Seminar

Interstellar gas

Katarina RoskarSupervisor: Dr. Tomaz Zwitter

Faculty of Mathematics and Physics, Department of Physics

9.4.2003

Contents

1 Introduction 2

2 Atomic Hydrogen 32.1 The Neutral Hydrogen - H I Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2.1.1 The Neutral Hydrogen 21-cm Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.1.2 The Neutral Hydrogen Lyman α Line . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 The Ionized Hydrogen - H II Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3 Molecular Gas 93.1 The Formation of Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103.2 Molecular Hydrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103.3 Observation of Molecular Gas Clouds - Radio Spectroscopy . . . . . . . . . . . . . . . . . . . 103.4 Sources of Gas Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

4 Planetary Nebulae 12

5 Supernova Remnants 14

6 The Hot Corona of the Milky Way 15

7 Conclusion 16

1

Abstract

In this seminar we discuss the main characteristics of interstellar gas in the galactic disk; its constitution,distribution (source) in the interstellar space, physical properties and methods of observation. Interstellargas consists mainly of hydrogen, and for that reason we restrict ourselves mainly to the properties of theatomic and molecular state of hydrogen. We also take an insight into gas regions connected with differentstages of stellar evolution and the larger gas envelope that surrounds the Milky Way.

1 Introduction

The mass of gas in interstellar space is a hundred times larger than that of dust, and nine times smaller thanthat of Milky way’s stars (it constitutes about 10% of the mass of the Milky way). Like interstellar dust,gas is concentrated in a thin disc in the galactic plane. The thickness of the gas layer is about twice thatof the dust, or about 200pc1. The Milky Way is a spiral galaxy and interstellar gas is therefore especiallyconcentrated in the spiral arms. The average density of interstellar gas in the disc is 1atom/cm3 = 1015gasatoms/km3 (about 1013 times larger than the density of dust), but the distribution is very inhomogeneous.Typically, there are denser regions, a few parsecs in size, where the densities may be 10-100atoms/cm3.Although there is more gas, it is less easily observed, since the gas does not cause general extinction of light(for more about extinction see for example [1]). In the optical region, it can only be observed on the basisof a small number of spectral lines.

The existence of interstellar gas began to be suspected in the first decade of the 20th century, when in1904 Johannes Hartmann observed that some absorption lines in the spectra of certain binary stars were notDoppler shifted by the motions of the stars like the other lines. It was concluded that these absorption lineswere formed in gas clouds in the space between the stars and the Earth. In some stars there were severalcomponents of the same line apparently formed in clouds moving with different velocities. The strongestabsorption lines in the visible region are those of neutral sodium in singly ionized calcium. In the ultravioletregion, the lines are more numerous. The strongest one is the hydrogen Lyman α line.

On the basis of the optical and UV lines, it has been found that many atoms are ionized in interstellarspace. This ionization is mainly due to ultraviolet radiation from stars and, to some extent, to ionizationby cosmic rays. Since the density of interstellar matter is very low, the free electrons only rarely encounterions, and the gas remains ionized.

About 30 elements have been discovered by absorption line observations in the visible and ultravioletregion. With a few exceptions, all elements from hydrogen to zinc (atomic number 30) and a few additionalheavier elements have been detected (see Table (1) and Figure (1)).

Like in the stars, most of the mass is hydrogen (about 70%) and helium (almost 30%). On the otherhand, heavy elements are significantly less abundant than in the Sun and other population I stars2. It isthought that they have been incorporated into dust grains, where they do not produce any absorption lines.The element abundances in the interstellar medium (gas and dust) would then be normal, although theinterstellar gas is depleted in heavy elements. This interpretation is supported by the observation that inregions where the amount of dust is smaller than usual, the element abundances in the gas are closer to theSolar.

11pc = 3.0857 · 1016m2Relatively young stars, containing a larger fraction of elements other than hydrogen and helium, that move almost circularly

in the galactic plane. The Sun is a Population I type star.

2

Atomicnumber

Name Chemicalsymbol

Interstellarabundance

Solarabundance

Abundanceratio

1 Hydrogen H 1 000 000 1 000 000 1.00

2 Helium He 85 000 85 000 1

3 Lithium Li 0.000 051 0.001 5a 0.034

4 Beryllium Be < 0.000 070 0.000 012 < 5.8

5 Boron B 0.000 074 0.004 6a 0.016

6 Carbon C 74 370 0.20

7 Nitrogen N 21 110 0.19

8 Oxygen O 172 660 0.26

9 Fluorine F – 0.040 –

10 Neon Ne – 83 –

11 Sodium Na 0.22 1.7 0.13

12 Magnesium Mg 1.05 35 0.030

13 Aluminium Al 0.0013 2.5 0.000 52

14 Silicon Si 0.81 35 0.023

15 Phosphorus P 0.021 0.27 0.079

16 Sulfur S 8.2 16 0.51

17 Chlorine Cl 0.099 0.45 0.22

18 Argon Ar 0.86 4.5 0.19

19 Potassium K 0.010 0.11 0.094

20 Calcium Ca 0.000 46 2.1 0.000 22

21 Scandium Sc – 0.001 7 –

22 Titanium Ti 0.000 18 0.055 0.0032

23 Vanadium V < 0.003 2 0.013 < 0.25

24 Chromium Cr < 0.002 0.50 < 0.004

25 Manganese Mn 0.014 0.s26 0.055

26 Iron Fe 0.28 25 0.011

27 Cobalt Co < 0.19 0.032 < 5.8

28 Nickel Ni 0.006 5 1.3 0.005 0

29 Copper Cu 0.000 64 0.028 0.023

30 Zinc Zn 0.014 0.026 0.53a Meteoritic abundance

Table 1: Element abundances in the interstellar medium in the direction of ζ Ophiuchi and in the Sun. The abundances aregiven relative to that of hydrogen, which has been defined to be 1 million. The last column gives the ratio of the abundancesin the interstellar medium and in the Sun.[2]

2 Atomic Hydrogen

2.1 The Neutral Hydrogen - H I Regions

Regions where atomic hydrogen is predominantly neutral are known as H I regions (in contrast to H IIregions where hydrogen is partly ionized).

2.1.1 The Neutral Hydrogen 21-cm Line

In extremely diffuse interstellar gas of neutral hydrogen the forbidden transition3 of neutral hydrogen canoccur. The spins of the electron and proton in the neutral hydrogen atom in the ground state may be eitherparallel or antiparallel. The energy difference between these two energy levels will give rise to a spectralline at the wavelength of 21.049cm with t1/2 = 11 · 106yrs (see Figure (2)). The existence of the line wastheoretically predicted by Hendrick van de Hulst in 1944, and was first observed by Harold Ewen and EdwardPurcell in 1951. Studies of this line have revealed more about the properties of the interstellar medium than

3The state of an electron cannot change arbitrarily; transitions are restricted by selection rules. Most probable are theelectric dipole transitions, all other transitions are much less probable - forbidden.

3

Figure 1: Element abundances in the interstellar medium in the direction of ζ Ophiuchi (dark blue) and in the Sun (orange).The abundances are given relative to that of hydrogen in logarithmic scale. The oscillation of abundances can be explained asit is for stars like the Sun.

Figure 2: The origin of the forbidden hydrogen 21-cm line. The spins of the electron and proton may be either parallel orantiparallel. The energy of the former state is slightly larger. The wavelength of a photon corresponding to a transition betweenthese states is 21cm.

any other method - one might even speak of a special branch of 21-cm astronomy. The spiral structure androtation of the Milky Way and other galaxies can also be studied by means of the 21-cm line.

Usually the hydrogen 21-cm line occurs in emission. Because of the large abundance of hydrogen, it canbe observed in all directions in the sky. An example of an observed 21-cm line profile is shown in Figure (3).For the 21-cm line hν/k = 0.07K, and thus hν/kT � 1 for all relevant temperatures. One may thereforeuse the Rayleigh-Jeans approximation of the Planck’s law for blackbody radiation:

Bν(T ) =2ν2kT

c2, (1)

where Bν is the intensity of radiation at a frequency ν and temperature T . We introduce τν , the opticalthickness of the medium (e.g. an interstellar cloud) at frequency ν which is defined by

ανdr = dτ (2)

αν is the absorption coefficient or opacity at frequency ν. In certain directions in the Milky Way there isso much hydrogen along the line of sight that the 21-cm line is optically thick, τν � 1. If τν � 1, the

4

Figure 3: An example of a hydrogen 21-cm emission line profile in the galactic plane at longitude 83◦ and latitude 52◦.[3] Eachpeak corresponds to a component. Altogether three distinct components were found searching trough a number of directions(different longitudes l and latitudes b). The horizontal axis gives radial velocity according to the Doppler formula (because thebroadening of the 21-cm spectral line is always due to gas motions either within the cloud (turbulence) or of the cloud as awhole). The vertical axis gives the temperature obtained assuming that at some wavelength λ the flux density Fλ on the surfaceof the star obeys Planck’s law.

time between collisions is 400 years on the average4, whereas the time for radiative transitions is 11 millionyears, the distribution of atoms in different levels is almost completely a result of mutual collisions of theatoms only. Thus will the temperature that gives the observed population of energy levels be the same asthe kinetic temperature. In that case the temperature we get from (the approximation of) the Planck’s lawis equal to the kinetic temperature of the gas.

The distance to a source can not be obtained directly from the observed emission. Thus one can onlystudy the number of hydrogen atoms in a cylinder with a 1-cm2 base area extending from the observer tooutside the Milky Way along the line of sight. This is called the projected or column density and is denotedby N . One may also consider the column density N(ν)dν of atoms with velocities in the interval (v,v + dv).

It can be shown that if the gas is optically thin (τν � 1), the temperature (obtained from Planck’slaw) in a spectral line is directly proportional to the column density N of atoms with the correspondingradial velocity. Hence, if the diameter L of a cloud along the line of sight is known, the gas density can bedetermined from the observed line profile:

n = N/L . (3)

The diameter L can be obtained from the apparent diameter, if the distance and shape of the cloud areassumed known.

The distances of clouds can be determined from their radial velocities by making use of the rotationof the Milky Way. Thus if the observed peaks in the 21-cm line profiles (Figure (3)) are due to individualclouds, their distances and densities can be obtained. Since radio observations are not affected by extinction,it has been possible in this way to map the density distribution of neutral hydrogen in the whole galacticplane. The latest resulting distribution[4] is shown in Figure (4). The hydrogen 21-cm line may also occur inabsorption, when the light from a bright distant radio source, e.g. a quasar5, passes through an interveningcloud.

The same cloud may give rise to both an absorption and an emission spectrum. In that case, thetemperature, optical thickness and hydrogen content of the cloud can all be derived.

4This is the average value for hydrogen gas in the disk; n ≈ 1atom/cm3, T ≈ 100K.5Quasar is an abbreviation for quasistellar radio source. Some prefer the use of the designation QSO (quasistellar object)

since not all quasars emit radio radiation. They are known for very large redshifts an extremely large luminosities (1038-1041W ).Most radio quasars are point sources (the emitting region is not larger than a few light days), others have double structure.

5

Figure 4: Neutral hydrogen column density plot in atoms/cm2 in the galactic plane (lower picture showing the enlargedcentral region).[4] The position of the Sun (at ≈ 8.5kpc) and the galactic centre are marked.

6

2.1.2 The Neutral Hydrogen Lyman α Line

The conditions in interstellar space are such that all hydrogen atoms are in the ground state with n = 1.The energy needed for the transition of the neutral hydrogen atom from the ground state to the firstexited state is 10.21eV . Such photons are rare in the interstellar space. Nevertheless the Lyman α lineis a strong absorption line, the strongest interstellar absorption line in the ultraviolet region, whereas theBalmer absorption lines (in the visible part of the spectrum), which arise from the excited initial staten = 2, are unobservable6. This absorption line corresponds to the transition of the electron in the hydrogenatom from a state with principal quantum number n = 1 to one with n = 2; λ = 121.6nm (equivalentwidth7 more than 1nm) shown on Figure (5). Ultraviolet observations have provided an excellent way of

Figure 5: Interstellar absorption lines in the ultraviolet spectrum of ζ Ophiuchi. The strongest line is the hydrogen Lyman αline. The roman number I after the letter denotes the ground state, II the first exited state, and so on. The observations weremade with Copernicus satellite.[5]

studying interstellar neutral hydrogen. Observations have to be made with the use of satellites becauseEarth’s atmosphere absorbs these wavelenghts. The first observations of the interstellar Lyman α line weremade from a rocket already in 1967. More extensive observations comprising 95 stars were obtained by theOAO 2 satellite. The distances of the observed stars are between 100 and 1000 parsec.

Comparison of the Lyman α observations with observations of the 21-cm line have been especiallyuseful. The distribution of neutral hydrogen over the whole sky has been mapped by means of the 21-cmline. However, the distances to nearby hydrogen clouds are difficult to determine from these observations.In the Lyman α observations one usually knows the distance to the star in front of which the absorbingclouds must lie.

The average gas density within about 1kpc of the Sun derived from the Lyman α observations is about0.7atoms/cm3. Because the interstellar Lyman α line is so strong, it can be observed even in the spectra ofvery nearby stars. For example, it has been detected by the Copernicus satellite in the spectrum of Arcturus,whose distance is only 11pc. The deduced density of neutral hydrogen between the Sun and Arcturus is0.02 − 0.1atoms/cm3. Thus the Sun is situated in a clearing in the interstellar medium (see Figure(4)),where the density is less than 1/10 of the average density.

If a hydrogen atom in its ground state absorbs radiation with wavelength smaller than 91.2nm (which isthe boundary wavelength of the Lyman series), it will be ionized. Knowing the density of neutral hydrogen,one can calculate the expected distance a 91.2-nm photon can propagate before being absorbed by theionization of a hydrogen atom. Even in the close neighbourhood of the Sun, where the density is exceptionallylow, the mean free path of a 91.2-nm photon is only about 1pc and that of a 10-nm photon a few hundredparsec. Thus only the closest neighbourhood of the Sun can be studied in the extreme ultraviolet (XUV)spectral region.

6The Balmer lines are strong in stellar atmospheres with temperatures of about 10, 000K, where a large number of atomsare in the first excited state.

7The equivalent width of a spectral line is defined so that the observed line and the totally dark rectangular line of this widthhave the same area; same total absorption.

7

2.2 The Ionized Hydrogen - H II Regions

In many parts of space, hydrogen does not occur as neutral atoms, but is ionized. This is particular aroundhot stars, which radiate strongly in the ultraviolet. If there is enough hydrogen around such a star, it willbe visible as an emission nebula of partly ionized hydrogen. Such nebulae are known as H II regions (seeFigures (6) and (7)). A typical emission nebula is the Great nebula in Orion, M42. In the middle of the

Figure 6: The Great Nebula Orion (M42, NGC 1976) can be found just below and to the left of the easily identifiable belt ofthree stars in the popular constellation Orion. This fuzzy patch contains one of the closest stellar nurseries, lying at a distanceof about 1500 light years. The red region on the left consists of nebulae designated M42 and M43 (NGC 1977) and contains thebright Trapezium open cluster. The dark regions are opaque dust clouds in front of the nebula. The blue region on the right isa nebula primarily reflecting the light from bright stars inside. Recent observations of the Orion Nebula by the Hubble SpaceTelescope have located solar system-sized star-forming regions.[6]

Figure 7: The Lagoon Nebula (M8, NGC 6523) in Sagittarius (is named ”Lagoon” for the band of dust seen to the left ofthe open cluster’s center). This H II region contains many hot stars and young stars that have not yet entered the hydrogenburning phase. They show their high temperature by their blue glow. Small, round dark nebulae, globules, are also visibleagainst the bright background. These are gas clouds in the process of condensation into stars. The Lagoon Nebula lies about5000 light-years away and is about 100 light-years across. [6]

nebula is a group of four hot stars known as the Trapezium. The Trapezium stars emit strong ultravioletradiation, which keeps the gas nebula partly ionized.

Unlike a star, a cloud of partly ionized gas has a spectrum dominated by few narrow emission lines. Thecontinuous spectrum of H II regions is weak. In the visible region there are strong hydrogen Balmer emissionlines (for historical reasons usually denoted Hα (n = 3→ n = 2), Hβ (n = 4→ n = 2), Hγ (n = 5→ n = 2),etc.). These are formed when a hydrogen atom recombines into an excited state and subsequently returnsto the ground state via a sequence of radiative transitions. Typically, a hydrogen atom in a H II regionremains ionized for several hundred years. Upon recombination it stays neutral for some months, beforebeing ionized again by a photon from a nearby star.

The number of recombinations per unit time and volume is proportional to the product of the densitiesof electrons and ions:

nrec ∝ neni . (4)

8

In completely ionized hydrogen, ne = ni, and hence:

nrec ∝ n2e (5)

Most recombinations will include the transition n = 3 → 2 i.e. will lead to the emission of a Hα photon.Thus the surface brightness of an nebula in the Hα line will be proportional to the emission measure EM ,

EM =∫

n2edl , (6)

where the integral is along the line of sight through the nebula.The ionization of a helium atom requires more energy than that of a hydrogen atom, and thus regions

of ionized helium are formed only around the hottest stars. In these cases, a large H II region will surrounda smaller He+ or He2+ region. The helium lines will then be strong in the spectrum of the nebula.

Although hydrogen and helium are the main constituents of clouds, their emission lines are not alwaysstrongest in the spectrum. At the beginning of the last century, it was even suggested that some strongunidentified lines in the spectra of nebulae were due to the new element nebulium. However, in 1927, Ira S.Bowen showed that they were forbidden lines of ionized oxygen and nitrogen, O+, O2+ and N+. Forbiddenlines are extremely difficult to observe in the laboratory, because their transition probabilities are so smallthat at laboratory densities the ions are de-excited by collisions before they have had time to radiate. Inextremely diffuse interstellar gas, collisions are much less frequent, and thus there is a chance that an excitedion will make the transition to a lower state by emitting a photon.

Because of interstellar extinction of light, only the nearest H II regions can be studied in visible light.At infrared and radio wavelengths, much more distant regions can be studied. The most important lines atradio wave lengths are recombination lines of hydrogen and helium; thus the hydrogen transition betweenenergy levels n = 110 and n = 109 at 5.01GHz has been much studied. These lines are also importantbecause with their help, radial velocities, and hence (using the galactic rotation law), distances of H IIregions can be determined, just as for neutral hydrogen.

The physical properties of H II regions can also be studied by means of their continuum radio emission.The intensity of the radiation is proportional to the emission measure EM defined in equation (6). H IIregions also have a strong infrared continuum emission. This is thermal radiation from dust inside thenebula.

H II regions are formed when a hot star begins to ionize its surrounding gas. The ionization steadilypropagates away from the star. Because neutral hydrogen absorbs ultraviolet radiation so efficiently, theboundary between the H II region and the neutral gas is very sharp. In a homogeneous medium the H IIregion around a single star will be spherical, forming a so-called Stromgren sphere. The temperature of aH II region is higher than that of the surrounding gas, and it, therefore tends to expand. After millions ofyears, it will eventually merge with the general interstellar medium.

3 Molecular Gas

The first interstellar molecular gases were discovered in 1937-1938, when molecular absorption lines werefound in the spectra of some stars. Three simple diatomic molecules were detected: Methylidyne CH, itspositive ion CH+ and cyanogen CN. A few other molecules were later discovered by the same method inthe ultraviolet. Thus molecular hydrogen H2 was discovered in the early 1970’s, and carbon monoxide CO,which had been discovered by radio observations, was also detected in the ultraviolet. Molecular hydrogenis the most abundant interstellar molecule, followed by carbon monoxide. About 120 different moleculeshave been identified.

Since the chemical bond of molecules is sensitive to high-energy radiation and high temperatures moleculesare found in cool astronomical environments and where dust is abundant. Dust shields the molecules fromthe stellar UV radiation, which would otherwise destroy them.

9

3.1 The Formation of Molecules

The molecules are formed through a complicated network of chemical reactions inside the clouds where theyare found. The most important type of reaction is that between a positive ion and a neutral molecule inthe gas phase8, for example O+ + H2 → OH+ + H. Positive ions exist everywhere in space, because theyare constantly being created when fast cosmic-ray particles collide with atoms and molecules and tear offelectrons in the process. In lack of long-range electrostatic interaction, reactions between two neutrals aregenerally much slower, but may nevertheless work in some cases. It is also clear that the surfaces of theinterstellar dust grains have the ability to act as catalysts for certain reactions.

3.2 Molecular Hydrogen

The detection and study of molecular hydrogen has been one of the most important achievements of UVastronomy. Molecular hydrogen has a strong absorption band at λ = 105nm, which was first observed in arocket experiment in 1970 by George R. Carruthers, but more extensive observations could only be madewith the Copernicus satellite. The observations showed that a significant fraction of interstellar hydrogen ismolecular, and that this fraction increases strongly for denser clouds with higher extinction. In clouds withvisual extinction larger than one magnitude essentially all the hydrogen is molecular. Hydrogen moleculescan only be formed on the surface of interstellar grains, which thus act as a chemical catalyst.

It is of interest to know whether gas and dust are well mixed or whether they form separate clouds andcondensations. UV observations have provided a reliable way of comparing the distribution of interstellargas and dust. The amount of dust between the observer and a star is obtained from extinction of stellarlight. Furthermore, the absorption lines of atomic and molecular hydrogen in the ultraviolet spectrum ofthe same star can be observed. Thus the total amount of hydrogen (atomic and molecular) between theobserver and the star can also be determined. Observations indicate that the gas and dust are well mixed.

3.3 Observation of Molecular Gas Clouds - Radio Spectroscopy

Absorption lines can only be observed if there is a bright star behind the molecular gas cloud. Because ofthe large dust extinction, no observations of molecules in the densest clouds can be made in the optical andultraviolet spectral regions. Thus only radio observations are possible for these objects, where molecules areespecially abundant.

Radio spectroscopy signifies an immense step forward in the study of interstellar molecules. In theearly 1960’s it was still not believed that there might be more complicated molecules than diatomic onesin interstellar space. It was thought that the gas was too diffuse for molecules to form and that any thatformed would be destroyed by ultraviolet radiation. The first molecular radio line, the hydroxyl radicalOH, was discovered in 1963. Many other gas molecules have been discovered since then. In 1993, over 80molecules had been detected, the heaviest one being the 13-atom molecule HC11N.

Molecular lines in the radio region may be observed either in absorption or emission. Radiation fromdiatomic molecules like CO (see Figure (8)) may correspond to three kinds of transitions.

Electronic transitions correspond to changes in the electron cloud of the molecule. These are like thetransitions in single atoms, and their wavelengths lie in the optical or UV region. Vibrational transitionscorrespond to changes in the vibrational energy of the molecule. Their energies are usually in the infraredregion. Most important for radio spectroscopy are the rotational transitions, which are changes of therotational energy of the molecule. Molecules in their ground state do not rotate, i.e. their angular momentumis zero, but they may be excited and start rotating in collisions with other molecules. For example, carbonsulfide CS returns to its ground state in a few hours by emitting a millimetre region photon.

A number of interstellar gas molecules are listed in Table (2). Many of them have only been detectedin the densest clouds (mainly the Sagittarius B2 cloud at the galactic centre), but others are very common.

8The reason why these reactions play such an important role is that the net charge carried by the ion gives the ability topolarize and electrostatically attract the electrons in the neutral reactant.

10

Figure 8: Radio map of the distribution of carbon monoxide 13C16O in the molecular cloud near the Orion Nebula. Thecurves are lines of constant intensity.[7]

Molecule Name Year of discoveryDiscovered in the optical and ultraviolet region:

CH methylidyne 1937

CH+ methylidyne ion 1937

CN cyanogen 1938

H2 hydrogen molecule 1970

CO carbon monoxide 1971Discovered in the radio region:

OH hydroxyl radical 1963

NH3 ammonia 1968

H2O water 1969

H2CO formaldehyde 1969

CO carbon monoxide 1970

HCN hydrogen cyanide 1970

CS carbon monosulfide 1971

SiO silicon monoxide 1971

SO sulfur monoxide 1973

(CH3)2O dimethyl ether 1974

HCOOH formic acid 1975

C2H5OH ethanol 1975

HC11N cyanopentacetylene 1981

HCCNC isocyanoacetylene 1991

H2CCCC cumulene carbene 1991

Table 2: Some molecules observed in the interstellar medium.[2]

The most abundant molecule H2 cannot be observed at radio wavelengths, because it has no suitable spectrallines9. The next most abundant molecules are carbon monoxide CO the hydroxyl radical OH and ammoniaNH3, although their abundance is only a small fraction of that of hydrogen. However, the masses ofinterstellar clouds are so large that the number of molecules is still considerable. (The Sagittarius B2 cloudcontains 1027l of ethanol C2H5OH.)

As a rule of the thumb, the more complex a molecule, the smaller its relative concentration in theinterstellar clouds. Some molecules were not known on Earth at the time of their detection in space andrequired real detective work. Moreover, at the physical conditions prevailing the interstellar clouds, not allrotational energy levels are populated and therefore far from all lines are observable. Larger, asymmetricmolecules can have considerably more complicated spectra and the intensity of each line will generallydecrease with increasing molecular complexity. Identification of complex molecules is a delicate process thatrequires very careful evaluation of large number of matching lines in conjunction with a solid knowledge of

9Because the molecule is very light, the first excited rotational level is as high as 500K above the ground rotational state.Thus it does not emit radio radiation under normal interstellar conditions.

11

the physical conditions in the observed source. This delays their discovery (see Table (2)).

3.4 Sources of Gas Molecules

Most of the molecules in Table (2) have only been detected in dense molecular clouds occurring in connectionwith H II regions. Almost every molecule yet discovered has been detected in Sagittarius B2 near the galacticcentre. Another very rich molecular cloud has been observed near the H II region Orion A, i.e. the well-known Orion Nebula M24 (shown on Figure (6)).

Inside the actual H II region there are no molecules, since they would be rapidly dissociated by the hightemperature and strong ultraviolet radiation. Three types of molecular sources have been found near H IIregions (see Figure(9)):

large gas and dust envelopes around the H II region The large envelopes have been discovered pri-marily by CO observations. OH and H2CO have also been detected. Like in the dark nebulae, thegas in these molecular clouds is probably mainly molecular hydrogen. Because of the large size anddensity (n = 103 - 104molecules/cm3) of these clouds, their masses are very large, 103 or even 106

solar masses (Sgr B2). They are among the most massive objects in the Milky Way. The dust inmolecular clouds can be observed on the basis of its thermal radiation. Its peak falls at wavelengthsof 10-100µm, corresponding to a dust temperature of 30-100K.

small dense clouds inside these envelopes

very compact OH and H2O maser sources Some interstellar clouds contain maser sources. In these,the emission lines of OH, H2O and SiO may be many million times stronger then elsewhere. Thediameter of the radiating regions is only about 5-10AU10. The conditions in these clouds are suchthat radiation in some spectral lines is amplified by stimulated emission as it propagates through thecloud. Hydroxyl and water masers occur in connection with dense H II regions and infrared sources,and appear to be related to the formation of protostars. In addition, maser emission (OH, H2O andSiO) occurs in connection with Mira variables11 and some red supergiant stars12. The maser emissioncomes from a gas molecule and dust envelope around the star, which also gives rise to an observableinfrared excess.

4 Planetary Nebulae

Bright regions of ionized gas do not occur only in connection with newly formed stars, but also around starsin late stages of their evolution. The planetary nebulae are gas shells around small hot blue stars. The starsthat begin to pulsate because of instabilities in the stage of helium burning, may eject the whole atmosphereinto space. A gas shell expanding at 20-30km/s will be formed around a small and hot (50, 000 - 100, 000K)star, the core of the original star.

The planetary nebulae were given their name in the 19th century, because certain small nebulae visuallylooked like planets such as Uranus. The apparent diameter of the smallest known planetary nebula is onlya few arc seconds, whereas the largest ones may be one degree in diameter (the Helix Nebula is quite large,shown on Figure (10)). The expanding gas in planetary nebulae is ionized by ultraviolet radiation from thecentral star, and its spectrum contains many of the same bright emission lines as that of an H II region.The brightest emission lines are often due to forbidden transitions. For example see Figure (11).

101AU = 1.49597870 · 1011m11Mira variables are red supergiants with large light variation that has a period of 100-500days. They are losing gas in a

steady stellar wind. Teff ≈ 2000K.12Supergiants (M ≈ 30solar masses) are the biggest, brightest, and most luminous stars. Surface temperature of a red star

is ≈ 3000K.

12

Figure 9: Infrared map of the central part of the Orion Nebula. In the lower part are the four Trapezium stars. Aboveis an infrared source of about 0.5′ diameter, the so-called Kleinmann-Low nebula (KL). BN is an infrared point source, theBecklin-Neugebauer object. Other infrared sources are denoted IRS. The large crosses indicate OH masers and the small crossesH2O masers. On the scale of Figure (6) this region would be less than a millimetre in size. [2]

Figure 10: The Helix Nebula (NGC 7293) is estimated to be a mere 450 light-years from the Sun, in the direction of theconstellation Aquarius. The Helix Nebula is the closest example of a planetary nebula created at the end of the life of aSun-like star. The outer gasses of the star expelled into space appear from our vantage point as if we are looking down a helix.The remnant central stellar core, destined to become a white dwarf star, glows in light so energetic it causes the previouslyexpelled gas to fluoresce. In this colour image the nebula glows red in the light of nitrogen and hydrogen atoms energized bythe ultraviolet radiation from the central star. The main rings themselves, though faint, have an angular size about half that ofthe full Moon and span about 1.5 light-years. Because the nebula is so close, it is a prime subject for study by astronomers.[6]

Planetary nebulae are generally much more symmetrical in shape than most H II regions, and theyexpand more rapidly. For example, the well-known Ring Nebula in Lyra (M57) has expanded visibly inphotographs taken at 50-year intervals. In a few ten thousand years, the planetary nebulae disappear in thegeneral interstellar medium and their central stars cool to become white dwarfs13.

The total number of planetary nebulae in the Milky Way has been estimated to be 50, 000. About 1, 000have actually been observed.

13The final stage of a star of mass 0.08solar mass ≤ M ≤ 0.26solar mass. Has no internal sources of energy. All the hydrogenhas been burnt to helium. It is a degenerate star; its radius decreases as the mass increases.

13

Figure 11: This planetary nebula’s simple, graceful appearance is thought to be due to perspective - our view from planetEarth looking straight into what is actually a barrel-shaped cloud of gas shrugged off by a dying central star. The greencolour of the central parts of the Ring Nebula in Lyra is due to the forbidden lines of doubly ionized oxygen at λ = 495.9 andλ = 500.7nm. The red colour of the outer parts is due to the hydrogen Balmer α line (λ = 656.3nm) and the forbidden lines ofionized nitrogen (λ = 654.8nm, λ = 658.3nm). Dark, elongated structures can also be seen near the nebula’s edge. The RingNebula is about one light-year across and 2,000 light-years away in the northern constellation Lyra.[6]

5 Supernova Remnants

Massive stars end their evolution in a supernova explosion. The collapse of the stellar core leads to theviolent ejection of the outer layers, which then remain as an expanding gas cloud.

About 120 supernova remnants (SNR’s) have been discovered in the Milky Way. Some of them areoptically visible as a ring or an irregular nebula (e.g. the Crab Nebula; see Figure (12)), but most aredetectable only in the radio region (because radio emission suffers no extinction). These differences are due

Figure 12: The Crab Nebula (M1, NGC 1952) in Taurus is the remnant of a supernova explosion. The nebula is also a strongradio source. Its energy source is the central rapidly rotating neutron star, pulsar, which is the collapsed core of the originalstar and that emits visible flashes of light.[6]

to the different emission processes in H II regions and in SNR’s. In an H II region, the radio emission is free-free radiation from the hot plasma. In a SNR it is synchrotron radiation from relativistic electrons movingin spiral orbits around the magnetic field lines. The synchrotron process gives rise to a continuous spectrumextending over all wavelength regions. For example, the Crab Nebula looks blue in colour photographsbecause of optical synchrotron radiation.

In the Crab Nebula, red filaments are also visible against the bright background. Their emission isprincipally in the hydrogen Hα line. The hydrogen in SNR is not ionized by a central star as in the H IIregions, but by the ultraviolet synchrotron radiation.

The supernova remnants in the Milky Way fall into two classes. One type has a clearly ringlike structure(e.g. Cassiopeia A or the Veil Nebula in Cygnus; see (13) and (14)); another is irregular and bright at themiddle (like the Crab Nebula). In the remnants of the Crab Nebula type, there is always a rapidly rotatingpulsar at the centre. This pulsar provides most of the energy of the remnant by continuously ejecting

14

Figure 13: Pictured above is the west end of the Veil Nebula known technically as NGC 6960 but less formally as the Witch’sBroom Nebula. The rampaging gas gains its colours by impacting and exciting existing nearby gas. The supernova remnant liesabout 1400 light-years away towards the constellation of Cygnus. This Witch’s Broom actually is over three times the angularsize of the full Moon. The bright blue star 52 Cygnus is unrelated to the ancient supernova.[6]

Figure 14: Cataloged as NGC 6992, these glowing filaments of interstellar shocked gas are part of a larger spherical supernovaremnant known as the Cygnus Loop or the Veil Nebula - expanding debris from a star which exploded over 5,000 years ago.[6]

relativistic electrons into the cloud. The evolution of this type of SNR reflects that of the pulsar and forthis reason has a timescale of a few ten thousand years.

Ringlike SNR’s do not contain an energetic pulsar; their energy comes from actual supernova explosion.After the explosion, the cloud expands at a speed of 10, 000 - 20, 000km/s. About 50-100 years after theexplosion the remnant begins to form a spherical shell as the ejected gas starts to sweep up interstellar gasand to slow down in its outer parts. The swept-up shell expands with a decreasing velocity and cools until,after about 100, 000 years, it merges into the interstellar medium.

6 The Hot Corona of the Milky Way

As early as 1956, Lyman Spitzer showed that the Milky Way has to be surrounded by a large envelope ofvery hot gas. Almost two decades later The Copernicus satellite, whose scientific programme was directedby Spitzer, found evidence for this kind of gas, which began to be called galactic coronal gas, in analogywith the solar corona. The satellite observed emission lines of e.g. five times ionized oxygen (O VI), fourtimes ionized nitrogen (N V) and triply ionized carbon (C IV). The formation of these lines requires a hightemperature (100, 000 - 2, 000, 000K), and a high temperature is also indicated by the broadening of thelines.

Galactic coronal gas is distributed through the whole Milky Way and extends several thousand parsecsfrom the galactic plane. Its density is only of the order of 10−3atoms/cm3 (recall that the mean density inthe galactic plane is 1atom/cm3). Thus coronal gas forms a kind of background sea, from which the denserand cooler forms of interstellar matter, such as neutral hydrogen and molecular clouds, rise as islands.

In the early 1980’s the IUE satellite also detected similar coronae in the Large Magellanic Cloud and in

15

the spiral galaxy M100. Coronal gas is probably quite a common and important form of matter in galaxies.Supernova explosions are probably the source of both coronal gas and its energy. When a supernova explodes,it forms a hot bubble in the surrounding medium. The bubbles from neighbouring supernovae will expandand merge, forming a foamlike structure. In addition to supernovae, stellar winds from hot stars may providesome of the energy of the coronal gas.

7 Conclusion

Throughout this seminar we have considered the following five phases of interstellar gas:

very cool molecular gas clouds (mostly molecular hydrogen H2) T = 20K, n > 103/cm3

cool gas clouds (mostly neutral atomic H I regions) T = 100K, n = 20/cm3

warm neutral gas enveloping the cooler clouds T = 6000K, n = 0.05-0.3/cm3

hot ionized gas (mainly atomic H II regions around hot stars) T = 8000K, n > 0.5/cm3

very hot and diffuse ionized coronal gas, ionized and heated by supernova explosions T = 106K,

n = 10−3/cm3

To conclude, we should mention another very important phase of interstellar clouds, the one that is adjacentto the early phase of stellar evolution.

Interstellar clouds play a significant role in star formation. Stars are formed by the collapse of interstellarclouds (mostly located in the spiral arms of the Galaxy) that are massive enough for gravity to overwhelmthe pressure. The limiting mass is the Jeans mass. In a typical interstellar neutral hydrogen cloud, n =106atoms/m3 and T = 100K, giving the Jeans mass, 30, 000solar masses. It is thought that star formationbegins in clouds of a few thousand solar masses in diameters of about 10pc. Although the view that starsare formed in such a way is generally accepted, many details of the fragmentation process are still highlyconjectural[8]. Thus the effects of rotation, magnetic fields and energy input are very imperfectly known.Why a cloud begins to contract is also not certain; one theory is that passage through a spiral arm compressesclouds and triggers contraction. This would explain why young stars are predominantly found in the spiralarms of the Milky way and other galaxies. The contraction of an interstellar cloud might also be initiatedby a nearby expanding H II region or supernova explosion. Understanding the processes which lead to theformation of stars is one of the fundamental challenges in astronomy.

16

References

[1] K. Roskar, “Seminarska naloga Medzvezdni prah.” Dec 2002. 1

[2] H. Karttunen, P. Kroger, H. Oja, M. Poutanen, and K. Donner, Fundamental Astronomy. Springer-Verlag Berlin Heidelberg New York, 3rd ed., 1996. 1, 2, 9

[3] G. L. Verschuur and A. L. Peratt, “Galactic Neutral Hydrogen Emission Profile Structure,” The Astro-nomical Journal, vol. 118, pp. 1252–1267, Sep 1999. 3

[4] H. Nakanishi and Y. Sofue, “Three Dimensional HI Gas Distribution in the Milky Way Galaxy,” in 8thAsian-Pacific Regional Meeting, pp. 179–180, Dec 2002. 4, 4

[5] D. S. Morton, “Interstellar absorption lines in the spectrum of zeta Ophiuchi,” Astrophysical Journal,vol. 197, pp. 85–115, Apr 1975. 5

[6] “http://observe.arc.nasa.gov/nasa/gallery/image gallery/universe/universe neb.html.”. 6, 7, 10, 11, 12,13, 14

[7] M. L. Kutner, N. J. Evans, and K. D. Tucker, “A dense molecular cloud in the OMC-1/OMC-2 region,”Astrophysical Journal, vol. 209, pp. 452–457, Oct 1976. 8

[8] R. Klessen, “Towards a Consistent Theory of Star Formation,” in Astronomische Gesellschaft MeetingAbstracts, p. 1, 2000. 7

17