The Diffuse Galactic Light

The Diffuse Galactic Light (DGL) is a diffuse light observed in the UV-optical-NIR, even in the darkest regions of the sky. It results from starlight that scatters on dust grains, lighting up interstellar clouds. This is a well known effect close to bright stars (c.f. reflection nebula) but the same effect is also observed far away from bright sources, even in the diffuse interstellar medium that is shined-on by the global interstellar radiation field (the sum of the light from every stars). Diffuse Galactic light is now observed routinely, with telescope dedicated to low-surface brightness observations (e.g. DRAGONFLY), with specific camera on large telescope (e.g. Megacam on CFHT), and even by amateur astronomers.

Fig. 1. The Angel Nebula. Diffuse Galactic light observed by amateur astronomers. It is a diffuse interstellar cloud located close to the Ursa Major region, at a distance of about 200 pc.

There is now an increase of interest for the DGL for several reasons. First, it provides a new way to image the structure of the interstellar medium with the greatest details over large areas. Secondly, the DGL brings new informations on the properties of interstellar dust grains, and how they evolve on their journey towards star and planet formation. Finally, the DGL is considered as a potential contamination for cosmological experiments observing in the optical, like Euclid or Vera C. Rubin (ex-LSST). For such experiments that aim at measuring the shape of galaxies with high precision, the very anisotropic and filamentary DGL needs to be either avoided or subtracted from the observations.

Fig. 2. CFHT-Megacam (band i) observations of the diffuse Galactic light at high Galactic latitude. The wispy DGL resulting from starlight scattered on dust grains, reveals the interstellar medium structure but also produces a challenge to measure the shape of background galaxies.

Impact for Euclid

The coverage area of Euclid is so large (15 000 square degrees - Fig. 3) that a significant fraction of the data is expected to be affected by the DGL. The intervening interstellar medium will not only contaminate the extra-galactic emission, it will also absorb and reddened it. Therefore the wavelength dependence of the extinction (the sum of absorption and scattering), and its variations across the sky, must be known to some degree of precision in order to achieve the cosmological goals of these missions. Currently, the knowledge on interstellar dust is not at par with the task.

Fig. 3. The Euclid footprint displayed over the full sky as seen by Gaia. The regions within the blue lines (the Galactic and ecliptic planes) are excluded from the survey. The sky outside this regions is part of the Euclid wide survey, spanning about 15000 square degrees.

Fig. 4. The all-sky dust emission as measured in the far-infrared and sub-millimeter by the Planck and IRAS satellites. This model of dust emission is used to predict DGL contamination for Euclid but it has a low resolution (5') and assumes a constant extinction curve.

Currently, the dust extinction models available that cover the full Euclid observing area are based on far-infrared / sub-millimeter dust emission observations, with an angular resolution of about 5 arcmin, which must be compared with the resolution of Euclid of 0.2 arcsec. In addition these modelling of the optical extinction relies on a constant extinction curve over the whole sky. The spectrum of the dust albedo, responsible for the scattering in the optical, is rather poorly constrained from astronomical observations. Only a handful of studies have been dedicated to this topic, most of them in bright star forming regions. Significant variations are seen in this small sample and it is very difficult at this point to make any precise predictions about the level of scattered light optical experiments like Euclid will be detecting.

Interstellar dust evolution

These informations on the dust extinction spectrum and its variations across the sky, keys to the success of cosmological experiments, are the exact same that are fundamental for interstellar medium physics. Indeed, over the last 10 years we discovered that interstellar dust grains evolve even in the diffuse ISM, away from star forming regions, revealing its reactive nature. The low brightness, diffuse interstellar medium was thought to be rather homogeneous in terms of grain properties and well explained by current models. That was before the advent of the Planck data, which shattered this image of a relatively simple, quiet and static medium.

Confronting Planck data with extinction estimates, the Planck Collaboration showed that the emission-to-extinction predicted by standard grain models is wrong by a factor of 2 to 3 (Planck Collab. XXIX. 2016). They showed that the dust luminosity is independent of its temperature, whereas its opacity decreases with increasing temperature (Planck Collab. XI. 2014). This was counter-intuitive regarding what we thought we knew about diffuse ISM dust.

Later detailed studies of the dust emission compared to dust extinction in the visible revealed that interstellar dust is evolving at earlier stages than what was thought before. While significant changes in dust opacities were unexpected, significant variations in the gas-to-dust mass ratio were no less so. Until recently, this ratio, often referred to as the Bohlin’s ratio, was believed to be canonical. However, the latest observations showed that this ratio is rather higher than previously thought by 20 to 60% and subject to local variations (e.g. Nguyen et al. 2018). These new observational constraints raise fundamental questions about how solid particles evolve in space, and how matter is assembled to later form planetesimals. Understanding this coagulation process is fundamental for interstellar medium physics but also to evaluate masses and star formation rates in galaxies as dust emission is often the only tracer available. Characterising these variations could also be paramount to the success of the Euclid mission.

PhD subject

The PhD subject proposed here will consist in the analysis of the DGL from state of the art observations of cosmological fields, obtained with Megacam on the Canada-France-Hawai telescope, like the one presented in Figure 2. These observations are pre-cursors of the upcoming Euclid observations. They are used in many aspects of the Euclid data processing and analysis preparation. One fundamental aspect, the topic of this thesis, is to know more about the DGL as a contaminating emission but also as a tracer of interstellar dust and structure.

More specifically, the goal of the PhD is to characterize grain properties variations in the diffuse ISM by confronting their thermal emission from mid-IR to submm with their ability to efficiently scatter optical starlight. Variations in the C abundance locked in grains, changes in their composition or size distribution are degenerate for these two tracers (Ysard et al. 2015). On the other hand, it is clearly established that scattering is extremely sensitive to the grain exact size and surface composition (e.g. Ysard et al. 2016, Ysard et al. 2018).

To model the observed DGL, the starting point will be The Heterogeneous dust Evolution Model for Interstellar Solids (THEMIS, Jones et al. 2017). Taking into account realistic variations in the dust properties and implementing them in the THEMIS optical property calculator, the PhD student will make a 3D cirrus cloud model to fit the observations from the optical to far-IR and thus trace variations in the dust emission, extinction, and scattering properties. This will then be used to predict optical scattering for Euclid using all-sky dust emission data, and hopefully propose separation strategies to reveal the extra-galactic signal on the largest area possible.