Discovery and characterization of substellar companions in stellar clusters using karhunen-loève image processing of hst data

  1. Strampelli, Giovanni
Dirigée par:
  1. Antonio Aparicio Juan Directeur
  2. Laurent Pueyo Co-directeur/trice
  3. Massimo Robberto Co-directeur/trice

Université de défendre: Universidad de La Laguna

Fecha de defensa: 14 avril 2021

Jury:
  1. Ramón Jesús García López President
  2. Carlo Manara Secrétaire
  3. Manuela Zoccali Rapporteur
Département:
  1. Astrofísica

Type: Thèses

Teseo: 652755 DIALNET lock_openRIULL editor

Résumé

Stars form in regions of space where sufficient quantities of matter accumulate, reaching densities high enough to eventually overcome all possible agents able to prevent gravitational collapses, such as thermal and radiative pressure, magnetic field, and turbulence. While more detailed analysis is always needed to better understand and constrain key properties of newborn protostars and star formation in general, binarity and multiplicity, in general, are widely accepted as one of the principal channels for star formation, with the inevitable implication that dynamical evolution is an essential part of early stellar evolution (Reipurth et al. 2014). Indeed, studies of binaries and multiple systems during the protostellar phase and later during the pre-main sequence (PMS) phase are fundamental because they offer the perfect stage to observe and analyze the results of fragmentation of molecular clouds, giving meaningful insight into the physics of the processes at play during these phases. The main focus of this thesis consists of the development and application of a pipeline able to recover faint companions embedded in the host star point spread function in HST/ACS and WFC3-IR images. To achieve this goal, the method proposed by Soummer et al. (2012) to perform PSF subtraction using principal component analysis (PCA) will serve as the basic stone from which the whole project is built. When a suitable library of reference PSF is assembled, the PCA is carried out and a Karhunen-Loeve Image Processing (KLIP) analysis of these references is used to create an orthogonal basis of eigenimages, on which the science image is then projected. Three papers so far have been the outcome of this project: • Strampelli et al. (2020a): introduce StraKLIP, the pipeline developed thanks to the collaboration between the ULL (Spain), the STScI (USA), and the Johns Hopkins (USA), to detect and characterize faint close-in astronomical signals around bright primaries taking advantage of the KLIP algorithm just introduced. • Strampelli et al. (2020b): present a first scientific application of StraKLIP on the HST/WFC3-IR data on the ONCin filter F130N and F139M. • Strampelli et al. (2020c): focus on the HST/ACS visible data of the same ONC, presenting a new PSF photometry for the cluster and a second application of the StraKLIP pipeline itself. One of the key characteristics of StraKLIP is that it is designed to work with sets of wide-field imaging observations not specifically tailored for HCI analysis. From this point of view, it represents a new tool that can be applied to virtually any large image dataset, either from space (HST, JWST, WFIRST-Nancy GraceRoman Telescope) or from the ground (e.g. LSST-Vera Rubin Observatory). In Strampelli et al. (2020c), the new ACS photometry has been combined with the precedent WFC3-IR catalog from Robberto et al. (2020), linking together two HST Treasury programs (GO-10246 and GO-13826, P.I. M. Robberto) in one unique catalog of sources for the ONC, spanning a wide range of filters from visible bands to infrared (i.e. filters F435W, F555W, F658N, F775W, F850LP, F130N, and F139M). Thanks to this wide variety of filters available, a Bayesian approach with MCMC strategy have been adopted to estimate three fundamental parameters through SED fitting: mass, age, and extinction. Three different IMF: Kroupa (Kroupa et al. 2001), Chabrier single and system (Chabrier et al. 2003), and two different reddening laws (Rv=3.1 and 5) have been tested. As representative values for this cluster, we find a median mass between 0.15 - 0.21 Msun, a median extinction Av between 1.64-2.42 magnitudes, and a median age between 1.6-1.8 Myr. A more in-depth analysis of the age distribution for stars in the cluster unveiled the presence of at least two different populations with an average age of ~ 0.85 and 1.6 Myr, strongly supporting the idea that star formation in the ONC was neither just one singular event nor a process diluted over a prolonged period, but instead is best described by at least two discrete and sequential episodes of star formation over the first few Myr (Beccari et al. 2017). Taking advantage of the available F658N filter in the survey, the accretion luminosity (Lacc) and the mass accretion rate (dMacc) have been estimated for a sample of ~700 bonafide cluster sources, providing one of the most extensive catalogs of mass accretion rates in the ONC up to-date. The dMacc - Mstar relationship exhibits two different regimes. On one side, we observe that more massive stars (Mstar > 0.3 Msun) display a rather shallow linear relation that agrees with that predicted in the context of centrally irradiated accretion disks around solar-mass stars with an active accretion layer (Mohanty et al. 2005, Hartmann et al. 2006). On the other, lower mass stars (Mstar ≤ 0.3 Msun) can be in turn divided in two different populations as well. If we extend the line fitting the higher mass stars, we see that it nicely matches a population of low mass objects with high accretion rates. There is, on the other hand, a second branch, departing at Mstar ~0.3 Msun, displaying a steeper relation between dMacc and Mstar. These two low mass populations also display age differences: one appears younger with a higher accretion rate, the other more evolved with a lower mass accretion rate. A decay of the mass accretion rate vs. age is not surprising, but in this case, it suggests the coexistence in the ONC of two different young populations. Moreover, an analysis of the extinction of these objects suggests that the older sources typically lie in regions less embedded, i.e., more evolved having been exposed for a longer time to the ionizing flux from the OB stars. For what concerns binarity in the ONC, combining both ONC Treasury programs, a new catalog of 119 binaries has been compiled, spanning a range of separations 0.12-1.95''. As a result, once completeness and chance alignment arguments are considered, a new value of the multiplicity fraction for the cluster is established in the range of Mp = 0.015-1.6 Msun, Mc = 0.004-1 Msun: MFonc=13% ± 1%. Compared to other star-forming regions, the MFonc is ~ 2 times smaller than e.g. Taurus, while it is still compatible with the value observed in the field over a similar range of masses and separations. Also, an overall bottom-heavy q distribution is found for the binaries in the cluster, very close to the value observed for field binaries in a similar range of masses and separations. The binary separation distribution shows two clearly distinct groups of binaries distinguishable by their respective separation: close (separation ≤ 0.6'') and a wide (separation > 0.6''). Moreover, an analysis of the distribution of wide over close binaries as a function of the distance from the center of the cluster, reveals a clear sign of dynamical evolution in its past. Overall, these results hint at the fact that the ONC binaries may represent a good template for the typical population of field binaries, supporting the hypothesis that the ONC may be regarded as a most typical nearby star-forming region in the Milky Way.