Abstract :
[en] Often quoted for their crucial role in the ecology of galaxies, massive stars remain mysterious.
The exact process leading to their formation, their stability or their stellar winds are so
many questions which are still unanswered. Since the lifetime of these objects is too short to allow
them to move far away from their birth places, the large majority of massive stars are located in
young open clusters or in OB associations. The observations of massive stars in these locations
thus provide with promising information on their formation mechanisms. One of the most interesting
properties, in this context, is their multiplicity. Indeed, the proportion of multiple systems,
the values of orbital periods, of eccentricities, or of mass ratios are parameters directly linked to
the formation mechanism and to the dynamical interactions occurring during the earliest stages of
their existence. Moreover, constraining this multiplicity with a sufficient accuracy allows to obtain
a better determination of the nature and physical parameters of massive stars.
The present work is devoted to the study of the O star populations located in the young open cluster
NGC2244 and in the surrounding Mon OB2 association as well as in several OB associations of
the Cygnus complex. We establish not only the multiplicity of about thirty stars but also their
fundamental parameters such as their masses or their N content. However, the complexity of
observed spectra of multiple systems makes these objects difficult to investigate. Their analysis
thus requires the development and the utilization of mathematical tools such as the "disentangling"
which allows to separate the spectra of the components from the observed ones or such as the
Doppler tomography which aims at mapping the wind interactions seen in some binary systems.
The first part of this dissertation is dedicated to the development and to the characterization of
these mathematical methods whilst the second part focuses on the study of populations of massive
stars. We refine the orbital and physical parameters of systems already known as multiple. In this
context, the analysis of the two components of LZ Cep system reveals modified abundances for
the secondary star, confirming a mass transfer from the secondary towards to primary. We also
detect for the first time a third star in the HD150136 system. This object, composed of an O3, an
O5.5 and an O6.5, constitutes a new test to apply these mathematical methods and thus to better
constrain the physical properties of the system as well as those of each component. Moreover, we
find, in the population of NGC2244, of Mon OB2 and of the Cygnus OB associations, six new
binary systems and derive the orbital solutions for five of them. These results show that NGC2244
hosts only one long-period binary and none with a short period whilst four short-period and no
long-period systems are detected in the Cygnus complex. This obvious lack of short-period system
in NGC2244 contrasts with the O star populations in other young open clusters such as NGC6231
or IC 2944. Besides the multiplicity, the fundamental parameters such as the N content also allows
us to, notably, adapt the so-called Hunter-diagram to galactic O stars. Moreover, we have obtained
photometric data for six O-type stars in NGC2244 and Mon OB2 with the CoRoT satellite. These
data of unprecedented quality allow us to detect the presence, in the O star light curves, of red noise
which is supposed to originate from either the sub-surface convection zone, the granulation or an
onset of clumping in the winds of the stars. We also highlight the existence of non-radial pulsations
in the light curve of Plaskett’s Star as well as a hot spot located between the two components of
that system. In addition, numerous frequencies, extracted from the CoRoT light curves, reveal, for
some of the observed stars, solar-like oscillations or beta Cephei-like pulsations. This analysis thus
gives a first observational constraint on the bright end of the massive star instability strip.
