Support one of our Projects:
- Nonprofit Adopt a Star - a program to support research on stars that NASA
is searching for planets.
- Whole Earth Telescope - an international collaboration of astronomers
at telescopes around the globe.
- Time-series CCD photometer - a portable instrument specifically designed
for our observations.
- Open Source White Dwarf Code - a computer code to model the
evolution and pulsations of white dwarfs.
- Asteroseismology Metacomputer - a commodity-hardware parallel
computer designed to run our models.
- Evolutionary Computing - an analysis method based on a genetic algorithm for objective global fitting.
Overview: Our
basic research goal is to observe and study the internal structure
and composition of white dwarf stars, the remnants of a nuclear fusion furnace
that once turned hydrogen into helium and energy, a process which still
powers stars like the Sun. An unexpected circumstance allows us to probe
their structure: some of these stars vibrate in a periodic manner that
sends seismic waves deep through their interior and brings information to
the surface. We see this manifested as complex periodic variations in
their brightness, which we can study and analyze, much as seismologists
study the inner structure of the earth using earthquakes. White dwarfs
once supported steady nuclear fusion, and would again if hydrogen were
injected into them. We essentially have a working fusion laboratory
to study, one that we must understand in detail if we are
ever to master this clean sustainable energy source and duplicate the
process on this planet.
We can determine the internal structure of pulsating white dwarfs using the
techniques of high speed photometry to observe their variations in brightness
over time, and then matching these observations with a computer model which
behaves the same way. The parameters of the model are chosen to correspond
one-to-one with the physical processes that give rise to the variations, so
a good fit to the data gives us confidence that our model reflects the actual
physics of the stars themselves. In the past decade, the observational
requirements of white dwarf seismology have been satisfied by the
development of the Whole Earth
Telescope (WET) -- an informal collaboration of astronomers at
observatories around the globe who cooperate to produce nearly continuous
time-series photometry of white dwarfs for up to 14 days at a time. This
instrument has provided a wealth of seismological data on the different
varieties of pulsating white dwarf stars.
We are working to establish a permanent WET endowment, which would support
the general operating expenses of this instrument through secure investments.
Ideally, this fund would allow for two runs of the WET each year, and provide
for travel expenses to send astronomers to remote observatories and to bring
international collaborators to headquarters for logistical support.
In an effort to bring the analysis of WET data to the level of
sophistication demanded by the observations, we are developing
a model-fitting method based on a
genetic algorithm. The underlying ideas for genetic algorithms were
inspired by Charles Darwin's notion of biological evolution through
natural selection. The basic idea is to solve a problem by evolving
the best solution from an initial set of random guesses. The computer
model provides the framework within which the evolution takes place, and
the individual parameters controlling it serve as the genetic building
blocks. Observations provide the selection pressure. In practice, this
method is much more efficient than other comparably global techniques.
We have had unprecedented success from the application of this method
to helium-atmosphere pulsating (DBV) white dwarfs, and we are currently
working to extend the method to the other types of pulsators. The hot
DOV white dwarfs promise to serve as interesting probes of neutrino
physics, while we hope to test the theory of stellar crystallization
with the cool massive DAV white dwarf BPM 37093 (known informally as
the "Diamond
in the sky").
Although extremely effective and objective in their application, genetic
algorithms still require a very large amount of computer time because they
involve running thousands of complex models for each set of observations.
To make this approach practical, we designed and built a
specialized computer -- a
collection of 64 minimal PCs connected by a network, which can run our
models in parallel about 60 times faster than any one of them by itself.
Our initial application of this new method to a well-observed pulsating
white dwarf demonstrated that our models are very sensitive to the central
composition, and allowed us to measure the astrophysically-important
(C + He → O) nuclear fusion reaction rate
with much greater precision than is possible in terrestrial laboratories.
The potential of this approach to probe interesting physics is clear. What
we can accomplish by applying it to other classes of objects is limited
only by the computational resources that we can devote to each problem.
Fortunately, the modular design of our computer is conducive to expansion,
and the off-the-shelf hardware has become both faster and less expensive
in the three years since we built the original. We are currently raising
funds to build a new parallel
computer based on our proven design, and to duplicate it at two
collaborating institutions, so we have the computational power necessary
to probe the other classes of white dwarf stars using this new method.
In an environment of dwindling resources, scientific investigations are
facing competitive stresses that are beginning to separate scientists into
two camps: those who guard their techniques jealously for fear of being
rendered obsolete, and those who embrace the true spirit of scientific
inquiry and share their results and resources freely with both colleagues
and competitors without prejudice. In some areas of astronomy, this
cultural split is beginning to hinder scientific progress.
The study of pulsating white dwarfs requires a special kind of instrument
capable of high speed imaging. When studying phenomena that change
rapidly, we do not have the luxury of increasing our exposure time to
improve the signal. Our instrument must be highly efficient even with
short exposures. We also need high timing precision to determine the
beginning and duration of each exposure accurately. Most CCD cameras
cannot obtain data continuously -- there is a dead time between exposures
when the detector is busy reading out the previous image. The time
required varies from a few seconds to a few minutes. We need an instrument
with essentially zero dead time, so we can record the rapidly variable
phenomena without interruption.
OpenWD is an attempt to
initiate an open, world-wide collaboration among experts who study white
dwarf stars to develop a complete, cutting-edge computer code to model the
evolutionary and pulsational characteristics of white dwarf stars. The
project is beginning with a complete FORTRAN code from the public domain
to serve as the basis for updates and additions from collaborators
who want to contribute modules that include their own prescriptions for
various physical ingredients. This distributes the workload among
scientists who have a stake in seeing their own work included, and
facilitates comparison between competing ideas by providing a common base
for all of the modules. It also helps to ensure a future where it will be
easier for all astronomers to embrace the collaborative spirit.
Most observatories do not provide high-speed imaging systems which satisfy
these special requirements. We are raising funds to build a high speed CCD
imaging instrument, based on a commercial CCD camera, optimized to observe
pulsating white dwarfs and other dynamic phenomena. The design is based on
an existing system, known as Argos,
at McDonald Observatory. The portable nature of this instrument will allow
us to send it wherever it is needed for Whole Earth Telescope observations
(particularly at Asian longitudes), and eventually duplicate it where
funds are available.
