Image: Hubble Space Telescope/WFC3,
Credits: A. M. Ghez, M. R. Morris, A. Stolte
Circumstellar disc survival in Milky Way starburst clusters
The second major question of the starburst cluster project was:
Can discs survive in starburst clusters?
How does the disc survival rate compare to moderate
star-forming regions such as Orion?
With both proper motion membership and mid-infrared observations at 3.5 micrometer
wavelength, in the thermal infrared regime, we were able to estimate the fraction
of circumstellar discs in each cluster. In the Arches and Quintuplet clusters,
we find disc rates of approximately 4% to 9% at the respective cluster ages of
2.5 and 4 Myr. Almost all of the objects with enhanced thermal emission
were confirmed as cluster members. This implies that circumstellar discs
were able to survive the harsh cluster environment over extended periods
of several millions of years. Alternatively, there must be a mechanism
that allows dense circumstellar discs to form at a later stage during
the cluster evolution.
Disc survival is unexpected in starburst clusters
In the dense environment of starburst clusters, circumstellar discs
are exposed to the intense UV radiation from high-mass neighbours,
to stellar winds, and to close-by encounters that might lead to
gravitational disruption of circumstellar material.
On first glance, the very low fractions of only 4-9% of discs
seem to confirm that discs in starburst clusters dissolve more
quickly than in moderate star-forming environments.
However, the disc host stars had masses of 3-15 solar masses!
The strong UV radiation of these massive stars are expected
to disrupt circumstellar material on timescales of less than 1 Myr
even without the aid of their supermassive neighbours.
This disruption timescale is much shorter than the cluster lifetime.
Disc survival or refueling?
Given our discovery of discs in both Galactic center clusters,
we discuss the possibility that the circumstellar material is
freshly refueled by binary mass transfer
(Stolte et al. 2015).
Hence, we speculate that these discs might be secondary discs
instead of native, primordial survivors. This would then be the
first case where the binary interaction and mass flow between
two massive stars was directly observed.
Whether or not this scenario can explain the observed thermal
radiation needs to be confirmed by future instruments and
telescopes with even substantially higher resolution than we were
able to achieve.
The first discs discovered in the Quintuplet cluster
The figures show the infrared colour-colour diagram and the proper
motion plane of stars in the central field of the Quintuplet cluster.
About 20 sources show thermal emission at L-band (3.8 micrometers),
which is evidenced by their locations (diamonds) to the right side
of the main sequence
cluster population. With one exception, all disc candidates are proper
members of the Quintuplet cluster, as can be seen from their location
inside the member selection circle in the proper motion plane (right
two figures). A detailed description of the figures can be found in
Stolte et al. 2015.
Location of stars with circumstellar material in the Quintuplet & Arches clusters
The maps below display the sources with thermal infrared emission
evidencing circumstellar material on the 2.2 micrometer (K-band) image.
All circles denote disc candidates. Most of the sources are ver faint
and hard to see in the 2-micron image. In the Arches cluster, some discs
were known from one of our previous investigations with the Keck telescope
on Mauna Kea, Hawai'i (Stolte et al. 2010). The discovery of discs in the
Arches cluster center led us to cover the extended areas of both clusters
out to their approximated tidal radius, such that most of the discs around
intermediate-mass could be detected in the survey.
The disc fraction in comparison to young Milky Way clusters
The figure below shows the disc fraction in nearby young star clusters
and in the starburst clusters investigated in the survey. The black circles
indicate young, rich clusters where only the high- and intermediate-mass
stars of types O, B, or A could be included in the analysis. These disc
host stars all have masses of more than 2 to 3 times the Sun.
Blue circles, on the other hand, include nearby young clusters where the
disc fraction was predominantly measured from lower-mass stars, similar
to our Sun or even smaller (T Tauri stars). The three starburst clusters
where we were able to measure the disc fraction from thermal emission
are shown in red. NGC 3603 YC has the highest disc fraction, which is
expected at its very young age.
All starburst clusters display much lower disc fractions
than their lower-mass counterparts in the solar neighbourhood.
It is not yet clear if the low disc fraction is a selection effect
because only high-mass stars could be investigated in the distant
starburst clusters, for which a low disc fraction is expected.
How much the dense environment with its strong stellar winds,
UV radiation field and interactions between stars in starburst clusters
affects the disc fraction remains to be investigated with simulations.
Comparison to NGC 3603
In contrast to the low disc fractions of the Arches and Quintuplet clusters,
a disc fraction of about 30% is found in NGC 3603
(Stolte et al. 2004,
Harayama et al. 2008) from mid-infrared L-band emission and Halpha
as disc indicators. A possible increase in the disc fraction from 20% in
the cluster core to 40% in the cluster outskirts is observed in seeing-limited
VLT/ISAAC observations
(Stolte et al. 2004). These disc fractions include
both pre-main sequence stars down to 2 solar masses and high-mass stars
up to 20 solar masses. If only the OB stars are considered, the disc
fraction decreases to 12% in the cluster centre and 25% at
larger radii
(Stolte et al. 2004). Especially the central disc fraction
from L-band excess sources of only 12% is similar to the Arches fraction
of circumstellar disc candidates. With NGC 3603 being slightly younger
than the Arches
(Kudryavtseva et al. 2012,
Martins et al. 2008), a higher fraction of primordial circumstellar discs
is expected. The low disc fraction around OB stars in both
cluster cores suggests that discs around the mass segregated high-mass
stars are depleted rapidly either by their own UV radiation field or
in the presence of external radiation and stellar encounters in the dense
cluster cores. If, however, the discs in the Arches and the Quintuplet
are of a secondary origin, e.g. from mass transfer in tight binary systems,
the discs would not originate from the survival of primordial circumstellar
material and the disc fraction would be a function of the fraction of
high-mass mass-transfer binaries in each cluster core. In view of a
secondary disc origin from mass transfer
(Stolte et al. 2015), it is interesting to note that
approximately 30% of the pre-main sequence stars in NGC 3603 are located
on a secondary sequence in the colour-magnitude diagram, indicating that
they are near-equal mass binaries
(Stolte et al. 2004). Although the densest
of these systems are good candidates for mass transfer, it is currently not
clear whether the mass stream would be sufficient to build up a dusty
disc with strong infrared emission.