The origin of thick disks and their connection with thin disks
Most disk galaxies, when seen edge on, show a typical morphological structure with two disks of different thicknesses: the thick and the thin disks. Theoretical studies have proposed different formation scenarios to explain the observed thick-disk properties: 1) in-situ formation at high redshift from turbulent gas (Brook et al. 2004), 2) dynamical heating of stars formed in a preexisting thinner disk, which becomes thicker (e.g., Quinn et al. 1993; Di Matteo et al. 2011), and 3) accretion of stars from satellites through galaxy mergers (Abadi et al. 2003). However, observational works have been alternatively supporting one or another formation scenario, and the origin of thick disks is still a matter of debate. The advent of integral-field spectroscopy (IFS) has brought important advances in this field. Our recent IFS studies, supported by our recent results from simulations, revealed that different scenarios do not have to be mutually exclusive and thick disks can form via combinations of them (see below).
Observations of edge-on galaxies.
Thick disks in lenticular galaxies. My PhD thesis “Unveiling the chemo-dynamical properties and origin of thick disks in galaxies” focused on the origin of thick disks, aiming at solving the controversy associated with their formation scenarios. Thick disks are the thicker and fainter stellar envelopes of the bright thin disks located in the midplane of disk galaxies.
From deep IFS data from MUSE at VLT from the Fornax 3D survey (Sarzi et al., 2018), during part of my PhD thesis, I performed a two-dimensional analysis of the kinematic and stellar-population properties of three edge-on lenticular galaxies in the Fornax cluster. I mapped with unprecedented spatial resolution the stellar kinematics and populations of these galaxies. I extracted the mean stellar age, metallicity and [Mg/Fe]-abundance ratio, using the full-spectral-fitting technique pPXF (Cappellari & Emsellem, 2004; Cappellari, 2017), coupled to MILES single-stellar-population (SSP) models (Vazdekis et al., 2015). The structure and stellar populations of the three galaxies showed global differences between each other: one is overall very old and shows the most complex structure in the central region, with a nuclear disk and a boxy bulge, while the other two show young stars in their thin disk. Since these galaxies are located at different distances from the Fornax cluster center, with the oldest galaxy located in the core, possible evolutionary paths in relation to their environment were discussed. In spite of the global differences between the three galaxies in the sample, their thick disks share a composite star-formation history.
Furthermore, a structural decomposition allowed the reconstruction of the star-formation history of the individual components. The star-formation history of their thick disks showed the presence of a significant fraction of younger chemically-distinct stars, formed in a satellite galaxy. My study provided the first observational signs of accretion in stellar populations of thick disks and uncovered a complex formation scenario combining an in-situ formation of the oldest main component at high redshift, with the later accretion of younger and more metal-poor populations (Pinna et al., 2019, A&A, 623, A19; Pinna et al., 2019, A&A, 625, A95).
During my postdoctoral phase, I achieved a more complete picture of the origin of thick disks and their connection with thin disks, investigating also disk galaxies with different morphological types other than lenticular, and in other environments not as dense as a galaxy cluster. My recent simulation-based studies have revealed the secrets of the thick/thin-disk evolutionary connection and quantified the role of accreted stars in the buildup of thick disks.
The thick disk of a massive Sb galaxy. Using a very similar approach as in my PhD thesis, we analyzed deep MUSE data of one massive Sb, NGC 5746, where we also found a complex and old thick disk made up of different stellar populations: a more metal-rich component, probably formed in situ, and a more metal-poor one, accreted from a satellite. Also in this galaxy, we found signs of a past merger about 8 Gyr ago, that might have contributed a large fraction of thick-disk stars and fresh gas for the extended star formation in the younger thin disk. This important merger, with a large impact on the growth of thin and thick disks, did not modify galaxy morphology. The massive satellite fueled thin-disk star formation without causing significant vertical dynamical heating and, more interestingly, did not trigger the formation of a classical bulge. This important result challenges models of hierarchical assembly of galaxies, where mergers drive the morphological transformation from later to earlier galaxy types (Martig, Pinna et al., 2021, MNRAS, 508, 2458).
Capturing the first phase of the inside-out formation of a thinner disk embedded in a thicker disk…One more galaxy, NGC 3501, a low-surface-brightness Scd, was analyzed by my student Natascha Sattler during her BSc thesis (University of Heidelberg), defended in April 2021 and later leading to a published paper (Sattler, Pinna et al., 2023, MNRAS, 520, 3066). Only one disk component was identified in a previous morphological decomposition (Comerón et al. 2018). However, its stellar-population maps reveal, apart from a thicker disk, mainly young and relatively metal poor, also a younger, smaller metal-rich thinner disk in the inner region of the midplane. We propose that this young galaxy, with not yet clearly distinct thin and thick disks, is being observed in an early stage with a thick disk already in place, and a thin disk in the first stages of its formation. This work shows, for the first time, the first stages of the formation of a thin metal-rich disk in the inner region of a late-type spiral galaxy, embedded in a thicker disk
The data used for this work corresponds to the ESO program “Uncovering the origin of thick disks with MUSE” (098.B-0662), which I led as PI while I was a PhD student. The program was allocated 9 hours of observation with MUSE (in a period, P98, the number of requested nights on the telescope UT4 exceeded the available time by a factor of 18.9).
Thick disks in dusty star-forming galaxies. During her MSc thesis «Thick and thin disc formation and evolution in eight late-type galaxies from their two-dimensional stellar population analysis», my student Natascha Sattler (University of Heidelberg) analyzed eight more edge-on galaxies, five of them showing a clear morphological distinction between a thick and a thin disk. This is somehow a different galaxy sample from what was previously analyzed, since these galaxies are low-mass, intensively star-forming and quite dusty galaxies, most of them showing offplane ionized gas. This is the first systematic stellar-population study of thick disks in galaxies of this type with integral-field spectrocopy. We used MUSE data from the ESO Data Release of the Program «Deep MUSE IFU Observations of Nearby Edge-On Galaxies» (Comerón et al. 2019). Our results showed that these galaxies are overall young, with very young thin disks but also relatively young thick disks, just slightly older than their corresponding thin disks. Interestingly, these galaxies are also overall metal poor (subsolar metallicities), with the thin disks slightly more metal rich than the thick disks. Star-formation histories point to a surprisingly slow thick-disk formation, extended to very recent times. This is supported by previous findings of extraplanar star formation and signs of gas accretion (Rautio et al. 2022). These results suggest a different thick-disk formation scenario from previous publications, with a late in-situ formation and growth, probably fueled by gas accretion. This work led to one publication which has been recently submitted (Sattler, Pinna et al. 2024).
![Light-weighted age (left), metallicity (middle) and [Mg/Fe]-abundance (right) maps for the full sample. For the galaxies with
distinct morphological thick and thin discs, grey dashed lines mark the regions above and below which the thick disc dominates the
vertical surface-brightness profiles](https://francescapinna.es/wp-content/uploads/2024/10/lw_maps-637x1024.png)
distinct morphological thick and thin discs, grey dashed lines mark the regions above and below which the thick disc dominates the
vertical surface-brightness profiles.
The GECKOS survey.
GECKOS is a new international collaboration (about 60 members from all around the world, P.I. J. van de Sande) with the main goal of tracing the mass assembly and accretion history of Milky Way-mass disk galaxies. GECKOS targets a representative sample of 35 edge-on galaxies with deep MUSE observations, to reconstruct their histories from gas and stellar properties. While observations are still ongoing, the first results are very promising. With one of the main focuses being the origin of thick disks and their connection with thin disks, I have been a member since the preparation of the ESO-VLT large program observing proposal (1110.B-4019, subm. 25/03/2022), and I am currently co-heading the stellar-population efforts in the team, as well as the studies on the origin of stellar disks. I am also leading two major science projects which will use the entire galaxy sample and unveil trends of thick and thin disks with galaxy star-formation rates (SFRs): “Uncovering the origin of thick disks from their star-formation histories” and “Evolution of galaxies with one-disk structure”. GECKOS results have so far shown that galaxies with higher SFRs display little differences between morphological thick and thin disks. They are still forming their thick disks which are relatively young and formed slowly. More quiescent galaxies show a clear-cut distinction between stellar population properties of thick and thin disks, pointing to the traditional early and fast formation of thick disks, and later and slower formation of thin disks (Pinna et al., in prep.).
Simulations
Thick-disk properties and origin in AURIGA simulations. Integral-field spectroscopy observational samples have been historically very limited since deep data are required to extract the stellar populations of faint structures such as thick disks. Simulations offer large samples of galaxies and allow us to go back in time to the moment when different structures formed. I have recently used simulations to complement my observational work on thick disks and conclude on their formation. This work, including a total of three papers, establishes a bridge between observations and theory, answering some of the puzzling questions about galaxy evolution.
I analyzed stellar populations of thick disks in 24 spiral galaxies from the AURIGA simulations, projected edge on. I produced mock age, metallicity and [Mg/Fe] maps seen edge on, which show similar thick disks to the observed ones but with a larger variety of stellar-population properties. Mapping properties at different redshifts shows the time of formation of thick and thin disks, and the time of mergers when these are present. With an initial component formed already thick at high redshift, plus a (often significant) fraction of accreted stars, and with mergers often also fueling in-situ growth with accreted gas, thick disks result from the interplay of in-situ and ex-situ processes. The distribution of accreted stars and the quantification of the accreted mass fraction show that minor mergers have a significant impact on thick disks (almost insignificant in most thin disks). The balance between those processes varies from a galaxy to another (Pinna et al., 2024, A&A, 683, A236). Mock stellar-population maps are publicly available to the community in the AURIGA webpage.
In the second part of the work, I extracted the star-formation histories and average chemical evolution of the 24 thick disks. I also tracked star particles to check where they came from, and split the star-formation histories into in-situ and ex-situ components. This confirms the hints found in observations, indicating that thick-disk star-formation histories result from the interplay between internal evolution, with the chemical enrichment driven by the formation of subsequent generations of stars, and external processes such as galaxy mergers, contributing more stars and gas with different metallicity and chemical abundances (Pinna et al., in prep.).
Chemical bimodality of stellar disks. The well-known bimodal distribution of Milky-Way disk stars in the [α/Fe] -metallicity plane is often used to define thick and thin disks. In external edge-on galaxies, there have been attempts to identify this type of bimodality using integral-field spectroscopy (IFS) data. However, for unresolved stellar populations, observations only contain integrated information, making these studies challenging. I used the same AURIGA sample to assess the ability to recover chemical bimodalities in IFS observations of edge-on galaxies. I considered integrated stellar-population properties from previously obtained mock maps, and investigated how the distribution of stars in the [Mg/Fe] – [Fe/H] plane is affected by edge-on projection and spatial binning. Distributions are observed as continuous (but mostly bimodal) distributions from high [Mg/Fe] and low [Fe/H], to lower [Mg/Fe] values and higher [Fe/H]. The overlap in [Fe/H] is small, and different [Mg/Fe] components show up as peaks instead of sequences. The larger the spatial bins, the narrower the [Mg/Fe] – [Fe/H] distribution. We have also assessed the correspondence of chemical bimodalities with the distinction between geometric thick and thin disks. Integrated properties of geometric thick and thin disks in mock maps do mostly segregate into different regions of the [Mg/Fe] – [Fe/H] plane. In bimodal distributions, they correspond to the two distinct peaks. Our results show that this approach can be used for bimodality studies in future IFS observations of edge-on external galaxies (Pinna et al., in press).
Disk dynamical heating.
Part of my work is dedicated to one of the proposed formation scenarios for thick disks: the dynamical heating of pre-existing thinner disks, whose stars would have acquired larger velocity dispersions some time later than their formation. The shape of the disk Stellar Velocity Ellipsoid (SVE), with the vertical, radial and azimuthal components of the stellar velocity dispersion as semi axes, has been often used as an indicator of the predominant disk dynamical heating mechanisms (e.g. Merrifield et al. 2001, Binney & Schönrich, 2016). A more spheroidal SVE was traditionally associated with isotropic heating agents such as encounters with giant molecular clouds or galaxy-galaxy interactions, while flatter SVE were traditionally associated to mechanisms perturbating motions of stars mainly in the disk plane, such as spiral density waves. Furthermore, a clear trend of the SVE shape (measured as the ratio between the vertical and radial velocity dispersion) with the galaxy Hubble type was found by Gerssen & Shapiro Griffin (2012). With earlier-type galaxies displaying more isotropic SVEs, three-dimensional agents would be the main contributors to the heating of their disks. On the contrary, the flatter ellipsoids in disks hosted by late-type galaxies were due to the action of radial agents, most likely the spiral density waves.
For my MSc thesis “Unveiling the sources of disk heating in spiral galaxies”, I analyzed the stellar velocity dispersions for a sample of 29 intermediate-inclination disk galaxies from the CALIFA integral-field spectroscopy (IFS) survey. I used the emcee Python implementation of the MCMC (Markov chain Monte Carlo) minimization method to fit the three velocity dispersion components (i.e. the SVE shape). These were projected to fit the observed velocity dispersion along the line of sight (LOS), using an analytical model based on the axisymmetric thin-disk approximation and exponential shape of the velocity-dispersion radial profile (Gerssen et al., 1997; Weijmans et al., 2008). My detailed analysis of the radial profiles and maps of the velocity dispersions showed a much larger complexity and variety of different patterns. This suggested that such analytical model was too simplistic and in fact provided good fits only for a small subsample of seven galaxies with smooth and regular exponential profiles. We analyzed the shape of the SVE of this subsample as a function of Hubble type, but no clear trend was found. Both the complexity in the velocity dispersion profiles and maps and the lack of a trend of the SVE with the Hubble type were found also in a sample of 26 galaxies from zoom-in cosmological simulations from Martig et al. (2012), clearly pointing to a much more complicated scenario than what was previously proposed in the literature. This work was presented orally in Cozumel (Mexico) in April 2016 at the international conference “The interplay between local and global processes in galaxies”.
I followed up on this topic during my PhD and revisited the SVE-Hubble type relation using an exhaustive compilation of observational measurements from the literature, and interpreted the results with the help of the same numerical simulations used in my MSc thesis. No clear trend was found between the SVE shape and galaxy morphology. Especially for late-type galaxies, all kinds of shapes seem to be possible for the SVE. This points to a quite complicated picture where different disk-heating agents can be predominant in the same Hubble type, while the same agents can act in different galaxy types. This complex scenario was validated via the analysis of simulations back in time to high redshift, showing a large variety of peculiar cases. This work led to my first published paper as first author (Pinna et al., 2018, MNRAS, 472, 2697).
Later on, during my first postdoctoral phase (POP), I proposed and supervised the Summer Grant “Assessment and improvement of an analytical model to fit the SVE of disks in galaxies” (student: P. Rodríguez-Beltrán). We used the same approach used in my MSc thesis to fit MUSE data, much deeper and with higher spatial resolution than data from CALIFA, of ten disk galaxies from the Fornax 3D Survey. This work showed that the high complexity of disks in the Fornax Cluster, nicely shown in the high-resolution velocity dispersion maps (Iodice et al., 2019), require more powerful and sophisticated techniques, such as dynamical models based on orbit superposition (Zhu et al., 2018), to fit the individual velocity dispersions. I am currently leading the analysis of the disk SVE in disk galaxies of the Fornax 3D sample, which is mostly being carried out by my BSc student Ernesto Quintana Ojeda (Unversidad de la Laguna). He is using the results from the population-orbit superposition method (Zhu et al., 2020; Ding et al., 2023) extracted by the Fornax 3D team. These results allow to associate ages and metallicities to velocity dispersions and thus to analyze trends of the SVE with stellar populations. This will provide me and the community a view of dynamical evolution of galaxies at large scale.
Finally, I was also part of the project “Local variations of the SVE”, using AURIGA high-resolution zoom-in cosmological simulations to show the importance of peculiarities and spatial variations of individual velocity dispersions and their ratios, adding new insights to my previous work. The connection between local and global kinematic properties is revealed to be very important when extracting general trends on the SVE. (Walo et al. – incl. Pinna – 2021; Walo, Pinna, et al., 2022).