Introduction

Interstellar Dust: A Key To Understanding Galaxy Evolution

Understanding galaxy evolution is one of the main objectives of observational cosmology, as it allows mapping the history of the Universe, from the dark ages to the present times (e.g. Madau & Dickinson, 2014; Buat, 2015). At the center of this evolution lies the InterStellar Medium (ISM). This complex intertwining of ionized, atomic and molecular gas phases mixed with dust grains, fills the volume of a galaxy, ultimately leading to star formation (SF), by gravitational collapse (e.g. Klessen & Glover, 2016). Although accounting for only $\simeq \text{1}\%$ of its mass, dust is an essential component of the ISM. It consists of solid particles ($\text{0.3}\text{nm}\lesssim \text{radius}\lesssim \text{0.3}\mu m$) made out of the available heavy elements, predominantly arranged in silicate and carbonaceous compounds (e.g. Draine, 2003a). These grains, absorbing and scattering starlight, have a radical impact on a galaxy.

• They have a nefarious role of obscuring our direct view of star formation. They normally re-radiate $\simeq \text{25}\%$ of the stellar power in the InfraRed (IR; cf. Table A.4), and up to $\simeq \text{99}\%$ in ultraluminous IR galaxies (e.g. Bianchi et al., 2018; Clements et al., 1996).
• Dust is an essential ingredient of star formation, as it contributes to (e.g. Li & Greenberg, 2003):
  • radiatively evacuating the gravitational energy of collapsing molecular clouds;
  • shielding the molecules from starlight, which protect them from destruction and reduces their ionization fraction, allowing the formation of protostellar cores.
• In addition, grains are responsible for the heating of the gas in PhotoDissociation Regions (PDR), by the photoelectric effect (Draine, 1978; Kimura, 2016).
• They are also catalysts of numerous chemical reactions, including the formation of the most abundant molecule in the Universe, H₂ (Gould & Salpeter, 1963; Bron et al., 2014).
• Elongated grains tend to align with the magnetic field. Polarized extinction (in the visible) and emission (in the IR) by grains are therefore one of the most popular tools to study the magnetic field in the ISM (e.g. Planck Collaboration et al., 2016b,d; Guillet et al., 2018).

Dust physics is characterized by the great complexity of its make-up, as the number of ways to combine elements to build interstellar solids is virtually limitless. Most of the progress in this field thus relies on empirical constraints: observations and laboratory experiments on cosmic dust analogs. Our current knowledge of interstellar dust (ISD) properties is however hampered by several factors (Galliano et al., 2018, for a review). First, observations of interstellar regions are always the superimposition, along the line of sight and within the telescope beam, of a range of physical conditions: (i) intensity and hardness of the InterStellar Radiation Field (ISRF); (ii) gas density; and (iii) presence of shocks. Consequently, since we can never accurately recover the 3-Dimensional (3D) structure of a region, several degeneracies between the grain constitution and their excitation prevent a unique solution. Second, the grain constitution ¹ is known to evolve under the effects of ISRF and gas density (e.g. Draine, 2009; Jones et al., 2013; Ysard et al., 2015). It is thus likely that, in addition to variations of excitation conditions, ISD observables are coming from a combination of altered grain mixtures. Finally, the derivation of precise dust properties, even from observations towards a uniform, uncontaminated region, is limited by an incomplete spectral coverage and by instrumental uncertainties.

The Relevance of Nearby Galaxies

Due to their proximity, ISM regions of our own galaxy, the Milky Way (MW), can be observed with the finest linear resolution. MW studies have consequently laid the ground for the development of physical dust models (Draine, 2003a, for a review). They are however limited by the small range of environmental conditions they span.

• There are no really massive star-forming regions in the MW. The only Super Star Cluster (SSC; $M_{\star }\gtrsim 10^5{M}_{\odot }$; e.g.  Johnson, 2001; Dowell et al., 2008) is Westerlund 2 in RCW 49 (Moffat et al., 1991).
• As in most spiral galaxies, there is a radial gradient of metallicity (the mass fraction of elements heavier than He, noted $Z$; e.g. Asplund et al., 2009). It however ranges narrowly ($0.7Z_{\odot }\lesssim Z\lesssim 2Z_{\odot }$;  Henry & Worthey, 1999).
• The massive central black hole, Sgr A${}^{\star }$ of the Galactic center is relatively passive (Mezger et al., 1996) compared to Active Galactic Nuclei (AGN), such as NGC 1068 (e.g. Le Floc’h et al., 2001).

MW studies are also limited by the confusion along the sightline, as we are seeing the projected material of the entire disk. Finally, distances of interstellar clouds can be difficult to estimate for the same reason, although 3D maps of the MW are becoming more precise (e.g. Lallement et al., 2018).

In contrast, nearby galaxies (closer than $\simeq \text{100}$ Mpc;  Galliano et al., 2018) represent an under-tapped population with several potentials. First, they harbor a wider range of environmental parameters, allowing us, in particular, to probe dust in extreme conditions.

• Nearby galaxies, especially Blue Compact Dwarfs (BCD), contain numerous SSCs (e.g. O’Connell et al., 1994; Johnson et al., 2000; Martín-Hernández et al., 2005). They are ideal laboratories to understand the impact of massive star formation on the ISM.
• Extremely low-metallicity objects, such as I Zw 18 ($Z\simeq 1∕35Z_{\odot }$;  Izotov et al., 1999), allow us to study dust in environments where the chemical enrichment resembles primordial galaxies.
• AGNs have a radical impact on the ISM of their host galaxies, heavily processing the grains (e.g. their crystalline fraction; Spoon et al., 2006). In addition, bright AGNs can be used to study grains in absorption (e.g. Spoon et al., 2002; Mason et al., 2007).

Second, face-on galaxies, observed at high Galactic latitude, provide clearer sightlines than in the MW. Finally, the lower linear resolution we can reach in nearby objects ($\simeq \text{100}$ pc to 1 kpc in the IR) is the ideal length scale to adopt a statistical description of the distribution of clouds and stars ², whereas detailed, parsec-scale MW studies are left to the nearly impossible task of inferring the precise geometry of each single cloud and the position in space of the surrounding stars.

Scope of the Manuscript

Most of my career until now has been devoted to studying the dust properties and their evolution in nearby galaxies. I have chosen to focus the present manuscript on the following directions.

• It is primarily a synthesis of my achievements in this field. The goal is not to repeat the content of my publications, but to put them in perspective with other studies. For that reason, the presentation of my results occupy only a small fraction of the manuscript.
• It also provides a review of our current understanding of ISD, and outlines the most promising directions this field should explore during the next decade.
• Finally, I have tried to compile introductory material, concepts and figures that could be useful to students starting in this field, but that are otherwise scattered across the literature. Unless otherwise noted in the caption, the figures presented here are original and are licensed under CC BY-SA 4.0 ³.

It is divided as follows.

Chap. 1

provides a reminder of the main concepts at the foundation of ISD physics.

Chap. 2

gives a general introduction about the most reliable observational evidences we have about ISD, and the current models attempting at synthesizing them.

Chap. 3

reviews dust properties in the MW and nearby galaxies, and the way they are constrained.

Chap. 4

reviews evidences of dust evolution and models accounting for their formation and destruction at the scale of a galaxy.

Chap. 5

is a more original take on my methodological approach, motivated by some epistemological concepts.

Chap. 6

is a summary of what we have learned about ISD in the past decade and what we should do during the next one.