Research Interests

Back to my home page

Black Holes and Elliptical Galaxies

I have been studying the central regions of galaxies in order to understanding their evolutionary history. The fist step involves measuring the mass of any central black hole. Together with the Nuker Team , we have found that nearly all galaxies contain a central supermassive black hole. Furthermore, the mass of the black hole strongly correlates with various galaxy properties. I have been working on a Black Hole Webpage that describes the data and results. In order to measure the mass of the black hole, we use a sophisticated orbit-based model to represent the galaxy. These models provide one of the most general solutions for how stars can orbit in a galaxy. This modeling code allows us to not only measure the central black hole accurately, but also to determine how the stars orbit throughout the galaxy. Both of these relate to how the galaxy formed and evolved.

The plot on the right represents the paths of stars as they orbit in a galaxy that contains a black hole. Each color is a different star.


HETDEX stands for the Hobby-Eberly Telescope Dark Energy Experiment. Dark energy is one of the most important quantities of our Universe, but we have essentially no idea as to what it is. Along with Gary Hill and Phillip MacQueen, we have designed and are building an instrument capable of accurately measuring potential evolution of dark energy. The experiment is described here . We will use the HET to conduct a large redshift survey, detecting nearly 1 million galaxies over a huge volume. These results will determine the expansion history of the universe using baryonic oscillations, thereby determing the evolution of dark energy.

Evolution of Elliptical Galaxies

I have been measuring the Fundamental Plane for field ellipticals out to z around 1.0 using data gathered for the DEEP project, on the Keck telescope. The data show strong luminosity evolution: around two magnitudes at z=1.0. This amount of galaxy brighening suggests a rather recent formation epoch--around z=1.5--in contrast to results for cluster galaxies, which suggest formation epochs before z=3.0. These observations will provide a strong constraint on evolutionary models for galaxy formation.

Dynamics of Globular Clusters

Dynamics of globular clusters continue to be one of my main areas of interest. Tad Pryor, Ted Williams, and I have pioneered the use of an imaging Fabry-Perot (FP) for acquiring large velocity samples in globular cluster (both Galactic and Magellenic). This work is a continuation of my thesis work, where I developed the observing strategies and analysis techniques for using the FP. Current techniques, such as fibers and multi-slit spectrographs, suffer from crowding problems in dense regions since light from a single bright star will dominate the neighboring starlight. The two-dimensional information of the FP allows one to properly apportion the light. This technique is the most efficient for obtaining velocity information in crowded regions, as we have increased the number of stars with measured velocities in the central regions by factors of 10--100 with only 3--4 hours of observing time per cluster. I have now begun collaborations with other astronomers who are interested in using this technique.

With the velocity measurements from the FP, I have been developing new techniques which can determine the mass density profile non-parametrically. The technique takes estimates of the velocity dispersion and the surface brightness profile and, after inversion through the Abel integrals, uses the Jeans equation, assuming isotropy, to provide a non-parametric estimate of the mass density and mass-to-light (M/L) ratio as a function of radius. Applying this technique to many clusters, we have found increases in the M/L in the central and the outer regions. The central increase is explained through mass segregation and, for the first time, we are able to directly estimate the heavy remnant population. The increase in the outer parts can be explained with having a population of low mass stars at those radii. The advantage of using the non-parametric techniques is that we are now able to put strong constraints on models, either N-body or Fokker-Plank, and we can directly estimate the present-day mass function. The mass functions for all of the clusters studied show a significant number of objects with a mass around 0.7~M$_\odot$. If we assume these are primarily white dwarfs, we can use their numbers to place constraints on the initial mass function.

From the two-dimensional FP data, I am able to measure a velocity map, pixel-by-pixel, using the integrated light of the cluster, providing an accurate measure of any rotation. For most of clusters, there is a clear indication of rotation in the inner 0.5 parsecs. We also have a measure of the rotation at 3 parsecs, and the amplitude of the rotation is generally the same. Although not dynamically significant, the flat rotation curve has not been seen in standard models and N-body codes, since solid body rotation is expected. For M15, we have measured a significant increase in the rotation amplitude at small radii. This increase is not expected from evolutionary models and may be the result of a central mass concentration. A $1000\Msun$ black hole is consistent with both the rotation and velocity dispersion profiles.

Adaptive Optics Studies

Adaptive optics will quickly become one of the most important astronomical tools. I have recently been involved in two AO projects: kinematic observations for core-collapse clusters and M/L gradient in late-type galaxies.

Globular Cluster Systems

Globular cluster systems around early-type galaxies provide an excellent opportunity to study the evolutionary and formation history of the host galaxy. Working with Markus Kissler-Patig we have used the radial velocities of the clusters around M87 to measure its spin parameter. The results demonstrate significant rotation in the outer halo of M87, consistent with simulations that include a merger of two massive objects; it is difficult to create such a high spin parameter from accretion by random infall of nearby smaller galaxies.

Kinematics of Galaxy Clusters

I have also worked with T.C. Beers in the area of clusters of galaxies. At the time, this was an area that lacked rigorous statistical analysis and I worked mainly on developing the statistical tools to properly analyze the cluster data. We wrote a statistical package, ROSTAT, which applied a variety of estimators for location, scale, confidence bands, and goodness-of-fit to a specific distribution, among many other tests, to a given dataset. This package was distributed among the astronomical community and has received considerable use and praise from the users. Some of the subroutines which are more heavily used are the biweight estimator for location and scale and the bootstrap technique for determining confidence bands.

I analyzed velocity measurements from clusters of galaxies to determine both the significance of cD galaxy velocity offsets and the existence of bound populations around cD galaxies. Using robust statistical techniques we showed that a much smaller percentage of cD galaxies exhibited significant velocity offsets or bound populations than had previously been reported. This difference was due to the more robust statistical analysis and we encouraged a better understanding in the astronomical community of current statistical techniques.

To study positional data from galaxy clusters, I implemented an adaptive kernel technique (Silverman, B.W. 1986, Density Estimation for Statistics and Data Analysis, Chapman and Hall, London) for density estimation, which allowed us to better determine the significance of substructure. In addition, we used clustering algorithms, incorporating both velocity and positional data, to provide an estimate of sub-clumps within a cluster. When applied to Abell 400, these techniques showed that the cluster is best modeled as two groups which are presently merging. This result has an important effect on estimates of the velocity dispersion of the cluster since the dispersion was significantly higher when using all the velocities from the whole system as compared to calculating the dispersions separately for the two groups. The inferred M/L was a factor of four lower if the subclumps were considered separately. We have looked at other galaxy clusters and, after applying the adaptive kernel to determine the membership of particular sub-groups, we have calculated robust velocity dispersions and locations in order to better understand the present-day kinematical states of the clusters. We found that if substructure is present, but ignored, the derived velocity dispersion may be overestimated by a factor of two.