astronomy. On this dispute correspond to the dwarf galaxies, such as hinged the question of the magnitude of those that surround the Milky Way Galaxy the universe: were we confined to a single, but of which only a dozen or so remain, limited stellar system that lay embedded the rest having presumably been accreted alone in empty space, or was our Milky by the main galaxy. Such computer Way Galaxy just one of millions of galax- models, called “n-body simulations,” are ies that pervaded space, stretching 168 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies beyond the vast distances probed by our seen from northern latitudes.) Moreover, most powerful telescopes? How this ques- the irregular shapes of the objects and tion arose, and how it was resolved, is an their numerous hot blue stars, star clusters, important element in the development of and gas clouds did indeed make them our prevailing view of the universe. resemble the southern Milky Way Galaxy. Up until 1925, spiral nebulae and their The American astronomer Harlow related forms had uncertain status. Some Shapley, noted for his far-reaching work scientists, notably Heber D. Curtis of the on the size and structure of the Milky Way United States and Knut Lundmark of Galaxy, was one of the first to appreciate Sweden, argued that they might be the importance of the Magellanic Clouds remote aggregates of stars similar in size in terms of the nature of spiral nebulae. to the Milky Way Galaxy. Centuries ear- To gauge the distance of the Clouds, he lier the German philosopher Immanuel made use of the period-luminosity (P-L) Kant, among others, had suggested much relation discovered by Henrietta Leavitt the same idea, but that was long before of the Harvard College Observatory. In the tools were available to actually mea- 1912 Leavitt had found that there was a sure distances and thus prove it. During close correlation between the periods of the early 1920s astronomers were divided. pulsation (variations in light) and the Although some deduced that spiral neb- luminosities (intrinsic, or absolute, bright- ulae were actually extragalactic star nesses) of a class of stars called Cepheid systems, there was evidence that con- variables in the Small Magellanic Cloud. vinced many that such nebulae were local Leavitt’s discovery, however, was of little clouds of material, possibly new solar practical value until Shapley worked out a systems in the process of forming. calibration of the absolute brightnesses of pulsating stars closely analogous to the The Problem of the Cepheids, the so-called RR Lyrae variables. Magellanic Clouds With this quantified form of the P-L relation, he was able to calculate the distances It is now known that the nearest external to the Magellanic Clouds, determining galaxies are the Magellanic Clouds, two that they were about 75,000 light-years patchy irregular objects visible in the from Earth. The significance of the skies of the Southern Hemisphere. For Clouds, however, continued to elude sci- years, most experts who regarded the MagÂ entists of the time. For them, these objects ellanic Clouds as portions of the Milky still seemed to be anomalous, irregular Way Galaxy system separated from the patches of the Milky Way Galaxy, farther main stream could not study them away than initially thought but not suffi- because of their position. (Both cient to settle the question of the nature Magellanic Clouds are too far south to be of the universe. Galaxies | 169

Novae in the S Andromeda. If they were ordinary novae, Andromeda Nebula then M31 must be millions of light-years away, but then the nature of S Andromeda An unfortunate misidentification ham- became a difficult question. At this vast pered the early recognition of the distance its total luminosity would have northern sky’s brightest nearby galaxy, to be immense—an incomprehensible out- the Andromeda Nebula, also known as put of energy for a single star. M31. In 1885 a bright star, previously Completion of the 254-cm (100-inch) invisible, appeared near the centre of telescope on Mount Wilson in 1917 M31, becoming almost bright enough to resulted in a new series of photographs be seen without a telescope. As it slowly that captured even fainter objects. More faded again, astronomers decided that it novae were found in M31, mainly by must be a nova, a “new star,” similar to Milton L. Humason, who was an assistant the class of temporary stars found rela- at the time to Edwin P. Hubble, one of the tively frequently in populous parts of the truly outstanding astronomers of the day. Milky Way Galaxy. If this was the case, it Hubble eventually studied 63 of these was argued, then its extraordinary bright- stars, and his findings proved to be one of ness must indicate that M31 cannot be the final solutions to the controversy. very far away, certainly not outside the local system of stars. Designated S The Scale of the Andromeda in conformity with the pat- Milky Way Galaxy tern of terminology applied to stars of variable brightness, this supposed nova At the same time that spiral nebulae were was a strong argument in favour of the being studied and debated, the Milky Way hypothesis that nebulae are nearby Galaxy became the subject of contentious objects in the Milky Way Galaxy. discussion. During the early years of the By 1910, however, there was evidence 20th century, most astronomers believed that S Andromeda might have been that the Milky Way Galaxy was a disk- wrongly identified. Deep photographs shaped system of stars with the Sun near were being taken of M31 with the Mount the centre and with the edge along a thick Wilson Observatory’s newly completed axis only about 15,000 light-years away. 152-cm (60-inch) telescope, and the This view was based on statistical evi- astronomers at the observatory, especially dence involving star counts and the spatial J.C. Duncan and George W. Ritchey, were distribution of a variety of cosmic objects— finding faint objects, just resolved by the open star clusters, variable stars, binary longest exposures, that also seemed to systems, and clouds of interstellar gas. All behave like novae. These objects, how- these objects seemed to thin out at dis- ever, were about 10,000 times fainter than tances of several thousand light-years. 170 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

This conception of the Milky Way variables of this type in any given glob- Galaxy was challenged by Shapley in ular cluster have the same apparent 1917, when he released the findings of his brightness. If all RR Lyrae variables have study of globular clusters. He had found the same intrinsic brightness, then it that these spherically symmetrical groups follows that differences in apparent of densely packed stars, as compared brightness must be due to different dis- with the much closer open clusters, were tances from Earth. The final step in unusual in their distribution. While the developing a procedure for determining known open clusters are concentrated the distances of variables was to calcu- heavily in the bright belt of the Milky late the distances of a handful of such Way Galaxy, the globular clusters are for stars by an independent method so as to the most part absent from those areas enable calibration. Shapley could not except in the general direction of the con- make use of the trigonometric parallax stellation Sagittarius, where there is a method, since there are no variables close concentration of faint globular clusters. enough for direct distance measurement. Shapley’s plot of the spatial distribution However, he had recourse to a technique of these stellar groupings clarified this devised by the Danish astronomer Ejnar peculiar fact: the centre of the globular Hertzsprung that could determine dis- cluster system—a huge almost spherical tances to certain nearby field variables cloud of clusters—lies in that direction, (i.e., those not associated with any partic- some 30,000 light-years from the Sun. ular cluster) by using measurements of Shapley assumed that this centre must their proper motions and the radial veloc- also be the centre of the Milky Way ity of the Sun. Accurate measurements Galaxy. The globular clusters, he argued, of the proper motions of the variables form a giant skeleton around the disk of based on long-term observations were the Milky Way Galaxy, and the system is available, and the Sun’s radial velocity thus immensely larger than was previ- could be readily determined spectro- ously thought, its total extent measuring scopically. Thus, by availing himself of nearly 100,000 light-years. this body of data and adopting Shapley succeeded in making the Hertzsprung’s method, Shapley was able first reliable determination of the size of to obtain a distance scale for Cepheids in the Milky Way Galaxy largely by using the solar neighbourhood. Cepheids and RR Lyrae stars as distance Shapley applied the zero point of the indicators. His approach was based on Cepheid distance scale to the globular the P-L relation discovered by Leavitt and clusters he had studied with the 152-cm on the assumption that all these variables (60-inch) telescope at Mount Wilson. have the same P-L relation. As he saw it, Some of these clusters contained RR this assumption was most likely true in Lyrae variables, and for these Shapley the case of the RR Lyrae stars, because all could calculate distances in Galaxies | 171 a straightforward manner from the P-L that, if the Milky Way Galaxy was so relation. For other globular clusters he immense, then the spiral nebulae must made distance determinations, using a lie within it. His conviction was rein- relationship that he discovered between forced by two lines of evidence. One of the brightnesses of the RR Lyrae stars these has already been mentioned—the and the brightness of the brightest red nova S Andromeda was so bright as to stars. For still others he made use of suggest that the Andromeda Nebula most apparent diameters, which he found to be certainly was only a few hundred light- relatively uniform for clusters of known years away. The second came about distance. The final result was a catalog of because of a very curious error made by distances for 69 globular clusters, from one of Shapley’s colleagues at Mount which Shapley deduced his revolutionary Wilson Observatory, Adrian van Maanen. model of the Milky Way Galaxy—one that not only significantly extended the limits The van Maanen Rotation of the galactic system but that also displaced the Sun from its centre to a location During the early 20th century, one of the nearer its edge. most important branches of astronomy Shapley’s work caused astronomers was astrometry, the precise measure- to ask themselves certain questions: How ments of stellar positions and motions. could the existing stellar data be so Van Maanen was one of the leading wrong? Why couldn’t they see something experts in this field. Most of his determi- in Sagittarius, the proposed galactic cen- nations of stellar positions were accurate tre, 30,000 light-years away? The reason and have stood the test of time, but he for the incorrectness of the star count made one serious and still poorly under- methods was not learned until 1930, when stood error when he pursued a problem Lick Observatory astronomer Robert J. tangential to his main interests. In a Trumpler, while studying open clusters, series of papers published in the early discovered that interstellar dust pervades 1920s, van Maanen reported on his dis- the plane of the Milky Way Galaxy and covery and measurement of the rotation obscures objects beyond only a few thou- of spiral nebulae. Using early plates taken sand light-years. This dust thus renders by others at the 152-cm (60-inch) Mount the centre of the system invisible opti- Wilson telescope as well as more recent cally and makes it appear that globular ones taken about 10 years later, van clusters and spiral nebulae avoid the Maanen measured the positions of sev- band of the Milky Way. eral knotlike, nearly stellar images in the Shapley’s belief in the tremendous spiral arms of some of the largest-known size of the local galactic system helped to spiral nebulae (e.g., M33, M101, and M51). put him on the wrong side of the argu- Comparing the positions, he found dis- ment about other galaxies. He thought tinct changes indicative of a rotation of 172 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies the spiral pattern against the background subconsciously, and this subtle effect of surrounding field stars. In each case, manifested itself in prejudicing the deli- the rotation occurs in the sense that the cate measurements, or (2) possibly the spiral arms trail. The periods of rotation first set of plates was the problem. Many were all approximately 100,000 years. of these plates had been taken in an Angular motions were about 0.02 second unconventional manner by Ritchey, who of arc per year. swung the plate holder out of the field Shapley seized the van Maanen whenever the quality of the images was results as evidence that the spirals had to temporarily poor because of atmospheric be nearby; otherwise, their true space turbulence. The resulting plates appeared velocities of rotation would have to be excellent, having been exposed only impossibly large. For example, if M51 is during times of very fine seeing; however, rotating at an apparent rate of 0.02 sec- according to some interpretations, the ond of arc per year, its true velocity would images had a slight asymmetry that led be immense if it is a distant galaxy. to a very small displacement of star Assuming that a distance of 10,000,000 images compared with nonstellar images. light-years would lead to an implausibly Such an error could look like rotation if large rotation velocity of 12,000 km/sec not recognized for what it really was. In (7,456 miles/sec), Shapley argued that, if a any case, the van Maanen rotation was more reasonable velocity was adopted— accepted by many astronomers, includ- say, 100 km/sec (62 miles/sec)—then the ing Shapley, and temporarily sidetracked distances would all be less than 100,000 progress toward recognizing the truth light-years, which would put all the spirals about galaxies. well within the Milky Way Galaxy. It is unclear just why such a crucial The Shapley-Curtis Debate measurement went wrong. Van Maanen repeated the measures and obtained the The nature of galaxies and scale of the same answer even after Hubble demon- universe were the subject of the Great strated the truth about the distances to Debate, a public program arranged in the spirals. However, subsequent work- 1920 by the National Academy of Sciences ers, using the same plates, failed to find at the Smithsonian Institution in any rotation. Among the various hypoth- Washington, D.C. Featured were talks by eses that science historians have proposed Shapley and the aforementioned Heber as an explanation for the error are two Curtis, who were recognized as spokes- particularly reasonable ideas: (1) possibly men for opposite views on the nature of the fact that spiral nebulae look like they spiral nebulae and the Milky Way Galaxy. are rotating (i.e., they resemble familiar This so-called debate has often been cited rotational patterns that are perceivable in as an illustration of how revolutionary nature) may have influenced the observer new concepts are assimilated by science. Galaxies | 173

It is sometimes compared to the debate, willing to concede that there might be centuries before, over the motions of the two classes of novae, yet, because he con- Earth (the Copernican revolution); how- sidered the Milky Way Galaxy to be small, ever, though as a focal point the debate he underestimated their differences. The about Earth’s motion can be used to van Maanen rotation also entered into define the modern controversy, the Shapley’s arguments: if spiral nebulae Shapley-Curtis debate actually was much were rotating so fast, they must be within more complicated. the Milky Way Galaxy as he conceived it. A careful reading of the documents For Curtis, however, the matter provided involved suggests that, on the broader less of a problem: even if spiral nebulae topic of the scale of the universe, both did rotate as rapidly as claimed, the small men were making incorrect conclusions scale of Curtis’s universe allowed them to but for the same reasons—namely, for have physically reasonable speeds. being unable to accept and comprehend The Shapley-Curtis debate took place the incredibly large scale of things. near the end of the era of the single- Shapley correctly argued for an enormous galaxy universe. In just a few years the Milky Way Galaxy on the basis of the P-L scientific world became convinced that relation and the globular clusters, while Shapley’s grand scale of the Milky Way Curtis incorrectly rejected these lines of Galaxy was correct and at the same time evidence, advocating instead a small that Curtis was right about the nature of galactic system. Given a Milky Way spiral nebulae. Such objects indeed lie Galaxy system of limited scale, Curtis even outside Shapley’s enormous Milky could argue for and consider plausible Way Galaxy, and they range far beyond the extragalactic nature of the spiral neb- the distances that in 1920 seemed too vast ulae. Shapley, on the other hand, for many astronomers to comprehend. incorrectly rejected the island universe theory of the spirals (i.e., the hypothesis Hubble’s Discovery of that there existed comparable galaxies Extragalactic Objects beyond the boundaries of the Milky Way Galaxy) because he felt that such objects During the early 1920s Hubble detected would surely be engulfed by the local 15 stars in the small, irregular cloudlike galactic system. Furthermore, he put object NGC 6822 that varied in luminosity, aside the apparent faint novae in M31, and he suspected that they might include preferring to interpret S Andromeda as Cepheids. After considerable effort, he an ordinary nova, for otherwise that object determined that 11 of them were in fact would have been unbelievably luminous. Cepheid variables, with properties indis- Unfortunately for him, such phenomena— tinguishable from those of normal called supernovae—do in fact exist, as Cepheids in the Milky Way Galaxy and in was realized a few years later. Curtis was the Magellanic Clouds. Their periods 174 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies ranged from 12 to 64 days, and they were M31 at a meeting in 1924, he did not com- all very faint, much fainter than their plete his research and publish the results Magellanic counterparts. Nevertheless, for this conspicuous spiral galaxy until they fit a P-L relation of the same nature five years later. as had been discovered by Leavitt. While the Cepheids made it possible Hubble then boldly assumed that the to determine the distance and nature of P-L relation was universal and derived an NGC 6822, some of its other features estimate for the distance to NGC 6822, corroborated the conclusion that it was a using Shapley’s most recent (1925) ver- separate, distant galaxy. Hubble discov- sion of the calibration of the relation. This ered within it five diffuse nebulae, which calibration was wrong, as is now known, are glowing gaseous clouds composed because of the confusion at that time mostly of ionized hydrogen, designated over the nature of Cepheids. Shapley’s H II regions. (H stands for hydrogen and calibration included certain Cepheids in II indicates that most of it is ionized; H I, globular clusters that subsequent investi- by contrast, signifies neutral hydrogen.) gators found to have their own fainter P-L He found that these five H II regions had relation. (Such Cepheids have been des- spectra like those of gas clouds in the ignated Type II Cepheids to distinguish Milky Way Galaxy system—e.g., the Orion them from the normal variety, which are Nebula and Eta Carinae. Calculating their referred to as Type I.) Thus, Hubble’s dis- diameters, Hubble ascertained that the tance for NGC 6822 was too small: he sizes of the diffuse nebulae were normal, calculated a distance of only 700,000 similar to those of local examples of giant light-years. Today it is recognized that H II regions. the actual distance is closer to 2,000,000 Five other diffuse objects discerned light-years. In any case, this vast distance— by Hubble were definitely not gaseous even though underestimated—was large nebulae. He compared them with globular enough to convince Hubble that NGC clusters (both in the Milky Way Galaxy 6822 must be a remote, separate galaxy, and in the Magellanic Clouds) and con- much too far away to be included even in cluded that they were too small and faint Shapley’s version of the Milky Way Galaxy to be normal globular clusters. Convinced system. Technically, then, this faint neb- that they were most likely distant galaxies ula can be considered the first recognized seen through NGC 6822, he dismissed external galaxy. The Magellanic Clouds them from further consideration. Modern continued to be regarded simply as studies suggest that Hubble was too hasty. appendages to the Milky Way Galaxy, Though probably not true giant globular and the other bright nebulae, M31 and clusters, these objects are in all likelihood M33, were still being studied at the Mount star clusters in the system, fainter, smaller Wilson Observatory. Although Hubble in population, and probably somewhat announced his discovery of Cepheids in younger than normal globular clusters. Galaxies | 175

The Dutch astronomer Jacobus Because M31 is much larger than the Cornelius Kapteyn showed in the early field of view of the 152- and 254-cm (60- 20th century that statistical techniques and 100-inch) telescopes at Mount could be used to determine the stellar Wilson, Hubble concentrated on four luminosity function for the solar neigh- regions, centred on the nucleus and at bourhood. (The luminosity function is a various distances along the major axis. curve that shows how many stars there The total area studied amounted to less are in a given volume for each different than half the galaxy’s size, and the other stellar luminosity.) Eager to test the unexplored regions remained largely nature of NGC 6822, Hubble counted unknown for 50 years. (Modern compre- stars in the galaxy to various brightness hensive optical studies of M31 have been limits and found a luminosity function conducted only since about 1980.) for its brightest stars. When he compared Hubble pointed out an important and it with Kapteyn’s, the agreement was puzzling feature of the resolvability of excellent—another indication that the M31. Its central regions, including the Cepheids had given about the right dis- nucleus and diffuse nuclear bulge, were tance and that the basic properties of not well resolved into stars, one reason galaxies were fairly uniform. Step by step, that the true nature of M31 had previously Hubble and his contemporaries piled up been elusive. However, the outer parts evidence for the fundamental assumption along the spiral arms in particular were that has since guided the astronomy of resolved into swarms of faint stars, seen the extragalactic universe, the uniformity superimposed over a structured back- of nature. By its bold application, astron- ground of light. Current understanding of omers have moved from a limiting this fact is that spiral galaxies typically one-galaxy universe to an immense vast- have central bulges made up exclusively ness of space populated by billions of of very old stars, the brightest of which are galaxies, all grander in size and design too faint to be visible on Hubble’s plates. than the Milky Way Galaxy system was Not until 1944 did the German-born once thought to be. astronomer Walter Baade finally resolve the bulge of M31. Using red-sensitive The Distance to the plates and very long exposures, he man- Andromeda Nebula aged to detect the brightest red giants of this old population. Out in the arms there In 1929 Hubble published his epochal exist many young, bright, hot blue stars, paper on M31, the great Andromeda and these are easily resolved. The bright- Nebula. Based on 350 photographic est are so luminous that they can be seen plates taken at Mount Wilson, his study even with moderate-sized telescopes. provided evidence that M31 is a giant The most important of Hubble’s dis- stellar system like the Milky Way Galaxy. coveries was that of M31’s population of 176 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

Cepheid variables. Forty of the 50 vari- measured for the inner parts of M31 by ables detected turned out to be ordinary spectrographic work, he calculated (on Cepheids with periods ranging from 10 the basis of the distance derived from the to 48 days. A clear relation was found Cepheids) that M31’s mass must be about between their periods and luminosities, 3.5 billion times that of the Sun. Today and the slope of the relation agreed with astronomers have much better data, those for the Magellanic Clouds and which indicate that the galaxy’s true total NGC 6822. Hubble’s comparison indi- mass must be at least 100 times greater cated that M31 must be 8.5 times more than Hubble’s value, but even that value distant than the Small Magellanic Cloud clearly showed that M31 is an immense (SMC), which would imply a distance of system of stars. Furthermore, Hubble’s two million light-years if the modern estimates of star densities demonstrated SMC distance was used (the 1929 value that the stars in the outer arm areas of employed by Hubble was about two times M31 are spread out with about the same too small). Clearly, M31 must be a distant, density as in the Milky Way Galaxy sys- large galaxy. tem in the vicinity of the Sun. Other features announced in Hubble’s paper were M31’s population of bright, The Golden Age of irregular, slowly varying variables. One Extragalactic Astronomy of the irregulars was exceedingly bright; it is among the most luminous stars in Until about 1950, scientific knowledge of the galaxy and is a prototype of a class of galaxies advanced slowly. Only a very high-luminosity stars now called Hubble- small number of astronomers took up Sandage variables, which are found in galaxy studies, and only a very few tele- many giant galaxies. Eighty-five novae, scopes were suitable for significant all behaving very much like those in the research. It was an exclusive field, rather Milky Way Galaxy, were also analyzed. jealously guarded by its practitioners, Hubble estimated that the true occur- and so progress was orderly but limited. rence rate of novae in M31 must be about During the decade of the 1950s, the 30 per year, a figure that was later con- field began to change. Ever-larger optical firmed by the American astronomer telescopes became available, and the Halton C. Arp in a systematic search. space program resulted in a sizable Hubble found numerous star clusters increase in the number of astronomers in M31, especially globular clusters, 140 emerging from universities. New instru- of which he eventually cataloged. He mentation enabled investigators to explore clinched the argument that M31 was a galaxies in entirely new ways, making it galaxy similar to the Milky Way Galaxy possible to detect their radio, infrared, by calculating its mass and mass density. and ultraviolet emissions and eventually Using the velocities that had been even radiation at X-ray and gamma-ray Galaxies | 177 wavelengths. Whereas in the 1950s there optical appearance of galaxy images on was only one telescope larger than 254 photographic plates, galaxies are divided cm (100 inches)—and only about 10 into three general classes: ellipticals, spi- astronomers conducting research on rals, and irregulars. Hubble subdivided galaxies worldwide—by the year 2000 the these three classes into finer groups. number of large telescopes had grown In The Hubble Atlas of Galaxies immensely, with 12 telescopes larger than (1961), the American astronomer Allan R. 800 cm (300 inches), and the number of Sandage drew on Hubble’s notes and his scientists devoted to galaxy study was in own research on galaxy morphology to the thousands. By then, galaxies were revise the Hubble classification scheme. being extensively studied with giant Some of the features of this revised arrays of ground-based radio telescopes, scheme are subject to argument because Earth-orbiting optical, X-ray, ultraviolet, of the findings of very recent research, and infrared telescopes, and high-speed but its general features, especially the computers—studies that have given rise coding of types, remain viable. A descrip- to remarkable advances in knowledge and tion of the classes as defined by Sandage understanding. The tremendous progress is given here, along with observations in both theoretical and observational concerning needed refinements of some work has led many to say that the turn of of the details. the 21st century happened during the “golden age” of extragalactic astronomy. Elliptical Galaxies

Types of galaxies These systems exhibit certain characteristic properties. They have complete All galaxies are not spirals like our own rotational symmetry; i.e., they are figures Milky Way. Some are elliptical systems. of revolution with two equal principal Others, the irregulars, have no definite axes. They have a third smaller axis that shape at all. Astronomers have devised is the presumed axis of rotation. The sur- systems of classification that encompass face brightness of ellipticals at optical all these types. wavelengths decreases monotonically outward from a maximum value at the Principal Schemes of centre, following a common mathemati- Classification cal law of the form:

−2 Almost all current systems of galaxy I = I0( r/a +1 ) , classification are outgrowths of the initial scheme proposed by the American where I is the intensity of the light, I0 is astronomer Edwin Hubble in 1926. In the central intensity, r is the radius, and a Hubble’s scheme, which is based on the is a scale factor. The isophotal contours 178 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies exhibited by an elliptical system are simi- regions than in the outer parts. The major lar ellipses with a common orientation, axes sometimes do not line up either; each centred on its nucleus. No galaxy of their position angles vary in the outer this type is flatter than b/a = 0.3, with b parts. Finally, astronomers have found and a the minor and major axes of the that a few ellipticals do in fact have small elliptical image, respectively. Ellipticals numbers of luminous O and B stars as contain neither interstellar dust nor well as dust lanes. bright stars of spectral types O and B. Many, however, contain evidence of the Spiral Galaxies presence of low-density gas in their nuclear regions. Ellipticals are red in Spirals are characterized by circular sym- colour, and their spectra indicate that metry, a bright nucleus surrounded by a their light comes mostly from old stars, thin outer disk, and a superimposed spiral especially evolved red giants. structure. They are divided into two par- Subclasses of elliptical galaxies are allel classes: normal spirals and barred defined by their apparent shape, which is spirals. The normal spirals have arms that of course not necessarily their three- emanate from the nucleus, while barred dimensional shape. The designation is spirals have a bright linear feature called En, where n is an integer defined by a bar that straddles the nucleus, with the arms unwinding from the ends of the bar. n = 10( a− b)/a. The nucleus of a spiral galaxy is a sharp- peaked area of smooth texture, which can A perfectly circular image will be an E0 be quite small or, in some cases, can make galaxy, while a flatter object might be an up the bulk of the galaxy. Both the arms E7 galaxy. (As explained above, elliptical and the disk of a spiral system are blue in galaxies are never flatter than this, so colour, whereas its central areas are red like there are no E8, E9, or E10 galaxies.) an elliptical galaxy. The normal spirals Although the above-cited criteria are are designated S and the barred varieties generally accepted, current high-quality SB. Each of these classes is subclassified measurements have shown that some into three types according to the size of the significant deviations exist. Most ellipti- nucleus and the degree to which the spiral cal galaxies do not, for instance, exactly arms are coiled. The three types are denoted fit the intensity law formulated by Hubble; with the lowercase letters a, b, and c. There deviations are evident in their innermost also exist galaxies that are intermediate parts and in their faint outer parts. between ellipticals and spirals. Such sys- Furthermore, many elliptical galaxies tems have the disk shape characteristic of have slowly varying ellipticity, with the the latter but no spiral arms. These inter- images being more circular in the central mediate forms bear the designation S0. Galaxies | 179

S0 Galaxies which has a double, almost rectangular bulge around a central nucleus. Another These systems exhibit some of the prop- type of peculiar S0 is found in NGC 2685. erties of both the ellipticals and the This nebula in the constellation Ursa spirals and seem to be a bridge between Major has an apparently edge-on disk these two more common galaxy types. galaxy at its centre, with surrounding Hubble introduced the S0 class long after hoops of gas, dust, and stars arranged his original classification scheme had in a plane that is at right angles to the been universally adopted, largely because apparent plane of the central object. he noticed the dearth of highly flattened objects that otherwise had the properties Sa Galaxies of elliptical galaxies. Sandage’s elabora- tion of the S0 class yielded the These normal spirals have narrow, tightly characteristics described here. wound arms, which usually are visible S0 galaxies have a bright nucleus because of the presence of interstellar that is surrounded by a smooth, feature- dust and, in many cases, bright stars. less bulge and a faint outer envelope. Most of them have a large amorphous They are thin; statistical studies of the bulge in the centre, but there are some ratio of the apparent axes (seen projected that violate this criterion, having a small onto the sky) indicate that they have nucleus around which is arranged an intrinsic ratios of minor to major axes in amorphous disk with superimposed faint the range 0.1 to 0.3. Their structure does arms. NGC 1302 is an example of the nor- not generally follow the luminosity law of mal type of Sa galaxy, while NGC 4866 is elliptical galaxies but has a form more representative of one with a small nucleus like that for spiral galaxies. Some S0 sys- and arms consisting of thin dust lanes on tems have a hint of structure in the a smooth disk. envelope, either faintly discernible arm- like discontinuities or narrow absorption Sb Galaxies lanes produced by interstellar dust. Several S0 galaxies are otherwise pecu- This intermediate type of spiral typically liar, and it is difficult to classify them with has a medium-sized nucleus. Its arms are certainty. They can be thought of as pecu- more widely spread than those of the Sa liar irregular galaxies (i.e., Irr II galaxies) variety and appear less smooth. They or simply as some of the 1 or 2 percent of contain stars, star clouds, and interstellar galaxies that do not fit easily into the gas and dust. Sb galaxies show wide dis- Hubble scheme. Among these are such persions in details in terms of their galaxies as NGC 4753, which has irregular shape. Hubble and Sandage observed, for dust lanes across its image, and NGC 128, example, that in certain Sb galaxies the 180 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies arms emerge at the nucleus, which is often arms that are open, with relatively large quite small. Other members of this sub- pitch angles. The arms, moreover, are class have arms that begin tangent to a lumpy, containing as they do numerous bright, nearly circular ring, while still others irregularly distributed star clouds, stellar reveal a small, bright spiral pattern inset associations, star clusters, and gas clouds into the nuclear bulge. In any of these cases, known as emission nebulae. the spiral arms may be set at different pitch As in the case of Sb galaxies, there angles. (A pitch angle is defined as the are several recognizable subtypes among angle between an arm and a circle centred the Sc systems. Sandage has cited six on the nucleus and intersecting the arm.) subdivisions: (1) galaxies, such as the Hubble and Sandage noted further Whirlpool Galaxy (M51), that have thin deviations from the standard shape estab- branched arms that wind outward from a lished for Sb galaxies. A few systems tiny nucleus, usually extending out about exhibit a chaotic dust pattern superim- 180° before branching into multiple seg- posed upon the tightly wound spiral ments, (2) systems with multiple arms arms. Some have smooth, thick arms of that start tangent to a bright ring centred low surface brightness, frequently on the nucleus, (3) those with arms that bounded on their inner edges with dust are poorly defined and that span the lanes. Finally, there are those with a large, entire image of the galaxy, (4) those with smooth nuclear bulge from which the a spiral pattern that cannot easily be arms emanate, flowing outward tangent traced and that are multiple and punctu- to the bulge and forming short arm seg- ated with chaotic dust lanes, (5) those ments. This is the most familiar type of with thick, loose arms that are not well Sb galaxy and is best exemplified by the defined—e.g., the nearby galaxy M33 (the giant Andromeda Galaxy. Triangulum Nebula)—and (6) transition Many of these variations in shape types, which are almost so lacking in remain unexplained. Theoretical models order that they could be considered of spiral galaxies based on a number of irregular galaxies. different premises can reproduce the Some classification schemes, such as basic Sb galaxy shape, but many of the that of the French-born American astron- deviations noted above are somewhat omer Gerard de Vaucouleurs, give the mysterious in origin and must await more last of the above-cited subtypes a class of detailed and realistic modeling of galac- its own, type Sd. It also has been found tic dynamics. that some of the variations noted here for Sc galaxies are related to total lumi- Sc Galaxies nosity. Galaxies of the fifth subtype, in particular, tend to be intrinsically faint, These galaxies characteristically have a while those of the first subtype are very small nucleus and multiple spiral among the most luminous spirals Galaxies | 181 known. This correlation is part of the feature external to the bar. SBa galaxies justifi cation for the luminosity classifi - have bright, fairly large nuclear bulges cation discussed below. and tightly wound, smooth spiral arms that emerge from the ends of the bar or SB Galaxies from a circular ring external to the bar. SBb systems have a smooth bar as well as The luminosities, dimensions, spectra, and relatively smooth and continuous arms. distributions of the barred spirals tend to In some galaxies of this type, the arms be indistinguishable from those of normal start at or near the ends of the bar, with spirals. The subclasses of SB systems exist conspicuous dust lanes along the inside in parallel sequence to those of the latter. of the bar that can be traced right up to There are SB0 galaxies that feature a the nucleus. Others have arms that start large nuclear bulge surrounded by a disk- tangent to a ring external to the bar. In like envelope across which runs a SBc galaxies , both the arms and the bar luminous featureless bar. Some SB0 sys- are highly resolved into star clouds and tems have short bars, while others have stellar associations. The arms are open in bars that extend across the entire visible form and can start either at the ends of image. Occasionally there is a ringlike the bar or tangent to a ring.

Barred spiral galaxy NGC 1300. NASA, ESA, and The Hubble Heritage Team (STScI/AURA) 182 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

Irregular Galaxies the galaxies differently. A notable example of one such system is that of de Most representatives of this class consist Vaucouleurs. This scheme, which has of grainy, highly irregular assemblages of evolved considerably since its inception luminous areas. They have neither notice- in 1959, includes a large number of codes able symmetry nor an obvious central for indicating different kinds of morpho- nucleus, and they are generally bluer in logical characteristics visible in the colour than are the arms and disks of spi- images of galaxies. The major Hubble ral galaxies. An extremely small number galaxy classes form the framework of de of them, however, are red and have a Vaucouleurs’s scheme, and its subdivision smooth, though nonsymmetrical, shape. includes different families, varieties, and Hubble recognized these two types stages. The de Vaucouleurs system is so of irregular galaxies, Irr I and Irr II. The detailed that it is more of a descriptive Irr I type is the most common of the irreg- code for galaxies than a commonly used ular systems, and it seems to fall naturally classification scheme. on an extension of the spiral classes, Galaxies with unusual properties beyond Sc, into galaxies with no discern- often have shorthand names that refer to ible spiral structure. They are blue, are their characteristic properties. Common highly resolved, and have little or no examples are: nucleus. The Irr II systems are red, rare objects. They include various kinds of D: Galaxies with abnormally large, chaotic galaxies for which there appar- distended shapes, always found ently are many different explanations, in the central areas of galaxy clus- including most commonly the results of ters and hypothesized to consist galaxy-galaxy interactions, both tidal dis- of merged galaxies. tortions and cannibalism; therefore, this S: Seyfert galaxies, originally recog- category is no longer seen as a useful way nized by the American astronomer to classify galaxies. Carl K. Seyfert from optical spec- Some irregular galaxies, like spirals, tra. These objects have very are barred. They have a nearly central bar bright nuclei with strong emis- structure dominating an otherwise chaotic sion lines of hydrogen and other arrangement of material. The Large MagÂ common elements, showing ellanic Cloud is a well-known example. velocities of hundreds or thousands of kilometres per second. Other Classification Most are radio sources. Schemes and Galaxy Types N: Galaxies with small, very bright nuclei and strong radio emission. Other classification schemes similar to These are probably similar to Hubble’s follow his pattern but subdivide Seyfert galaxies but more distant. Galaxies | 183

Q: Quasars, or QSOs, small, extremely a way to measure their distances. In an luminous objects, many of which earlier section, it was explained how are strong radio sources. Quasars astronomers first accomplished this apparently are related to Seyfert exceedingly difficult task for the nearby and N galaxies but have such galaxies during the 1920s. Until the late bright nuclei that the underlying decades of the 20th century, progress galaxy can be detected only with was discouragingly slow. Even though great difficulty. increased attention was being paid to the problem around the world, a consensus There are also different schemes used was not reached. In fact, the results of for extremely distant galaxies, which we most workers fell into two separate see in their youth. When a very distant gal- camps, in which the distances found by axy is examined with a very large telescope, one were about twice the size of the other’s. we see its structure as it was when the light For this reason, shortly after its launch was emitted billions of years ago. In such into Earth orbit in 1990, the Hubble Space cases, the distinctive Hubble types are not Telescope (HST) was assigned the special so obvious. Apparently, galaxies are much task of reliably determining the extra- less well organized in their early years, and galactic distance scale. Led by the these very distant objects tend to be highly Canadian-born astronomer Wendy irregular and asymmetrical. Although spe- Freedman and the American astronomer cial classification schemes are sometimes Robert Kennicutt, the team used a con- used for special purposes, the general siderable amount of the HST’s time to scheme of Hubble in its updated form is measure the properties of the Cepheid the one most commonly used. variable stars in a carefully selected set of galaxies. Their results were intermediate The external galaxies between the two earlier distance scales. With subsequent refinements, the scale Beyond the Milky Way stretches the vast of distances between the galaxies is now universe filled with billions of galaxies. on fairly secure footing. Astronomers have measured their dis- The HST distance scale project estab- tances and determined their properties lished the scale of distances for the nearby to a degree that would have seemed universe. Establishing the distances to unbelievable only a few decades ago. galaxies over the entire range of present observations (several billion light-years) The Extragalactic is an even more difficult task. The process Distance Scale involved is one of many successive steps that are all closely tied to one another. Before astronomers could establish the Before even the nearby galaxy distances existence of galaxies, they had to develop measured by the HST can be established, 184 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies distances must first be determined for a distances are measured by using PopulaÂ number of galaxies even closer to the tion II stars, such as RR Lyrae variables Milky Way Galaxy, specifically those in or luminous red giants. the Local Group. For this step, criteria are Beyond the Local Group are two used that have been calibrated within the nearby groups for which the P-L relation Milky Way Galaxy, where checks can be has been used: the Sculptor Group and made between different methods and the M81 Group. Both of these are small where the ultimate criterion is a geomet- clusters of galaxies that are similar in size ric one, basically involving trigonometric to the Local Group. They lie at a distance parallaxes, especially those determined of 10 to 15 million light-years. by the Hipparcos satellite. These distance One example of an alternate method criteria, acting as “standard candles,” are to the Cepheid P-L relationship makes then compared with the HST observa- use of planetary nebulae, the ringlike tions of galaxies beyond the Local Group, shells that surround some stars in their where other methods are calibrated that late stages of evolution. Planetary nebu- allow even larger distances to be gauged. lae have a variety of luminosities, This general stepwise process continues depending on their age and other physi- to the edge of the observable universe. cal circumstances; however, it has been The Local Group of galaxies is a con- determined that the brightest planetary centration of approximately 50 galaxies nebulae have an upper limit to their dominated by two large spirals, the Milky intrinsic brightnesses. This means that Way Galaxy and the Andromeda Galaxy. astronomers can measure the bright- For many of these galaxies, distances can nesses of such nebulae in any given be measured by using the Cepheid P-L galaxy, find the upper limit to the appar- law, which has been refined and made ent brightnesses, and then immediately more precise since it was first used by calculate the distance of the galaxy. This American astronomer Edwin Hubble. For technique is effective for measuring dis- instance, the nearest external galaxy, the tances to galaxies in the Local Group, in Large Magellanic Cloud, contains thou- nearby groups, and even as far away as sands of Cepheid variables, which can be the Virgo cluster, which lies at a distance compared with Cepheids of known dis- of about 50 million light-years. tance in the Milky Way Galaxy to yield a Once distances have been estab- distance determination of 160,000 light- lished for these nearby galaxies and years. This method has been employed groups, new criteria are calibrated for for almost all galaxies of the Local Group extension to fainter galaxies. Examples that contain massive-enough stars to of the many different criteria that have include Cepheids. Most of the rest of the been tried are the luminosities of the members are elliptical galaxies, which do brightest stars in the galaxy, the diame- not have Cepheid variables; their ters of the largest H II regions, supernova Galaxies | 185 luminosities, the spread in the rotational Supercluster). Radial velocities cannot velocities of stars and interstellar gas (the give reliable distances beyond a few bil- Tully-Fisher relation), and the luminosi- lion light-years, because, in the case of ties of globular clusters. All of these such galaxies, the observed velocities criteria have difficulties in their applica- depend on what the expansion rate of the tion because of dependencies on galaxy universe was then rather than what it is type, composition, luminosity, and other now. The light that is observed today was characteristics, so the results of several emitted several billion years ago when methods must be compared and cross- the universe was much younger and checked. Such distance criteria allow smaller than it is at present, when it might astronomers to measure the distances to have been expanding either more rapidly galaxies out to a few hundred million or more slowly than now. light-years. To find the distances of very distant Beyond 100 million light-years galaxies, astronomers have to avail them- another method becomes possible. The selves of methods that make use of expansion of the universe, at least for the extremely bright objects. In the past, immediate neighbourhood of the Local astronomers were forced to assume that Group (within one billion light-years or the brightest galaxies in clusters all so), is almost linear, so the radial velocity have the same true luminosity and that of a galaxy is a reliable distance indicator. measuring the apparent brightness of the The velocity is directly proportional to brightest galaxy in a distant cluster will the distance in this interval, so once a gal- therefore give its distance. This method axy’s radial velocity has been measured, is no longer used, however, as there is too all that must be known is the constant of much scatter in the brightness of the proportionality, which is called Hubble’s brightest galaxies and because there are constant. Although there still remains reasons to believe that both galaxies and some uncertainty in the correct value of galaxy clusters in the early universe were Hubble’s constant, the value obtained by quite different from those of the present. the HST is generally considered the best The only effective way found so far current value, which is very near 25 km/ for measuring distances to the most- sec (15 m/sec) per one million light-years. distant detectable galaxies is to use the This value does not apply in or near the brightness of a certain type of supernova, Local Group, because radial velocities called Type Ia. In the nearby universe measured for nearby galaxies and groups these supernovae—massive stars that are affected by the Local Group’s motion have collapsed and ejected much of their with respect to the general background of material explosively out into interstellar galaxies, which is toward a concentration space—show uniformity in their maximum of galaxies and groups of galaxies cen- brightnesses; thus, it can be assumed that tred on the Virgo cluster (the Local any supernovae of that type observed in a 186 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies very distant galaxy should also have the The mass of a typical large spiral is about same luminosity. Recent results have 500,000,000,000 Suns. strongly suggested that the universe’s In the late 20th century it became expansion rate is greater here and now clear that most of the mass in galaxies is than it was in the distant past. This not in the form of stars or other visible change of the expansion rate has impor- matter. By measuring the speed with tant implications for cosmology. which stars in spiral and elliptical galaxies orbit the centre of the galaxy, one can Physical Properties of measure the mass inside that orbit. Most External Galaxies galaxies have more mass than can be accounted for by their stars. Therefore, Like stars, galaxies display staggering there is some unidentified “dark matter” differences. Studies of their physical that dominates the dynamics of most properties can reveal their origins in the galaxies. The dark matter seems to be early universe. distributed more broadly than the stars in galaxies. Extensive efforts to identify this Size and Mass dark matter have not yet been satisfac- tory, though the detection of large The range in intrinsic size for the exter- numbers of very faint stars, including nal galaxies extends from the smallest brown dwarfs, was in some sense a by- systems, such as the extreme dwarf galax- product of these searches, as was the ies found near the Milky Way that are discovery of the mass of neutrinos. It is only 100 light years across, to giant radio somewhat frustrating for astronomers to galaxies, the extent of which (including know that the majority of the mass in their radio-bright lobes) is more than galaxies (and in the universe) is of an 3,000,000 light-years. Normal large spiral unknown nature. galaxies, such as the Andromeda Galaxy, have diameters of 100,000 to 500,000 Luminosity light-years. The total masses of galaxies are not The external galaxies show an extremely well known, largely because of the uncer- large range in their total luminosities. tain nature of the hypothesized invisible The intrinsically faintest are the extreme dark halos that surround many, or possibly dwarf elliptical galaxies, such as the Ursa all, galaxies. The total mass of material Minor dwarf, which has a luminosity of within the radius out to which the stars or approximately 100,000 Suns. The most gas of a galaxy can be detected is known luminous galaxies are those that contain for many hundreds of systems. The range quasars at their centres. These remark- is from about 100,000 to roughly ably bright superactive nuclei can be as 1,000,000,000,000 times the Sun’s mass. luminous as 2,000,000,000,000 Suns. The Galaxies | 187 underlying galaxies are often as much Some ellipticals formed almost all their as 100 times fainter than their nuclei. stars during the first few billion years, Normal large spiral galaxies have a lumi- while others may have had a more com- nosity of a few hundred billion Suns. plicated history, including various periods of active star formation related to Age the merging together of smaller galaxies. In a merging event the gas can be com- Even though different galaxies have had pressed, which enhances the conditions quite different histories, measurements necessary for new bursts of star forma- tend to suggest that most, if not all, galax- tion. The spirals and the irregulars, on ies have very nearly the same age. The the other hand, have been using up their age of the Milky Way Galaxy, which is materials more gradually. measured by determining the ages of the oldest stars found within it, is approxi- Composition mately 13 billion years. Nearby galaxies, even those such as the Large and Small The abundances of the chemical ele- Magellanic Clouds that contain a multi- ments in stars and galaxies are remarkably tude of very young stars, also have at least uniform. The ratios of the amounts of the a few very old stars of approximately that different elements that astronomers same age. When more distant galaxies observe for the Sun are a reasonably good are examined, their spectra and colours approximation for those of other stars in closely resemble those of the nearby gal- the Milky Way Galaxy and also for stars axies, and it is inferred that they too must in other galaxies. The main difference contain a population of similarly very old found is in the relative amount of the stars. Extremely distant galaxies, on the primordial gases, hydrogen and helium. other hand, look younger, but that is The heavier elements are formed by stel- because the “look-back” time for them is a lar evolutionary processes, and they are significant fraction of their age; the light relatively more abundant in areas where received from such galaxies was emitted extensive star formation has been taking when they were appreciably younger. place. Thus, in such small elliptical galax- It seems likely that all the galaxies ies as the Draco system, where almost all began to form about the same time, when the stars were formed at the beginning of the universe had cooled down enough for its lifetime, the component stars are matter to condense, and they all thus nearly pure hydrogen and helium, while started forming stars during nearly the in such large galaxies as the Andromeda same epoch. Their large differences are a Galaxy there are areas where star forma- matter not of age but rather of how they tion has been active for a long time (right proceeded to regulate the processing of up to the present, in fact), and there inves- their materials (gas and dust) into stars. tigators find that the heavier elements 188 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies are more abundant. In some external where I is the surface brightness (or the galaxies as well as in some parts of the stellar density) at position r, r is the radial

Milky Way Galaxy system, heavy ele- distance from the centre, and Ie and re are ments are even more abundant than in constants. This expression, introduced by the Sun but rarely by more than a factor the French-born American astronomer of two or so. Even in such cases, hydrogen Gerard de Vaucouleurs, is an empirical and helium make up most of the constitu- formula that works remarkably well in ent materials, accounting for at least 90 describing the spheroidal components of percent of the mass. almost all galaxies. An alternative formula, put forth by Edwin Hubble, is of the form Structure −2 I = I0( r/a + 1) , The spiral arms of some galaxies are the most notable part of their structure. where I is the surface brightness, I0 is the However, there are other no less impor- central brightness, r is the distance from tant pieces of a galaxy from a halo of old the centre, and a is a scaling constant. stars to its interstellar gas. Either of these formulas describes the structure well, but neither explains it. The Spheroidal Component A somewhat more complicated set of equations can be derived on the basis of the Most and perhaps all galaxies have a mutual gravitational attraction of stars spheroidal component of old stars. In the for one another and the long-term effects ellipticals this component constitutes all of close encounters between stars. These or most of any given system. In the spi- models of the spheroidal component rals it represents about half the constituent (appropriately modified in the presence stars (this fraction varies greatly accord- of other galactic components) fit the ing to galaxy type). In the irregulars the observed structures well. Rotation is not spheroidal component is very inconspic- an important factor, since most elliptical uous or, possibly in some cases, entirely galaxies and the spheroidal component absent. The structure of the spheroidal of spiral systems (e.g., the Milky Way component of all galaxies is similar, as if Galaxy) rotate slowly. One of the open the spirals and irregulars possess a skel- questions about the structure of these eton of old stars arranged in a structure objects is why they have as much flatten- that resembles an elliptical. The radial ing as some of them do. In most cases, distribution of stars follows a law of the measured rotation rate is inadequate the form to explain the flattening on the basis of a model of an oblate spheroid that rotates (−3.33{[r/re]¼ − 1}) I = Ie10 , around its short axis. Some elliptical Galaxies | 189 galaxies are instead prolate spheroids that found galaxies that have extensive arms rotate around their long axis. (extending around the centre for two or more complete rotations) and those that The Disk Component have a chaotic arm structure made up of many short fragments that extend only Except for such early-type galaxies as S0, 20° or 30° around the centre. All spiral SB0, Sa, and SBa systems, spirals and arms fit reasonably well to a logarithmic irregulars have a flat component of stars spiral of the form described in chapter 1, that emits most of their brightness. The The Milky Way Galaxy. disk component has a thickness that is approximately one-fifth its diameter (this Gas Distribution varies, depending on the type of stars being considered). The stars show a radial If one were to look at galaxies at wave- distribution that obeys an exponential lengths that show only neutral hydrogen decrease outward; i.e., the brightness gas, they would look rather different from obeys a formula of the form their optical appearance. Normally the gas, as detected at radio wavelengths for neu- log I = −kr, tral hydrogen atoms, is more widely spread out, with the size of the gas component where I is the surface brightness, r is the often extending to twice the size of the opti- distance from the centre, and k is a scal- cally visible image. Also, in some galaxies a ing constant. This constant is dependent hole exists in the centre of the system both on the type of the galaxy and on its where almost no neutral hydrogen occurs. intrinsic luminosity. The steepness of the There is, however, enough molecular hydro- outward slope is greatest for the early gen to make up for the lack of atomic Hubble types (Sa and SBa) and for the hydrogen. Molecular hydrogen is difficult least-luminous galaxies. to detect, but it is accompanied by other molecules, such as carbon monoxide, which Spiral Arms can be observed at radio wavelengths.

The structure of the arms of spiral galax- Galaxy Clusters ies depends on the galaxy type, and there is also a great deal of variability within Galaxies tend to cluster together, some- each type. Generally, the early Hubble times in small groups and sometimes in types have smooth, indistinct spiral arms enormous complexes. Most galaxies have with small pitch angles. The later types companions, either a few nearby objects have more-open arms (larger pitch or a large-scale cluster; isolated galaxies, angles). Within a given type there can be in other words, are quite rare. 190 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

Types of Clusters many as 10,000 galaxies, which are concentrated toward the cluster centre. There are several different classification schemes for galaxy clusters, but the sim- Distribution plest is the most useful. This scheme divides clusters into three classes: groups, Clusters of galaxies are found all over the irregulars, and sphericals. sky. They are difficult to detect along the Milky Way, where high concentra- Groups tions of the Galaxy’s dust and gas obscure virtually everything at optical wave- The groups class is composed of small lengths. However, even there clusters compact groups of 10 to 50 galaxies of can be found in a few galactic “windows,” mixed types, spanning roughly five mil- random holes in the dust that permit lion light-years. An example of such an optical observations. entity is the Local Group, which includes The clusters are not evenly spaced in the Milky Way Galaxy, the Magellanic the sky; instead, they are arranged in a Clouds, the Andromeda Galaxy, and way that suggests a certain amount of about 50 other systems, mostly of the organization. Clusters are frequently dwarf variety. associated with other clusters, forming giant superclusters. These superclusters Irregular Clusters typically consist of 3 to 10 clusters and span as many as 200 million light-years. Irregular clusters are large loosely struc- There also are immense areas between tured assemblages of mixed galaxy clusters that are fairly empty, forming types (mostly spirals and ellipticals), voids. Large-scale surveys made in the totaling perhaps 1,000 or more systems 1980s of the radial velocities of galaxies and extending out 10,000,000 to revealed an even-larger kind of structure. 50,000,000 light-years. The Virgo and It was discovered that galaxies and gal- Hercules clusters are representative of axy clusters tend to fall in position along this class. large planes and curves, almost like giant walls, with relatively empty spaces Spherical Clusters between them. A related large-scale structure was found to exist where there Spherical clusters are dense and consist occur departures from the velocity-distance almost exclusively of elliptical and S0 relation in certain directions, indicat- galaxies. They are enormous, having a ing that the otherwise uniform linear diameter of up to 50,000,000 light- expansion is being perturbed by large years. Spherical clusters may contain as concentrations of mass. One of these, Galaxies | 191 discovered in 1988, has been dubbed “the having envelopes that extend out to radii Great Attractor.” as large as one million light-years. Many of them have multiple nuclei, and most Interactions Between are strong sources of radio waves. The Cluster Members most likely explanation for cD galaxies is that they are massive central galactic Galaxies in clusters exist in a part of the systems that have captured smaller clus- universe that is much denser than aver- ter members because of their dominating age, and the result is that they have gravitational fields and have absorbed several unusual features. In the inner the other galaxies into their own struc- parts of dense clusters there are very few, tures. Astronomers sometimes refer to if any, normal spiral galaxies. This condi- this process as galactic cannibalism. In tion is probably the result of fairly this sense, the outer extended disks of cD frequent collisions between the closely systems, as well as their multiple nuclei, packed galaxies, as such violent inter- represent the remains of past partly actions tend to sweep out the interstellar digested “meals.” gas, leaving behind only the spherical One more effect that can be traced to component and a gasless disk. What the cluster environment is the presence remains is in effect an S0 galaxy. of strong radio and X-ray sources, which A second and related effect of galaxy tend to occur in or near the centres of interactions is the presence of gas-poor clusters of galaxies. These will be dis- spiral systems at the centres of large cussed in detail in the next section. irregular clusters. A significant number of the members of such clusters have Galaxies as a Radio Source anomalously small amounts of neutral hydrogen, and their gas components are Some of the strongest radio sources in smaller on average than those for more the sky are galaxies. Most of them have a isolated galaxies. This is thought to be peculiar morphology that is related to the result of frequent distant encounters the cause of their radio radiation. Some between such galaxies involving the dis- are relatively isolated galaxies, but most ruption of their outer parts. galaxies that emit unusually large A third effect of the dense cluster amounts of radio energy are found in environment is the presence in some large clusters. clusters—usually rather small dense clusters—of an unusual type of galaxy Radio Galaxies called a cD galaxy. These objects are somewhat similar in structure to S0 gal- The basic characteristics of radio galax- axies, but they are considerably larger, ies and the variations that exist among 192 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies them can be made clear with two examples. The other notable example of a radio The first is Centaurus A, a giant radio galaxy is Virgo A, a powerful radio source structure surrounding a bright, peculiar that corresponds to a bright elliptical gal- galaxy of remarkable morphology desig- axy in the Virgo cluster, designated as nated NGC 5128. It exemplifies a type M87. In this type of radio galaxy, most of of radio galaxy that consists of an opti- the radio radiation is emitted from an cal galaxy located at the centre of an appreciably smaller area than in the case immensely larger two-lobed radio source. of Centaurus A. This area coincides in In the particular case of Centaurus A, the size with the optically visible object. extent of the radio structure is so great Virgo A is not particularly unusual except that it is almost 100 times the size of the for one peculiarity: it has a bright jet of central galaxy, which is itself a giant gal- gaseous material that appears to emanate axy. This radio structure includes, besides from the nucleus of the galaxy, extending the pair of far-flung radio lobes, two other out approximately halfway to its faint sets of radio sources: one that is approxi- outer parts. This gaseous jet can be mately the size of the optical galaxy and detected at optical, radio, and other (e.g., that resembles the outer structure in X-ray) wavelengths; its spectrum sug- shape, and a second that is an intense gests strongly that it shines by means of small source at the galaxy’s nucleus. the synchrotron mechanism. Optically, NGC 5128 appears as a giant About the only condition that can elliptical galaxy with two notable charac- account for the immense amounts of teristics: an unusual disk of dust and gas energy emitted by radio galaxies is the surrounding it and thin jets of interstellar capture of material (interstellar gas and gas and young stars radiating outward. stars) by a supermassive object at their The most plausible explanation for this centre. Such an object would resemble whole array is that a series of energetic the one thought to be in the nucleus of events in the nucleus of the galaxy the Milky Way Galaxy but would be far expelled hot ionized gas from the centre more massive. In short, the most probable at relativistic velocities (i.e., those at type of supermassive object for explain- nearly the speed of light) in two opposite ing the details of strong radio sources directions. These clouds of relativistic would be a black hole. Large amounts of particles generate synchrotron radiation, energy can be released when material is which is detected at radio (and X-ray) captured by a black hole. An extremely wavelengths. In this model the very large hot high-density accretion disk is first structure is associated with an old event, formed around the supermassive object while the inner lobes are the result of from the material, and then some of the more-recent ejections. The centre is still material seems to be ejected explosively active, as evidenced by the presence of from the area, giving rise to the various the nuclear radio source. radio jets and lobes observed. Galaxies | 193

Another kind of event that can result to be distant galaxies and quasars, while in an explosive eruption around a nuclear others were relatively nearby objects, black hole involves cases of merging gal- including neutron stars (extremely dense axies in which the nuclei of the galaxies stars composed almost exclusively of “collide.” Because many, if not most, gal- neutrons) in the Milky Way Galaxy. axy nuclei contain a black hole, such a A substantial number of the X-ray gal- collision can generate an immense amount axies so far detected are also well-known of energy as the black holes merge. radio galaxies. Some X-ray sources, such as certain radio sources, are much too X-ray Galaxies large to be individual galaxies but rather consist of a whole cluster of galaxies. Synchrotron radiation is characteristically emitted at virtually all wavelengths Clusters of Galaxies as Radio at almost the same intensity. A synchro- and X-ray Sources tron source therefore ought to be detectable at optical and radio wave- Some clusters of galaxies contain a wide- lengths, as well as at others (e.g., infrared, spread intergalactic cloud of hot gas that ultraviolet, X-ray, and gamma-ray wave- can be detected as a diffuse radio source lengths). For radio galaxies this does or as a large-scale source of X-rays. The seem to be the case, at least in circum- gaseous cloud has a low density but a very stances where the radiation is not high temperature, having been heated by screened by absorbing material in the the motion of the cluster’s galaxies through source or in intervening space. it and by the emission of high-energy X-rays are absorbed by Earth’s atmo- particles from active galaxies within it. sphere. Consequently, X-ray galaxies The form of certain radio galaxies in could not be detected until it became clusters points rather strongly to the pres- possible to place telescopes above the ence of intergalactic gas. These are the atmosphere, first with balloons and “head-tail” galaxies, systems that have a sounding rockets and later with orbiting bright source accompanied by a tail or tails observatories specially designed for that appear swept back by their interaction X-ray studies. For example, the Einstein with the cooler more stationary intergalac- Observatory, which was in operation tic gas. These tails are radio lobes of ejected during the early 1980s, made a fairly gas whose shape has been distorted by complete search for X-ray sources across collisions with the cluster medium. the sky and studied several of them in detail. Beginning in 1999, the Chandra Quasars X-ray Observatory and other orbiting X-ray observatories detected huge numbers of An apparently new kind of radio source emitters. Many of the sources turned out was discovered in the early 1960s when 194 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies radio astronomers identified a very small up to 100 times as much radiation as an but powerful radio object designated 3C entire galaxy. It is a complex mixture of 48 with a stellar optical image. When they very hot gas, cooler gas and dust, and obtained the spectrum of the optical particles that emit synchrotron radiation. object, they found unexpected and at first Its brightness often varies over short unexplainable emission lines superim- periods—days or even hours. The galaxy posed on a flat continuum. This object underlying the brilliant image of a quasar remained a mystery until another similar may be fairly normal in some of its prop- but optically brighter object, 3C 273, was erties except for the superficial large-scale examined in 1963. Investigators noticed effects of the quasar at its centre. Quasars that 3C 273 had a normal spectrum with apparently are powered by the same the same emission lines as observed in mechanism attributed to radio galaxies. radio galaxies, though greatly redshifted They demonstrate in an extreme way (i.e., the spectral lines are displaced to what a supermassive object at the centre longer wavelengths), as by the Doppler of a galaxy can do. effect. If the redshift were to be ascribed With the gradual recognition of the to velocity, however, it would imply an causes of the quasar phenomenon has immense velocity of recession. In the come an equally gradual realization that case of 3C 48, the redshift had been so they are simply extreme examples of a large as to shift familiar lines so far that process that can be observed in more they were not recognized. Many more familiar objects. The black holes that are such objects were found, and they came thought to inhabit the cores of the quasar to be known as quasi-stellar radio sources, galaxies are similar to, though more abbreviated as quasars. explosive than, those that appear to occur Although the first 20 years of quasar in certain unusual nearer galaxies known studies were noted more for controversy as Seyfert galaxies. The radio galaxies fall and mystery than for progress in under- in between. The reason for the differences standing, subsequent years finally saw a in the level of activity is apparently related solution to the questions raised by these to the source of the gas and stars that are strange objects. It is now clear that qua- falling into the centres of such objects, sars are extreme examples of energetic providing the black holes with fuel. In the galaxy nuclei. The amount of radiation case of quasars, evidence suggests that emitted by such a nucleus overwhelms an encounter with another galaxy, which the light from the rest of the galaxy, so causes the latter to be tidally destroyed only very special observational techniques and its matter to fall into the centre of can reveal the galaxy’s existence. the more massive quasar galaxy, may A quasar has many remarkable prop- be the cause of its activity. As the mate- erties. Although it is extremely small rial approaches the black hole, it is greatly (only the size of the solar system), it emits accelerated, and some of it is expelled by Galaxies | 195 the prevailing high temperatures and Andromeda Galaxy drastically rapid motions. This process probably also explains the impressive but The great spiral Andromeda Galaxy (M31) lower-level activity in the nuclei of radio is the nearest external galaxy (except for and Seyfert galaxies. The captured mass the Magellanic Clouds, which are com- may be of lesser amount—i.e., either a panions of the Milky Way Galaxy, in smaller galaxy or a portion of the host which Earth is located). The Andromeda galaxy itself. Quasars are more common Galaxy is one of the few visible to the in that part of the universe observed to unaided eye, appearing as a milky blur. It have redshifts of about 2, meaning that is located about 2,480,000 light-years they were more common about 10 or so from Earth; its diameter is approximately billion years ago than they are now, which 200,000 light-years; and it shares various is at least partly a result of the higher characteristics with the Milky Way sys- density of galaxies at that time. tem. It was mentioned as early as 965 CE, in the Book of the Fixed Stars, by the Gamma-Ray Bursters Islamic astronomer s- ūfī, and rediscov- ered in 1612, shortly after the invention of In the 1970s a new type of object was the telescope, by the German astronomer identified as using orbiting gamma-ray Simon Marius, who said it resembled the detectors. These “gamma-ray bursters” light of a candle seen through a horn. For are identified by extremely energetic centuries astronomers regarded the bursts of gamma radiation that last only Andromeda Galaxy as a component of seconds. In some cases the bursters are the Milky Way Galaxy—i.e., as a so-called clearly identified with very distant galax- spiral nebula much like other glowing ies, implying immense energies in the masses of gas within the local galactic bursts. Possibly these are the explosions system (hence the misnomer Andromeda of “hypernovae,” posited to be far more Nebula). Only in the 1920s did the American energetic than supernovae and which astronomer Edwin Powell Hubble deter- require some extreme kind of event, such mine conclusively that the Andromeda as the merging of two neutron stars. was in fact a separate galaxy beyond the Milky Way. Notable galaxies and The Andromeda Galaxy has a past galaxy clusters involving collisions with and accretion of other galaxies. Its peculiar close compan- These galaxies and galaxy clusters are ion, M32, shows a structure that indicates some of the most astronomically impor- that it was formerly a normal, more massive tant. They range from the Small galaxy that lost much of its outer parts Magellanic Cloud to the vast Great and possibly all of its globular clusters to Attractor. M31 in a past encounter. Deep surveys of 196 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies the outer parts of the Andromeda Galaxy Coma Cluster have revealed huge coherent structures of star streams and clouds, with proper- The Coma cluster is the nearest rich clus- ties indicating that these include the ter of galaxies; it contains thousands of outer remnants of smaller galaxies systems. The Coma cluster lies about 33 “eaten” by the giant central galaxy, as million light-years away, about seven well as clouds of M31 stars ejected by the times farther than the Virgo cluster, in strong tidal forces of the collision. the direction of the constellation Coma

A representation of galaxies distributed on the surfaces of what are thought to be enormous bubblelike voids. The pie-shaped segment of the sky contains about 1,000 galaxies, all located within 300 million light-years from the Earth. The Coma cluster of galaxies, lying near the middle of the segment, seems to occur where several voids intersect. M.J. Geller and J.P. Huchra, Smithsonian Astrophysical Observatory Galaxies | 197

Berenices. The main body of the Coma reasonably ascribed to the known stellar cluster has a diameter of about 2.5 107 populations. A similar situation exists for light-years, but enhancements above the every rich cluster that has been examined background can be traced out to a super- in detail. When Swiss astronomer Fritz cluster of a diameter of about 2 108 Zwicky discovered this discrepancy in light-years. Ellipticals or S0s constitute 1933, he inferred that much of the Coma 85 percent of the bright galaxies in the cluster was made of nonluminous matter. Coma cluster; the two brightest ellipticals The existence of nonluminous matter, or in Coma are located near the centre of the “dark matter,” was later confirmed in the system and are individually more than 10 1970s by American astronomers Vera times as luminous as the Andromeda Rubin and W. Kent Ford. Galaxy. These galaxies have a swarm of smaller companions orbiting them and Cygnus A may have grown to their bloated sizes by a process of “galactic cannibalism” like Cygnus A is the most powerful cosmic that hypothesized to explain the super- source of radio waves known, lying in the giant elliptical cD systems. northern constellation Cygnus about The spatial distribution of galaxies in 500,000,000 light-years (4.8 × 1021 km [3 x rich clusters such as the Coma cluster 1021 miles]) from Earth. It has the appear- closely resembles what one would expect ance of a double galaxy. For a time it was theoretically for a bound set of bodies thought to be two galaxies in collision, moving in the collective gravitational but the energy output is too large to be field of the system. Yet, if one measures accounted for in that way. Radio energy is the dispersion of random velocities of the emitted from Cygnus A at an estimated Coma galaxies about the mean, one finds 1045 ergs per second, more than 1011 times that it amounts to almost 900 km per the rate at which energy of all kinds is emit- second (500 miles per second). For a gal- ted by the Sun. The source of the energy axy possessing this random velocity of Cygnus A remains undetermined. along a typical line of sight to be gravitationally bound within the known Great Attractor dimensions of the cluster requires Coma to have a total mass of about 5 1015 solar The Great Attractor is a proposed con- masses. The total luminosity of the Coma centration of mass that influences the cluster is measured to be about 3 1013 movement of many galaxies, including solar luminosities; therefore, the mass- the Milky Way. In 1986 a group of astron- to-light ratio in solar units required to omers observing the motions of the Milky explain Coma as a bound system exceeds Way and neighbouring galaxies noted by an order of magnitude what can be that the galaxies were moving toward the 198 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

Hydra-Centaurus superclusters in the light-years from Earth, and the SMC lies southern sky with velocities significantly 190,000 light-years away. The LMC and different from those predicted by the SMC are 14,000 and 7,000 light-years in expansion of the universe in accordance diameter, respectively, and are smaller with the Hubble law. One possible expla- than the Milky Way Galaxy, which is nation for this perturbation in the Hubble about 140,000 light-years across. flow is the existence of the so-called Great The Magellanic Clouds were formed Attractor—a region or structure of huge at about the same time as the Milky Way mass (equivalent to tens of thousands of Galaxy, approximately 13 billion years galaxies) exerting a gravitational pull on ago. They are presently captured in orbits the surrounding galaxies. It is estimated around the Milky Way Galaxy and have that the Great Attractor would have a experienced several tidal encounters with diameter of about 300 million light-years each other and with the Galaxy. They and that its centre would lie about 147 contain numerous young stars and star million light-years away from Earth. clusters, as well as some much older stars. The Magellanic Clouds serve as excellent Magellanic Clouds laboratories for the study of very active stellar formation and evolution. With the The Magellanic Clouds are two satellite Hubble Space Telescope it is possible for galaxies of the Milky Way Galaxy, the astronomers to study the kinds of stars, vast star system of which Earth is a minor star clusters, and nebulae that previously component. These companion galaxies could be observed in great detail only in were named for the Portuguese navigator the Milky Way Galaxy. Ferdinand Magellan, whose crew discovered them during the first voyage around M81 Group the world (1519–22). The Magellanic Clouds are irregular The M81 group of more than 40 galaxies galaxies that share a gaseous envelope is found at a distance of 12 million light- and lie about 22° apart in the sky near years from Earth, one of the nearest the south celestial pole. One of them, the galaxy groups to the Local Group (the Large Magellanic Cloud (LMC), is a lumi- group of galaxies that includes the Milky nous patch about 5° in diameter, and the Way Galaxy). The dominant galaxy in other, the Small Magellanic Cloud (SMC), the M81 group is the spiral galaxy M81. measures less than 2° across. The MagÂ Much like the Andromeda and Milky ellanic Clouds are visible to the unaided Way galaxies, M81 is of Hubble type Sb eye in the Southern Hemisphere, but they and luminosity class II. cannot be observed from the northern There are two subgroups in the M81 latitudes. The LMC is about 160,000 group: one group is associated with Galaxies | 199

M81 and another is associated with the in interstellar space prevent nearly all spiral galaxy NGC 2403. These two sub- visible light emitted by external galaxies groups are moving toward each other. from reaching Earth. The total mass of the M81 group has been Maffei I is a large elliptical galaxy. determined from the motion of galaxies At about 3,000,000 light-years’ distance, within it to be 1 1012 solar masses. M81 has it is close enough to belong to what is a mass of 6.7 1011 solar masses. called the Local Group of galaxies, of The M81 group also has a few galax- which the Milky Way Galaxy is a mem- ies with classifications similar to those of ber. Maffei II has a spiral structure and is galaxies in the Local Group, and it was about three times farther away than noticed by some astronomers that the Maffei I. linear sizes of the largest H II regions (which are illuminated by many OB stars) Virgo A in these galaxies had about the same intrinsic sizes as their counterparts in Virgo A (catalog numbers M87, and the Local Group. This led American NGC4486, ) is a giant elliptical galaxy in astronomer Allan Sandage and the the constellation Virgo whose nucleus German chemist and physicist Gustav provides the strongest observational Tammann to the (controversial) tech- evidence for the existence of a black nique of using the sizes of H II regions as hole. Virgo A is the most powerful known a distance indicator, because a measure- source of radio energy among the thou- ment of their angular sizes, coupled with sands of galactic systems comprising knowledge of their linear sizes, allows an the so-called Virgo cluster. It is also a inference of distance. powerful X-ray source, which suggests the presence of very hot gas in the gal- Maffei I and II axy. A luminous gaseous jet projects outward from the galactic nucleus. Both The two galaxies Maffei I and II are rela- the jet and the nucleus emit synchrotron tively close to the Milky Way Galaxy but radiation, a form of nonthermal radia- were unobserved until the late 1960s, tion released by charged particles that when the Italian astronomer Paolo are accelerated in magnetic fields and Maffei detected them by their infrared travel at speeds near that of light. Virgo radiation. Studies in the United States A lies about 50 million light-years from established that the objects are galaxies. the Earth. Lying near the border between the con- In 1994 the Hubble Space Telescope stellations Perseus and Cassiopeia, they obtained images of Virgo A that showed are close to the plane of the Milky Way, a disk of hot, ionized gas about 500 where obscuring dust clouds light-years in diameter at a distance of 200 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies about 60 light-years from the galaxy’s Virgo Cluster centre. The disk’s gases are revolving about the nucleus at a speed of about The Virgo cluster is the closest large clus- 550 km per second, or about 1.9 million ter of galaxies; it is located at a distance km (1.2 million miles) per hour, a velocity of about 5 × 107 light-years in the direction so great that only the gravitational pull of the constellation Virgo. More than 2,000 of an object with a mass six billion times galaxies reside in the Virgo cluster, scat- that of the Sun would be capable of tered in various subclusters whose largest holding the disk together. This super- concentration (near the famous system massive object could occupy a region M87 [Virgo A]) is about 5 × 106 light-years as small as the galactic nucleus only if in diameter. Of the galaxies in the Virgo it were a black hole. Gravitational cluster, 58 percent are spirals, 27 percent energy released by gas spiraling down are ellipticals, and the rest are irregulars. into the black hole produces a beam of Although spirals are more numerous, the electrons accelerated almost to the four brightest galaxies are giant ellipticals, speed of light; the bright gaseous jet among them Virgo A. Calibration of the that emanates from Virgo A is thought absolute brightnesses of these giant ellip- to be radiation from this beam of ticals allows a leap to the measurement of electrons. distant regular clusters. Appendix: Other Stars and Star Clusters

cattered throughout the sky are a Big Dipper (Ursa Major); however, the Smyriad of stars and clusters of stars. two are three light-years apart and thus Some, such as Antares and Spica, are are not gravitationally bound to each prominent in the night sky or have other. The ability to separate the dim star unusual properties, such as Mira Ceti or Alcor from Mizar 0.2° away with the Geminga. Others, including HD 209458 unaided eye may have been regarded by or 61 Cygni, have been prominent in the the Arabs (and others) as a test of good history of astronomy. vision. The pair have also been called the Horse and Rider. 61 Cygni Aldebaran In 1838, German astronomer Friedrich Wilhelm Bessel obtained a distance of Aldebaran (Arabic: “The Follower”) is a 10.3 light-years for 61 Cygni, the first star reddish giant star in the constellation whose distance from Earth was measured. Taurus. Aldebaran (also called Alpha The European Space Agency satellite Tauri) is one of the 15 brightest stars, with Hipparcos made much more accurate an apparent visual magnitude of 0.85. Its distance measurements than ground- diameter is 44 times that of the Sun. It is based telescopes had accomplished and accompanied by a very faint (13th magni- obtained a distance to 61 Cygni of 11.4 tude) red companion star. Aldebaran lies light-years. The star is a visual binary, the 65 light-years from Earth. The star was components of which revolve around once thought to be a member of the each other in a period of 659 years, and is Hyades cluster, but in fact Aldebaran is located in the northern constellation 85 light-years closer to Earth. Aldebaran Cygnus. They are of fifth and sixth was probably named “The Follower” magnitudes. because it rises after the Pleiades cluster of stars. Alcor Algol Alcor (Arabic: “Faint One”) is a star with apparent magnitude of 4.01. Alcor makes Algol, or Beta Persei, is the prototype of a a visual double with the brighter star class of variable stars called eclipsing Mizar in the middle of the handle of the binaries, the second brightest star in the 202 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies northern constellation Perseus. Its appar- seems to come from a Greek phrase ent visual magnitude changes over the meaning “rival of Ares” (i.e., rival of the range of 2.1 to 3.4 with a period of 2.87 planet Mars) and was probably given days. Even at its dimmest it remains read- because of the star’s colour and ily visible to the unaided eye. The name brightness. probably derives from an Arabic phrase meaning “demon,” or “mischief-maker,” Barnard’s star and the Arabs may have been aware of the star’s variability even before the Barnard’s star is the third nearest star to invention of the telescope. the Sun (after Proxima Centauri and The first European astronomer to Alpha Centauri’s A and B components note the light variation was the Italian considered together), at a distance of Geminiano Montanari in 1670; the English about 6 light-years. It is named for Edward astronomer John Goodricke measured Emerson Barnard, the American astrono- the cycle (69 hours) in 1782 and suggested mer who discovered it in 1916. Barnard’s partial eclipses of the star by another star has the largest proper motion of any body as a cause, a hypothesis proved known star—10.25 seconds of arc annu- correct in 1889. The comparatively long ally. It is a red dwarf star with a visual duration of the eclipse shows that the magnitude of 9.5; its intrinsic luminosity dimensions of the two stars are not negli- is only 1/2,600 that of the Sun. gible in comparison with the distance Because of its high velocity of between them. A third star, which does approach, 108 km (67 miles) per second, not take part in the eclipses, revolves Barnard’s star is gradually coming nearer about the other two with a period of the solar system and by the year 11,800 1.862 years. will reach its closest point in distance— namely, 3.85 light-years. The star is of Antares special interest to astronomers because its proper motion, observed photographi- Antares is a red, semiregular variable cally between the years 1938–81, was star, with apparent visual magnitude thought to show periodic deviations of about 1.1, the brightest star in the zodia- 0.02 seconds of arc. This “perturbation” cal constellation Scorpius and one of the was interpreted as being caused by the largest known stars, having several hun- gravitational pull of two planetary com- dred times the diameter of the Sun and panions having orbital periods of 13.5 10,000 times the Sun’s luminosity. It has a and 19 years, respectively, and masses of fifth-magnitude blue companion. Antares about two-thirds that of Jupiter. However, (also called Alpha Scorpii) lies about 600 this finding has not been supported by light-years from the Earth. The name results from other methods of detection. Appendix: Other Stars and Star Clusters | 203

BETA LyRAE

Beta Lyrae is an eclipsing binary star, the two component stars of which are so close together that they are greatly distorted by their mutual attraction; they exchange material and share a common atmosphere. Beta Lyrae is a member of a class of binary systems known as W Serpentis stars. It is of about third magnitude and lies in the northern constellation Lyra. The variable character of Beta Lyrae was discovered in 1784 by the English amateur astronomer John Goodricke . Its period of about 13 days is increasing by about 19 seconds per year, probably Debris disk surrounding the star Beta because the stars are steadily losing mass Pictoris, in an image gathered by the to a continually expanding gaseous ring European Southern Observatory’s 3.6-metre (140-inch) telescope at La Silla, Chile. The surrounding them. warping seen in the disk’s bright inner region may be indirect evidence for one or BETA PICTORIS more orbiting planets. European Southern Observatory The fourth-magnitude star Beta Pictoris is located 60 light-years from Earth in the southern constellation Pictor and is with the Hubble Space Telescope revealed notable for an encircling disk of debris the disk to be warped and its inner that might contain planets. The star is regions to be relatively clear. A likely, of a common type somewhat hotter and but not exclusive, explanation for these more luminous than the Sun. In 1983 it characteristics is that one or more extra- was discovered to be an unexpectedly solar planets exist there. The outer part strong source of infrared radiation of of the disk shows rings, possibly caused the character that would be produced by a passing star. In 2008, infrared by a disk of material surrounding the images showed a possible planet orbit- star. The disk was later imaged and ing Beta Pictoris. This planet is estimated found to have a width roughly 2,000 to have a mass eight times that of Jupiter times the Earth-Sun distance (2,000 and to be at a distance from Beta Pictoris astronomical units [AU]). Observations of 8 AU. 204 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

Canopus additional two component stars form an eclipsing binary system of red dwarfs Canopus, or Alpha Carinae, is the second revolving around each other in less than brightest star (after Sirius) in the night a day and orbiting the four main stars sky, with a visual magnitude of −0.73. in a period of 14,000 years. The system is Lying in the southern constellation 51.5 light-years from Earth. Carina, 310 light-years from Earth, Canopus is sometimes used as a guide in Cor Caroli the attitude control of spacecraft because of its angular distance from the Sun and Cor Caroli, which is also called Alpha the contrast of its brightness among Canum Venaticorum, is a binary star nearby celestial objects. The Syrian Stoic located 110 light-years from Earth in the philosopher Poseidonius (c. 135–50 BCE) constellation Canes Venatici and con- used sightings of Canopus near the hori- sisting of a brighter component (A) of zon in his estimation of the size of Earth. visual magnitude 2.9 and a companion (B) of magnitude 5.5. It is the prototype Capella for a group of unusual spectrum-variable stars that show strong and fluctuating Capella (also called Alpha Aurigae) is the absorption lines of silicon, chromium, sixth brightest star in the night sky and strontium, or certain rare earths. Europium the brightest in the constellation Auriga, apparently is concentrated around one with an apparent visual magnitude of magnetic pole, chromium around the 0.08. Capella (Latin: “She-Goat”) is a other. Cor Caroli (“Heart of Charles”) was spectroscopic binary comprising two named by Sir Edmond Halley for King G-type giant stars that orbit each other Charles II of England. every 104 days. It lies 42.2 light-years from Earth. Cygnus X-1

Castor The binary star system Cygnus X-1 is a strong source of X-rays and provided the Castor, which is also called Alpha first major evidence for the existence of Geminorum, is a multiple star having six black holes. Cygnus X-1 is located about component stars, in the zodiacal con- 7,000 light-years from Earth in the constellation Gemini. The stars Castor and stellation Cygnus. The primary star, HDE Pollux are named for the twins of Greek 226868, is a hot supergiant revolving mythology. Castor’s combined apparent about an unseen companion with a period visual magnitude is 1.58. It appears as a of 5.6 days. Analysis of the binary orbit bright visual binary, of which both led to the finding that the companion has members are spectroscopic binaries. An a mass greater than seven solar masses. Appendix: Other Stars and Star Clusters | 205

(The mass has been determined from a hot main-sequence star surrounded by subsequent observations to be nearly an immense ring or shell of gas that nine solar masses.) A star of that mass eclipses the primary for two years. should have a detectable spectrum, but the companion does not; from this and Eta Carinae other evidence astronomers have argued that it must be a black hole. The X-ray Eta Carinae, which is also called HomunÂ emission is understood as being due to culus Nebula, is a peculiar red star and matter torn from the primary star that nebula about 7,500 light-years from Earth is being heated as it is drawn to the in the southern constellation Carina and black hole. is now known to be a binary star system. It is one of a small class of stars called Delta Cephei luminous blue variables. The English astronomer Sir Edmond Halley noted it Delta Cephei is the prototype star of in 1677 as a star of about fourth magni- the class of Cepheid variables and is tude. In 1838 Sir John Herschel observed in the constellation Cepheus. Its appar- it as a first-magnitude star. By 1843 it had ent visual magnitude at minimum is 4.34 reached its greatest recorded brightness, and at maximum 3.51, changing in a regu- approximately −1 magnitude, or as bright lar cycle of about five days and nine hours. as the brightest stars. Unlike the common Its variations in brightness were discov- types of exploding stars called novae and ered in 1784 by the English amateur supernovae, it remained bright for sev- astronomer John Goodricke, and peri- eral years. From about 1857 it faded odic changes in radial velocity (now steadily, disappearing to the unaided eye attributed to pulsation) were established only about 1870. Since then it has varied in 1894. irregularly about the seventh magnitude. The nebula around the star was formed Epsilon Aurigae during its 19th-century brightening and is an expanding shell of gas and dust, The binary star system Epsilon Aurigae shaped like an hourglass with a disk at is of about third magnitude and has one its centre. of the longest orbital periods (27 years) In 2005 astronomers studying far- among eclipsing binaries. It is located an ultraviolet spectral observations of Eta estimated 2,000 light-years from Earth in Carinae made by spacecraft found that it the constellation Auriga. The primary is a binary star system with an orbital star is a yellow-white star about 200 times period of 5.52 years. Its A component has the size of the Sun. The secondary star, a temperature of about 15,000 K; its B once thought to be thousands of times component, about 35,000 K. The main the size of the Sun, is now believed to be star in Eta Carinae is about 100 times 206 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies more massive than the Sun. Its luminosity afterward, astronomers independently has been estimated as five million times confirmed the planet’s presence by that of the Sun. Flaring events produc- observing that HD 209458 changed in ing not only visible effects but also brightness with the same 3.5-day period X-ray, ultraviolet, and radio-wave effects predicted from the discovery data for the have been observed. It is expected to planet’s orbit. Although the planet could become a supernova in the next several not be seen directly, its passages between thousand years. its star and Earth provided important information about its physical properties Geminga and atmosphere not otherwise available. The extrasolar planet is about 1.3 times The isolated pulsar Geminga is about the size of Jupiter but has only two-thirds 800 light-years from Earth in the constel- of Jupiter’s mass. It orbits surprisingly lation Gemini and is unique in that close to the star—about 10 stellar radii. about 99 percent of its radiation is in the gamma-ray region of the spectrum. HR 8799 Geminga is also a weak X-ray emitter, but it was not identified in visible light (as a The star HR 8799 has the first extrasolar 25th-magnitude object) until nearly two planetary system to be seen directly in decades after its discovery in 1972. It was an astronomical image. HR 8799 is a the first pulsar not detected at radio young (about 60 million years old) main- wavelengths. It pulsates with a period of sequence star of spectral type A5 V 0.237 second, has a radius of about 10 located 128 light-years from Earth in the km (6 miles), and probably originated in constellation Pegasus. Observations of a supernova explosion about 300,000 this star taken by the Infrared AstroÂ years ago. nomical Satellite and the Infrared Space Observatory showed a disk of dust such HD 209458 as that expected in the last stages of planetary formation. In 2008 an international The seventh-magnitude star HD 209458, team of astronomers released images 150 light-years away in the constellation taken with the telescopes at the Keck and Pegasus, was the first star that had a Gemini North observatories of three planet detected by its transit across the planets orbiting HR 8799. Observations star’s face. The star, which has physical taken over the period 2004–08 showed characteristics similar to those of the Sun, that the planets moved with the star and was shown in late 1999 to have a planet therefore were not background objects. by detection of the planet’s gravitational The planets range in mass from 7 to 10 effects on the star’s motion. Shortly times that of Jupiter and orbit between Appendix: Other Stars and Star Clusters | 207

3.6 and 10.2 billion km (2.2 and 6.3 billion miles) from HR 8799. These planets are gas giants with temperatures of about 900 to 1,100 kelvins (600 to 800 °C, or 1,200 to 1,500 °F).

HyADES

The Hyades is a cluster of several hundred stars in the zodiacal constellation Taurus. As seen from Earth, the bright star Aldebaran appears to be a member of the cluster, but in fact Aldebaran is much closer to the Earth than the Hyades’ distance of about 150 light-years. Five genuine members of the group are Composite image of Kepler’s Nova, or Kep- visible to the unaided eye. Their name ler’s Supernova, taken by the Chandra X-ray (Greek: “the rainy ones”) is derived from Observatory. NASA, ESA, R. Sankrit and W. the ancient association of spring rain Blair, Johns Hopkins University with the season of their heliacal (near dawn) rising. at the time as evidence of the mutability KEPLER’S NOvA of the stars.

Kepler’s Nova was one of the few super- MIRA CETI novae (violent stellar explosions) known to have occurred in the Milky Way Galaxy. Mira Ceti, which is also called Omicron Jan Brunowski , Johannes Kepler’s assis- Ceti, was the fi rst variable star (apart from tant, fi rst observed the phenomenon in novae) to be discovered, lying in the south- October 1604; Kepler studied it until ern constellation Cetus, and the prototype early 1606, when the supernova was no of a class known as long-period variables, longer visible to the unaided eye. At its or Mira stars. There is some evidence that greatest apparent magnitude (about -2.5), ancient Babylonian astronomers noticed the exploding star was brighter than its variable character. In a systematic study Jupiter. No stellar remnant is known to in 1638, a Dutch astronomer, Phocylides exist, though traces of nebulosity are Holwarda , found that the star disappeared observable at the position of the super- and reappeared in a varying cycle of about nova. Like Tycho’s Nova, Kepler’s served 330 days. It thus acquired the name Mira 208 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies

(Latin: “MiracÂulous”). Its brightness varies amateur astronomer J.P.M. Prentice, in from cycle to cycle, but generally it is about the northern constellation Hercules. It magnitude 3 at maximum light and mag- reached an apparent visual magnitude of nitude 9 at minimum. Mira is a binary; the 1.4 and remained visible to the unaided red giant primary has a faint bluish white eye for months. At its centre was found an companion. In 2006 the ultraviolet satel- eclipsing binary pair of small stars, revolv- lite observatory Galaxy Evolution Explorer ing around each other with a period of 4 discovered that Mira had shed material hours and 39 minutes. One component is into a cometary tail 13 light-years in length. a rapidly spinning white dwarf star accret- Mira is about 350 light-years from Earth. ing material from its companion.

Mizar Nova Persei

Mizar, also called Zeta Ursae Majoris, was Nova Persei, or GK Persei, was a bright the first star found (by the Italian astrono- nova that attained an absolute magnitude mer Giovanni Battista Riccioli in 1650) to of .2. Spectroscopic observations of the be a visual binary—i.e., to consist of two nova, which appeared in 1901, provided optically distinguishable components important information about interstellar revolving around each other. Later, each of gas. The shell thrown off by the explod- the visual components was determined to ing star was unusually asymmetrical, and be a spectroscopic binary; Mizar is actually a bright nebulosity near the star appeared a quadruple star. Apparent visual magni- to be expanding incredibly fast, at practi- tudes of the two visual components are cally the speed of light. This apparent 2.27 and 3.95. Set in the middle of the Big speed is thought to have been an effect Dipper’s handle, Mizar (Arabic: “Veil,” or of reflection within a preexisting dark neb- “Cloak”) makes a visual double with the ula around the star. From this phenomenon, fainter Alcor (Arabic: “Faint One”); how- sometimes called a light echo, it is possible ever, the two are three light-years apart and to calculate the distance of the nova from thus are not gravitationally bound to each Earth, about 1,500 light-years. other. The ability to separate the dim star Alcor from Mizar 0.2° away with the unaided Omega Centauri eye may have been regarded by the Arabs (and others) as a test of good vision. Omega Centauri (NGC 5139) is the brightest globular star cluster. It is located Nova Herculis in the southern constellation Centaurus. It has a magnitude of 3.7 and is visible to Nova Herculis, or DQ Herculis, was one the unaided eye as a faint luminous patch. of the brightest novae of the 20th century, Omega Centauri is about 16,000 light- discovered Dec. 13, 1934, by the British years from Earth and is thus one of the Appendix: Other Stars and Star Clusters | 209 nearer globular clusters. It is estimated to average distance of 253 million km (157 contain several million stars; several million miles). Pollux is the brightest star hundred variables have been observed in with a known extrasolar planet. it. There is some evidence for a black hole at the centre of Omega Centauri that is Praesepe 40,000 times as massive as the Sun. The English astronomer John Herschel in the Praesepe, which is also known as the 1830s was the first to recognize it as a star Beehive, is an open cluster of about 1,000 cluster and not a nebula. stars in the zodiacal constellation Cancer and is located about 550 light-years from Pleione Earth. Visible to the unaided eye as a small patch of bright haze, it was first dis- Pleione is a star in the Pleiades, thought tinguished as a group of stars by Galileo. to be typical of the shell stars, so called It was included by Hipparchus in the ear- because in their rapid rotation they throw liest known star catalog, c. 129 BCE. off shells of gas. In 1938 sudden changes The name Praesepe (Latin: “Cradle,” in the spectrum of Pleione were attrib- or “Manger”) was used even before uted to the ejection of a gaseous shell, Hipparchus’ time. The name Beehive is which by 1952 had apparently dissipated. of uncertain but more recent origin. Pleione is a blue-white star of about the fifth magnitude. Some astronomers con- Procyon jecture that it may have been brighter in the past; it would then have made a sev- Procyon is the brightest star in the north- enth bright star in the Pleiades cluster, ern constellation Canis Minor (Latin: which is named for seven mythological “Lesser Dog”) and one of the brightest in sisters. the entire sky, with an apparent visual magnitude of 0.41. Procyon lies 11.4 light- Pollux years from Earth and is a visual binary, a bright yellow-white subgiant with a faint, Pollux is the brightest star in the zodiacal white dwarf companion of about the 10th constellation Gemini. A reddish giant magnitude. The name apparently derives star, it has an apparent visual magnitude from Greek words for “before the dog,” in of 1.15. The stars Castor and Pollux are reference to the constellation. named for the mythological twins. Pollux is also called Beta Geminorum and is 33.7 Ras Algethi light-years from Earth. In 2006, a planet, Pollux b, was discovered. Pollux b has Ras Algethi, which is also called Alpha nearly three times the mass of Jupiter, Herculis, is a red supergiant star, whose orbits Pollux every 590 days, and is at an diameter is nearly twice that of Earth’s 210 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies orbit. It lies in the constellation Hercules discovered in the direction of the con- and is of about third magnitude, its stellation Scorpius. Detected in 1962, its brightness varying by about a magnitude X-radiation is not only strong but, like every 128 days. It is 380 light-years from other X-ray sources, quite variable as well. Earth. The name comes from an Arabic Its variability exhibits two states, one at phrase meaning “the kneeler’s head,” higher output with great variability on a referring to the Arabic name of the time scale of minutes and another at constellation. lower output with the variability corre- spondingly lessened. Regulus Scorpius X-1 was observed in visible light for the first time in 1966. Optically Regulus, or Alpha Leonis, is the brightest it is much less impressive, bluish in star in the zodiacal constellation Leo and colour and appearing only faintly. one of the brightest in the entire sky, hav- Scorpius X-1 is a close double star, ing an apparent visual magnitude of one component of which is optically about 1.35. It is 77 light-years from Earth. invisible—a neutron star. The X-rays are The name Regulus, derived from a Latin generated when matter from the optically word for king, reflects an ancient belief in visible, bluish hot star falls onto the the astrological importance of the star. neutron star. This matter is tremendously accelerated and crushed by the enor- Rigel mous gravity of the neutron star. Unlike the majority of binary X-ray sources, the Rigel (Beta Orionis) is one of the bright- visible member does not appear to be est stars in the sky, intrinsically as well very massive; it is only 42 percent the as in appearance. A blue-white super- mass of the Sun. The neutron star is 1.4 giant in the constellation Orion, Rigel is solar masses. Scorpius X-1 is about 9,000 about 870 light-years from the Sun and light-years from Earth. is about 47,000 times as luminous. A companion double star, also bluish white, S Doradus is of the sixth magnitude. The name Rigel derives from an Arabic term meaning S Doradus is a variable supergiant star in “the left leg of the giant,” referring to the the Large Magellanic Cloud. S Doradus figure of Orion. (and the Large Magellanic Cloud) is visible to viewers in the Southern HemiÂ Scorpius X-1 sphere in the constellation Dorado. It is one of the most luminous stars known, Scorpius X-1 is the brightest X-ray radiating more than 1,000,000 times as source in the sky and the first such object much energy as the Sun. Appendix: Other Stars and Star Clusters | 211

Spica occur during supernovae. Study of the evolving remnant continued into the 21st Spica (Latin: “Head of Grain”) is the century. brightest star in the zodiacal constellation Virgo and one of the 15 brightest in Tycho’s Nova the entire sky, having an apparent visual magnitude of 0.98. It is a bluish star; Tycho’s Nova (SN 1572) was one of the spectroscopic examination reveals Spica few recorded supernovae in the Milky to be a binary with a four-day period, its Way Galaxy. The Danish astronomer two components being of the first and Tycho Brahe first observed the “new star” third magnitudes, respectively. Spica lies on Nov. 11, 1572. Other European observ- about 250 light-years from Earth. ers claimed to have noticed it as early as the preceding August, but Tycho’s pre- Supernova 1987A cise measurements showed that it was not some relatively nearby phenomenon, Supernova 1987A was the first supernova such as a comet, but at the distance of the observed in 1987 (hence its designation) stars, and that therefore real changes and the nearest to Earth in more than could occur among them. three centuries. It occurred in the Large The supernova remained visible to Magellanic Cloud, a satellite galaxy of the unaided eye until March 1574. It the Milky Way Galaxy that lies about attained the apparent magnitude of 160,000 light-years distant. The super- Venus (about −4) and could be seen by nova originated in the collapse and day. There is no known stellar remnant subsequent explosion of a supergiant but only traces of glowing nebulosity. It star, and it is unique in that its progenitor is, however, a radio and X-ray source. In star had been observed and cataloged 2008 a team of international astronomers prior to the event. The fact that the super- used light from the original explosive giant was hotter than expected for an event reflected off nearby interstellar immediate progenitor led to important dust to determine that Tycho’s Nova improvements in supernova theory. A was a Type Ia supernova, which occurs burst of neutrinos that accompanied the when a white dwarf star accretes material star’s collapse was detected on Earth, from a companion star and that mate- providing verification of theoretical rial explodes in a thermonuclear reaction predictions of nuclear processes that that destroys the white dwarf. Glossary binary star A pair of stars in orbit nebulae A mass of interstellar gas around a common centre of gravity. and dust. black hole An area in space with an nova A star that brightens temporarily intense gravitational field whose while ejecting a shell explosively. escape velocity exceeds the speed photons Packets of radiation. of light. protostars A contracting mass of gas cloud cores The densest regions of that represents an early stage in the molecular clouds, where stars formation of a star, before nucleo- typically form. synthesis has begun. dwarf stars Low-luminosity stars. pulsars Neutron stars that emit pulses eclipsing binary Two close stars of radiation once per rotation. moving in an orbit so placed in quasars The abbreviation for quasi- space in relation to Earth that the stellar radio sources, which emit up light of one can at times be hidden to 100 times as much radiation as an behind the other. entire galaxy. extrasolar Revolving around stars other recombination The process by which than the Sun. the higher stage of ionization cap- interferometer An instrument that tures an electron, usually at low combines light waves from two or energies, into a high level of the ion. more different optical paths and thermal ionization The process at higher can be used to measure the angle temperatures where collisions between subtended by the diameter of a star atoms and electrons and the absorption at the observer’s position. of radiation tend to detach electrons kinematics The study of motion. and produce singly ionized atoms. light-year The distance that light waves z distances Distances above the plane travel in one Earth year. of the Galaxy. For Further Reading

Block, David L., and Kenneth C. the Black Holes That Drive Them. Freeman. Shrouds of the Night: London, England: Springer, 2007. Masks of the Milky Way and Our Kwok, Sun. The Origin and Evolution of Awesome New View of Galaxies. Planetary Nebulae. Cambridge, Berlin, Germany: Springer, 2008. England: Cambridge University Buta, Ronald J., et. al. The de Vaucouleurs Press, 2007. Atlas of Galaxies. Cambridge, Percy, John R. Understanding Variable England: Cambridge University Stars. Cambridge, England: Press, 2007. Cambridge University Press, 2007. Clark, Stuart. Galaxy: Exploring the Rees, Martin, ed. Universe. New York, Milky Way. New York, NY: Fall River NY: Dorling Kindersley, 2005. Press, 2008. Salaris, Maurizo, and Santi Cassisi. Coe, Steven R. Nebulae and How to Evolution of Stars and Stellar Observe Them (Astronomers’ Populations. West Sussex, England: Observing Guides). New York, NY: John Wiley & Sons, 2006. Springer, 2006. Sarazin, Craig L. X-Ray Emission from Eckart, Andreas. The Black Hole at the Clusters of Galaxies. Cambridge, Center of the Milky Way. London, England: Cambridge University England: Imperial College Press, 2005. Press, 2009. Freedman, Roger A., and William J. Schaaf, Fred. The Brightest Stars: Kaufmann, III. Universe: Stars and Discovering the Universe through Galaxies. New York, NY: W. H. the Sky’s Most Brilliant Stars. Freeman & Co., 2008. Hoboken, NJ: John Wiley & Gray, Richard O., and Christopher J. Sons, 2008. Corbally. Stellar Spectral Sparke, Linda S., and John S. Gallagher. Classification. Princeton, NJ: Galaxies in the Universe: An Princeton University Press, 2009. Introduction. Cambridge, England: Green, Simon F., and Mark H. Jones, eds. Cambridge University Press, 2007. An Introduction to the Sun and Stars. Stahler, Steven W., and Francesco Palla. Cambridge, England: Cambridge The Formation of Stars. Weinheim, University Press, 2004. Germany: Wiley VCH, 2005. Gribbin, John. Galaxies: A Very Short Wheeler, J. Craig. Cosmic Catastrophes: Introduction. New York, NY: Oxford Exploding Stars, Black Holes, University Press, 2008. and Mapping the Universe. New Kitchin, Chris. Galaxies in Turmoil: The York, NY: Cambridge University Active and Starburst Galaxies and Press, 2007. Index

A Book of the Fixed Stars, 195 Bright Star Catalogue, 57 Albireo, 54 brown dwarfs, 27, 35, 45, 52, 57, 59, 60, 65, Aldebaran, 46, 68 94–96, 186 Algol, 61, 112 Alpha Centauri, 13, 46, 49, 50, 53, 122 Altair, 46, 123 C Ambartsumian, Victor A., 117–118 Canopus, 46, 53, 122 Andromeda Galaxy, 18, 28, 29, 37, 121, 146, Capella, 46, 67, 68, 88 180, 184, 186, 187, 190, 195–196, 197, 198 carbon cycle, 86–87, 88, 92 Antares, 46, 79, 89 Cassiopeia A, 156, 159–160 Arcturus, 46, 68, 122, 123 Cassiopeia-Taurus Group, 36 astronomical unit, definition of, 60 Cepheid period-luminosity (P-L) law, 184 Cepheids, 29, 38, 72, 73, 74, 75, 76, 78, 112, 115, B 121, 124, 168, 170, 173, 174, 175, 176, 183, 184 Chandrasekhar limit, 101, 103 Baade, Walter, 28–29, 151 charge-coupled devices (CCDs), 132 Barnard’s star, 39, 50 Clark, Alvan, 126 Bell, Jocelyn, 106 cloud cores, 90–91 Bessel, Friedrich W., 61, 126 Coalsack, 138, 160 Beta Canis Majoris, 73, 74, 75 Coma Berenices, 110–111, 114, 196–197 Beta Pictoris, 67 Coma cluster, 196–197 Betelgeuse, 46, 54, 75, 79, 123 Cosmic Background Explorer, 44 big bang, 97, 98, 109, 148, 151, 165, 167 cosmic rays, 99, 102, 139, 155 binary/double stars, 46, 50–51, 52, 56, 63–67, Crab Nebula, 79, 99, 107, 157–158 69, 70, 71, 76–77, 78, 79, 80, 82, 93, 96, 101, Crab Pulsar, 106, 107 102, 104, 106, 107, 108, 109, 112, 118, 121, Cygnus A, 197 125, 126, 157, 169 Cygnus Loop (Veil Nebula), 158 eclipsing binary stars, 60, 61–62, 67–68, 72, Cysat, Johann, 162 73, 77, 80, 82, 112, 115 spectroscopic binary stars, 60, 61, 62, 63, 67, 124 D visual binary stars, 60–61, 67, 72 dark energy, 78, 101 black dwarfs, 104 dark matter, 27–28, 127, 186 black holes, 14, 22, 77–78, 79, 100–101, 108–109, Darquier, Augustin, 162 156, 192, 193, 194, 200 Delta Scuti, 73, 76 blue stragglers, 116 Deneb, 54, 123 bolometric magnitude, 55 density distribution of stars, 35–38 Index | 215 density-wave pattern/theory, 24, 37 79, 110, 112, 113, 114–117, 119, 120, 121, diffuse ionized gas, 129, 130, 158–159 122, 151, 165, 166, 170, 171, 173, 174, 176, Draper, Henry, 162 185, 195 “Gould Belt,” 36 E Great Attractor, 191, 195, 197–198 Great Rift, 160 echelle spectrograph, 133 Greenstein, Jesse L., 30 eclipsing binary stars, 60, 61–62, 67–68, 72, 73, Gum, Colin S., 142, 160 77, 80, 82, 112, 115 Gum Nebula, 142, 160 Eddington, Arthur, 72, 75, 76 Eddington limit, 59 Einstein, Albert, 107, 108 H elliptical galaxies, explanation of, 177–178 Halley, Edmund, 114 Epsilon Orionis, 58 Hawking, Stephen, 109 explosive variables, 73, 76–78 Helfer, H. Lawrence, 30 Henry Draper Catalogue, 57 F Herschel, John, 36, 120, 131 Fabry-Pérot interferometers, 133 Herschel, William, 131 51 Pegasi, 63, 122 Hertzsprung, Ejnar, 68, 70 flare stars, 74, 78–79, 112 Hertzsprung-Russell diagram, 68, 70, 92, 93 Fomalhaut, 67, 123–124 Hewish, Antony, 106 Ford, W. Kent, 197 Horsehead Nebula, 160–161 formation function, 34, 35 H II regions, 16, 90, 129, 130, 133–134, 136, Freedman, Wendy, 183 141–148, 149, 150, 158, 159, 160, 161, 174, 184, 199 G Hubble, Edwin, 177, 178, 179, 180, 182, 183, 184, 188, 195 galaxies Hubble’s constant, 185 external, 183–189 Hubble Space Telescope (HST) distance formation and evolution of, 16, 165–167 scale project, 183–184 as a radio source, 191–195 Hulse, Russell, 106–107 types of, 177–183 Huygens, Christiaan, 131, 143 galaxy clusters, 16–18, 189–191 Hyades, 71, 110, 113 giant/supergiant stars, 28, 29, 30, 45, 46, 50, hydrogen clouds (H I region), 24, 26, 56, 57, 58, 59, 62, 67, 68, 70, 71, 72, 75, 76, 140–141 78, 79, 81, 82, 85, 88, 89, 93, 94, 103, 104, 109, 113, 115, 118, 121, 123, 130, 135, 140, I 151, 152, 154, 175, 178, 184 Gliese, Wilhelm, 35 Index Catalogues (IC), 131–132 globular star clusters, 21, 23, 25, 29, 30, interstellar dust, 14, 16, 21, 24, 25, 27, 55, 127, 31, 32, 34, 36, 38, 43–44, 70, 71–72, 74, 76, 134–136, 138, 160, 161, 167, 171, 178, 179 216 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies interstellar gas, 14, 16, 25, 27, 90, 109, 127–129, mass of, 26–28 131, 136, 138, 141, 144, 148, 149, 150, 152, rotation of, 26 155, 161, 167, 169, 179, 185, 188, 191, 192 structure of, 22–25 interstellar medium, explanation of, 127 molecular clouds, 15, 16, 24, 89, 90, 91, 129, 133, irregular galaxies, explanation of, 182 134, 135, 136, 137–140, 142, 143, 144, 155, 159 Morgan, William W., 58 J moving star clusters, 113 Joy, Alfred, 118 N K nebulae chemical composition and physical Kamp, Peter van de, 40 processes of, 133–148 Kapteyn’s star, 40 classes of, 14–16, 129–131 Keenan, Philip C., 58 definition of, 127 Kennicutt, Robert, 183 early observations of, 131–132 Kepler, Johannes, 27 20th-century discoveries about, 132–133 Keplerian orbits, 27 neutron stars/pulsars, 65, 74, 77, 79, 99, 100, Kepler’s nova, 154, 156, 158 104–108, 156, 157, 158, 159, 193, 195 Kumar, Shiv, 95 New General Catalogue (NGC), 131, 132 North American Nebula, 161 L novae/supernovae, 14, 16, 25, 65, 71, 73, 76–78, Lagoon Nebula, 161 79, 94, 97, 99–102, 104, 106, 107, 108, 115, Lin, Chia-Chiao, 24 129, 130, 131, 133, 137, 143, 144, 154–156, Lindblad, Bertil, 26 157, 158, 159–160, 169, 171, 173, 176, 184, Local Group, 16, 29, 44, 120, 121, 146, 184, 185, 185, 195 190, 198, 199 O M OB and T associations, 90, 112, 117–119, 120, 121, 199 Maffei, Paolo, 199 Omega Centauri, 114, 117, 120 Maffei I and II, 199 Oort, Jan H., 26 Magellanic Clouds, 18, 71, 100, 117, 120, 121, open star clusters, 32, 34, 35, 38, 71, 72, 146, 147, 154, 182, 184, 187, 190, 195, 198 110–114, 118, 119, 120, 121, 124, 170, 171 magnetars, 108 Orion Nebula, 49, 114, 131, 132, 142, 143, 145, magnitude, stellar, explanation of, 53–54 147, 148, 161–162 Marius, Simon, 195 M81 group of galaxies, 198–199 Messier, Charles, 131 P Milky Way Galaxy Peiresc, Nicolas-Claude Fabri de, 131, 162 magnetic field of, 25–26, 137 perfect gas law, 83, 84 Index | 217 planetary nebulae, 16, 36, 71, 94, 129, 130–131, S 147, 148–154, 184 Pleiades, 70, 71, 96, 110, 112, 113, 121, 124, 141 Sagittarius A*, 22, 109 Polaris, 124 Saha, Meghnad N., 57 Population I stars, 28–29, 30, 31, 43, 70, 71, 72, Salpeter, Edwin E., 34 74, 75, 100, 112, 151 Sandage, Allan, 30, 35, 177, 179, 180, 199 Population II stars, 23, 25, 29, 30, 31, 32, 38, Schmidt, Bernhard, 132 43, 70, 72, 74, 75, 115, 151, 184 Schwarzschild, Karl, 108 Praesepe, 71, 72, 110 Schwarzschild radius, 108 Procyon, 46, 50 Seyfert, Carl K., 182 proton-proton cycle, 85–86, 87, 92 Seyfert galaxies, 182, 183, 194, 195 protostars, 49, 91, 92, 139 Shapley, Harlow, 21–22, 114 Proxima Centauri, 13, 49, 122 shell-source models, 88 pulsars/neutron stars, 65, 74, 77, 79, 99, 100, Shelton, Ian K., 100 104–108, 156, 157, 158, 159, 193, 195 Shu, Frank H., 24 pulsating variables, 73, 74–76 Sirius, 46, 50, 53, 57, 68, 81, 89, 104, 122, 125–126 solar motion, 40–44 Q solar wind, 46, 48, 81 quasars, 109, 183, 193–195 speckle interferometry, 67 spectroscopic binary stars, 60, 61, 62, 63, R 67, 124 spiral galaxies, explanation of, 178–181 Rayet, Georges-Antoine-Pons, 98 SS Cygni, 73, 77 Rayleigh, Lord, 81 star clusters Rayleigh scattering, 81 dynamics of, 119–120 R Coronae Borealis, 74, 78 in external galaxies, 120–122 red dwarfs, 56, 62, 68, 87, 122 globular clusters, 21, 23, 25, 29, 30, 31, 32, reflection nebulae, 16, 129, 141 34, 36, 38, 43–44, 70, 71–72, 74, 76, 79, 110, relativity, theory of, 107, 108 112, 113, 114–117, 119, 120, 121, 122, 151, 165, Reynolds, Ronald, 158 166, 170, 171, 173, 174, 176, 185, 195 Rigel, 46, 54, 89 OB and T associations, 90, 112, 117–119, 120, Ring Nebula, 162 121, 199 R Monocerotis, 162 open clusters, 32, 34, 35, 38, 71, 72, 110–114, rotating radio transients (RRATs), 108 118, 119, 120, 121, 124, 170, 171 RR Lyrae stars, 25, 29, 36, 42, 43, 44, 70, 73, stars 74, 76, 112, 114, 115, 116, 117, 121, 168, 170, atmosphere of, 80–82 171, 184 brightness and colour of, 53–54, 92–93, 115 Rubin, Vera, 197 categories of, 50 Russell, Henry Norris, 68 distances to, 49–52 RV Tauri, 75, 76, 115 end states of, 102–109 218 | The Milky Way and Beyond: Stars, Nebulae, and Other Galaxies formation and evolution of, 14, 89–102, U 139–140, 144, 152–153, 155 Ursa Major, 113, 123, 179 internal structure of, 82–88, 92–93 Ursa Minor, 37, 124, 186 masses of, 59–60 motion of, 38–40, 52–53 V positions of, 52–53 principal population types, 28–31 van Maanen rotation, 171–172 size and activity of, 45–49 van Rhijn, Pieter J., 31 temperatures of, 59, 92, 93, 95, 98, 99 van Rhijn function, 31, 32, 33, 34, 35 stellar associations, explanation of, 45, variable stars, explanation of, 73–80 110, 118 Vaucouleurs, Gerard de, 180, 182, 188 stellar luminosity function, 31–35 Vega, 46, 67, 123, 126 stellar winds, 46–49, 137, 143, 151, 152 Vela Pulsar, 105, 107 subgiant stars, 58, 68, 69 Virgo A, 192, 199–200 Sun, atmosphere and mechanism of energy Virgo cluster, 122, 184, 185, 192, 196, 199, 200 transport, 80–81, 85, 87–88 visual binary stars, 60–61, 67, 72 supergiant nebulae, 145–146 symbiotic stars, 154 W Wallerstein, George, 30 white dwarfs, 14, 34, 35, 36, 39, 46, 50, 62, 67, T 69, 70, 71, 72, 76, 77, 78, 84, 94, 101, Tammann, Gustav, 199 103–104, 105, 126, 153, 154, 167 Tarter, Jill, 95 Wilkinson Microwave Anisotropy Probe, 44 Taylor, Joseph, 106–107 Wolf, Charles-Joseph-Étienne, 98 Trifid Nebula, 131, 162 Wolf-Rayet stars, 97–98, 152 Trumpler, Robert J., 114 Tully-Fisher relation, 185 Z turbulence (in nebulae), 16, 136–137, 138 Zanstra, H., 152 Tycho’s nova, 154, 156, 158 Zwicky, Fritz, 197