It is now well known that spiral galaxies are in general embedded in massive
dark halos, as evidenced by the flat rotation curves of the cold HI gas
disks observed beyond their visible stellar disks.
However, early-type galaxies (ellipticals and S0s) lack such easily-observed
ubiquitous kinematical tracers, leaving the properties of their outer parts
(mass and stellar orbit distributions beyond an effective radius
still woefully unconstrained,
except for a relatively small sample of well-studied galaxies.
Knowledge of these properties would permit direct comparison with
spiral galaxies and with predictions from theories of galaxy formation.
- Large-R stellar kinematics: Dark halos
The information derivable from stellar kinematics at
R >> Reff
is manifold. One could first of all easily establish the
presence or absence of a large dark halo. So far, only a small
handful of ellipticals have been studied well enough to verify the
presence of dark matter, and those have been exclusively confined to
the brightest galaxies. One could also determine in more detail the
radial profile of the mass distribution. This offers a key
test of CDM hierarchical merger models (e.g., Navarro, Frenk & White
1997), which predict a specific functional form for , as well
as a tight correlation between the function's parameters (total mass
and concentration). In the outer parts of the galaxies
(~ 10 kpc), CDM predicts a dark matter density profile
~ r-2.5, while monolithic collapse models predict
~ r-2. Thus, by measuring at large radii, where the
complicating effects of baryonic dissipation are insignificant, we
will be able to make a clean test of the different models of galaxy
- Large-R stellar kinematics : Angular momenta
Another readily-determined quantity is the angular momentum of the
stars in galaxies' outer parts. Although the main bodies of
ellipticals are known to contain much less angular momentum than in
spirals of similar luminosity, some simulations of galaxy formation
(Zurek et al. 1988; Barnes 1992; Weil & Hernquist 1996) suggest that
large amounts of angular momentum resides in the peripheries of these
systems. These models are approaching the point of being able to make
robust and detailed predictions of the magnitude and direction of
angular momentum in ellipticals, and it is important to investigate
such properties observationally for a reasonably-sized sample of these
- Large-R stellar kinematics : Orbital structure
The orbital structure of the stars in the outer parts can also provide
clues about the formational histories of ellipticals. In the inner
parts, the strong effects of baryonic and dynamical dissipation will
have likely altered the stellar orbital distribution considerably, but
in the outer parts (R ~ 20 kpc ~ 5 Reff), the
dynamical times are long enough (~ 1-2 Gyr) to leave some
observable ``residue'' of the galaxy's formation (such as recent
mergers) in the form of kinematical substructure.
- PNe as kinematical tracers
It is clear theoretically that there is much to be learned by studying
the outer parts of ellipticals. Furthermore, deep surface
photometry of ellipticals has turned up clear evidence that there are
often distinct stellar components at large radii (e.g., Schombert
1986; Weil et al. 1997). The nature of these components will only
become apparent once their kinematics is measured. One of the most
promising techniques for making such measurements is by using
planetary nebulae (PNe) as dynamical tracers. PNe are abundant in all
galaxy types, and the fact that they emit most of their light at a
select few discrete wavelengths means that, with the use of
appropriate filters, they can be identified against the background
light to many Reff, and that their velocities can be easily
measured. Although globular clusters can be used for some of these
purposes, PNe are better employed because they are direct tracers of a
galaxy's bulk stellar distribution.
Previous observing programmes have examined the kinematics of PNe
around a few giant ellipticals using general-purpose instrumentation
(see figure below; Ciardullo et al. 1993; Arnaboldi et al. 1994, 1996, 1998;
Tremblay et al. 1995; Hui et al. 1995). The results have been
tantalising, suggesting that these galaxies have large amounts of
angular momentum misaligned with their inner stellar isophotes.
Unfortunately, these programmes have been hampered by a number of
observational difficulties, as well as by the inordinate amount of time
required to acquire PNe in sufficient numbers.
A sample of 433 planetary nebulae (PNe) overlaid on the stellar isophotes
of NGC 5128,
where Reff 5 kpc.
The PNe marked with circles have negative relative velocities, and
the PNe with crosses have positive velocities.
marks the position of the rotation axis.
With our proposed observations, we will be able to obtain data of
comparable quality for a much more representative sample of
From Hui et al. (1995).
(Click on figure to enlarge.)
With the advent of the Planetary Nebula Spectrograph (PN.S), all these
issues may finally be tackled head-on. The PN.S is a specialised
instrument, optimised for measuring PN kinematics, which is set to
become a World-leading facility for measuring the dynamics of the outer
parts of early-type galaxies. The instrument will be complementary to
the WHT spectrograph currently surveying the inner parts of
such galaxies (R < Reff; Bacon et al. 2001). Drawing on
these unprecedented dynamical studies, we will finally be able to
unravel the nature and origin of early-type galaxies.
For information about our first proposed observing programme with
the PN.S, see the next page.
For a comparison of PNe and globular clusters as
- PNe as distance indicators
A further use for observations of PNe in external galaxies is as
tertiary distance indicators.
The luminosity function (PNLF) of the PNe in any given galaxy has
been demonstrated to follow a rather universal functional form.
In particular, the PNLF cuts off at the bright end at an absolute magnitude
which is relatively invariant among different galaxies.
Just how reliable is the PNLF as a distance indicator? Don't
take our word for it, but look at what Robin Ciardullo, one of
the best-known workers in the field, has to say
Note that [O III] observations alone do not provide
sufficient discrimination against HII regions to prevent
contamination of the PNLF. The PNLF method can be
applied to spiral galaxies (read this Ap.J.
but only if auxiliary measurements are made. Distance work
with the basic PNS is
therefore limited to early-type galaxies.
Where do PNe distances fit in with other distance indicators?
Ciardullo provides the following
To what distance will reliable results be obtained?
The bright end of the
PNLF is unambiguously measured when the `flat' portion of the
luminosity function can be fitted. This requires completeness down
to approximately one magnitude below the brightest PNe (`completeness'
here means freedom from flux-dependent selection). According to
the simulations done this occurs for D=15 Mpc in 7 hours of observing
with a 4-meter-class telescope.
last modified by AJR, 3 July 2001