Less than a few hundred thousand years after the Big Bang, the temperature was high enough that cosmic gas consisted of protons, free electrons and light nuclei. Once the Universe cooled to about 3000 K, the electrons and protons were moving sufficiently slowly that they combined to form hydrogen atoms. With scattering of photons much reduced, they were able to move in straight lines indefinitely, and may be seen redshifted into the microwave part of the spectrum as the 2.7K CMB. So began the era of recombination, or so-called “dark ages” when the IGM became mostly neutral. Within the current cold Dark Matter model for the hierarchical formation of structure, mini-halos of mass ∼106M⊙ (Couchman & Rees 1986) provided the gravitational seeds for the first stars at z ≈ 20–30, ending the “dark ages” through re-ionization of the IGM. A comprehensive review of the astrophysical role of dark matter is provided by Jungman, Kamionkowski, & Griest (1996).

Galaxies formed as baryonic gas cooled in the centers of dark matter structures, from which galaxy mass built up via mergers of halos and proto-galaxies (White & Rees 1978; Davis et al. 1985). Since most present-day galaxies are relatively old, it follows that they formed at z ≥2. The timescale over which galaxies assembled remains unclear, particularly the bulges and disks which are the main components of present-day galaxies.

A detailed discussion of stellar atmospheres is beyond the scope of this book. Nevertheless, our means of studying the properties of hot massive stars relies upon our ability to properly interpret the stellar continuum and line information typically formed in the thin boundary layer between the unseen interior and effectively vacuum interstellar medium. An excellent monograph on the topic of stellar photospheres is provided by Gray (2005), whilst more advanced techniques are introduced by Mihalas (1978).

With respect to normal stars, our interpretation of hot, luminous stars is hindered by two effects. Firstly, the routine assumption of LTE breaks down for high-temperature stars, and particularly for supergiants, due to the intense radiation field, such that the solution of the statistical rate equations (non-LTE) is necessary. Secondly, the simplifying assumption of plane-parallel geometry is no longer valid for blue and red supergiants, so the scale heights of their atmospheres are no longer negligible with respect to their stellar radii. It is the combination of requiring non-LTE plus spherical geometry that has prevented the routine study of OB star atmospheres until recently.

LTE atmospheres

Effective temperatures of early-type stars, essential for subsequent determinations of radii and luminosities, are derived from a comparison between observed photometry and/or spectroscopy and models. Surface gravities also require comparison between observed line profiles and models.

LTE model atmospheres developed by Robert Kurucz during the 1970s and 1980s account very thoroughly for metal line blanketing and are widely employed for both early- and late-type stars.

Massive stars are born in interstellar clouds made up of molecular gas and dusty material. Most of these stars originate from GMCs with typically ∼105M⊙. Upon collapse, these lead to massive star clusters. Some massive stars are born separate from these massive concentrations of gas and dust in smaller clouds and end up in more compact star clusters. Truly isolated massive stars seem to be rare in our Galaxy. Indeed, de Wit et al. (2005) argue that only a few percent of massive stars are born away from clusters. All these massive stars are found highly concentrated towards the Galactic plane where current star formation is still proceeding, albeit at a relatively restrained rate at present.

Consider first an individual massive star. It is formed when gravitational instability overwhelms a cloudlet of gas and dust which then begins a process of collapse. The collapse brings more and more material to the central object in a process of heating and rapid accretion. This phase is very rapid and will only be observed at far-IR wavelengths as the central material aggregates, begins to heat up, and emits radiation or molecular emission at radio wavelengths. Very quickly sufficient material accumulates such that an individual object can be identified. As this material continues to heat from the continuous contraction and accretion of more gas and dust it takes on the characteristic of what is called a “hot core”, radiating also now at mid-IR wavelengths.

Several textbooks discuss the interstellar medium in depth, notably The Physics of the Interstellar Medium (Dyson & Williams 1997) and The Physics and Chemistry of the Interstellar Medium (Tielens 2005). This chapter focuses on aspects relevant to massive stars, namely the properties of interstellar dust, ionized nebulae surrounding individual O stars (giant HII regions are discussed later on), wind blown bubbles and ejecta nebulae around LBVs and W-R stars.

Diffuse gas in the ISM may be in a neutral or ionized form, of which 90% is in the form of hydrogen, either in an atomic, molecular, or ionized state. Cold (∼100 K), atomic hydrogen can be traced via the 21 cm (1420 MHz) hyperfine line, first predicted by van de Hulst (Bakker & van de Hulst 1945) and observed by Ewen & Purcell (1951). This provided the key means of mapping out the structure of the Milky Way and external galaxies. The Lyman series of neutral hydrogen can also be observed in the UV against a suitably hot background source.

Most of the cold (∼10 K) molecular ISM is in the form of H2. This molecule, however, does not emit at radio wavelengths. Since H2 is well correlated with carbon monoxide (CO), the CO emission at 1.3 and 2.6 mm is used as a proxy for H2.

Stellar winds are ubiquitous amongst massive stars, although the physical processes involved depend upon the location of the star within the H-R diagram. Mass-loss crucially affects the evolution and fate of a massive star (Chapter 5), while the momentum and energy expelled contribute to the dynamics and energetics of the ISM (Chapter 8). The interested reader is referred to the monograph by Lamers & Cassinelli (1999) on the topic of stellar winds, or Kudritzki & Puls (2000) for a more detailed discussion of mass-loss from OB and related stars.

The existence of winds in massive stars was first proposed by Beals (1929) to explain the emission line spectra of Wolf–Rayet stars. This gained observational support in the 1960s when the first rocket UV missions revealed the characteristic P Cygni signatures of massloss from CIV λ1550, SilV λ1400, and NV λ1240 in O stars (Morton 1967). A theoretical framework for mass-loss in hot stars was initially developed by Lucy & Solomon (1970) involving radiation pressure from lines, and refined by Castor, Abbott, & Klein (1975), thereafter known as CAK theory. The observational characteristics of stellar winds are velocity and density. The former can be directly observed, whilst the latter relies on a varying complexity of theoretical interpretation.

Radiation pressure

When a photon is absorbed or scattered by matter, it imparts its energy, hv, and momentum, hv/c, where h is Planck's constant and c is the velocity of light. Consequently, radiation is a very inefficient carrier of momentum.

Giant HII regions are extensive regions of ionized hydrogen (and other elements) powered by hot stars. Their high concentration of very massive stars often makes them appear as the optically most luminous structures in galaxies. Nearby giant HII regions allow us to study star formation and evolution in great detail and help us understand unresolved giant HII regions in distant galaxies.

Giant HII regions: definition and structural parameters

In this section we will concentrate on the properties of giant HII regions, such as the famous 30 Doradus nebula in the LMC. The focus of this book is on hot, massive stars, and we are interested in HII regions because their properties allow us to learn about the stars powering them. As opposed to smaller, Orion-like HII regions, giant HII regions are sufficiently rich in O stars that the entire upper end of the mass spectrum is sampled and an unbiased view of the hot-star population can be obtained.

Giant HII regions are among the most conspicuous objects in nearby late-type galaxies, in particular when observed in narrow emission lines, such as Hα. Systematic studies of nearby giant HII regions such as that by Kennicutt (1984) have established their fundamental properties. These HII regions typically have diameters of 100 pc or larger, densities of a few particles per cm3, and ionized gas masses of order 105M⊙. In Fig. 9.1 we show a well-known example, NGC 604 in the Triangle galaxy M33.

This monograph had its origins about a decade ago when it became apparent that the field of luminous hot stars was rapidly expanding in extent and depth and connections to extragalactic astrophysics, in particular starburst galaxies, were first recognized. At that time a decade had passed since the 1988 O stars and Wolf–Rayet stars NASA monograph of Conti and Underhill. Since then, there have been far reaching advances in the astrophysics of these stars and of star-forming galaxies, both locally and at high-redshift, together with the way each affect their surroundings. On the observational side, progress interpreting the spectra of luminous hot stars allows their physical parameters to be derived with unprecedented accuracy. Studies of their ubiquitous stellar winds plus their dependence upon the element abundances has provided the impetus for revised stellar evolutionary model calculations. For the first time, additional physics, such as rotational mixing and magnetic fields, is being considered.

The advent of the Hubble Space Telescope in 1990 provided us with a UV wavelength range of greatly increased sensitivity just where most of the energy of hot stars is emitted. The use of newly commissioned 8-m ground-based telescopes has opened the door to studies of more distant hot stars in external galaxies, with different initial abundances and star formation histories compared to the Milky Way. In addition, techniques for the identification of starforming galaxies at an epoch when the Universe was as little as one tenth of its present age have been developed, and exploited with these instruments.

The phrase “starburst” was used before in this volume, referring to a region with intense star formation. In this chapter we will define and discuss starbursts in more rigorous terms and relate them to the properties of the host galaxies. While luminous HII regions are often considered starbursts as well, starburst phenomena are more general and have profound cosmological implications. A major theme of this chapter, in contrast to the local focus of Chapter 9, will be the global interrelation between a starburst region and its environment. The starburst is nourished by the gas supply of the host, and it energizes the galaxy ISM with stellar winds and supernovae.

Definition of a starburst

One of the earliest uses of a term similar to “starburst” goes back to a seminal publication by Searle, Sargent, & Bagnuolo (1973). These authors added data points of many late-type galaxies to a theoretical two-color diagram which plots (B–V) against (U–B) color. Surprisingly, the galaxies had bluer colors than expected for a stellar population either evolving with steady, continuous star formation or with no star formation at all after the formation of the first stellar generation. This immediately suggested that some galaxies are observed in a special epoch: they are currently forming stars at a higher rate than previously, otherwise the bluer colors, indicative of an excess of hot stars, would not be understandable.