Introduction

A detailed knowledge of collisional excitation processes is important in various aspects of the study of interstellar clouds. Because local thermodynamic equilibrium rarely obtains in such environments, the diagnosis of physical conditions, such as temperature, particle density, and radiation density, requires a quantitative understanding of all microscopic processes (i.e. collisional excitation and deexcitation and radiative decay and absorption) which influence the excitation conditions. Quite often, only rotational excitation of simple molecules need be considered. For instance, dense, cold cloud gas is studied primarily by observing millimeter and submillimeter emission features arising from transitions between different rotational levels of molecules in their ground electronic and vibrational states. In diffuse clouds, simple diatomics such as H2, CN and C2 are observed through electronic absorption transitions involving different rotational levels of the ground vibrational and electronic state. Even some atomic species such as C and C+ are observed in various fine structure states. This information is important for the diagnosis of diffuse clouds. Observable emission from vibrationally excited molecules arises in hotter gas or in regions exposed to a strong ultraviolet radiation field, but the collisional excitation of vibrational states is not well understood for the appropriate temperature range. However, the observed emissions due to the decay of collisionally excited fine structure levels of atomic species, such as C and C+ can be used to investigate clouds.

Second, collisional excitation of atomic and molecular species is always followed by spontaneous radiative emission leading to a loss of energy from the medium which is an important cooling process of the interstellar gas.

The present volume has been designed to be a self-contained introduction to the field of molecular astrophysics. It can serve as the text for a one semester postgraduate course concerning that subject exclusively or as a supplementary text in a postgraduate course on the interstellar medium. It can also be used by research astronomers, atomic and molecular physicists, chemists, and atmospheric scientists who have interests in weakly ionized plasmas and the physical and chemical processes which occur in them and who wish to become familiar with recent work in molecular astrophysics.

Many of the articles concern theoretical studies and modelling. Part I, consisting of two chapters written by observers, provides a general description of the astronomical context within which much of the remainder of the volume should be considered. The contribution by Per Friberg and Åke Hjalmarson reviews briefly our understanding of the global physical properties of the Galactic interstellar medium, describes the wide range of conditions within the dark molecular clouds (those having visual extinctions greater than about 1) and the distribution in the Galaxy of the dark clouds, and compares the chemical contents of different regions in various dark clouds. Leo Blitz has written about observations of molecules in other galaxies including some with active nuclei.

The study of chemistry in diffuse molecular clouds (those having a visual extinction of about 1 or less) constitutes the subject of Part II and is perhaps the most fundamental area in molecular astrophysics.

Introduction

Our understanding of the structure of the interstellar medium has progressed enormously in the last 15 years largely as a result of the development of radio emission and ultraviolet absorption observations. From CO radio surveys we now know that about one-half of the interstellar gas is in the form of giant molecular clouds that are the formation sites of O and B stars, while measurements of highly excited gas, such as hot bubbles and expanding shells, indicate that a significant volume of the interstellar medium contains dilute, hot gas. The physical properties and dynamics of the interstellar gas are strongly influenced by the interaction of the massive O and B stars and supernovae. Supernova shells sweep up interstellar matter creating clouds and also impact them causing compression and fragmentation. The O and B stars produce large HII regions which expand and compress the interstellar gas. Thus on a theoretical basis we expect to find complex velocity fields, shock structures, and density inhomogeneities in the diffuse and translucent component of the interstellar clouds.

Today it is hard to imagine studying the diffuse interstellar clouds without relying on observations of molecular lines. In an excellent review article by Dalgarno and McCray (1972) on the heating, ionization and cooling processes of HI regions, these authors stated that ‘The processes leading to interstellar molecule formation are only partially understood …’. The central role of molecules as both important components and physical probes was just becoming evident.

Introduction

Astrophysical environments offer chemical modelers the unusual challenge of predicting reaction rate coefficients for temperature ranges well outside those commonly encountered in the laboratory (250–2000 K). The direct extrapolation of thermal laboratory data involves implicit, often incorrect, assumptions about the dynamics of chemical reactions at extreme temperatures. Furthermore, the internal energy distributions of constituent atoms and molecules in astrophysics are frequently nonthermal, and direct application of thermal reaction rate data may not be appropriate. An understanding of the physics of chemical reactions is essential for accurate modeling (Dalgarno 1985). Elucidation of the role of various forms of energy in chemical reactions – translational, vibrational, rotational, and fine structure – now affords improved predictions of rate coefficients for astrophysical situations.

In the low-temperature environment of cold interstellar clouds, the chemical kinetics are dominated by reactions whose potential energy surfaces are without barriers. Such reactions are characterized by large exothermicities and long-range electrostatic interactions that suppress chemical barriers that may occur as collision systems approach short distances. Current models of cold interstellar cloud chemistry illustrate the importance of exothermic ion–molecule reactions (cf. Dalgarno and Black (1976), Black and Dalgarno (1977), van Dishoeck and Black (1986)). Many neutral systems are also likely to react rapidly at very low temperatures and may be important in the chemical kinetics of interstellar clouds (Graff 1989). The following section discusses dynamical characteristics of neutral reactions that are likely to be fast at low temperatures.

The cloudy structure of our Galaxy

The interstellar medium (ISM) in our Galaxy is in a complex state. The temperature and the density vary by about five and ten orders of magnitude respectively. The medium is exposed to cosmic rays and starlight and contains magnetic fields. Its average density is about 10−24 g cm−3, or about one hydrogen atom per cubic centimeter, corresponding to 0.025 M⊙ pc−3. Thus on average one solar mass of the medium is contained in 40 pc3. About 75% of the medium is hydrogen, about 25% is helium, and the remaining part (about 2%) is in heavier atoms. A large fraction of the heavier elements has condensed out as dust grains having an average density of about 0.001 M⊙ pc−3.

The local kinetic gas temperatures range from 10 K to 106 K. These temperatures correspond to energies which are much lower than those typical of cosmic rays. The lower temperatures correspond to energies which are less than those typical for starlight. Cosmic rays and starlight drive processes including ionization; hence, the ionization structure sometimes is very far from that in a gas in thermal equilibrium at the gas kinetic temperature. The energy density of cosmic rays is ≈0.5 eV cm−3, of Galactic magnetic fields ≈0.2 eV cm−3 (≈10−6G), and of diffuse starlight ≈0.5 eV cm−3. Hence there is a rough equipartition of energy among these components. The description of the interaction of matter with radiation is complicated by the huge variation of opacities for different wavelengths.

Introduction

When low and intermediate mass stars evolve off the main sequence, they become red giants. During the first ascent up the red giant branch, the stars lose mass at a relatively modest rate (∼10−9 M⊙ yr−1. Dupree (1986)). However, during the second ascent up the red giant branch, the asymptotic giant branch (AGB), stars lose mass at a much greater rate; up to 10−4 M⊙ yr−1 (Zuckerman 1980, Jura 1986a, Olofsson 1985). These outflows from AGB stars are very cold (T < 1000 K), and they contain large amounts of dust and molecules. The chemistry in these outflows is the topic of this chapter (see also Olofsson (1987), Omont (1987), Millar (1988)).

Stars derive most of their luminosity from nuclear reactions that occur in their interiors. In the red giant phase, material from the interior is mixed with that at the surface and the composition of the photosphere of a red giant can be very different from the star's initial composition. In most cases, it appears that in the atmospheres of the red giants, the major elements are still hydrogen and helium, even though material is mixed from the interiors to the surface. However, the next most common elements, carbon, nitrogen and oxygen, may have very nonstandard compositions, and this can greatly affect the chemistry. In particular, the ratio [C]/[O] is critical. In the usual thermodynamic description for material in cool stellar atmospheres, we expect CO, because of its very high binding energy, to contain as much carbon and oxygen as possible.

Introduction

Giant interstellar clouds are the most massive chemical ‘factories’ in our Galaxy containing around 80 molecules presently identified (neglecting isotopic variants) and ranging in complexity from H2 and CO to large saturated molecules such as ethanol, CH3CH2OH, and highly unsaturated cyanopolyynes including the 13-atom chain HC11N. Millimetre and sub-millimetre observations of interstellar molecules allow one to probe the densities, temperatures, and dynamics of interstellar clouds and can give information on the initial conditions for star formation. It is also of interest to understand the chemistry of these molecules since chemical kinetic modelling can be used together with observational data to constrain uncertain parameters such as elemental abundances and the cosmic ray ionisation rate. An understanding of deuterium fractionation in interstellar molecules can be used to determine the D/H ratio and thus has a bearing on cosmological models of the origin of the universe.

In recent years models of increasing chemical, physical, and computational complexity have been developed to study molecular formation and destruction in various astronomical regions. Models of interstellar cloud chemistry can be divided roughly into two classes: (1) steady-state models, in which chemical abundances are calculated through solving a coupled system of non-linear algebraic equations; and (2) time-dependent models, in which the variations of abundances as a function of time are followed through solving a coupled system of stiff, non-linear, first-order, ordinary differential equations. Traditionally, steady-state models have been the dominant tool for studying chemistry in diffuse clouds which, in the absence of shocks, reach steady-state well within their lifetimes.

Introduction

The first unambiguous detection of molecules in galaxies other than the Milky Way was that of OH in absorption toward the nuclear continuum sources in NGC 253 and M82 (Weliachew 1971). Subsequently, all of the species detected in the Milky Way that show strong emission or absorption lines have now been detected in other galaxies. Although extragalactic molecules have been observed in regions of the spectrum other than the radio, notably the vibrational–rotational lines of H2 in the infrared (Thompson, Lebofsky and Rieke 1978), observations remain largely the domain of radioastronomers. Furthermore, the largest fraction of all extragalactic molecular observations have been carried out using the J = 1−0 transition of CO, because it is the easiest transition to detect in most galaxies.

Initial progress in the field was slow, hampered by the simultaneous requirements of large bandwidth, high sensitivity and high angular resolution. The technical requirements were largely overcome in the early 1980s and the field has now burgeoned into one where developments are occurring so fast, on so many fronts that it is impossible in a short review to do justice to all of them. Accordingly, I have chosen a small subset of what I think are among the most interesting recent developments in the study of molecules in other galaxies with the aim of focussing on a few problems of wide astrophysical interest. A good introduction to the field is the comprehensive review by Morris and Rickard (1982).

Introduction

There is an increasing body of evidence for a population of large molecules in the interstellar medium. Large molecules have been suggested as the source of near infrared continuum radiation and the near infrared emission bands observed in reflection nebula, planetary nebula, HII regions and active galaxies (Duley and Williams 1981, Léger and Puget 1984, Allamandola, Tielens and Barker 1985, d'Hendecourt et al. 1986, Barker, Allamandola and Tielens 1987). Large molecules have also been proposed as the carriers of the diffuse interstellar bands (vander Zwet and Allamandola 1985, Léger and d'Hendecourt 1985, Crawford, Tielens and Allamandola 1985). The large molecules proposed have between 30 and 100 atoms with suggested abundances in the range 10−7−10−6 relative to hydrogen.

Mathis, Rumpl and Nordsieck (1977; MRN) determined a size distribution for grains that would fit the extinction measured to many sources. However, existing extinction measurements do not extend far enough into the ultraviolet to infer the population of the smallest grains. An extrapolation of the MRN distribution to very small grains having between 30 and 100 atoms gives an abundance roughly the same as that proposed for the large molecules, suggesting that the large molecules may be an extension of the interstellar grain distribution.

It appears that a population of small grains or large molecules exists in much of the interstellar medium. We will use the terms large molecules and small grains interchangeably to refer to clusters of between 30 and 100 atoms.

Introduction

The wide variety of ionized and neutral molecules detected in diffuse and dense interstellar clouds are mainly synthesized in the gas phase by many sequential and parallel positive ion–neutral reactions. The product positive ions may then be neutralized by proton transfer or charge transfer reactions with ambient molecules and/or via dissociative positive ion–electron recombination reactions and/or (as has recently been proposed) via mutual neutralization (positive ion–negative ion recombination) reactions involving large negatively-charged molecules. The basic gas phase ion chemistry was outlined in the early 1970s (Solomon and Klemperer 1972, Herbst and Klemperer 1973, Dalgarno 1975, Dalgarno and Black 1976) and since that time more sophisticated quantitative ion–chemical models have been developed to describe the routes to, and relative abundances of, the increasing number of complex molecules detected in interstellar clouds (Leung, Herbst and Huebner 1984, Millar and Nejad 1985, Dalgarno 1986, van Dishoeck and Black 1986, Millar, Leung and Herbst 1987). These models require the input of a large amount of kinetic data which (ideally) has been obtained at the low temperatures pertaining to interstellar clouds. In particular, the rate coefficients and products of many positive ion–neutral reactions and recombination reactions are required. Hence, a good deal of effort has been made to design and exploit laboratory experiments to provide these data (Huntress 1977, Smith and Adams 1981a, Anicich and Huntress 1986, Adams and Smith 1987a, Rowe 1988, Adams and Smith 1988a). This has led to a better understanding of the molecular processes that can occur under the extreme conditions of low pressure and low temperature in quiescent interstellar clouds.

Introduction

Fast shocks destroy molecules. Whereas the compressive and heating effects of slow shocks profoundly alter the chemical composition of a molecular gas but leave its molecular nature intact, interstellar shocks travelling faster than about 50 km s−1 result in the complete and very rapid dissociation of any pre-existing molecules by collisional processes (see Figure 3 in McKee, Chernoff and Hollenbach 1984), and if the shock velocity exceeds 70 km s−1, the atomic dissociation products are then largely ionised. Shocks faster than 80 km s−1 generate enough ultraviolet radiation to destroy molecules by photoionisation before they even reach the shock front (Hollenbach and McKee 1989).

Dissociative shocks may be present in the interstellar medium wherever gas is moving supersonically at velocities greater than 50 km s−1. Such velocities of bulk motion may result from the outflow of material from young stellar objects, from stellar winds, or from supernova explosions. Clearly, dissociative shocks are intimately associated with stars: they may be generated during the birth, throughout the life and by the explosive death of stars. Emission from fast shocks may serve as a tracer of active star formation in the Galaxy.

Despite the extremely destructive effects of fast shocks, molecules can reform behind a passing dissociative shock before the shocked gas has cooled to the temperature of the ambient, unshocked medium. Free electrons produced behind the shock front initiate the process by catalysing H2 formation. The rapid return of a dissociated shocked plasma to the molecular phase is a striking demonstration of the tendency toward molecule formation in the interstellar medium.