Observations of H2 line emission have revealed higher-than-expected gas temperatures in a number of photodissociation fronts. We discuss the heating and cooling processes in photodissociation regions. Observations of NGC 2023 are compared to a theoretical model in which there is substantial gas at temperatures T = 500 – 1000K heated by photoelectric emission and collisional de-excitation of H2. In general the model successfully reproduces the observed H2 line emission from a wide range of energy levels. The observed [SiII]34.8µm emission appears to indicate substantial depletion of Si in NGC 2023.

Introduction

A significant fraction of the ultraviolet radiation emitted by massive stars impinges on the molecular gas associated with star formation. The resulting photodissociation regions (PDRs) therefore play an important role in re-processing the energy flow in star-forming galaxies. Modelling these PDRs is therefore an important theoretical challenge, both to test our understanding of the physical processes in interstellar gas, and to interpret observations of star-forming galaxies.

It is frequently the case that the illuminating star is hot enough to produce an H II region, in which case the photodissociation region is bounded on one side by an ionization front, and on the other by cold molecular gas which has not yet been appreciably affected by ultraviolet radiation. The hv < 13.6eV photons propagating beyond the ionization front raise the fractional ionization, photo-excite and photo-dissociate the H2, and heat the gas via photoemission from dust and collisional de-excitation of vibrationally-excited H2.

Most of the diffuse interstellar medium is cold, but it must harbor pockets of hot gas to explain the large observed abundances of molecules like CH+, OH and HCO+. Because they dissipate locally large amounts of kinetic energy, MHD shocks and coherent vortices in turbulence can drive endothermic chemical reactions or reactions with large activation barriers. We predict the spectroscopic signatures in the H2 rotational lines of MHD shocks and vortices and compare them to those observed with the ISO-SWS along a line of sight through the Galaxy which samples 20 magnitudes of mostly diffuse gas.

The trigger of hot chemistry in the cold diffuse medium

The large observed abundances of CH, CH+, HCO+, and OH in the (mainly cold) diffuse medium (T ≈ 50 K) imply that activation barriers and endothermicities of several 103 K are overcome. Pockets of hot gas must therefore exist. Large ion-neutral drift speeds, of several km s−1, can equally contribute to triggering certain ion-neutral reactions in cold gas, such as the endothermic reaction C+ + H2 which forms CH+ (ΔE/k=4640 K). Two phenomena, operating at very different scales, are able to reproduce the observed abundances. These are MHD shocks (Flower & Pineau des Forêts, 1998 and references therein) and intense vortices, thought to be responsible for a large fraction of the viscous dissipation of supersonic turbulence (Joulain et al. 1998).

Molecular hydrogen is the most abundant constituent in interstellar space, but is difficult to observe in its most common form. Because H2 has no dipole moment, it does not have allowed rotational or vibrational transitions and therefore in its cold, unexcited state it has no radio or infrared spectral features. Therefore the most sensitive method for detecting cold H2 is through its allowed electronic bands, which lie in the far-ultraviolet portion of the spectrum. Previous UV instruments have provided some information on far-UV spectra of cold H2, but all were limited in either throughput or spectral resolving power, or both. The current FUSE mission has a combination of high throughput and moderately high spectral resolution, and is providing information on molecular hydrogen in interstellar regimes that have been previously unexplored. This review summarizes previous UV observations of H2 and then gives an overview of early FUSE results, with an emphasis on H2 in translucent clouds.

Introduction

The mass in the galactic interstellar medium is dominated by molecular hydrogen, which begins to outweigh atomic hydrogen in diffuse clouds and is expected to become completely dominant for translucent and dense clouds; i.e. clouds having visual extinctions greater than about Av = 2 magnitudes. Even in diffuse clouds, H2 is important, playing a crucial role in cloud chemical and physical processes.

Despite the importance of understanding H2 and its distribution, physics, and chemistry, relatively little direct information is available because of the obtuse spectroscopic properties of the molecule.

Two observable quantities have been calculated by using the data blocks which provide all details of interstellar UV absorption in H2 gas from the electronic ground state of H2 into 6 electronically excited states and fluorescence emission back into bound and continuum states of the electronic ground state. Both quantities describe details happening in the edges of interstellar H2 abundances where UV radiation is very efficiently shielded and HI gas is produced by fluorescent radiative dissociation. The first is the fluorescence spectrum of H2 in the wavelength range between 1350 and 1700 Å. Comparisons with published spectra show very nice agreement of the resolved features. The second is the velocity distribution of HI produced in fluorescent radiative dissociation. Subject of this comparison are the observed 21 cm spin-flip line profiles of the CNM and WNM (Cold and Warm Neutral Medium) HI species. The profile of the HI velocity dispersion, compared with the “narrow” WNM component, has been obtained at ≈ 9 km/s FWHM which is slightly below the statistical average, and this profile was found to be a molecular property of the H2 gas rather than a function of intensity of the incident UV field or the temperature. The required long tail (“wide” Gaussian component) is provided by all LTE models.

The rest of the paper is devoted to an outline of a dark matter model which fulfills the condition of statistical equilibrium in an active gaseous interphase between the interstellar UV field and the constituent dark matter mass.

We present a complete study of the H2 infrared emission, including the pure rotational lines, of the proto Planetary Nebulae CRL 618 with the ISO SWS. A large number of lines are detected. The analysis of our observations shows: (i) an OTP ratio very different from the classical value of 3, probably around 1.76-1.87; (ii) a stratification of the emitting region, and more precisely different regions of emission, plausibly located in the lobes, in an intermediate zone, and close to the torus; (iii) different excitation mechanisms, collisions and fluorescence.

Introduction

CRL 618 is one of the few clear examples of an AGB star in the transition phase to the Planetary Nebula stage: a Proto Planetary Nebula (PPN). It has a compact HII region created by a hot central C-rich star, and is observed as a bipolar nebula at optical, radio and infrared wavelengths. The expansion velocity of the envelope is around 20 kms−1, but CO observations show the presence of a high-velocity outflow with velocities up to 300 kms−1 (Cernicharo et al.1989). High-velocity emission in H2 is also detected (Burton & Geballe 1986). The high velocity wind and the UV photons from the star perturb the circumstellar envelope producing shocks and photodissociation regions (PDRs) which modify the physical and chemical conditions of the gas (Cernicharo et al.1989, and Neri et al.1992). Clumpiness within the visible lobes and low-velocity shocks being the remnant of the AGB circumstellar envelope are also proposed by Latter et al.(1992).

A brief history of observations of H2 and HD molecules is presented. The properties of H2 and HD that make observations of them a uniquely powerful diagnostic probe are pointed out. The interpretations of the observations in the ultraviolet, near infrared and infrared made of a diverse range of astrophysical objects are discussed and the influence of H2 and HD on their evolution is described.

Introduction

The hydrogen molecule occupies a central place in astrophysics. There are many reasons. The hydrogen molecule was the first neutral molecule to be formed in the Universe and it played a crucial role in the collapse of the first cosmological objects. It is the most abundant molecular species in the Universe, able to survive in hostile environments and found to exist in diverse astronomical objects ranging from planets to quasars. Hydrogen molecules have been detected by the absorptions and emissions at ultraviolet and infrared wavelengths to which they give rise. Molecular hydrogen has specific radiative and collisional properties that make it a diagnostic probe of unique capability.

Ultraviolet absorption

The first detection of H2 beyond the solar system was accomplished by Carruthers (1970) who employed a rocket-borne ultraviolet spectrometer and detected absorption bands of the Lyman system of H2 looking towards the star ξ Persei. More extensive data at much higher spectral resolution were obtained with a spectrometer aboard the Copernicus satellite (Spitzer and Jenkins 1975).

We present recent results from a program devoted to the study of small scale structure in translucent molecular clouds, using dust as a tracer. Several methods have been employed: i) statistical analysis of stellar fields, ii) studies of background galaxies; iii) searches for time variations of the extinction and reddening. For each method, we summarize the principles, the type of constraints provided (scales, sensitivity) and present the results obtained so far. We conclude by some prospects concerning direct studies of the distribution of H2 itself.

Motivations

Originally, we wished to constrain in a direct way the level at which the penetration of the stellar UV and visible radiation is enhanced by structure effects (cf Boissé 1990). Studies of spatial variations of dust extinction and reddening offer a powerful way to address this question. Indeed, in the presence of density fluctuations, the analysis of extinction in stellar fields directly provides a measure of the effective opacity (Thoraval et al. 1997). Further, in the visible, the required observations are easy to perform and provide an excellent spatial resolution.

Another motivation comes from the detection of very small scale structure both in atomic (Frail et al. 1994) and molecular gas (Moore & Marscher 1995). In the assumption of a uniform dust to gas ratio, one should observe corresponding variations of the amount of dust, resulting in local variations of the extinction.

Polycyclic aromatic hydrocarbons (PAHs) could play an important role in interstellar chemistry. In particular, it is important to evaluate their possible contribution to the formation of H2. To address this question, recent laboratory results and new observations are presented. Although still preliminary, these first results are very encouraging. First, the photodissociation of PAHs isolated in ion traps and exposed to UV light involves the loss of pairs of hydrogen atoms which are likely to form H2 molecules. Second, the PAH and H2 emission observed in the photodissociation region associated with the young stellar object S106-IR was found to coincide at some positions. This suggests a coupling between the interstellar PAH and H2 populations. More results are expected in the near future.

Introduction

The Infrared Space Observatory (ISO) has recently showed the ubiquity of the emission features at 3.3, 6.2, 7.7, 8.6, 11.3 and 12.7 µm in the interstellar medium (First ISO Results 1996, Boulanger 1999). Amongst the carriers for these features, polycyclic aromatic hydrocarbons (PAHs) are the best candidates as far as an excitation mechanism and a reasonable spectral agreement are concerned. A lot of infrared spectroscopy has been performed in the laboratory since the initial proposal by Léger & Puget (1984) and Allamandola et al. (1985) to find the laboratory species whose spectrum match the interstellar spectrum (Szczepanski & Vala 1993, Joblin et al. 1995, Cook et al. 1998, Allamandola & Hudgins 1999, Hudgins & Allamandola 1999, etc…).

We discuss several recent observational results according to which a significant fraction of the Galactic dark matter halo may exist in the form of dense self-gravitating clumps of H2.

We model the large scale mass distribution in the Milky Way and derive in a self-consistent way the associated gravitational potential. The basic constraint in our model is that the shape of the gaseous Galactic halo, observable from synchrotron radiation, γ-rays, soft X-ray background and from H I lines with high-velocity dispersion needs to be explained by the mass distribution.

The resulting mass model has a flat rotation curve. In the solar vicinity surface density, volume density and gravitational acceleration Kz are consistent with all known constraints. We find that the distributions of HI gas and dark matter are closely related to each other. Furthermore, the mass distribution implies a co-rotation of disk and halo for radii R > 5 kpc.

Our analysis supports strongly the hypothesis, that the gaseous halo of the Milky Way traces dark matter in the form of dense self-gravitating H2 condensations as indicated from an analysis of the γ-ray background observed with EGRET (Shchekinov et al., these proceedings). In addition we find evidence for an extended non-baryonic dark matter halo, which co-exists with the baryonic component. We derive for the Milky Way a baryonic mass fraction of 12% in close agreement with cosmological predictions.

H2 and HD molecules provide the cooling needed for the fragmentation and collapse of the first structures in the universe. In this review, we describe the main chemical and physical processes occurring in the primordial gas after the recombination epoch. We also highlight the areas where improvements in the determination of reaction rates and excitation coefficients are necessary to reduce the remaining uncertainties in the predictions of the numerical models. The interaction of primordial molecules with the CBR and the role of H2 and HD cooling in the early universe are discussed. Finally, we comment on the results of recent simulations of the fragmentation and collapse of primordial clouds, with an emphasis on the typical mass scale of the first objects.

Introduction

The formation of the first stellar objects in the universe is a fascinating, yet little understood process. Although we have now observational data on bright quasars and galaxies out to redshifts of about 5 (corresponding to 109 yr after the Big Bang) and on the density fluctuations at redshifts about 1000 (or an age of ∼ 106 yr), there is no direct evidence as to when and how the first structures formed. This unique epoch and the nature of the primordial objects define what has been called the end of the “Dark Ages” (Rees 1999).

According to Big Bang cosmology, there must have been an epoch in the history of the universe during which the original gas mixture was altered by the manufacture of heavy elements inside stars.

Two topics of relevance for H2 formation in the interstellar medium are considered: (i) the interaction of H and H-H with a model-graphite surface (Coronene: C24H12), and (ii) H− formation by charge transfer in the interaction of H with a model-silicate surface (MgO{100} representing forsterite: Mg2SiO4{100}). The first topic is related to the frequently invoked Langmuir-Hinshelwood and Eley-Rideal mechanisms for H2 formation near carbonaceous zones of interstellar dust grains. Ab initio calculations based on Density Functional Theory are used. The second topic proposes a new scenario in which the efficient production of H− ions would subsequently enable the formation of H2 via associative detachment. It stems from recent work of the authors on charge transfer between neutral atoms and ionic insulators.

Introduction

The mechanism of H2 formation in the interstellar medium (ISM) is still an open problem. Owing to the temperature and density conditions existing in the ISM, 3-body recombination and radiative association processes in the gas phase are unable to account for actual H2 abundances. The existence of dust particles in the ISM has attracted the attention of astrophysicists as plausible catalysts or mediators of H-H recombination in space (Hollenbach and Salpeter (1970), Hollenbach and Salpeter (1971)). Current knowledge of interstellar dust particles (IDPs) indicates that they have both a carbonaceous and a silicate composition. This has in particular stimulated the investigation of the role graphitic bonds may have on H2 formation as a result of elementary interactions between H atoms and graphite-like surfaces or platelets.

The bulk of the molecular gas in spiral galaxies is under the form of cold H2, that does not radiate and is only suspected through tracer molecules, such as CO. All tracers are biased, and in particular H2 could be highly underestimated in low metallicity regions. Our knowledge is reviewed of the H2 content of galaxies, according to their types, environment, or star-forming activities. The HI and CO components are generally well-mixed (spiral arms, vertical distribution), although their radial distributions are radically different, certainly due to radial abundance gradients. The hypothesis of H2 as dark matter is discussed, as well as the implications on galaxy dynamics, or the best perspectives for observational tests.

How to observe H2 in galaxies?

The bulk of molecular hydrogen in a galaxy is cold, around 10–20K, and therefore invisible. The first rotational level, accessible only through a quadrupolar transition, is more than 500 K above the fundamental. The presence of H2 is inferred essentially from the CO tracer. The carbon monoxide is the most abundant molecule after H2; its dipolemoment is small (0.1 Debye) and therefore CO is easily excited, the emission of CO(1–0) at 2.6mm (first level at 5.52K) is ubiquitous in the Galaxy.

The H2/CO conversion ratio

To calibrate the H2/CO ratio, the most direct and natural is to compare the UV absorption lines of CO and H2 along the same line of sight (Copernicus, e.g. Spitzer & Jenkins 1975; ORFEUS, cf Richter et al., this conference).

This book gathers all contributions to the International Conference on H2in Space held in Paris, France, on September 28- October 1st, 1999. The attendance was 106 participants from 16 countries. The goal was to gather together representatives of three communities: experts in the physics of the molecule, including experimentalists, observers of the interstellar medium, in particular of the warm H2 detected in infrared lines, and theoreticians studying H2 formation and cooling in astrophysical objects, from the early universe to the present galaxies.

The electronic structure of the H2 molecule has been well studied in the past, but it was shown that recent progress has been made on the theory of highly excited and long-range Rydberg states, both from calculations of line-strengths and comparison with experiments. New and more accurate data are now available for the rates of collisional excitation of rovibrational transitions by neutrals, as well as protons or electrons. It has been known for several decades that interstellar H2 is formed on dust grains, but, until now, the efficiency of this process was poorly known. Recent experiments have reproduced this formation process on silicates and amorphous carbon and shown that the efficiency is strongly dependent on temperature. In particular, the mobility of H atoms on grains is much lower than previously thought. Laboratory experiments with trapped ions and nanoparticles have opened new avenues of investigation. The relationship of H2 with PAHs is being addressed both in the laboratory and by means of astronomical observations.

We review laboratory studies of the formation of molecular hydrogen on surfaces and in conditions of astrophysical interest. Theoretical analysis and predictions are given on how experimental results can shed light on actual physico-chemical processes occuring in the interstellar medium. Preliminary measurements of H atom sticking are also shown.

Introduction

Molecular hydrogen is the most abundant molecule in the Universe. In space it plays two main roles that render it of incomparable importance:

1. – once formed it becomes a very efficient coolant that increases the rate of collapse of interstellar clouds contributing to shape galaxies and to regulate their dynamics;

2. – once ionized by UV photons or by cosmic rays H2 enters and triggers all reactions schemes that form most of molecular species in the gas phase.

Molecular hydrogen has always been a source of puzzles for astrophysicists in spite of its simplicity. The mechanism of its formation in the extreme conditions encountered in the interstellar medium is certainly one of them. When a sufficient abundance of electrons and ions exists, ion-atom reactions may be quite effective in producing H2 in the gas phase (Stancil & Dalgarno 1998; when, on the contrary, physical conditions allow almost only the presence of neutral hydrogen atoms, the radiative association of two of them is highly inefficient and cannot explain the observed abundances. The main reason is that the time-scale for the release of a consistent fraction of the formation energy (4.5 eV) through the emission of a photon via forbidden transitions is too long and the proto-molecule that is formed in a repulsive state almost invariably breaks up.

We describe spectro-imaging observations of the bright western ridge of the supernova remnant IC443 obtained with the ISOCAM circular variable filter (CVF) on board the Infrared Space Observatory (ISO). The CVF data show that the 5 to 14 µm spectrum is dominated by the pure rotational lines of molecular hydrogen (v = 0–0, S(2) to S(8) transitions). We compare the data to a new time-dependent shock model.

Introduction

The supernova remnant IC443 is a prime example of the interaction of a supernova blast wave with an ambient molecular cloud. On optical plates, IC443 appears as an incomplete shell of filaments (Fig. 1) with a total extent of about 20 arcmin, i.e. ∼ 9pc for an adopted distance of 1,500 pc. The shock generated by the supernova explosion, that occurred (4–13) × 103 years ago, encountered nearby molecular gas which is mainly found along a NW-SE direction across the face of the optical shell. IC443 has been the subject of numerous studies from X-rays, visible, infrared to radio wavelengths (e.g., Mufson et al. 1986 and references therein). Studies of the interaction between the shock and the ambient molecular gas were done by observing molecular hydrogen in the rotational–vibrational transitions (Burton et al. 1988, 1990 – see Fig. 1 – and Richter et al. 1995a), in the pure rotational S(2) transition (Richter et al. 1995b).

We derive and discuss the strong dependence on metallicity of the CO to H2 conversion factor X = N(H2)/Ico = 12.2 – 2.5log[O]/[H] appropriate to extragalactic objects, as well as the weaker dependence found for such objects from interferometer measurements.

Introduction

The difficulty of directly observing molecular hydrogen (H2), the major constituent of the interstellar medium in galaxies, and ways of doing so indirectly are reviewed elsewhere in this volume (Combes 2000). Usually, H2 cloud properties are derived by extrapolation from more easily conducted CO observations. For instance, observed CO cloud sizes and velocity widths yield total molecular gas masses under the assumption of virial equilibrium. However, in extragalactic systems especially, this method is beset by pitfalls (see Israel, 1997, hereafter Is97) and requires high linear resolutions (i.e. use of interferometer arrays). More seriously, the fundamental assumption of virialization appears to be false. As individual components (‘clumps’) have velocities of only a few km s−1 and CO complex sizes are 50–100 pc, crossing times are comparable to CO complex lifetimes of only a few times 107 years or less (Leisawitz et al. 1989; Fukui et al. 2000; see also Elmegreen 2000). As equilibrium cannot be reached in a single crossing time or less, the virial theorem is not applicable to such complexes. Indeed, the elongated and interconnected filamentary appearance of many large CO cloud complexes do not suggest virialized systems (see also Maloney 1990).