Introduction

Observations of infrared line emission from molecular hydrogen in astronomical sources have gone from the novel to the commonplace in a time that is short relative to most timescales for the advancement of astrophysical knowledge. In the face of the current onslaught of observations of H2, it may seem surprising that only a dozen years ago, when the 2 µm lines were first detected, it is reported to have taken their discoverers several months to identify the emitting species. Excited H2 continues to be detected in new and surprising places. In the Galaxy, H2 line emission is found at interfaces between young stellar winds and the interstellar medium, where it first was discovered by Gautier et al. (1976), in reflection nebulae (Gatley et al. 1987), in supernova remnants (e.g., Burton et al. (1988)), even including the Crab Nebula (Graham, Wright, and Longmore 1989), in planetary and proto-planetary nebulae (Treffers et al. 1976), and in the nucleus (Gatley et al. 1984). Beyond the Galaxy H2 line emission is found in Seyfert, as well as starburst galaxies (Thompson, Lebofsky, and Rieke 1978, Joseph et al. 1986, Fischer et al. 1987), in individual HII regions of normal spiral galaxies (Israel et al. 1989), and in interacting and merging galaxies (e.g., Joseph et al. 1986). Closer to home, H2 line emission has been detected very recently in the aurora of Jupiter (Trafton et al. 1988). In all of these examples, because of the unique physical conditions which must be satisfied in order that its infrared line emission be detectable, H2 lines provide important information about environments that are difficult to study by other means.

Introduction

Ultraviolet radiation is a crucial ingredient in any theory of interstellar chemistry. In the interplay of molecule formation and destruction processes, ultraviolet photons adopt a multiple role, destroying neutral species on the one hand, while creating chemically reactive ions and depositing thermal energy on the other. It has long been recognized (e.g. Stief et al. (1972)), that dust in a cloud's outer layers attenuates ambient Galactic ultraviolet starlight, thereby enhancing the survival of molecules against photodestruction. Unfortunately, however, the degree of attenuation is sensitive to the grain scattering properties, which are not well determined at ultraviolet wavelengths (Sandell and Mattila 1975, Leung 1975, Whitworth 1975, Bernes and Sandqvist 1977, Sandell 1978, Flannery, Roberge, and Rybicki 1980). Since even a small amount of ultraviolet radiation has profound consequences in dark regions, the chemical and ionization balance of such regions has remained uncertain.

The early studies of dust shielding may have been overly pessimistic about uncertainties, however, as noted by Chlewicki and Greenberg (1984a,b). This is due in part to the existence of strong constraints on grain properties that follow from secure observational data, and also to the discovery of the chemical consequences of the ultraviolet emission associated with gas-cosmic ray interactions (Prasad and Tarafdar (1983); see also Chapter 16). The interaction produces an ultraviolet field in clouds which, at great depths, destroys molecules more rapidly than cosmic rays or attenuated starlight. As a result, the role of starlight is restricted to a relatively narrow region near a cloud's surface, where the effects of uncertainties in grain properties are moderate.

Introduction

Shock waves are ubiquitous in the interstellar medium (ISM) because efficient radiative cooling allows interstellar gas to cool to temperatures low enough that the sound speed is small compared to the velocities of disturbances in the ISM, such as cloud–cloud collisions, bipolar outflows, expanding HII regions, and supernova explosions. Shock waves in dense molecular gas are almost always radiative: The relative kinetic energy of the shocked and unshocked gas is converted into radiation, and since the radiating gas is dense, it is very bright. Because much of the mass in molecular clouds is obscured by dust, the emission from shocks provides a powerful probe of energetic activity occurring in these clouds. In particular, stars inject large amounts of energy into their surroundings in the process of formation, giving rise to bipolar outflows with velocities in excess of 100 km s−1, characteristic of stellar escape velocities (Lada 1985). Intense maser emission in the 1.35 cm line of water is also observed to be associated with newly formed stars, particularly massive stars, with velocities of tens to hundreds of kilometers per second (Genzel 1986). Understanding the structure and spectrum of the shocks associated with these high velocity flows in dense molecular gas is thus a prerequisite for unraveling the complex physical processes attending the birth of stars.

Early studies of shocks in molecular clouds assumed that the neutrals and ions were tied together into a single fluid, and that the shock front was an abrupt transition on the scale of the molecular mean free path (e.g., Field et al. (1968), Hollenbach and McKee (1979)).

Introduction

Interstellar chemistry began to be studied in a fairly serious way when, in the late 1960s and early 1970s, it was demonstrated that a wide variety of molecules existed in dense molecular clouds. At first the main effort was in identifying the main chemical routes by which molecules were formed and destroyed. It was realized that, even in dark molecular clouds where starlight is excluded, cosmic rays may penetrate and cause ionizations which drive a chemistry which would otherwise ‘run down’. This chemistry would, therefore, be largely one of positiveions and molecules. This early recognition met with great success and – although the level of ionization in molecular clouds remains uncertain – the detection of interstellar ions such as HCO+ and N2H+ is strong support for positive ionneutral molecule chemistry. Models of interstellar chemistry involving hundreds or even thousands of reactions are now routinely studied: some of these reactions may be important.

These early studies, understandably, concentrated on the chemistry. They deliberately made the dynamics as simple as possible. Thus, uniform density and temperature were usually invoked, in geometrically convenient shapes such as semi-infinite slabs, or spheres. Steady-state calculations were often performed, without a full consideration of the applicability of steady-state. Later studies showed that it might take around 30 million years to achieve steady-state in molecular clouds, and it was realised that such extended periods might not be available in interstellar clouds.

At the same time, detailed observations were indicating that molecular clouds were far from simple objects.

Historical perspective

The study of interstellar chemistry started, appropriately, about 60 years ago. In 1926, Eddington discussed in his remarkable Bakerian Lecture the possibility of molecule formation and absorption in dark nebulae. At that time, only atomic species had been identified in interstellar space through their narrow absorption lines superposed on the spectra of background stars. In the next decade, several new interstellar features were detected which could indeed be ascribed to molecules: CH, CH+ and CN.

In spite of this early success, no other molecule was found in the interstellar gas for the next 25 years, until OH was detected in 1963 by its radio emission lines. In the next two decades, more than 70 different interstellar molecules were identified by centimeter and millimeter wavelength techniques. However, these radio emission line studies were mostly concerned with dense and dark clouds, whereas the early absorption line observations probed much more diffuse gas.

Although a wide variety of interstellar molecules has now been detected in dark clouds, still only a handful of molecules has been found in diffuse clouds. The molecules H2, HD, OH and CO were discovered in the 1970s by their absorption lines in the ultraviolet through rocket experiments and by the Copernicus satellite. Since the detection of C2 in 1977 by ground-based techniques, however, no new molecule has firmly been identified in diffuse clouds. The list of molecules sought but not detected is considerably longer and includes such interesting species as NH, HC1, NaH, MgH, H2O and C3 (see van Dishoeck and Black (1988a) for a recent summary).

Introduction

Over the past ten years, observations of the far-infrared and submillimeter fine structure emission from neutral oxygen and carbon, and singly ionized carbon and silicon, have revealed the presence of a dense (nH ≤ 103 cm−3), intermediate temperature (60 ≥ Tgas ≥ 1500 K) component of the interstellar medium in which a substantial fraction of the gas is atomic (see Table 14.1). In addition, observations of atomic fine structure emission from regions which are mostly molecular have helped to define the nature of the shocks which frequently accompany star formation and to challenge long-held beliefs about the structure and chemistry of molecular clouds.

Emission from atomic and ionic species with ionization potentials less than that of hydrogen arises predominantly in diffuse HI clouds, the photodissociated surfaces of molecular clouds, and the warm gas downstream of passing shock waves, generated deep within molecular clouds by the outflows from newly formed stars. The latter two regions, of primary interest in this contribution, are characterized by gas temperatures of 60–2000 K. Since H and He have no lowlying levels which can effectively cool the gas in this temperature range, the fine structure transitions of OI (63, 146 µm), C+ (158 µm), and Si+ (35 µm) generally dominate the cooling and thus serve as important diagnostics of these regions. An indication of the cooling power in just the [OI] 63 µm and [CII] 158 µm lines is given in Table 14.2; between 0.1 and 1% of the total luminosity of these sources escapes in these two lines.

Introduction

The high observed column densities of CH+, one of the first identified (Douglas and Herzberg 1941) interstellar molecules, and of CO apparently indicate that existing static, equilibrium models do not provide adequate descriptions of the natures of diffuse molecular interstellar clouds. (See Chapter 3.) It has been argued that velocity structures in lines formed in such clouds provide evidence for the existence of shocks in them (e.g. Crutcher (1979), but see the detailed assessment by Langer in Chapter 4). If such shocks do exist, they will drive the production of detectable column densities of a number of chemical species.

The chemistry in shocked gas can be exceptionally rich since many reactions which, because they are endothermic or have activation barriers, are unimportant in cool, static gas, can proceed in shocked gas. For instance, the endothermic reactions C+ + H2 → CH+ + H (Elitzur and Watson 1978a) and S+ + H2 → SH+ + H (Millar et al. 1986) can initiate hydrogen abstraction sequences in shocked gas but are unimportant in static, cool diffuse clouds. A neutral–neutral sequence (Aannestad 1973) which is of no relevance to low temperature chemistry but which plays a major role in shock chemistry is O + H2 → OH + H; OH + H2 → H2O + H. The fractional abundances of CH+ and OH are high in some diffuse cloud shocks, and SH+ may serve as a diagnostic of shocks.

Collisionally induced rotational excitation of molecular hydrogen can also occur in diffuse cloud shocks.

Statement of the problem

Most studies of light scattering in planetary rings have assumed layers which are many particles thick, plane parallel, and homogeneous. However, real rings may be thin, vertically warped, and clumpy. We have developed a ray tracing code which calculates the light scattered by an arbitrary distribution of particles. This approach promises to clarify a number of puzzling observations of the Saturnian and Uranian rings.

(1) Many studies have concluded that Saturn's rings are many particles thick (e.g. Lumme et al. 1983), whereas dynamical calculations predict that optically thick rings should be physically thin (Wisdom & Tremaine 1988 and references therein). Lumme et al. argue that the particles in Saturn's B Ring fill only 2% of the volume of the ring, while Wisdom and Tremaine predict a filling factor of 20% or more.

The claim that Saturn's rings are thick is based on their observed opposition surge, a rapid brightening (0.3 mag in the V band) which occurs at phase angles below about 1.5°. The surge is attributed to particles covering their own shadows near opposition. Shadowing can occur either between discrete particles, or within the surface structure of a particle. The range in phase angle over which the brightening takes place is proportional to the volume filling factor of the ring or surface. Thus the very narrow opposition effect of Saturn's rings implies a very porous ring, unless individual particles backscatter extremely strongly.

Simulation

A videomovie, showing a simulation of a disc galaxy perturbed by a companion, has been made. It is easy to see some time-dependent phenomena in the movie that would have been hard to discover on paper plots.

The videomovie shows results from a 2-dimensional N-body simulation with 60 000 particles using a polar coordinate code. The disc is self-gravitating and surrounded by an inert spherical halo. It has Mestel's density distribution, which gives a flat rotation curve, and the companion is represented by a point mass which passes on an initially parabolic orbit.

The initially stable disc evolves as follows: A very sharply defined material arm quickly appears. After some time a counter arm is formed due to self-gravitation. Soon after the appearance of the counter arm, a third arm forms from the remains of the first material arm. This third arm has a higher radial velocity than the counter arm which makes it pass, or expand, through the latter. Some time later, after the passage of the companion, a density wave pattern, much more long lived than the material arms, is formed. The density waves first appear between the material arms giving the impression of a fork in the arm, such as has been observed in several galaxies. As the material arms are decaying when this happens, a fork would be a short lived phenomenon.

Introduction

In two contexts, this talk revisited an old idea – that any concentrated mass orbiting within a shearing disc tends to create an inclined and sometimes quite intense wavelike wake amidst the material that flows past it. More exactly, as actually delivered, this talk began with a fairly lengthy review of such swing-amplified wake-making in the gorgeous global setting of Zang's V = const disc, and only then did it proceed to my two main contexts.

Polarisation in the shearing sheet

One context was the idealized shearing sheet composed of thousands of identical softened mass points, with which I have been conducting a long series of numerical experiments in recent years. As I illustrated with slides and even with a homemade (!) movie, the typical impression there is one of an ever-changing kaleidoscope of “spiral” features, very much suggesting some sort of a recurrent strong instability in a system that really ought not to have any. Appearances aside, it turns in fact that those features are no instabilities. Instead, they are logically just superpositions of the separate wakes of the many random particles, each of which in this sense keeps acting simultaneously as both an aggressor and a victim. It is the collection of these wakes that forever keeps shifting in location and appearance as the individual particles constantly drift in this shear flow, but each particle in turn always carries along a wake of its own.

Introduction

High resolution (≤ 10 km s−1) spectra of a selection of symbiotic stars have been obtained using the Manchester Echelle Spectrograph and IPCS on the 2.5 m Isaac Newton Telescope on La Palma, as a preliminary to a high resolution survey of all known symbiotic stars now being conducted from the INT and ESO. In several cases, the [OIII] 5007 Å region shows complex structure, probably originating in extended outflows. However, the Hα line in many objects shows a well-known double-peaked profile (see e.g. Anderson et al. 1980). This is very reminiscent of those associated with dwarf novae, where observations through eclipse indicate that the emission originates in the accretion disc surrounding the white dwarf component of the semi-detached binary (see e.g. King, this volume).

If this were also the case with symbiotic systems, then theoretical modelling of the line profiles would enable us to constrain the all-important binary parameters. Our preliminary aim, however, is to determine whether such line profiles can be reconciled with emission from accretion discs at all. Anderson et al. (1980) concluded that the case still remained ambiguous.

The model

In order to construct theoretical line profiles from accretion discs, we have adapted the optically thick disc model of Home & Marsh (1986) producing double-peaked profiles which have a deep “V”-shaped central reversal. These match the observed profiles more closely than earlier (optically thin) models (e.g. Smak 1969).

Introduction

There is growing evidence for circumstellar discs associated with young stellar objects (YSOs). Motivated by observational evidence suggesting that these discs produce significant luminosity, LD ∼ L⋆, and have moderate masses, MD ∼ M⋆, (Adams, Lada & Shu 1988), we explore the possibility that the accretion mechanism ultimately owes its origin to the growth of spiral gravitational instabilities. As a start, we study the growth and structure of linear, global, gravitational disturbances in star/disc systems.

The physics of m = 1 modes

For simplicity, we take the unperturbed discs to be infinitesimally thin and in centrifugal equilibrium; we characterize the surface density and temperature profiles in the disc as power-laws in radial distance from the star. Since the potential well of the star dominates that of the disc everywhere except near the disc's outer edge, the rotation curve is nearly Keplerian throughout most of the disc's radial extent.

Our study concentrates on modes with azimuthal wave number m = 1, since these modes can be global in extent and may also be the most difficult modes to suppress in unstable protostellar discs. Modes with m = 1 correspond to elliptic streamlines (i.e. eccentric particle orbits), which play a unique role in Keplerian potentials, a fundamental point explicitly recognized by Kato (1983). In an exactly Keplerian potential, circular streamlines of zero pressure are neutrally stable to kinematic perturbations that make them ellipses.

Observations

While NGC 1068 has received much attention in recent years, little is known of the large-scale dynamics of the ionized gas in this nearby Seyfert galaxy. We have used the Hawaii Imaging Fabry-Perot Interferometer (HIFI, Bland & Tully 1989) at the CFHT to obtain detailed spectrophotometry at 65 km s−1 FWHM resolution for Hα and the [NII]λλ6548, 6583 lines. The final maps are derived from ∼ 100 000 fits to spectra taken at 0.4″ increments over a 200″ field-of-view. The flux-weighted Hα + [NII] velocity field is presented in Figure 1.

Deep images of NGC 1068 reveal an outer θ-shaped ring lying roughly east-west which encompasses a visibly bright, inner disc with diameter ∼ 20 kpc (230″), orthogonal to the outer ring. The Hα line flux is dominated by a luminous, elliptic ring of HII regions with diameter 3 kpc and major axis 45°. The “3 kpc ring” is aligned with an oval, bar-like distortion recently discovered at λ2µ (Scoville et al. 1988). The inner disc is marked by high concentrations of atomic and molecular gas (∼ 1010M⊙) which is thought to fuel the rapid star formation that characterizes the 3 kpc ring (Scoville et al. 1983). Most of the ring bolometric luminosity (1.5 × 1011L°), which is comparable to that of the active nucleus, emerges in the far-infrared as re-radiation from dust heated by young stars (Telesco et al. 1984).

From Figure 1, the large-scale disc appears to undergo flat rotation with V(Rmax) = 170/ sin i where Rmax ≈ 30″.

Introduction

A two stage mechanism for fuelling AGN is proposed which makes use of stellar dynamical and gas dynamical instabilities (Shlosman et al. 1989). First, a stellar bar sweeps the interstellar material inwards as a consequence of the gas losing angular momentum to the bar. In a second stage, the gaseous disc accumulated in the nuclear region of the host galaxy, goes bar unstable again and the material flows further in. The main criterion for the occurence of the second instability is that the mass of the gaseous disc must exceed some critical fraction of the total mass of the host galaxy. This critical mass fraction is of the order of a tenth or so according to our estimates. The inflowing gas may eventually join a viscosity-driven accretion disc if a black hole was already present or lead to its formation. If the host galaxy is relatively gas rich, but the disc formed during the first stage does not exceed the critical mass, or if the inflow of gas is halted around resonances, a nuclear starburst may follow. This mechanism may explain the association of bars and rings with nuclear activity and the dichotomy between AGN and starburst nuclei.

Gas in galaxies

The atomic and molecular gas content of spiral galaxies is observed to be in the range 108 – 1010 M⊙ and to peak at Hubble types Sb-Sbc (Haynes & Giovanelli 1984, Verter 1987).

Introduction

Large-scale magnetic fields could play an important role in the dynamics of astrophysical discs. Here we report some results showing how the structure of non-axisymmetric magnetic fields is affected by differential rotation. A turbulent disc is likely to be surrounded by a gaseous corona. We shall study in particular how the field structure in the disc is affected by surrounding gas.

We are interested in the question of the origin of galactic magnetic fields. It appears that an appreciable fraction of galactic fields are bisymmetric, i.e. the field in alternate spiral arms is in opposite directions (Sofue et al. 1988). This poses a problem, since on general grounds one expects that non-axisymmetric fields should be destroyed by differential rotation on a fairly short timescale. This difficulty would be avoided if it could be shown that a non-spherically symmetric distribution of turbulent diffusivity could actually lead to dynamo generation of non-axisymmetric fields, as was suggested by Rädler (1983) and Skaley (1985). For this reason, and also in order to avoid the uncertainties involved in assuming some specific model, we have not included an α-effect in these computations. Unfortunately, we have not found any growing field modes yet, but the decaying modes are of some interest in their own right.

Method and results

The evolution of the magnetic field is governed by the induction equation. We solve for the eigenmodes in a system consisting of a high-conductivity disc embedded in a low-conductivity corona using the so-called Bullard-Gellman formalism. For details of the model and numerical methods, see Donner & Brandenburg (1989) and references therein.