To date, I have authored and co-authored more than 40 publications, 6 of which are first-author. You can check out my full
record on ADS.
Here is a list of selected publications that best represent my current research areas:
ACCESS I: An Optical Transmission Spectrum of GJ 1214b Reveals a Heterogeneous Stellar Photosphere
In this work, we used Magellan/IMACS in order to study the atmosphere of the
exoplanet GJ1214b, which is the most studied sub-Neptune to date. Past studies have shown that the atmosphere
has clouds, which mutes all the spectral signatures of the atmosphere, leaving us with a flat transmission
spectrum in many wavelenghts. In the optical, however, the transmission spectrum that we observed not only shows
a decreasing slope, but also an overall offset with respect to precise infrarred observations. No exoplanet atmosphere
model alone can explain all the data, but adding bright heterogeneities in the stellar photosphere to those atmospheric models
beautifully do. This shows that heterogeneities in the stellar photosphere can generate signals in the transmission spectra of
exoplanets, which in turn will be very important for future searches of optical atmospheric signatures of planets around M-stars such as
GJ1214. We must not forget that we observe planets orbiting stars, which have their own complex photospheres.
In this letter, we use two ideas in order to predict the C/O ratios in the atmospheres of warm giant planets. The first one, introduced
in Oberg et al. (2011) is that, because in the outer parts of protoplanetary disks (outwards of about 2 AU) important molecules such as
H2O, CO2 and CO condense out, the solid material will get richer in oxygen than the gas and, thus, the C/O ratio of the solids will be
smaller than the C/O ratio of the gas (and of the star). The second one is due to the recent work by Thorngren et al. (2016) in which
it is estimated that the envelopes of these warm giant planets are heavily metal enriched, which is a signature of the important ammount
of solids these atmospheres accreted during their formation. The consequence of this is that we predict that most of the envelopes of these
giant exoplanets should have C/O ratios which are smaller than that of the star, because the ammount of carbon and oxygen accreted from the
solid material overwhelms that accreted from the gas. Because most stars have C/O less than 1, this implies these planets should have
C/O ratios less than 1 as well. If hot exoplanets form in similar ways as their warm counterparts (which is likely at least for a fraction of
them), this implies that water formation should be ubiquitous in them.
In this article, we present the discovery of a dense (with a density of about 2 times the density of water) hot-Jupiter orbiting a star which
is similar to our own Sun. The planet is among the most dense hot-Jupiters discovered to date, with a likely explaination being the large
ammount heavy elements in the planet (which we estimate to be around 110 earth-masses). The planet is about 30 percent more massive than
Jupiter, but 10 percent smaller, and the level of irradiation that it receives from its host star is just above the limit in which we
believe inflation mechanisms start to be important for these hot, big exoplanets. The planet was discovered using photometry from the K2
mission, and radial-velocities from the FEROS spectrograph.
The HATSouth survey is a project which aims at detecting planets from sub-Neptune to Jupiter sized in the southern hemisphere, using
three arrays of automated small telescopes around the world: one array is located at Las Campanas Observatory (Chile), another one
is located at the High Energy Stereoscopic System (HESS) Observatory in Namibia (Africa), and the other is located in the Siding
Spring Observatory (SSO, Australia). Using these arrays, in this paper we report the discovery of six "inflated" hot-Jupiters, all
of which are very interesting for follow-up observations, including possible atmospheric studies. In order to confirm the nature of these
planets, we use photometric and spectroscopic measurements, but here we make use for the first time in the HATSouth survey of high resolution
imaging using AstraLux, a lucky imaging camera mounted on the 3.6m New Tecnology Telescope (NTT) at La Silla Observatory, which allowed us to
obtain nearly diffraction limited high resolution imaging from the ground. In order to analyze that data, code was released along this paper
that generates contrast curves for any image at hand, which is useful for constraining the brightness of possible companions around the planet
hosting star which might mimic or distort the transit signal.
In this article, we report the discovery of a planet which is only 2 times the size of Earth, but 16 times more massive. Although the
errors on our mass measurement are large (6 earth-masses), this is a very interesting planet because of many reasons. First, the planet
falls just where two-layer models of planetary structure predict planets made of pure rock to fall. Of course, these models are too simple
to explain most planets, and is likely that this planet is not made of pure rock but, instead, has either large fractions of water and/or
a thick hydrogen envelope. Second, taking its radius and mass at face value, planets as small and massive as this one are rare. However,
further data is needed in order to better constrain the mass of the planet, which is currently very uncertain. The discovery was made using
data from the K2 mission, and observations made with the HARPS spectrograph mounted on the ESO La Silla 3.6m telescope. Code (named
EXONAILER) is released that allows to model transit lightcurves and radial velocities simultaneously, in an easy and efficient way.
Limb-darkening, the apparent darkening of a star thowards the limb, is a very significant effect on the modelling of transit lightcurves,
which we use to extract several parameters of the system including the planet-to-star ratio, which in turn gives us the planetary radius
if we know the radius of the star. Past works (see below) have shown that it is better to model this effect directly from the transit lightcurve
rather than trusting in stellar models to model the effect, because the levels of precision we are currently achieving are high enough that
small differences between our stellar models and the actual stellar atmospheres make significant impact on the retrieved planetary parameters.
However, different parametrizations exist to model limb-darkening and, to date, it has not been explored which law is the optimal to use in
different settings. In this work, we explore which laws are the best ones to use, which we note strongly depends on the stellar parameters of the
host star, the noise level of the lightcurve and the number of in-transit points. We release code to decide for a given lightcurve which is the
optimal law to use.
Limb-darkening, the apparent darkening of a star thowards the limb, is an evident effect observed in transit lightcurves, which are distorted
because of it. A failure in modelling this effect, thus, might introduce biases, which have not been studied in detail. Here, we
ask whether (1) we understand this effect well enough, and (2) whether it matters that we do at the precision level we are currently working
with. Here we revised how limb-darkening is actually calculated from stellar atmosphere models, and find several inconsistencies both
theoretical (i.e., how models are interpreted) and technical (i.e., how models are actually handled). On top of this, even correcting them
stellar models are not up for the challenge. We conclude this by comparing limb-darkening coefficients extracted directly from Kepler
high-precision lightcurves to the expectations from model atmospheres, where we observed there are significant biases between them. This
answered the first question: no, we don't understand limb-darkening well enough. The second question is answered via simulations, and it is
proven that it does matter given our current high precision photometric measurements. Code is released along the paper which one can use
to generate limb-darkening coefficients for arbitrary response functions, releasing for the first time the detailed calculations involved
in generating them.
In this work, we used Magellan/IMACS for the first time in order to study the atmosphere of the
exoplanet WASP-6b, via the technique of transmission spectroscopy, validating this instrument as capable of
performing these high-precision spectrophotometric measurements at optical wavelengths. We found that the longer the wavelength is,
the smaller the observed radius of the planet was, which is a signature that hazes or condensates in the atmosphere of the exoplanet
can give rise to.