• Ammassi globulari
• Buchi neri di massa intermedia
• Dinamica collisionale

Globular Clusters (GCs) are among the most studied objects in astronomy. They historically were regarded as a single-burst stellar population, as opposed to galaxies, which show
evidence of a more complex star formation history. Such apparent simplicity led astronomers
to regard them as the ideal stellar evolutionary laboratory, while in the field of dynamics, the
truncated-Maxwellian King (1966) models were generally accepted as good fits to the surface
brightness profiles of most Galactic GCs. In the last decade, much of this long-standing confidence in GC simplicity was challenged by improved observations. The Hubble Space Telescope
produced accurate HR diagrams, which for some GCs can be explained only by multiple stellar
populations (Gratton et al. 2004; Bedin et al. 2004; Piotto et al. 2005, 2007). Exotic objects
such as blue stragglers, X-ray sources, and pulsars proved ubiquitous, likely the result of an
interplay between cluster dynamics and stellar evolution (e.g. Belczynski et al. 2006; Shara &
Hurley 2006; Hut 2006). High-resolution imaging of GC cores revealed central density cusps
at odds with King-model expectations of a flat core (Noyola & Gebhardt 2006, 2007). On the
other hand, direct N-body numerical simulations allowed to simulate the dynamics of GCs with
an almost realistic number of stars and dynamical ingredients such as binaries, tidal mass-loss
and a spectrum of stellar masses.
In this context, this Thesis is focused on a particular new ingredient in GC dynamics: Intermediate Mass Black Holes (IMBHs). IMBHs are elusive objects the existence of which is an
intriguing issue in its own right, for the consequences it would have on the seeding of super-massive black holes, on explaining Ultra Luminous X-ray Sources (ULXs), and on modeling
potential astrophysical sources of gravitational radiation.
An approach stressing model-independence, non-parametric statistical tools and extensive
data visualization is followed throughout, and is a distinctive feature of this Thesis. A catalogue
of GC structural parameters (luminosity, fraction-of-light radius and average surface brightness)
is obtained from a model-independent spline-smoothing algorithm applied to GC surface brightness profiles. The parameters thus obtained, together with other properties from the literature, are
extensively explored using data-visualization techniques appropriate for multivariate data-sets
(Pasquato & Bertin 2008, 2010). Tools such as cluster analysis, quantile-quantile plots, kernel
density estimation, and conditioning plots can lead to the discovery of a number of interesting
features, usually hidden to previous research. A relation between deviations from the GC fundamental plane and the slope of central cusps in the surface brightness profile is found (Pasquato &
Bertin 2008). If such cusps originate from IMBHs, this would point to a global effect of IMBHs
on the GC fundamental plane. On the other hand, cuspy profiles appear naturally in simulated
GCs evolved beyond core-collapse even without an IMBH (Trenti et al. 2010).
In this Thesis we contribute to the development of a new method to look for IMBHs in GCs,
based on the effects on mass segregation predicted from N-body simulations with a realistic number of stars. An IMBH is expected to reduce the amount of mass segregation
observed in relaxed GCs. The method is applied to two GCs using HST archival data. NGC
2298 is shown to be an unlikely host to an IMBH (Pasquato et al. 2009), while M10 is more
promising but requires a quantitative determination of the stellar binary fraction to allow a conclusion (Beccari et al. 2010). The model-independent calculation of GC structural parameters
presented in this Thesis is an integral part of the framework I devised to compare simulations
and observations on an equal footing.
N-body simulations of GCs with binaries and a realistic mass spectrum are run to core-collapse and beyond and analyzed as if they were observed GC data-sets (Trenti et al. 2010). The
surface brightness profile of main-sequence stars does not undergo deep core collapse, because
the collapse of dark remnants and/or binaries provide energy to the system. King model fits
to simulated post-core collapse GCs are shown to produce unstable results with respect to GC
structural parameters, lending further support to the non-parametric approach introduced here.