Stellar Structure & Evolution

Philipp Podsiadlowski (homepage)
Anthony Lynas-Gray

The evolution of single and binary stars

We have several projects studying the various phases in the evolution
of stars, starting with the basic formation process to their final
phase, the planetary-nebula or supernova stage.

We are particularly interested in applying stellar evolution theory
to stars in binaries to see how binary interactions can affect the
structure and evolution of stars, for example, to explain the origin
of stars with chemical anomalies (e.g., barium stars) or to explain
the non-spherical, often bipolar morphologies of many planetary nebulae.
In these studies, we use a variety of analytical and numerical tools
(specially modified stellar structure codes, hydrodynamical codes in
one to three dimensions) and maintain active collaborations with
various theoretical and observational groups around the world.

At present, the two major applications of this work are the study
of pre-supernova evolution in interacting binaries and X-ray binaries.

Supernovae

There are two fundamentally different types of supernovae:
thermonuclear explosions which occur when a white dwarf is pushed
above its maximum mass of 1.4 solar masses (the Chandrasekhar limit)
and which leads to the complete disruption of the star (referred to
as a "Type Ia supernova"), and core-collapse supernovae which take
place when the core of a massive star has exhausted all of its nuclear
fuel. In the latter case, the core of the star collapses to leave a
compact remnant, a neutron star or in some cases a black hole. A small
fraction of the energy released in the collapse is deposited in the
envelope, leading to an explosion and the bright, spectacular display
we refer to as a supernova. The appearance of the supernova, however,
depends sensitively on the pre-supernova structure of the envelope and
hence the star's evolutionary history in a binary. The various
binary interactions (mass loss, mass accretion and merging) can
fundamentally change the structure of a massive star and may thereby
account for many of the observational supernova sub-classes.

We are particularly interested in applying this theory to two of the
most interesting nearby supernovae of our generation, SN 1987A and SN
1993J. SN 1987A was the first naked-eye supernova since Kepler's
supernova in 1604 and is a highly anomalous supernova. Contrary to
what had been predicted, the progenitor was a blue supergiant instead
of a red supergiant, and is surrounded by a spectacular, but very
complex, nebula consisting of three bright rings (seen most clearly
with recent HST images) and has various chemical anomalies in its
envelope. At the moment, these anomalies can be best explained if the
progenitor originally was in a binary and has merged with its
companion in the not-too-distant past. We are actively modelling all
the detailed physical processes involved in the merging of two
massive stars, in particular the dynamical evolution of the system in
the merger phase, the associated anomalous nucleosynthesis and the
dynamical ejection of part of the envelope, presumably producing the
triple-ring nebula around the supernova.

The progenitor of SN 1993J on the other hand seems to have been
a star mainly consisting of helium and heavier elements with a very
small hydrogen-rich envelope. This again suggests that it was a member
of a binary and that it lost most of its hydrogen-rich envelope by
mass transfer. In both cases, future observations and theoretical
calculations will be required to confirm or refute the suggested
binary connections.

Type Ia supernovae have recently been used as cosmological distance
candles to measure the curvature of the Universe. At face value, the
results suggest an accelerating Universe, a dramatic break from the
previous picture. However, these results do not take into account
possible evolutionary changes in the population of Type Ia supernova
progenitors. Considering that there is no agreement on what the
progenitor systems of these supernovae actually are, this seriously
weakens the cosmological argument. We are presently studying
various types of binary systems proposed as progenitors for Type Ia
supernovae and model the evolution of the population of these
progenitors since the early Universe, using binary population
synthesis methods.

Hypernovae are a rare, new supernova type, first identified in 1998,
which are much more energetic than a normal supernova and are probably
caused by the collapse of a rapidly rotating helium star to a
stellar-mass black hole. Some hypernovae are observationally known to
be related to gamma-ray bursts, the most violent and most energetic
events known in the Universe. Again our main interest is in
understanding what type of stellar system can produce the progenitors
of these spectacular events.

The formation and evolution of X-ray binaries and millisecond
pulsars

Recent observations of the space velocities of pulsars (rapidly
rotating neutron stars) have shown that when neutron stars are born in
a supernova they receive very large kicks (presumably due to an
asymmetry in the supernova explosion). This has important effects on
the orbital parameters and the Galactic distribution of neutron stars
in binaries. We are particularly interested in so-called low-mass
X-ray binaries (LMXB), where a low-mass star with the mass of the Sun
or less transfers matter to a neutron star, which as a result becomes
a luminous X-ray source. While supernova kicks are important for
understanding the formation of LMXBs in the Galactic disc, in dense
stellar environments like globular clusters, other processes may be
more important. For example, in globular clusters, LMXBs may form as a
result of the tidal capture of a neutron star by a normal
star. However, whether this is a viable process depends on the
response of the normal star when the tidal energy is deposited inside
the star (the tidal energy can be a large fraction of the binding
energy of the star). During the LMXB phase, it is generally believed
that the neutron star is spun up by accretion of matter, leaving a
millisecond pulsar once the X-ray phase has ended. However,
statistical comparisons between millisecond pulsars and LMXBs suggest
that there are either too many millisecond pulsars relative to the
number of LMXBs or that the duration of the LMXB phase has been
overestimated by a large factor. This latter, more likely possibility
may be understood by irradiation effects which can dramatically change
the structure of the irradiated normal star. In particular, the
secondary may become inflated which leads to accelerated evolution and
a shorter duration of the LMXB phase. The details of this process
depend, however, on the circulation inside the secondary caused by the
one-sidedness of the irradiation in a binary. This is an important
problem which we are actively studying at the moment, developing both the
theoretical framework and the numerical tools to tackle this problem.

One of the most important recent discoveries in this field, which
indeed may challenge the above paradigm for LMXBs, is the realization
that many stars in so-called "low-mass" X-ray binaries originate from
much more massive progenitors (e.g., Cyg X-2, Cyg X-3). Based on our
recent calculations, it now seems that a large fraction, if not the
majority, of low-mass X-ray binaries may actually belong to a much
more massive, previously largely ignored class of intermediate-mass
X-ray binaries, and that standard textbooks on the subject need to be
rewritten. Apart from modelling individual systems, we use binary
populations synthesis techniques to model the population of X-ray
binaries (with U.S. collaborators) and keep active collaborations with
observational groups to test our predictions and improve our modelling
efforts.

The formation of planets around pulsars (indeed, the first
planets outside the solar system were discovered around pulsars) is a
related problem of much current interest. Many massive X-ray binaries
are predicted to merge in the near future. As a result of a complete
merger, the neutron star sinks to the centre of the massive star
forming a new hybrid object, generally referred to as a Thorne-Zytkow
object (TZO). While no such TZO has yet been discovered, recent
theoretical calculations predict that these objects should show
anomalous surface abundances of many chemical elements which should be
easily detectable spectroscopically. We are planning to use these
predicted anomalies to search for these objects observationally.

Stellar Atmospheres

A knowledge of stellar abundances is crucial to our understanding of
stellar and galactic evolution. While most stars have solar abundances
or are metal-deficient with respect to the Sun, more pronounced
abundance anomalies are also found. A few stars, for instance, have
photospheres composed of 99% helium and 1% carbon (by numbers) with
all other elements (including hydrogen) present only in trace
amounts. Detailed abundance studies of stars in external galaxies
(beyond the Magellanic Clouds) are becoming feasible using the Hubble
Space Telescope and 10 metre class ground-based facilities. Improved
techniques for the analysis of stellar spectra are under development;
these take advantage of better computer hardware, numerical methods
and (most important of all) recent advances in atomic and molecular
physics. Synthesis of whole spectral regions with inclusion of all
likely spectral features is now a viable method of approach.

Subdwarf-B (sdB) star research is of interest to investigations of
stellar mass-loss, as these stars appear to represent extreme cases of
near complete envelope loss during the later stages of stellar
evolution. If most stars evolve through the sdB star stage, following
evolution up the giant branch, understanding the envelope loss which
results in sdB star formation is crucial to understanding the late
stages of stellar evolution; it would also be the key to understanding
chemical evolution of a galaxy or star cluster over several stellar
generations. A better understanding of sdB star evolution is needed
in order to determine the contribution of sdB stars to the observed
ultraviolet upturn in giant elliptical galaxies with the view to its
calibration as an age indicator. Following the exciting discovery of
low-amplitude pulsation in seventeen sdB stars, the techniques of
asteroseismology can be used to constrain stellar evolution models
through a determination of envelope structure. The intention is to
establish sdB star envelope structure by determining the dispersion
relation of acoustic or gravity waves in surface layers, where wave
scattering can be accurately computed. If sdB stars can only form in
binary systems, those that appear to be single stars should have as
yet undetected cool (M-dwarf or brown dwarf) companions; a search for
these using the adaptive optics of the Gemini (North) Telescope has
been proposed.