VVV gigapixel mosaic of the Milky Way's bulge (credit: ESO)
Research
The VVV Survey
The VISTA Variables in the Vía Láctea (VVV) is one of the six on-going ESO public surveys carried out with the VISTA telescope on Cerro Paranal, Chile. VVV is the first and only near-infrared time-domain survey of the central parts of our Galaxy. The use of near-infrared light enables VVV to tackle the problem of extinction due to interstellar dust concentrated around the Galactic plane, which prevents optical surveys from seeing the Milky Way's inner bulge and the far side of its disk. VVV also opens the time-domain over the inner Galaxy by observing it in up to a hundred epochs over seven years, enabling us to detect variable stars such as RR Lyrae stars and Cepheids, and use them as tracers for mapping the three dimensional structure of the Milky Way.
Since its start in 2010, VVV has been revolutionizing Galactic astronomy, but the biggest results are yet to come.
The VVV data challenge
The VVV Survey produces near-infrared time-series photometry for a billion objects, among which about a million are expected to be variable. Post-processing the photometric data (including tasks like source merging), finding the variable sources, and extracting accurate parameters is not only a grand job for a data analyst, but the sheer quantity of data makes it especially challenging. This challenge is the current hardest training ground in preparation for the data tsunami of the next generation synoptic surveys like LSST, but it's already not an easy wave to surf. In the past few years, I have been working on this task, preparing the pipeline for VVV's variability analysis, and analyzing hundreds of millions of light-curves, providing science-ready data products for the VVV Science Team (and for myself). It is not only a great game (with a steep learning curve) for one who enjoys programming, but it is also very rewarding to be the first one to see the final outcome of endless observations, to feel the results of the work of many to take its final form under my hands.
Unveiling the Milky Way's bulge
From the analysis of infrared photometric data of 2MASS and VVV, it was found that the
Milky Way's ancient central region, the bulge is barred, and has a peanut or rather
X-like shape traced by old red giant (red clump) stars. Such a prominent structure had
been already observed in case of external galaxies. From further kinematic studies
of these stars, we understood that this barred X-shaped structure rotates like a solid
body. The current theoretical interpretation of these observational findings is that the
bulge is in fact a so-called pseudo-bulge, formed from a more ancient disk via buckling
instabilities, although there are a couple of chemo-dynamical properties yet to be
explained.
At the same time, it has also been found that if we use even older stellar tracers, such
as RR Lyrae stars, to probe the 3D structure of the bulge, we do not find the same
features. The spatial structure of the underlying ancient stellar population is not even
barred: it's just a spheroid, similar what we see in the bulges of many external galaxies
that show no sign of boxy / peanut features. The emerging picture is that the Milky Way
has a composite bulge, where an even more ancient spheroid lies under the more prominent
pseudo-bulge. According to theory, such spheroids, together with the halo, form directly
from the hierarchical merging of protogalactic fragments, but it has not yet been clearly
demonstrated to be the early formation mechanism of the Milky Way. It is very important
to further study this hypothesis, especially in view of the predictions of the lambda CDM
paradigm of current cosmology. We are working on the extension of the current RR Lyrae-based
3D maps of the bulge using data from the VVV Survey, and in parallel with that, we are also
making significant efforts to establish a complete observational picture of the chemo-dynamical
properties of bulge RR Lyrae stars, similarly to what has been done earlier with the red clump
stars.
Globular clusters in the bulge are other important relics that should have witnessed its formation and early evolution history. They are poorly studied objects because of the observational challenges posed by extinction. By the analysis of their variable star content, we hope to understand whether bulge clusters commonly stand out from halo ones (as suggested by their various anomalies) by wearing fingerprints of distinct evolution histories. With this aim, we systematically explore the most obscured bulge clusters in the framework of the VVV Survey.
The other side of the Milky Way: the VVV Galactic Cepheid Program
The face-on map of the Milky Way is being drawn since 21 cm radio surveys of Galactic neutral hydrogen started almost 60 years ago, but is still far from completion. There is no consensus on the global spiral arm structure of the Milky Way. Maps obtained from different methods notoriously show significant differences, and most of the features on the nice-looking PR face-on maps are merely an extrapolation of a few actual measurements, with a lot of fantasy added. The far side of the Milky Way's disk has been mapped by measuring density peaks in the radial velocity distribution of neutral and ionized hydrogen gas or by mapping giant molecular clouds (GMCs) using CO lines or masers. A significant weakness of the velocity mapping methods is, particularly at large distances, that they rely on assumptions about the global kinematic properties of the Galaxy, and parameters like the distance to the Galactic center, and uncertainties in these quantities heavily bias the distance measurements. Also, these methods are blind towards the Galactic center and anticenter. The use of stellar tracers such as classical Cepheids would be therefore highly desirable, but it has been historically confined to the near side of the disk due to long-standing observational challenges posed by interstellar extinction. The VVV survey has changed this situation: we have been discovering distant classical Cepheid candidates all the way up to the far edge of the disk using VVV data. Spectroscopic follow-up observations are underway, in order to unambiguously verify their nature. We expect to detect several hundred long-periodic (young) bona-fide classical Cepheids, which will be used to accurately trace the spiral arms in the far side of the Galaxy, and its possible satellites that lie beyond.
The VVV Templates Project
Although the VVV Survey is a treasure trove for the discovery of variable stars, it also presents us with a great challenge: unlike in the optical, infrared photometry does not allow us to make an easy distinction between the type (that is, physical origin) of variability of the stars based on their light-curves alone. This is because the IR light-curves in general are much more featureless than optical ones, and also because IR light-curves of many types of variable stars were scarcely observed before (or not observed at all). The solution is to derive accurate classifiers for the automated classification of IR light-curves of stars discovered by the VVV Survey using machine learning methods (since the dataset is very big), and for that, establish a database of IR light-curve templates of variables of many different kinds, in order to be used as a "training set" for the machine learning. We have been making a large observational effort to obtain the necessary data, and the coverage of the VVV survey area by other, optical surveys such as OGLE also gives an enormous contribution to the case by providing templates directly in the VVV database. I have been working together with a team of mathematicians and computer scientists to tackle the problem, and our results are very promising.
Main collaborators
Márcio Catelan (IA-PUC, MAS, Chile)
Manuela Zoccali (IA-PUC, MAS, Chile)
Dante Minniti (UNAB, MAS Chile)
Wolfgang Gieren (UDEC, MAS, Chile)
Susana Eyheramendy (MAT-PUC, Chile)
Javier Alonso García (IA-PUC, MAS, Chile)
Andrés Jordán (IA-PUC, Chile)
Roberto Saito (U.F. de Sergipe, Brazil)
Oscar Gonzalez (ESO, Chile)
Sonia Duffau (IA-PUC, MAS, Chile)
Eamonn Kerins (U. of Manchester, UK)
Daniel Majaess (U. of Halifax, Canada)