Annual Report 2011 Max Planck Institute for Astronomy

60 III. Selected Research Areas
ther [α/Fe] nor [Fe/H], and are very different from the
α-old stars on near-circular orbits: going from stars on
near-circular to those on eccentric orbits, the vertical
scale height nearly doubles, and around [Fe/H] –0.6
the average rotational velocity halves while the vertical
velocity dispersion doubles. These differences
might, however, be naturally explained if the stars on
eccentric orbits formed through a gas-rich merger at
high redshift. It is expected that the disk from which
the stars form is not smooth and thin, but clumpy with
already a significant dispersion. Together with further
blurring over the long time since the merger, it might
well be that if any correlation with [α/Fe] nor [Fe/H]
existed, it is washed out by now. Even more so, an early-on
merger origin from an already hotter disk might
help explain the apparent jump between the average
orbital properties of G dwarfs on near-circular and eccentric
orbits.
Aside from understanding the origin of the Milky
Way disk(s), we are also using these G dwarfs as kinematic
tracers to infer the amount of dark matter in the
Solar Neighborhood, which in turn is a crucial ingredient
for experiments designed to detect dark matter particles.
This requires a precision determination of the
local gravitational potential Φ(R, z), yielding the total
mass density from which the accurately measured luminous
density is subtracted to recover the dark matter
density. The simplest way is to solve the vertical
Jeans equation d (νzσ2 z )/dz –νz dF (R0 , z)/dz for a
measured vertical density νz and vertical velocity dispersion
σz for a tracer (sub)population around the Solar
radius R0 .
However, after carefully measuring νz and σz for
sub-samples of G dwarfs with similar [α/Fe] and
[Fe/H], we find that the best-fit vertical Jeans solutions
do not yield a consistent underlying Galactic potential;
the α-young stars require a significant local dark matter
density, while the α-old can be fitted without dark
matter. We expect that this is, at least partly, because
the underlying assumption that vertical and radial motions
can be decoupled breaks down; indeed, whereas
this assumption requires that the velocity cross term
νR νz vanishes, initial challenging measurements
show that it is non-zero and varying with height with
an indication that the variation is stronger with increasing
[α / Fe].
To overcome this assumption on the velocity anisotropy,
we are looking into solutions of the axisymmetric
Jeans equations along curvilinear coordinates that
naturally allow this cross term to vary with height. We
also aim to fit Schwarzschild models that do not make
any assumption about the velocity anisotropy by numerically
integrating in an arbitrary gravitational potential
a representative set of orbits and weighting them
such that observed stellar positions and motions are
fitted simultaneously; the weighted orbital kinematics
yield the velocity anisotropy.
Super-massive black holes
In the last two decades, observational evidence has
mounted that Supermassive Black Holes (SMBH), defined
as black holes with mass, M , above 10 6 M 0 ,
lurk at the centres of most, if not all, massive galaxies.
Measuring M directly is difficult and, with the current
observational means, usually achievable only in nearby
(d 100 Mpc) galaxies. It requires spectroscopic data,
typically with sub-arcsecond angular resolution, as
well as careful dynamical modelling of the stars (or gas)
surrounding the SMBH. Yet, the number of galaxies with
directly measured M has grown considerably, reaching
70 as of 2012.
From this sample, it was found that the M are suprisingly
tightly correlated with several of their host galaxy
(bulge) properties, most prominently the stellar velocity
dispersion, σ, and luminosity, L. A SMBH comprises a
small fraction (on average approximately 1/500) of its
host mass, and accordingly wields significant gravitational
influence only in a tiny region of its host. Hence,
the correlations, also termed “Black Hole scaling relations”,
are not explained trivially. Instead, they are
thought to reflect a co-evolution of SMBHs and their
host galaxies, with the responsible astrophysical mechanisms
currently being subject of investigation. Apart
from their significance towards understanding the formation
of galaxies, as well as the origin and growth of
SMBHs, the scaling relations are nowadays widely used
to determine the M distribution and space density, including
evolution on cosmological timescales. Here, the
empirical scaling relations serve to predict M in a large
number of galaxies, or in galaxies too distant for a direct
measurement. Moreover, secondary (indirect) M estimators,
such as reverberation mapping (for active galaxies)
and the broadline-width method (quasars), are calibrated
by means of the locally-defined scaling relations.
Given the importance of the SMBH scaling relations
in current astrophysics, it is worth noting that their characterisation
is far from secure. The uncertainties pertain,
amongst others, to the universality with respect to host
morphology, the validity range and functional form at the
extreme mass ends, to consistency between scaling relations,
as well as the precision of and especially systematic
errors in the measured M and host galaxy properties.
Tackling the last of these mentioned, we currently work
with a new high-resolution, deep near-infrared data set,
obtained by the WIRCam imager on the Canada-France-
Hawaii Telescope and designed to improve the determination
of bulge luminosity, L bul , of SMBH host galaxies
with directly measured M . Near-infrared luminosity is
of special interest as it serves as a proxy for stellar mass,
and is hardly affected by the presence of dust. We use
these superior data to model bulges by careful and comprehensive
image decomposition. Simultaneously, we
establish the correlation of M with total host luminosity,
L tot . Our results imply that the M – L bul correlation