So far we have talked about stars in the solar neighborhood. But where did they come from, especially the luminous ones which were not here when the dinosaurs ruled the earth? Is there something between the stars?

The answer is yes. The interstellar medium is an excellent vacuum but there is about one atom per cm³ and a thousand or so specks of dust per cubic kilometer. Not much, but space is big.

[Earth atmospheric atom density at sea level is 10¹⁹ cm^-3.]

The gas is primarily H ("X") and He ("Y") and the dust probably H and anything left over such as C,N,O,Mg,Si,Fe... ("Z").

INTERSTELLAR DUST

About one percent of the mass of the interstellar medium is in the form of solid (frozen) material called interstellar dust or grains or sometimes cosmic dust. A typical grain might be 50 nm across (vision peaks at 555 nm) and weigh in at a billion atoms.

Why talk about dust before gas? Because the dust is what we "see" or more precisely obscures what we expect to see. If you look at the Milky Way you often see spaces that were originally suspected of being holes or tunnels in the galaxy. But how come all these "tunnels" point to earth? Clouds of dust between us and the rest of the milky way explains this and if you look just below kappa Crucis you can see the Coalsack, a remarkable "hole" in the Milky Way. The thing to know is that the predominant gas is completely transparent while dust is very efficient at blocking light.

DARK NEBULAE

The Milky Way has numerous dark patches and lanes caused by dust blocking the light behind them. In fact you could read a book by the light of the Milky Way if it were not for this obscuration. Some dark nebulae have been named, the Coalsack, the Emu, the Horsehead. The North America nebula is actually an opening in a dusty region. Many gaseous nebulae have associated dust lanes and sometimes small dark regions called Bok globules, possible protostar nurseries, overlie brighter nebulae. Evidently the interstellar medium has large density fluctuations.

REFLECTION NEBULAE

The tiny grains absorb about half the light that strikes them and scatters the rest. So if you look at a star through a cloud it will appear dim but if the star is to one side some of the scattered light can be seen. The Coalsack is actually a bit below the plane of the Milky Way and while it is pretty black compared with the stellar background it has about 10% the background brightness due to reflected Milky Way light. The brighter Pleiades stars are wrapped in dusty shrouds from a cloud of dust that they are travelling through. Most bright nebulae have some reflected light but often the gaseous emission swamps the reflection.

GENERAL OBSCURATION

More insidious is the dust that is just everywhere. It does not stand out, but you know that the farther away a star is the more dust the light must penetrate and the fainter the star will be than would be predicted by the inverse square law. So the distance modulus of a star will be overestimated unless we account for extinction. If A is the absorption in magnitudes at some wavelength then the distance modulus will become

m-M=5log(R/10)+A

and we might hope to extract some formula for A such as A=kR. Unfortunately, the distribution of dust is rather non uniform. But there is another way...

INTERSTELLAR REDDENING

The minute dust particles scatter blue light much more than red causing distant objects to appear much redder than they really are. That is, red light penetrates dust more easily than blue. (So infrared astronomy is good for looking through dusty regions.) This causes the color index of an object to increase, and we call the difference between the observed and intrinsic color index the color excess. CE=CI(observed)-CI(intrinsic)=CI^o-CI. By measuring the extinction of stars reddened by different amounts, we have found that for the V band and (B-V) color, A_V=3.2CE=3.2[(B-V)^o-(B-V)]. The 3.2 factor is an empirical value and is subject to change!

COLOR COLOR (TWO COLOR) DIAGRAMS

If we make a graph of (U-B) vs. (B-V) then the theoretical main sequence takes on a characteristic "S" shape. Due to the differential extinction between the two colors any reddening will displace the position of a star down and to the right on a "reddening line". Color color diagrams are rather useful for plotting clusters and associations since the reddening stands out like a sore thumb.

POLARIZATION

The light emitted by "ordinary" stars is unpolarized. But highly reddened stars show up to 10% polarization. We associate the polarization with the dust and suspect the dust particles are irregular and somehow align themselves along the galactic magnetic field. Now this is a problem in astronomy that needs further study!

R COR BOR STARS

There is a class of red irregular variable stars that suddenly dim in the visible and brighten in the infrared, just like what one would expect if the atmosphere became cool enough for "soot" to form. Perhaps this is what happens, and the soot driven into the interstellar medium. The dust has to come from somewhere!

NATURE OF THE GRAINS

The wavelength dependence of interstellar extinction, reflection, polarization, the existence of interstellar molecules, all give clues to the nature of the interstellar dust. The average interstellar extinction curve yields much information, including the fact that there must be a distribution of particle sizes. There is a gradual rise in the ultraviolet with a bump in the absorption curve at 200nm followed by a shallow dip then rise beyond the Lyman limit, 100nm where no light penetrates. The bump could be due to ~20nm graphite particles whose carbon bonds absorb at this wavelength. In the infrared we see absorption bands at 3.1 microns, possibly due to silicates and water ice, and 9.7 microns, possibly due to water ice.

We suspect refractory particles such as polycyclic aromatic hydrocarbons or PAH's, graphite and Si/SiOx/Mg/Si/Fe particles form in the outer layers of cool supergiants and are driven out by stellar winds.

Ices are another matter, these may form as a mantle around a refractory particle inside dense molecular clouds (see below).

INTERSTELLAR GAS

Most of the interstellar medium is composed of gas. But the gas is transparent over a broad spectral range despite the fact there are a trillion times more atoms than dust grains. Unless you know just where to look.

INTERSTELLAR ABSORPTION LINES

Most O and B stars exhibit very uncharacteristic sharp, often multiple, absorption lines of ionized calcium, the Ca II "H" and "K" resonance lines around 394nm. This is evidence that calcium, some or all of it in the first stage of ionization, lies between us and the star. The gas must be cold (or we would see emission lines) so there is no thermal doppler broadening as there would be in a star photosphere. And often the star shines through several clouds each with its own peculiar radial velocity. If the gas is cold why is it ionized? There is a lot of soft UV in interstellar space provided by early stars and very little Ca. If a Ca atom is ionized, recombination is unlikely because of the low gas density. So there are a lot of cold ionized Ca atoms along a kiloparsec of the interstellar medium. Where is the resonance fluorescence? It is there, just so feeble you cannot see it. You see the absorption because every atom along your line of sight to the star absorbs the spectrum of the star. Along any other line of sight the chances of an ion being excited and re-emitting in your direction are minute. Most Ca in the interstellar medium is singly ionized and in the ground state, just what is needed to produce what we see. Other strong absorption features include resonance lines of Ca I, Ti I, Ti II, Na I and even molecular lines of CN and CH. In the UV resonance lines of many metals are seen.

EMISSION NEBULAE

This may be what got you interested in astronomy. Only a fraction of the interstellar medium emits visible light but color photographs of these regions are stunning.

In the vicinity of early stars the copious UV flux shortward of the Lyman limit (91.1 nm) ionizes most atoms including H. (But not He whose ionization potential is a whopping 24.48 eV, almost twice that of H.) The hydrogen "uses up" these hard UV photons out to a distance that is a function of the gas density and the UV flux produced by the star. Beyond this distance -there is a fairly sharp boundary- the hydrogen remains neutral. Bengt Stromgren studied these nebulae and the region of ionization around young stars is known as a Stromgren sphere or H II region. The radius of the Stromgren sphere tells a lot about the UV flux of the star and density of the interstellar region.

It is usual to break the interstellar medium into the usual H I regions where H is not ionized and H II regions. The H II regions are quite visible through the recombination spectrum of hydrogen, and spectacular contrast is obtained by taking photographs through a filter that isolates a Balmer line such as H alpha.

Here is what happens: A Lyman continuum photon (wavelength shorter than 91.1 nm, energy greater than 13.6 eV) ionizes a hydrogen atom. The ion (proton in this case) recombines with an electron to some state with emission of a photon, a process we call photorecombination.

If recombination is back to the n=1 level, a Lyman photon is emitted that will be absorbed by some other H atom, not too interesting.

If the recombination is to the n=2 level with the emission of a Balmer continuum soft UV photon (wavelength less than 347 nm) the atom will decay to n=1 with the emission of a Lyman alpha photon, not enough to ionize H any more so the Lyman continuum photon has been essentially destroyed.

If recombination is to the n=3 level, the atom can decay to n=1 with the emission of a Lyman beta photon or to n=2 with the emission of H alpha followed by Lyman alpha.

Recombinations to higher levels give rise to more and more lines that make up the hydrogen recombination spectrum. The important thing to note is that the fluorescence process degrades the original UV flux and the volume of the Stromgren sphere is determined by how much Lyman flux the star produces and the H atom density.

Why not a "helium Stromgren sphere"? Sure, if the star is hot enough it will have enough UV with photon energy exceeding 24.48 eV you could define such a thing. It would probably be inside the H II region and show the He recombination spectrum. Yes, we do see such things but we don't give them a formal name.

Oxygen spheres? And others? Like the helium, the more easily ionized C, N, O and Ne show recombination lines just like one would expect. The important thing to note is that around a hot star the interstellar medium fluoresces and we see the beautiful phenomenon of an emission nebula.

Now for the unexpected. We expect to see recombination lines of O III and N II if we look hard enough but some of these lines are just too strong! I. S. Bowen noted that the very strong "Lyman alpha line" of He II at 30.378 nm happened to lie very close to a resonance line in O III. Now we expect most O III to be in the ground state but the He II line "pumps" the O III which then fluoresces by the light of this miraculous coincidence. So it looks like Nature gave us the first practical example of optical pumping. We call this the Bowen fluorescent mechanism.

But wait, there's more. Many strong lines just do not seem to be produced by any element in the lab. Like helium, whose lines were first seen in the solar spectrum (that's how the element got its name, it was found in the sun first), it seemed there was a new element nebulium. But actually the lines were from ordinary elements, just seen in extraordinary circumstances. I. S. Bowen provided the correct explanation of these lines in 1927: Some of the strongest lines seen in these emission nebula arise from forbidden transitions. Actually, a very special type of forbidden transition we call phosphorescence. See About Atomic Energy Levels. In brief, the lower energy levels of such atoms (ions) as N II, Ne III, S II all have the same parity and have level separations of several eV so [transitions] occur in the visible. The square braces around [transitions] is to get you used to the idea that we indicate forbidden lines with square brackets. If such an ion is excited by electron collision, how is it going to get back to the ground state? The transition is forbidden by the parity rule. We call such levels metastable but the term "forbidden" is not quite correct. It really should be "not very likely" but we keep using forbidden. The not very likely transitions can and do occur and such forbidden fluorescence is called phosphorescence. So why is this surprising? It isn't but remember, people like to see familiar things and these lines just do not appear in laboratory spectra. One could, in principle, produce these lines but you need a discharge tube several hundred AU long! Other places to see forbidden lines are the aurora and in certain phosphorescent powders (glow in the dark paint).

TEMPERATURES AND DENSITIES OF H I AND H II REGIONS

The kinetic temperature (kinetic energy divided by Boltzmann's constant) of a gas is related to the average speed of the gas particles. Away from hot stars in H I regions space is pretty cold. The atoms radiate into about a three degrees Kelvin heat sink but receive some energy from the stars of the Milky Way and (slowly) strive for equilibrium. Typical temperatures inferred from radio spectral line analysis indicates these regions are several hundreds of degrees K but with a wide range from rare (not dense) hot H I regions down to dense cold clouds which measure 20K.

The H II regions should be pretty hot but strangely enough forbidden radiation acts as a thermostat that cuts in strongly above 6000K. Ionization kicks energetic electrons (energy/k is temperature, an energy of 1 eV corresponds to 11604 K) which rapidly thermalize so heating the gas. The hot electrons also excite metastable states of atoms which, when they finally do radiate, simply remove the excitation energy -- the forbidden line cannot be reabsorbed -- and the hotter the gas the more efficient this removal process becomes. Most H II regions have temperatures measured from analysis of forbidden lines in the range 8000 to 10000 K.

The density is another matter. Nature abhors a pressure difference. So the hot regions tend to expand giving low density compared to the H I regions. Nature strives for pressure equilibrium, N₁T₁=N₂T₂, and generally succeeds. So the young stars sort of blow "bubbles" in the interstellar medium and the edges are compressed to higher densities than would be expected in equilibrium. Remember this when we consider star formation. Cold clouds on the other hand form regions of high density which, we will see, helps them cool further.

NEUTRAL HYDROGEN CLOUDS

Cold H I clouds are regions of relatively high density (N goes as 1/T) and that includes metals such as Ca, Na, etc whose resonance lines absorb the light from stars behind the cloud. These sharp interstellar lines are easily distinguished from the broad stellar lines and can be used to estimate the cloud density. Interestingly, the first recognition of these lines was from an analysis of the spectrum of a distant spectroscopic binary. Most of the lines shifted alternately up and down the spectrum as expected but a few lines did not shift. Often the interstellar lines are double or triple indicating more than one cloud along the line of sight. So where is the hydrogen?

THE 21 CM LINE

If you look at the ground (Lyman) level of hydrogen really hard you will find not one but two states separated in energy by [5.873 micro eV, 1420 MHz, 21.11cm].

The energy difference can be thought of in the following manner. The proton and electron each have a unit charge and "spin-half" and these "spinning charges" act like tiny electromagnets. When the electron and proton are bound they can either have their magnetic poles in the same direction (spins opposite or proton spin "up", electron spin "down") or opposed (spins in the same direction). Just like two bar magnets, if the poles are in the same direction they try to "flip" because "like poles repel". A hydrogen atom in its lowest energy state can be bumped by another hydrogen atom collisionally exciting the atom to the higher energy state. The radiative transition back to the ground state is strongly forbidden, an excited atom decays with a half life of 12 million years! The 21 cm line is strongly forbidden so once produced the galaxy is completely transparent to the radiation. But not radiotelescopes tuned to 21 cm!

The spin flip radiation was predicted by van de Hulst in Leyden during the Nazi occupation in 1944 then observed by Ewen and Purcell at Harvard in 1951. The 21 cm line is one of the most powerful tools the astronomer has for studying the distribution and properties of hydrogen in this and other galaxies.

Observations at 21 cm show that the gas is confined to a thin flat 100 pc sheet in (defining) the plane of the Milky Way. Cold clouds in the sheet have temperatures of order 100 K and densities of 50 atoms per cc with diameters from 1 to 10 pc. Cold clouds take up about two percent of the sheet. Warm clouds with temperatures of 2000 K and densities 1 atom per cc up to 5000 K (but still H I) take up about twenty percent.

INTERSTELLAR MOLECULES

We see interstellar absorption lines of CH, CH+ and CN but the big breakthrough in interstellar chemistry came in 1963 when the hydroxyl radical OH was detected after the characteristic frequencies were firmly established in the laboratory. The line can appear very strong when an OH region is in the vicinity of an infrared "pumping" source giving rise to maser emission, sort of a Bowen mechanism for radio sources. Five years later interstellar water vapor and ammonia were seen and since then over a hundred molecular species have been identified including common household items such as formaldehyde, acetone, ethyl and methyl alcohol.

Molecules are easily broken up by the dilute UV radiation that permeates space so it is no surprise that the molecules are associated with dense dust clouds which act as a shield. Moreover, we suspect the molecules are actually formed on the surface of dust grains but the whole process of grain growth and molecular formation is an astrochemical mystery.

CORONAL GAS

Space based ultraviolet observatories have discovered another component of the interstellar medium mostly found well above the plane of the galaxy, which show themselves through interstellar absorption resonance lines of five times ionized oxygen (O VI) at 103.2 and 103.8 nm. We infer temperatures of a million degrees K and densities of a few times 10^-3 atoms per cc with large variations on a small scale. Supernova remnants?

THE SOLAR NEIGHBORHOOD

The sun is located right in the middle of a hot bubble sometimes referred to as the Local Bubble. The bubble temperature is about a million K and density about 0.005 atoms per cc, and extends some 100 pc from the Sun. We hope this is unusual, and some suggest a recent (million years) supernova is responsible. There really should be more diffuse H I clouds in the Local Bubble. The Sun itself is inside a tenuous cloud with a density of about 0.1 atoms per cc and temperature of 10000 K, the Local Fluff. Towards the center of the galaxy about 20 pc distant there may be a more respectable warm cloud and the Local Fluff may be the edge of this cloud. We are still working on this!

MODELS OF THE INTERSTELLAR GAS

We have the following features to explain:

We see the interstellar medium is far from uniform, but we expect that the ISM strives for pressure equilibrium so the hot regions are rarified and the cold regions dense. That equilibrium is not exact can be seen in pictures of nebulae such as the Horsehead nebula where one type of cloud appears to be mixing with another. How do we make these various components?

If we put an O star in the intercloud medium we might get a Stromgren sphere one or two hundred pc across. No oproblem. After the star dies we are left with a hot H I cloud.

If the star becomes a cool red giant it may produce dust grains and blow dust into the interstellar medium so producing a region of dust. The dust can radiate and provide an additional cooling source for the region giving a cool dust cloud. Now for the tricky part, The dust can adsorb gas and so form hydrogen molecules and perhaps build up a shell of ice around the original grain. The spectrum of interstellar absorption suggests or at least does not preclude this. Furthermore, interstellar molecules are found in association with dust clouds. Larger grains radiate ever more efficiently and a dense cloud of dust can cool to tens of K, the cold clouds. These cold clouds can become giant molecular factories and giant molecular clouds can be the most massive objects in the galaxy.

Finally, if a star undergoes a supernova explosion a huge shock wave can plow through the interstellar medium heating the shock zone to millions of K which can be seen in radio and X-ray images. Examples are the Loop nebula, the Veil and Tycho's supernova remnant (SNR).