The ultraviolet Lyman and Werner absorption lines of H2 have been searched for in a number of high redshift quasar spectra, and detected unambiguously in at least 3 systems at redshifts z∼2. The lack of detectable H2 in most absorbers results from the strong selection in quasar studies against lines-of-sight with significant dust extinction. At high redshift, the ultraviolet radiation field is inferred to be higher than that observed in the local solar neighborhood, suggesting that vigorous star-formation is underway in these galaxies.

Introduction

Recent observations of the high redshift Universe, interpreted in the context of a new generation of computer simulated model Universes, are providing a clear picture of how large galaxies like the Milky Way formed. A number of different observations suggest that large galaxies were assembled from what appear at z = 2 – 3 to be several star-forming proto-galactic fragments (PGF's), widely distributed in space (Windhorst et al. 1994, Pascarelle et al. 1996ab, 1998; Steidel et al. 1996ab, Bechtold et al. 1998). Computer simulations suggest that initially small clumps of material collapsed at the intersection of sheets and filaments in the intergalactic medium, and began forming stars, and that eventually these clumps merged to form large galaxies (Haehnelt, Steinmetz & Rauch 1998, Steinmetz 1998 and references therein). Searches for the galaxies associated with damped Ly-α quasar absorbers show that at z ∼ 2 they are the same population of objects seen in the Hubble Deep Field faint galaxies and the Lyman dropout galaxies (Steidel et al. 1996ab; Bechtold et al. 1998).

We review the relevance of H2 molecules for structure formation in cosmology. Molecules are important at high–redshifts, when the first collapsed structures appear with typical temperatures of a few hundred Kelvin. In these chemically pristine clouds, radiative cooling is dominated by H2 molecules. As a result, H2 “astro–Chemistry” is likely to determine the epoch when the first astrophysical objects appear. We summarize results of recent three–dimensional simulations. A discussion of the effects of feedback, and implications for the reionization of the universe is also given.

Introduction

In current “best–fit” cosmological models, cold dark matter (CDM) dominates the dynamics of structure formation, and processes the initial density fluctuation power spectrum P(K) ∝ kn with n = 1 to predict n = 1 on large scales and n ≈ −3 on small scales (Peebles 1982). The r.m.s. density fluctuation σM then varies inversely with the mass-scale (σM ∝ M−2/3 for M » 1012M⊙, while the dependence is only logarithmic for M « 1012M⊙). The more overdense a region, the earlier it collapses, implying that the present structure was built from the bottom up, with smaller objects appearing first, and subsequently merging and/or clustering together to assemble the larger objects (Peebles 1980). The predicted formation epochs of “objects” (i.e. collapsed dark matter halos) with various masses in the so-called standard CDM cosmology (Bardeen et al. 1986) are shown in Figure 1.

Currently there are three quite different views about galaxy evolution, each one improving the previous state of knowledge:

1. (1) The older one (“ELS”) in which galaxies form by collapse early, quickly, and synchronously (during the “galaxy formation epoch”), ending the dynamically active period; subsequent galaxy evolution is merely a matter of stellar formation processes in a rigid potential.

2. (2) An alternative one (“SZ”) in which disks are viewed as forming inside out over an extended period of time. Galaxy evolution occurs without important internal dynamical instabilities.

3. (3) The slowly emerging picture, after 40 years of N-body simulations and the obvious evidences from recent high-z observations: galaxies evolve both dynamically and chemically over most of the Hubble time in a widely asynchronous way at different speeds, depending on the environment. The Hubble sequence, from late to early types, appears to represent a broad description of the general aging process.

Thus galaxies appear now as evolving structures over typical time-scales of order of 1 Gyr. A fundamental aspect of the micro-physics in galaxies is star formation and gas processes in which the H2 molecule must play a key role: indeed interstellar gas must first form H2 before being able to form stars, so star forming regions do trace molecules, although CO might not have been detected.

Most of the H2 in our Galaxy resides in the cold interiors of molecular clouds. The most reliable way to trace the H2 content of a molecular cloud is, in principle, to measure the distribution of dust through it. In this contribution we present a new observational approach that uses infrared dust extinction of starlight to construct high resolution maps of the distribution of dust (H2) inside molecular clouds over unprecedented ranges of cloud depth: 1 < Av < 40 magnitudes. We also present a comparison of our results with conventional molecular-line column density tracer C18O and conclude that for cloud depths of Av > 10 magnitudes this species is a very poor tracer of H2.

Introduction

Molecular clouds are the reservoirs of H2 in the Galaxy. They contain about half of the mass of the Interstellar Medium and hence an important fraction of the mass of the Galaxy. By far the most important characteristic of molecular clouds is that they are the nurseries out of which stars like our Sun were born. This creation process not only determines the origins of stars and planetary systems in our Galaxy but also regulates the structure and evolution of galaxies on the large scale. To understand star and planet formation is to understand how cold H2 clouds evolve.

We present near-infrared imaging of IC443, covering entire supernova remnant (50 diameter) from the Two Micron All Sky Survey (2MASS), which images are taken simultaneously in the J (1.25µm), H (1.65µm) and Ks (2.17µm) bands. Emission from IC443 was detected in all 3 bands from most of the optically bright parts of the remnant, revealing a shell-like morphology. These are the first near-infrared images that covers entire remnant. The color and structure are very different between the northeastern and southern parts. Bright J and H band emission from the northeast rim can be explained mostly by [Fe II] and the rest by hydrogen lines of Pβ and Br10. We also report ISO LWS observation of [O I] (63µm) for 11 positions in the northeast. Strong lines were detected and the strongest line is in the northeastern shell, where 2MASS image showed filamentary structure in J and H. In contrast, the southern ridge is dominated by Ks band light with knotty structure, and has weak J and H band emission. The shocked H2 line emission is well known from the sinus ridge produced by an interaction with dense molecular clouds. The large field of view and color of the 2MASS images show that the H2 emission extends to the east and the northeast. This H2 emission suggests that the interaction with the molecular clouds extends to the front side in the northeast.

Observations of the interstellar medium within 1 kpc of the Sun with the Copernicus satellite showed a value of the gas to dust ratio that varies by less than a factor of two from its average. The fraction of hydrogen that is molecular is well described by a steady state model that balances formation on grains with photo-destruction. However, in contrast to the local interstellar medium, both in quasar absorption line systems and in circumstellar disks around young stars, there appears to be relatively little H2. We particularly focus on estimating the amount of H2 in circumstellar disks around main sequence stars – the environment where planets form.

Introduction

One of the main results achieved with the Copernicus satellite was the systematic measurement of interstellar H and H2 within about 1 kpc of the Sun. It was found (Savage et al. 1977, Bohlin, Savage & Drake 1978) that the dust to gas ratio is uniform to within a factor of 2 of its average value with the mass in gas being approximately 100 times larger than the mass in dust. Also, the fraction of hydrogen that is molecular, [2N(H2)]/[N(H) + 2N(H2)], is well described by a standard steady state model (Hollenbach, Werner & Salpeter 1971). In this standard model, the H2 is formed on the surface of grains with a rate of about 3 × 1017 cm3 s−1 (Jura 1975) and destroyed by the absorption of ultraviolet photons with a rate near 5 × 10−11 s−1 when the gas is optically thin (Jura 1974).

We summarize the results of recent quantum mechanical calculations of cross sections and rate coefficients for the rovibrational excitation of H2 and HD by the principal perturbers, H, He, and H2. These results have been used to evaluate the rate of cooling of astrophysical media by H2 and HD molecules; these calculations are also described. The cooling of the primordial gas by rotational transitions of H2 is considered as a special case.

All the numerical results and related software are available from http://ccp7.dur.ac.uk/.

Introduction

Molecular hydrogen is recognized as a major contributor to the cooling of astrophysical media. Its role is all the more significant under conditions, such as those which prevailed in the primordial gas, where few other coolants were present; but H2 is also an important, sometimes the dominant coolant of low density interstellar gas, for kinetic temperatures T > 100 K. Interstellar gas can be heated to such temperatures by shock waves, by the dissipation of turbulence, or by absorbing energy from the local ultraviolet radiation field, as in photon-dominated regions.

Although the elemental abundance of deuterium is approximately 5 orders of magnitude less than that of hydrogen, it turns out that cooling by HD must often be taken into account, essentially for two reasons. First, chemical fractionation can, in media which are only partially molecular, enhance the abundance of HD, relative to that of H2.

We review recent H2 absorption line measurements in the diffuse interstellar medium, using FUV spectra from the Orbiting and Retrievable Far and Extreme Ultraviolet Spectrometer (ORFEUS). We investigate molecular hydrogen gas along lines of sight toward 5 stars in the Magellanic Clouds and toward 3 stars within the Milky Way. Molecular fractions in gas within the Magellanic Clouds are significantly lower than typically found in gas in the Milky Way, most likely caused by the lower dust content. The finding of H2 in a Galactic high-velocity cloud led us to speculate that the high-velocity gas in front of the Magellanic Clouds is part of the Galactic fountain. Sight lines toward the Galactic stars show well defined absorption by molecular hydrogen, deuterium and metals, allowing the study of physical and chemical conditions in the local interstellar gas in great detail.

Introduction

Molecular hydrogen is by far the most abundant molecule in the interstellar medium. Its measurement, however, is difficult: H2 has no permanent dipole moment and no radio emission is seen from H2, in striking contrast to the second most abundant molecule in the ISM, carbon monoxide (CO). For the study of the diffuse interstellar medium the FUV absorption spectroscopy is the only method to obtain information about the molecular hydrogen content, but satellites are required for this method, since the earth's atmosphere is opaque for radiation in the FUV domain.

Molecular hydrogen is a key species for the formation of the first luminous objects in the early universe. It is therefore crucial to understand the various physical processes leading to its formation and destruction and the feedbacks regulating this chemical network. Here we review both the radiative and SN-induced feedbacks and we assess the role of the objects relying on H2 for their collapse in the evolution of the reionization of the universe.

Introduction

At z ≈ 1100 the intergalactic medium (IGM) is expected to recombine and remain neutral until the first sources of ionizing radiation form and reionize it. Until recently, QSOs were thought to be the main source of ionizing photons, but observational constraints suggest the existence of an early population of pregalactic objects (Pop III hereafter) which could have contributed to the reheating, reionization and metal enrichment of the IGM at high redshift. In order to virialize in the potential well of dark matter halos, the gas must have a mass greater than the Jeans mass (Mb > MJ), which, at z ∼ 20 – 30 corresponds to very low virial temperatures (Tvir < 104 K). To have a further collapse and fragmentation of the gas, and to ignite star formation, additional cooling is required. It is well known that in these conditions the only efficient coolant for a plasma of primordial composition, is molecular hydrogen (Abel et al. 1997; Tegmark et al. 1997; Ciardi, Ferrara & Abel 2000 [CFA]).

Molecular hydrogen is formed on interstellar grains by two main processes. In the first, or Langmuir-Hinshelwood, mechanism, hydrogen atoms land on a grain and diffuse over the surface by either tunneling or hopping until they find each other. In the second, or Eley-Rideal, mechanism, hydrogen atoms landing on grains are fixed in position. Reaction occurs only when a gaseous hydrogen atom lands atop an adsorbed one. Based on new experimental results concerning the rate of diffusion of H atoms on interstellar-like surfaces, it is clear that the rate is significantly slower than estimated in the past. The range of temperatures over which diffusive formation of H2 occurs is correspondingly reduced although sites of strong binding can raise the upper temperature limit. The surface formation of molecules heavier than hydrogen is still not well understood.

Introduction

It is almost certain that H2 and a variety of other molecules are formed on the surfaces of low-temperature interstellar dust particles. On these surfaces, binding sites for adsorbates exist interspersed among regions of higher potential. On a grain of typical radius 0.1 µ there are roughly 106 such binding sites, onto which neutral gas-phase molecules stick with high efficiency. The binding energy, or energy required for desorption (ED), depends on the surface and on the adsorbate. For example, the binding energy of H atoms on olivine (a silicate-type material) has just been measured to be 372 K by Katz et al. (1999), who also measured the binding energy of H on amorphous carbon to be 658 K.

We describe the new capability provided by integral field spectroscopy for simultaneously mapping a wide range of shocked emission lines across outflows at high spatial resolution. We have used the MPE-3D near-IR integral field spectrometer on the AAT to carry out a detailed observational study of the physics of shocked H2 and [Fe II] excitation within individual bow shocks. Simultaneous measurement of line ratio variations with position across and along bow shocks will strongly constrain shock models in a number of outflow sources. In Orion, where broad H2 line widths had previously implied magnetically moderated C shocks, our higher resolution echelle observations of the H2 velocity profiles in two of the bullets (Tedds et al. 1999) contradict any steady-state molecular bow shock models. This suggests that instabilities or supersonic turbulence may be important in this case. 3D measurements of the corresponding H2 level populations will address this.

Introduction

The nature of molecular shocks, which play an important role in the processes of momentum and energy transfer within star forming molecular clouds (McKee 1989), is still uncertain (Draine & McKee 1993). In this paper we describe how new developments in integral field spectroscopy provide us with the opportunity to self-consistently distinguish between competing shock models. The Orion molecular cloud is the brightest known source of shocked H2 emission and as such has been the primary test bed for theoretical models.

This contribution summarizes experimental work which has been performed predominantly in our laboratory using ion guides and specific traps for studying ions, molecules and dust particles under astrophysical conditions. After a short reminder of the basics of the technique and a brief discussion of our newest device, the nanoparticle trap, we shall review experimental results for low temperature gas phase collisions with H2. In the last part we will summarize our present activities related to chemistry involving cold H atoms.

Introduction

Despite the fact that our knowledge on the role of hydrogen in space has significantly increased in recent years due to a combination of extensive new observations and astrophysical model calculations with fundamental theory and detailed innovative experiments, there are still many unsolved problems related to the interaction of H or H2 with ions, radicals, surfaces and also photons. The most obvious example is the formation of H2 itself; other examples include specific state-to-state cross sections, ortho-para transitions in H2, H-D isotopic scrambling, formation and destruction of the molecule, or the role of hydrogen clusters and anions. In addition to gas phase reactions we will discuss in this paper our most ambitious goal, the detection of catalytic formation of H2 molecules on an interstellar dust analogue localized in a cold trap.

Experimental: Ion guides and particle traps

Inhomogeneous RF or AC fields

From the point of view of experimental techniques, our research is predominantly based on the use of specific inhomogeneous, time-dependent, electrical fields, E0(r,t) = E0(r) · cos(Ωt).

Introduction

Dipole absorption to excited states of diatomic hydrogen lying above 13.6 eV is not usually considered in the discussion of interstellar photophysical processes. The purpose of this contribution is to provide a brief survey of these states, their structure and decay dynamics, and in particular of the theoretical methods used to describe them.

Above about 14.6 eV excitation energy the density of electronic states of H2 increases dramatically so that above 14.8 eV the spacing of successive electronic states becomes smaller than a vibrational quantum, and at an energy about 0.04 eV below the ionization potential (I.P. = 15.4254 eV) it becomes even smaller than a rotational quantum of energy. This means that the usual Born-Oppenheimer description of molecular structure becomes inadequate: rather than considering the rotational/vibrational motion of the nuclei as being slow and determined by the average field of the rapidly moving electrons, one must also take account of the opposite limit, corresponding to a rapidly rotating and vibrating ion core interacting with a highly excited, distant, and slowly orbiting electron. In terms of the level structure this means that for given electronic inversion symmetry (g/u) and electron spin (0/1) the electronic states n,(l),∧ with associated vibrational structure v,N and parity (– 1)p (p = 0, 1) are progressively reordered and eventually form Rydberg series. These series are appropriately labelled n, v+,N+ for each (l), N and parity (– l)p. l is the electron orbital quantum number which is is put into brackets because (albeit useful for book-keeping purposes) it is not always a good quantum number.

Introduction

The analysis of the spectra of astrophysical systems provides valuable information about their composition and dynamics. The purpose of this brief chapter is to introduce some basic concepts of atomic and molecular spectroscopy that are needed to appreciate the role played by spectra in astrophysics. The ideas developed in this chapter will be used in the study of stellar atmospheres, the interstellar medium (Vol. II), and in extragalactic astronomy (Vol. III).

Width of Spectral Lines

When a system makes a transition between two discrete energy levels E2 and E1 emitting a single photon, the frequency of the photon should be equal to ω = (E2 - E1)/ħ. Such a transition should lead to a sharp spectral line of infinite intensity and zero width. In reality, the frequency of the photon that is emitted is not precisely determined and the observed spectral line will have a finite width and intensity. The nature of the width of the spectral line contains important information about the state of the physical system.

The finite width of the spectral line can arise because of several reasons, among which three particular processes are of importance in astrophysics. To begin with, all energy levels (except the ground state) have a finite intrinsic width, that is, the energy of an excited state can be ascertained within only a finite accuracy ΔE2 around a mean value E2. This is because all excited states have a nonzero probability per second P for making a spontaneous transition to lower energy levels.