Research Activities

Spatially-resolved kinematics of distant galaxies

I am interested in galaxy formation and evolution, and in particular in how galaxies evolved during the past 8 Gyr (i.e., since z=1). During my thesis and later in the Large Programme IMAGES, we constructed one of the first sample of distant galaxies observed with 3D spectroscopy, using FLAMES/GIRAFFE at the VLT. Such observations allows us studying for the first time the spatially-resolved kinematics of distant galaxies (e.g., Flores et al. 2004; 2006; Puech et al. 2006). Spatially-resolved observations are now routinely obtained on 8m telescopes with, e.g., KMOS on the VLT.

Figure: Examples of morpho-kinematic observations from the IMAGES survey. The top panels show the HST/ACS color images of three z~0.6 galaxies (6 Gyr ago), while the bottom panels show the spatially-resolved kinematics obtained with FLAMES-GIRAFFE IFUs at the VLT. These three galaxies show disordered motions that are very unusual in the present-day Universe. (c)  ESO.

The IMAGES survey allowed measuring the evolution of galaxy scaling relations with increased accuracy. These relations are important to constrain galaxy formation models. Amongt the most useful relations are the Tully-Fisher relation between stellar mass and rotation velocity and the Fall relation between specific angular momentum and stellar mass. We showed that these relations are much more scattered than the relations established in the present-day Universe (Flores et al. 2006; Puech et al. 2010; 2019), which is due to the impact of the large fraction of galaxies with unvirialized motions. These data allowed the first study in distant galaxies of the baryonic Tully-Fisher relation (Puech et al. 2010) and of the evolution of the specific angular momentum (Puech et al. 2007).

Figure below: the baryonic Tully-Fisher relation between baryonic mass (stars + gas) and rotation velocity and the specific angular momentum vs. rotation velocity in z~0.6 galaxies (From Puech et al. 2010 & 2007 respectively). Blue dots represents rotating disks, green squares, perturbed rotators, and red triangles, galaxies with complex kinematics. Left panel: the black line represents the local relation, while the dash lines represent the distant relation determined suing rotating disks only. Right panel: the black dots represent local samples, while the black track represents the effect of a major merger in the jdisk vs Vmax plane.

Morpho-kinematic evolution of galaxies

One of the discovery established by the IMAGES survey is that 50% of galaxies 6 Gyr ago had peculiar morphology and/or kinematics, which is much more than in the present-day Universe (only a few percents; Hammer et al. 2009). In other words, half of the galaxies did not fit the classical Hubble sequence back then. These peculiar morpho-kinematic properties are mostly driven by major mergers (Puech et al. 2012) confirming earlier predictions using more indirect observations (Hammer et al. 2005).

Figure below: Evolution of the Hubble Sequence over the past 6 Gyr from Delgado-Serrano et al. (2010). Spiral galaxies were 2.3 times less abundant in the past, which is exactly compensated by a strong decrease by a factor 5 of Peculiar galaxies (on the left).

The role of galaxy major mergers

Half of local disk galaxies are predicted to be the remnants of gas-rich major merger that occurred over the past 9 Gyr (with baryonic mass ratios larger than 0.25), in remarkable agreement with state-of-the-art semi-empirical theoretical models and morpho-kinematic observations in the progenitors of local disks. The IMAGES survey allowed us to catch on-going gas-rich major mergers in very different phases (i.e., pre-fusion, fusion, and post-fusion/relaxation): all these mergers started at different epochs and are seen at different stages of evolution. The high fraction of morpho-kinematic peculiarities observed 6 Gyr ago is simply resulting from the very efficient combination of morphological and kinematic data which are sensitive to all the phases of gas-rich major mergers.

Figure below: observed Star Formation Rate vs. Time at which the hydrodynamical models best reproduce the observed morpho-kinematic observations in the IMAGES survey. The resulting sequence matches precisely what is expected theoretically from a gas-rich major merger, with a pre-fusion phase in which the two progenitors can be distinguished, a fusion phase during which the two galaxies merger, and a relaxation phase during which a disk can reform depending on the gas fraction at fusion time (and orbits).

Formation of present-day spiral galaxies

Modelling: Gas-rich major mergers are probably the dominant process by which the majority of present-day spiral galaxies acquired their structure. In line with this conclusion is the increasingly large number of present-day galaxies around which very faint streams of stars are detected. We indeed showed using hydrodynamical models that these streams are naturally  reproduced in gas-rich major mergers and several cases were specifically modeled (see NGC5907 Wang et al. 2012 and NGC4013, Wang et al. 2015).


Figure below: These images show the very deep observations and gas-rich major  merger models that reproduce the observations from Wang et al. 2012 and 2015, respectively (see links above). Note that the NGC5907 thin disk shown in the image is actually embedded in a thicker disk revealed by IR observations, which provides an even better match between observations and model. 

Demographics: Using the physics of gas-rich major mergers and the morpho-kinematic demographics of galaxies at z~0.6 from the IMAGES survey, one can directly link the properties of galaxies Gyr ago with those of present-day galaxies. This causal link allows us to predict the morphology of present-day galaxies (fractions of E/S0 and Spirals) within a few percents (Puech et al. 2019). This confirms quantitatively that the present-day Hubble sequence can be viewed as a vestige of recent merger events (Hammer et al. 2009).

Figure below: Prediction of the present-day galaxy morphology (E/S0 vs. Spirals determined according to their B/T) for the descendants of the IMAGES z ∼ 0.6 galaxies. Fractions are given relative to the overall population at given
redshift, while fractions indicated into brackets refer the sub-population of galaxies indicated in the directly-above row.

Using hydrodynamical simulations and initial conditions consistant with those expected in distant galaxies (i.e., large gas fractions), one can show represent most structures that can been seen in present-day galaxies, including pseudo-bulged and even relatively bulgeless galaxies (Sauvaget et al. 2017). Such structures were long thought to result exclusively from pure secular evolution; adopting realistic initial conditions in simulations show how major merger events do not prevent them to form along the virialization process of the merger remnant.

Figure below: Hubble Sequence constructed from virialized galaxy remnants as obtained from simulated major mergers with realistic initial conditions (thesis of Tabatha Sauvaget).

Formation of the Local Group

There are only two relatively massive galaxies in the local group, namely the MW and M31. We showed in 2007 that while M31 can be considered as a good representative case of the present-day intermediate-mass galaxy population, the MW, on which numerous simulations are nevertheless calibrated, appears to be smaller and less massive compared to other present-day galaxies. This was interpreted as an exceptionally quiet merger history for the MW. Further, we suggested in 2010, supported by hydrodynamical simulations, that M31 can be reproduced by a gas-rich major merger that happened about 5.5 Gyr ago. The models we presented can reproduce very well the large-scale properties of M31 such as its disk (rotation and warp), bulge, giant stream, and halo metallicity. M31 seems to be a very good example of the «rebuilding disk scenario» we proposed earlier from studying the evolution of the galaxy properties over the past 6 Gyr (see above).

Figure below: Major merger model reproducing the main properties of M31, including the giant stream.

We are now studying the consequences of a past gas-rich, major merger event at the location of M31 on the current properties of the Local Group. We found that this event could be a natural explanation of why dwarf galaxies around M31 and the MW orbit these galaxies within well-defined planes (which is predicted to be a very unusual event by cosmological simulations). These galaxies would be Tidal Dwarf Galaxies created in a tidal tail formed during the past merger at the M31 location and ejected through space right in the MW direction. The most massive TDGs could be the progenitors of the two Magellanic Clouds: while they were entering and interacting with the MW hot gas halo, they would have left behind a large stream of gas (because of the interplay between stellar feedback and ram-pressure stripping) that nicely reproduces for the first time the overall structure (in phase space) of the Magellanic Stream, the largest structure in the sky after the MW. We are working to refine this model and explore other consequences for the Local Group such as the content in dark matter of the local dwarf spheroidal galaxies.