The idea that the early Sun was surrounded by a rotating flattened nebula or disk out of which the planets formed has had a long history. However, a detailed application of the disk model to pre-main-sequence stars was not made until the seminal work of Lynden-Bell and Pringle (1974). These authors suggested that the excess emission of the low-mass, pre-main-sequence T Tauri stars could be powered by disk accretion; the extended dusty disk accounts for the excess infrared emission of T Tauri stars, while the hot gas predicted at the boundary layer between the star and disk produces the observed ultraviolet continuum emission. Lynden-Bell and Pringle further suggested that T Tauri disks could be quite massive, and might even outshine the central star in their early stages.

In retrospect, researchers in the field were not ready for these insights, partly due to the observational limitations of the time. Complicating the situation, ultraviolet and X-ray observations with the IUE and Einstein satellites in the late 1970s showed that young stars exhibit high-temperature chromospheric and coronal emission at much higher levels than observed on the Sun (e.g., Gahm et al. 1979; Cram et al. 1980; Walter & Kuhi 1981), undoubtedly as a result of solar-type magnetic activity. Thus, it was natural to assume that this excess optical and ultraviolet emission represented the extreme youthful limit of solar magnetic activity (Herbig 1970; Cram 1979; Calvet et al. 1983).

The topic of star and protoplanetary disk formation touches almost every area of astrophysics, from galaxy formation to the origin of the solar system. Our understanding of the early evolution of stars has advanced substantially in the last few years as a result of improved observational techniques, particularly in the infrared and radio spectral regions. Although many fundamental problems of star formation remain to be solved, so much has been learned in the last decade about pre-main-sequence accretion processes that an attempt to outline the emerging picture of low-mass star formation seems justified.

In this book I have tried to provide a discussion of accretion in early stellar evolution which can be used at a variety of levels: as an introduction to the subject for advanced graduate students; as a reference for researchers in star formation; and as an overview for scientists in other, related fields. The text assumes a basic familiarity with astronomical concepts and graduate-level physics, though I have made some effort to include some astronomical definitions and references to fundamental physical equations needed for my development. I have adopted a point of view close to that of my own research, which is generally near the interface between theory and observation, and so have tried to discuss basic physical concepts in relation to observational results. Many plausible and even aesthetically pleasing theories have been constructed which have failed to meet observational tests.

As described in the previous chapter, theory predicts that the protostellar envelope should collapse at near free-fall velocities to form the stellar core and disk. It is not easy to detect this collapse directly; on large scales, the infall velocities are small and difficult to isolate from the complex supersonic motions in the surrounding cloud, while freeze-out or other chemical effects remove some of the standard tracers in the inner envelope. In addition, the presence of high-velocity bipolar outflows cause further confusion. Nevertheless, an increasing body of evidence generally supports the rapid collapse model of protostar formation.

Infrared imaging and spectroscopy have provided the most broadly based indications of protostellar collapse. While dust emission does not directly measure infall motion, the presence of dust in the near environs of very young stars, as shown either directly in scattered light or inferred through detection of warm dust emission, demands a dynamical explanation. As the material is too cold to be thermally supported, it must be either falling in or flying out (the envelopes are mostly not in flattened disks, though somewhat flattened “toroids” are observed). The required infall rates are plausible from the collapse theories discussed in the previous chapter, while wind mass loss rates would have to be implausibly large (because the same density implies a larger mass flux for a larger velocity) – and, in any event, outflows have distinctive bipolar geometries, not toroidal expansion.

The remarkable eruptive FU Orionis objects found in star-forming regions are important to our understanding of protostellar accretion disk physics. The best-studied FU ors provide the clearest examples of the SED of an optically-thick accretion disk, with observations spanning a decade in wavelength or more. The very high accretion rates of these accretion disk systems imply that the MRI can easily operate through thermal ionization, at least out to radial distances of nearly an AU. In addition, the high temperatures mean that gaseous spectral lines are present which can be used to infer rotation, turbulence, disk surface mass ejection (Chapter 10), and even chromospheric activity. Clues to the magnitude of angular momentum transport can be derived from the timescales of variability of these systems. Finally, the eruptive behavior and high accretion rates for short periods of time provide an unexpected insight into how mass is added to stars during early stellar evolution.

Acceptance of the accretion disk model proposed by Paczynski (1976), Lin and Papaloizou (1985), and Hartmann and Kenyon (1985) was slowed by the optical appearance of FUors, with spectra similar to that of a G supergiant (except rapidly rotating, an otherwise unknown set of objects). An early model for FUors attempted to explain the substantial near-infrared excesses (Cohen & Woolf 1971; Rieke et al. 1972; Simon et al. 1972; Grasdalen 1973; Simon 1975) by invoking such rapid rotation that the equatorial regions were much more extended, and thus cooler, than the polar regions (Mould et al. 1978).

What do we mean by primeval? According to the Webster dictionary “Primeval: adj. [primaevus, from: primus first + aevum age] of or relating to the earliest ages (as of the world or human history)”. We will follow this definition and mostly discuss topics related to galaxies in the “early” Universe, whose limit we somewhat arbitrarily define at redshifts z ≳ 6, corresponding approximately to the first billion years (Gyr) after the Big Bang. In contrast the frequently employed adjective “primordial”, defined as “Primordial: adj. [primordialis, from primordium origin, from primus first + ordiri to begin] a) first created or developed b) existing in or persisting from the beginning (as of a solar system or universe) c) earliest formed in the growth of an individual or organ”, should not be used synonymously, for obvious reasons. Luckily “primeval” encompasses more than “primordial”, otherwise there would not be much in the way of observational aspects to discuss (now in 2006–2007) in these lectures!

If we follow the history of discoveries of quasars and galaxies over the last few decades it is indeed impressive to see how progress has been made in detecting ever-more-distant objects, increasing samples at a given redshift and their analysis and interpretation. During the last decade, approximately since the pioneering observations of the Hubble Deep Field in 1996 (Williams et al. 1996) and the spectroscopic studies of a large sample of star-forming galaxies at redshift 3 by Steidel and collaborators (Steidel et al. 1996), the observational limits have continuously been pushed further, reaching now record redshifts of z ∼ 7 (secure) (Iye et al. 2006) but maybe up to ∼ 10 (Pelló et al. 2004; Richard et al.

Introduction

Ultimately, the overwhelming majority of emission-line sources in the Universe are “galactic sources” - meaning discrete objects located within a particular galaxy (rather than some global property of a galaxy or some source not located in a galaxy). However, the most common of these, HII regions, are so ubiquitous that they are being covered elsewhere in this volume as the “baseline” source of emission lines. In addition, most of the other chapters are devoted to line emission either integrated over entire galaxies (or significant portions thereof) or from active galactic nuclei.

Given that coverage, I will focus this chapter primarily on “stellar” sources of line emission in the Milky Way other than HII regions - including young stellar objects, massive and/or evolved stars, and stellar remnants (planetary nebulae, supernova remnants, and accreting compact objects in binary systems). I will also put considerable emphasis on emission lines with rest wavelengths in the near-infrared waveband, due to the importance of this waveband for probing the dusty planar regions of the Milky Way where most of these sources are to be found.

In the sections below, I will begin with a review of important diagnostic optical emission lines and a more-detailed overview of key (rest-wavelength) infrared emission lines. I will then move on to “nebular” sources of emission lines (omitting HII regions, but including planetary nebulae and supernova remnants).

Star formation in our galaxy at present occurs in dense, cold clouds of molecular gas. The efficiency of star formation is generally low. While there are localized regions of high efficiency which produce bound star clusters, typical star formation efficiencies of nearby molecular cloud complexes are a few percent by mass. For a long time it was thought that this low efficiency of converting gas to stars was due to the slowing of gravitational collapse by magnetic fields. However, most nearby molecular clouds of significant mass harbor young stars, with typical ages of a few Myr, indicating that star formation is relatively rapid and that (at least local) molecular clouds are not long-lived, casting doubt on the importance of magnetic fields. The low efficiency of star formation is not due to the slowing gravitational collapse but to the disruption of molecular gas by stellar energy input, particularly from massive stars, which disperse clouds before all the mass can collapse.

The processes by which large molecular clouds fragment into molecular cloud cores, the precursors of protostars, are not yet well understood. Lower-density gas must be concentrated into dense regions, often filamentary in structure, with low(er) “turbulent” motions. Numerical simulations show that supersonic turbulence can create protostellar cores with the aid of rapid cooling and even thermal instability, which then gravitationally collapse.

Much of what we presently know or surmise about the physical processes involved in star formation is derived from the detailed study of a few nearby molecular cloud “nurseries”. Stars in the solar neighborhood are formed from the gravitationally induced collapse of cold molecular gas. Typical molecular gas clouds must contract by a factor of a million in linear dimensions to form a star. Because of this dramatic (and rapid) reduction in size, any small initial rotation of the star-forming cloud is enormously magnified by conservation of angular momentum during collapse. In this way a modestly rotating gas cloud produces a rapidly rotating object – a disk – in addition to a small, stellar core at the end of gravitational collapse. Probably most of the material of a typical star is accreted through its disk, with a small amount left behind to form planetary systems.

Advances in observational techniques spanning the electromagnetic spectrum have been essential in developing our present understanding of star formation. The launch of the Infrared Astronomy Satellite (IRAS) in 1983 led to the recognition that dusty disks are common around young stars. The ISO infrared satellite provided detailed mid-infrared spectra of many bright disks. The Spitzer Space Telescope, the latest in this line of infrared observatories, has now detected mid-infrared disk emission in very large samples of stars spanning a wide range of ages.

Introduction

Charles Fabry, who was born in 1867, specialized in optics and devised methods for the accurate measurement of interference effects. He worked with Alfred Pérot, during 1896–1906, on the design and uses of a device now known as the Fabry-Pérot interferometer, which was specifically designed for high-resolution spectroscopy, and is composed of two thinly silvered glass plates placed in parallel, producing interference due to multiple reflections.

In 1899 they described the Fabry-Pérot interferometer which enabled high-resolution observation of spectral features (Fabry & Pérot 1899). It was a significant improvement over the Michelson interferometer. The difference between the two lies in the fact that in the Fabry-Pérot design multiple rays of light reflected by the two plane surfaces are responsible for the creation of the observed interference patterns. The last sentence of the article reads We must emphasize the simplicity of the apparatus used and the ease with which it can be mounted at the telescope. When the silvering has been carefully selected, the interference apparatus does not cause the loss of much light and permits the study of objects of very feeble brightness.

Definition of a Fabry-Pérot interferometer

Basically, a Fabry-Pérot interferometer or etalon (from the French étalon, meaning “measuring gauge” or “standard”) is typically made of a transparent plate with two reflecting surfaces, or two parallel highly reflecting mirrors (technically the former is an etalon and the latter is an interferometer, but the terminology is often used inconsistently).

The discovery of many extrasolar planetary systems over the last decade, most with properties considerably different from those of the solar system, has revolutionized thinking about the processes of planet formation. An entire book could be devoted to the vast literature that has arisen on this complex topic. The aim of the present chapter is limited to outlining a few relevant astrophysical constraints on disk evolution and to sketch some of their potential implications for planet formation.

Our current lack of understanding of angular momentum transport in protoplanetary disks is a major obstacle to understanding how planets form. The presence or absence of a dead zone (§7.6) can strongly affect dust coagulation/planetesimal growth rates by piling up material over time; the activation or lack of same of the MRI will affect disk turbulence, which in turn affects the rates at which dust settles and grows in the disk. Astrophysical clues to these processes are meager at present, though there is evidence that settling and growth of small dust particles is at least somewhat limited by turbulence.

Dust disks within a few to 20 AU of the central star tend to “clear” (i.e., become undetectable in infrared excesses) on timescales of a few Myr. Individual systems exhibit a wide range of clearing timescales; some low-mass stars of ages ∼ 1 Myr have no detectable disks, while of order 5% of stars of ages ∼ 10 Myr exhibit substantial disk emission.

Initially it seemed like a good idea to revise this book, because so much has been learned about star and planet formation over the last ten years. It eventually became clear that it was a bad idea to revise this book, because so much has been learned about star and planet formation over the last ten years. By then I was halfway through and it was too late to back out.

I therefore beg the reader's indulgence for things I have left out or treated schematically. At some point in a project like this “the best is the enemy of the good”, as Voltaire apparently said; just give up and send it off. Perhaps there is some value in having a treatment that does not try to cover everything in an enormous tome, but instead provides accessible points of departure. As was the case for the first version, I hope that this will be a useful reference for non-specialists as well as a starting point for researchers entering the field.