Fundamental Questions in Astrophysics: Guidelines for Future UV Observatories

Abstract

Progress of modern astrophysics requires the access to the electromagnetic spectrum in the broadest energy range. The Ultraviolet is a fundamental energy domain since it is one of the most powerful tool to study plasmas at temperatures in the 3,000–300,000 K range as well as electronic transitions of the most abundant molecules in the Universe. Moreover, the UV radiation field is a powerful astrochemical and photoionizing agent.
The objective of this book is to describe the crucial issues that require access to the UV range.

Full text

FUNDAMENTAL QUESTIONS IN ASTROPHYSICS:
GUIDELINES FOR FUTURE UV OBSERVATORIES
Edited by:
ANAI.G´
OMEZ DE CASTRO and WILLEM WAMSTEKER
Reprinted from Astrophysics and Space Science
Volume 303, Nos. 1–4, 2006

Library of Congress Cataloging-in-Publication Data is available
ISBN 1-4020-4838-6 (hardbook)
ISBN 1-4020-4839-4 (eBook)
ISBN 978-1-4020-4838-6 (hardbook)
ISBN 978-1-4020-4839-4 (eBook)
Published by Springer,
P.O. Box 17, 3300 AA Dordrecht, The Netherlands.
Picture left: inserted as tribute to the late Willem Wamsteker who liked this image very much
Pictures right:
Top: Galaxy halo if the Universe reionized at redshift 15 or 6, by Kenji Bekki & Masashi Chiba, Tohuku University, Japan
Below: Sonic Point model of KiloHertz QPOs, by M. Colleman, F.K. Lamb & D. Psaltis
Below: Simulations of accretion disks, by J.F. Hawley, S.A. Balbus, J.M. Stone
Below: Simulations of the interaction of the accretion disk and the magnetized star inaTTauri System, by Brigitta von
Rekowsky & Axel Branderger
Bottom: Artist illustration of the evaporation of an exoplanet atmosphere, by European Space Agency and Alfred
Vidal-Madjar (Institut d’Astrophysique de Paris)
Printed on acid-free paper
All Rights Reserved
c
Springer 2006
No part of the material protected by this copyright notice may be reproduced or utilized in any form or
by any means, electronic or mechanical, including photocopying, recording or by any information storage
and retrieval system, without written permission from the copyright owner.
Printed in the Netherlands

TABLE OF CONTENTS
Foreword 1–2
M.A. Barstow and K. Werner / Structure and Evolution of White Dwarfs and their Interaction with the Local
Interstellar Medium 3–16
Isabella Pagano, Thomas R. Ayres, Alessandro C. Lanzafame, Jeffrey L. Linsky, Benjam´ın Montesinos and
Marcello Rodon`o / Key Problems in Cool-Star Astrophysics 17–31
Ana I. G´omez de Castro, Alain Lecavelier, Miguel D’Avillez, Jeffrey L. Linsky and Jos´e Cernicharo / UV
Capabilities to Probe the Formation of Planetary Systems: From the ISM to Planets 33–52
Boris T. G¨aansicke, Domitilla de Martino, Thomas R. Marsh, Carole A. Haswell, Christian Knigge, Knox S.
Long and Steven N. Shore / Ultraviolet Studies of Interacting Binaries 53–68
Willem Wamsteker, Jason X. Prochaska, Luciana Bianchi, Dieter Reimers, Nino Panagia, Andrew C. Fabian,
Claes Fransson, Boris M. Shustov, Patrick Petitjean, Phillipp Richter and Eduardo Battaner / The Need
for Ultraviolet to Understand the Chemical Evolution of the Universe, and Cosmology 69–84
Rosa M. Gonz´alez Delgado / Starbursts at Space Ultraviolet Wavelengths 85–102
Noah Brosch, John Davies, Michel C. Festou and Jean-Claude G´erard / A View to the Future: Ultraviolet Studies
of the Solar System 103–122
Wolfram Kollatschny and Wang Ting-Gui / Active Galaxies in the UV 123–132
Ana I. G´omez de Castro, Willem Wamsteker, Martin Barstow, Noah Brosch, Norbert Kappelmann, Wolfram
Kollatschny, Domitilla de Martino, Isabella Pagano, Alain Lecavelier des ´
Etangs, David Ehenreich,
Dieter Reimers, Rosa Gonz´alez Delgado, Francisco Najarro and Jeff Linsky / Fundamental Problems in
Astrophysics 133–145
Norbert Kappelmann and J¨urgen Barnstedt / Guidelines for Future UV Observatories 147–151
F. Najarro, A. Herrero and E. Verdugo / Massive stars in the UV 153–170

Astrophys Space Sci (2006) 303:1–2
DOI 10.1007/s10509-006-9175-z
Foreword
C
Springer Science +Business Media B.V. 2006
Modern astrophysics is a mature science that has evolved
from its early phase of discovery and classification to a
physics-oriented discipline focussed in finding answers to
fundamental problems ranging from cosmology to the origin
and diversity of life-sustainable systems in the Universe. For
this very reason, progress of modern astrophysics requires
the access to the electromagnetic spectrum in the broadest
energy range. The Ultraviolet is a fundamental energy do-
main since it is one of the most powerful tool to study plas-
mas at temperatures in the 3,000–300,000 K range as well
as electronic transitions of the most abundant molecules in
the Universe. Moreover, the UV radiation field is a powerful
astrochemical and photoionizing agent.
The impact of UV instruments in astronomical research
can be clearly traced through the considerable success of
the International Ultraviolet Explorer (IUE) observatory and
successor instruments such as the GHRS and STIS spectro-
graphs on-board the Hubble Space Telescope (HST), or the
FUSE satellite operating in the far UV (90–120 nm range).
Of particular importance has been access to high resolution
R40,000–100,000 spectra providing an ability to study
the dynamics of hot plasma and separate multiple galactic,
stellar or interstellar spectral lines. Furthermore, the GALEX
satellite is providing new exciting views of UV sources.
This book describes the fundamental problems in modern
astrophysics that cannot progress without easy and wide-
spread access to modern UV instrumentation. Three among
them should be stressed by its relevance:
1. Extrasolar planetary atmospheres and astrochemistry in
the presence of strong UV radiation fields.
2. Chemical evolution of the Universe and the diffuse bary-
onic content.
3. The physics of accretion and outflow: the astronomical
engines.
The volume contains a series of review articles that analyze
the scientific requirements for modern UV instrumentation.
The first article summarizes the science case for UV astron-
omy. After, several articles targeting the major research fields
of astrophysics from Solar System to cosmological research
are included. These articles analyze why and which UV in-
strumentation is required to make progress in each field. The
book ends with a summary of the UV instrumentation de-
manded by the community and a brief update on techno-
logical requirements. All articles in this volume have gone
through the peer-review system of the journal “Astrophysics
and Space Science.”
This book contains the thoughts and work of the Network
for UV astrophysics (NUVA) and the UV community at-
large. By the end of 2002, a group of European astronomers
coming from a broad range of areas: from fundamental as-
trophysical research to observational expertise in the optical,
UV and X-ray ranges, as well as space instruments develop-
ment teams, joined efforts to evaluate the need to develop new
UV instrumentation for the coming decade in order to achieve
some of the major scientific objectives of the astronomical
community and make full profit of the large astronomical
facilities planned for other spectral ranges; this group set the
seed for the Network for UltraViolet Astrophysics (NUVA).
At that time, STIS was working in HST and the Cosmic Ori-
gins Spectrograph (COS) was being built for replacement;
HST was thought to last till 2010/12, FUSE was working
nominally and GALEX was close to be launched. However,
the big space agencies had no plans for new UV spectroscopic
missions unless the large optical/UV telescope included in
the NASA Origins plan targeted to enter development phase
in 2015–2020 and to be launched in the second quarter of this
century. Unfortunately, on Friday 16 January 2004 NASA
informed that the planned shuttle mission to service and up-
grade HST (SM-4) including the substitution of STIS by
Springer

2Astrophys Space Sci (2006) 303:1–2
COS, had been cancelled. Just few months later STIS failed.
No access to high resolution UV spectroscopy has been fea-
sible during most of 2005 since FUSE resumed observations
in November 2005 after the failure of the third of the four
onboard reaction wheels in December 2004. The NUVA was
officially established in January 2004. A key objective of the
NUVA is to run an exploratory analysis to define the sci-
entific requirements for future UV observatories. This book
contains the first outcome of this work and its publication
has been funded by the NUVA as a part of its activities. The
NUVA is within the Optical Infrared Co-ordination Network
for Astronomy (OPTICON) funded by the European Com-
missions 6th Framework programme under contract number
RII3-Ct-2004-001566.
In these two years, some respected colleagues and dear
friends who actively promoted and enthusiastically colabo-
rated in this project have passed away.
Willem Wamsteker was the director of ESA’s IUE Ob-
servatory until the mission terminated in 1996. After his di-
rection, the IUE data archive became the first fully internet
driven astronomical archive. Willem was a promotor of the
NUVA and a key initiator of the World Space Observatory
UV (WSO/UV) project; a 1.7 m UV telescope equipped with
state of the art instrumentation providing a factor of 10 im-
provement on the high resolution spectroscopic capabilities
of HST/STIS. WSO/UV will be launched in 2010 becoming
the only UV spectroscopic facility available world-wide. The
project is driven by a broad international colaboration led
by Russia (ROSCOSMOS); there is a significant European
participation in the project. Willem Wamsteker was also co-
editor of “Astrophysics and Space Science” and deeply in-
volved in the edition of this volume.
Marcello Rodon´o was director of the Obsservatorio As-
trofisico di Catania and a very relevant member of the “cool
stars” community. He was keen off multi-wavelength studies
and an active promotor of the WSO/UV project.
Michel Festou was director of l’Observatoire de Besan¸con
and a passionate researcher. His work on UV spectroscopy
of comets is reknowned world-wide. He actively colabo-
rated with the NUVA in the identification of the UV facilities
needed to make progress on Solar System Research.
We would like to dedicate this book to their memory.
Finally, we would like to thank our OPTICON colleagues
and very especially John Davies (OPTICON Project Scien-
tist) and Gerry Gilmore (OPTICON P.I.) for their support.
We also acknowledge the support of the Universidad Com-
plutense de Madrid which is hosting and maintaining the
NUVA site (www.ucm.es/info/NUVA).
Prof. Ana In´es G´omez de Castro
NUVA, chair
Madrid, March 1st, 2006
Springer

Astrophys Space Sci (2006) 303:3–16
DOI 10.1007/s10509-006-2061-x
ORIGINAL ARTICLE
Structure and Evolution of White Dwarfs and their Interaction
with the Local Interstellar Medium
M.A. Barstow ·K. Werner
Received: 11 March 2005 / Accepted: 11 August 2005
C
Springer Science +Business Media B.V. 2006
Abstract The development of far-UV astronomy has been
particularly important for the study of hot white dwarf stars.
A significant fraction of their emergent flux appears in the far-
UV and traces of elements heavier than hydrogen or helium
are, in general, only detected in this waveband or at shorter
wavelengths that are also only accessible from space. Al-
though white dwarfs have been studied in the far-UVthrough-
out the past ∼25 years, since the launch of IUE,onlyafew
tens of objects have been studied in great detail and a much
larger sample is required to gain a detailed understanding of
the evolution of hot white dwarfs and the physical processes
that determine their appearance. We review here the current
knowledge regarding hot white dwarfs and outline what work
needs to be carried out by future far-UV observatories.
Keywords Ultraviolet astronomy ·Space missions ·White
dwarfs
1. Introduction
1.1. White dwarf stars and the local interstellar medium
in astrophysics
Many white dwarfs are among the oldest stars in the Galaxy.
Their space and luminosity distributions help map out the
history of star formation in the Galaxy and can, in princi-
ple, determine the age of the disk by providing an important
M.A. Barstow ()
Department of Physics and Astronomy, University of Leicester,
University Road, Leicester LE1 7RH, UK
e-mail: mab@star.le.ac.uk
K. Werner
Universit¨at T ¨ubingen, FRG, T¨ubingen
lower limit to the age of the Universe. Recently, it has been
suggested that cool white dwarfs may account for a substan-
tial fraction of the missing mass in the galactic halo (Op-
penheimer et al., 2001). White dwarfs are also believed to
play a role in the production of type Ia supernovae, through
possible stellar mergers or mass transfer. Although no can-
didate precursor system has yet been found, they are used
as “standard candles” to measure the distances of the most
remote objects in the Universe and to imply a non-zero value
for the cosmological constant. It is intriguing that such lo-
cal, low luminosity objects as white dwarf stars play a key
role in some of the most important cosmological questions
of our day, concerning the nature and fate of the Universe.
To understand and calibrate cosmologically important as-
pects of white dwarfs, such as their cooling ages, masses and
compositions, we require a thorough understanding of how
their photospheric compositions evolve. Atmospheric metal
abundances affect cooling rates and bias the determination
of temperature and surface gravity. Reliable masses can only
be derived from accurate effective temperature and surface
gravity measurements. Importantly, metals are difficult to de-
tect in cool white dwarfs but still play an important role in
cooling. Therefore, abundances measured in hotter stars pro-
vide important data as to what species may be present and in
what quantities.
Interstellar gas is a fundamental component of the Milky
Way and other galaxies. The local interstellar medium
(LISM) is particularly important because it is close enough
to us for detailed examination of its structure. Indeed, it is the
only part of the Universe, apart from the solar system, that we
can study directly, from the penetration of neutral particles
through the heliopause. Study of the composition of the local
interstellar gas tells us about the evolution of the Universe and
our galaxy. While modified considerably since the formation
of the solar system, the processes that have shaped the ISM
Springer

4Astrophys Space Sci (2006) 303:3–16
must have also affected the solar system and possibly even
the evolution of life. For example, the local radiation field
and gas properties of the LISM set the boundary conditions
for the heliosphere. Therefore, it regulates the properties of
the interplanetary medium and the cosmic ray fluxes through
the solar system (see Frisch et al., 2002), which may affect
the terrestrial climate (Mueller et al., 2003).
The relationship between white dwarfs and the LISM is
intimate. The process of producing white dwarfs in the disk
substantially enriches the content of the local ISM (Mar-
gio, 2001) and white dwarfs contribute significantly to the
total cosmic abundance of metals (Pagel, 2002). They also
provide a significant fraction of the total flux of ionizing ra-
diation within the LISM and will affect the ionization state
of the interstellar gas, even if the LISM is not in ionization
equilibrium. At the same time, white dwarfs may be accret-
ing material from the denser clouds in the ISM, which will
modify their photospheric abundances. Finally, white dwarfs
are ideal background sources for direct study of the LISM.
1.2. Physical structure and classification of white
dwarfs
Astronomers have known about the existence of white dwarf
stars for 150 years, since the discovery of a companion to the
brightest star in the sky, Sirius. Studying the regular wave-
like proper motion of Sirius, Bessel revealed the presence
of a hidden companion, with the pair eventually resolved by
Clark in 1862 and the orbital period revealed to be 50 years.
Later, the first spectroscopic study of the system by Adams
revealed a great enigma. While the temperature of Sirius B
was found to be higher than Sirius A, it was apparently 1000
times less luminous. This result could only be explained if
Sirius B had a very small radius, 1/100 Rand similar to
that of the Earth. However, with a known mass of ∼1Mfor
Sirius B, the implied density of 1400 Tonnes m−3was well
above that of any form of matter known to 19th and early
20th century physicists. Within a number of years, a handful
of stars similar to Sirius B were found, mostly in binaries.
The term “white dwarf” was coined later, on the basis of the
visual color and small size, when compared to other stars.
The answer to the riddle of their structure stems from the
development of the quantum statistical theory of the electron
gas by Enrico Fermi and Paul Dirac. When atoms in a mate-
rial are sufficiently close together, their most weakly bound
electrons move freely about the volume and can be consid-
ered to behave like a gas. Almost all the electrons in the gas
will occupy the lowest available energy levels. Any states
with the same energy are said to be “degenerate.” Hence, this
gas is known as the “degenerate electron gas.” Under condi-
tions of extreme pressure the electrons are forced to occupy
space much closer to the nuclei of the constituent atoms than
in normal matter, breaking down the quantised structure of
the energy levels. However, according to the Pauli Exclusion
Principle, no two electrons can occupy the same quantum
state, which has a finite volume in position-momentum space,
so there is a limit where a repulsive force arising from this, the
“electron degeneracy pressure”, resists further compression
of the material. Fowler (1926) showed that this pressure could
support a stellar mass against gravitational collapse and pro-
posed that this might explain the existence of white dwarfs.
Combining this insight with the equations of stellar structure,
Chandrasekhar (1931, 1935) determined the mass–radius re-
lation for white dwarfs and the maximum mass (1.4 M) that
electron degeneracy pressure could support against gravity,
the Chandrasekhar limit. As this Nobel Prize winning work
has been extended by subsequent developments, the basic
ideas remain unchanged. However, importantly, they have
also hardly been tested by direct observation.
Theoretical and observational study of stellar evolution
has placed white dwarfs as one possible end point of the pro-
cess. In general terms, all stars with masses below about 8
times that of the Sun will pass through one or more red giant
phases before losing most of their original mass to form a
planetary nebula. The remnant object, a white dwarf, is the
core of the progenitor star. In the absence of any internal
source of energy, the temperature of a white dwarf, after its
birth, is determined by how rapidly stored heat is radiated
into space. Estimates of white dwarf cooling times indicate
that it will take several billion years for the stars to fade to in-
visibility. Hence, white dwarfs are among the oldest objects
in the galaxy. Since, the galaxy is younger than the cool-
ing timescales; the lowest temperature (oldest) white dwarfs
yield a lower limit to its age.
The basis for understanding the nature of most stars is
analysis of their optical spectra and classification according
to the characteristics revealed. A number of physical pro-
cesses can alter the atmospheric composition of a white dwarf
as it cools. As noted by Schatzmann (1958), the strong grav-
itational field (log g∼8 at the surface) causes rapid down-
ward diffusion of elements heavier than the principal H or He
component. Hence, Schatzmann predicted that white dwarf
atmospheres should be extremely pure. Consequently, the
spectra should be devoid of most elements, showing signa-
tures of only hydrogen and, possibly, helium. White dwarfs
are thus divided into two main groups according to whether
or not their spectra are dominated by one or other of these
elements. The hydrogen-rich stars are given the classifica-
tion DA, while the helium-rich white dwarfs are designated
DO if HeII features are present (hotter than about 45,000 K)
and DB if only HeI lines are visible. Small numbers of hy-
brid stars exist, with both hydrogen and helium present.
In these cases, two classification letters are used, with the
first indicating the dominant species. For example, DAO
white dwarfs are mostly hydrogen but exhibit weak HeII
features.
Springer

Astrophys Space Sci (2006) 303:3–16 5
Fig. 1 Schematic description of
the production of H-rich and
He-rich branches of white dwarf
evolution
The above classification scheme applies only to white
dwarfs with temperatures above ∼10,000 K, when the H
and He energy levels are sufficiently populated above the
ground state to produce detectable features at visual wave-
lengths. At lower temperatures, although H and/or He may
be present these elements are no longer directly detectable.
Cool white dwarfs are divided into three main groups. DC
white dwarfs have continuous, completely featureless spec-
tra; DZ white dwarfs are He-rich and have only metal lines
visible; DQ white dwarfs show carbon features. Figure 1
summarises the current thoughts regarding the relationships
between the hot white dwarf groups along with the princi-
pal mechanisms that provide evolutionary routes between
them.
Gravity is not necessarily the only influence on photo-
spheric composition. It can be countered by radiation pres-
sure which acts outward to support heavy elements in the
atmosphere, a process termed “radiative levitation.” Another
mechanism that can mix elements that have settled out in
the stellar atmosphere is convection. If the convective zone
reaches down to the base of the atmosphere then heavy el-
ements can be dredged back up into the outer atmosphere.
A further complication is that material can also be accreted
from the interstellar medium.
1.3. The structure of the local interstellar medium
The Sun is embedded in the Local Interstellar Cloud (LIC), a
warm (T∼10,000 K), low-density (n∼0.1cm
−3) region,
which is observed in projection toward most, but not all,
nearby stars. Models of the LIC (Redfield and Linsky, 2000)
show the Sun located just inside the edge of the LIC in the di-
rection of the Galactic Center and toward the North Galactic
Pole (NGP). The LIC is surrounded by a hot 106K substrate
and appears to be part of the expanding cloud complex rep-
resenting the Loop I radio shell, which is either a supernova
remnant or a wind blown shell from the Sco-Cen Associa-
tion (Fig. 2). Indeed, evidence for shocks in the LIC within
the past 2 Myr is supported by EUV observations of ion-
ized interstellar HeII (Lyu and Bruhweiler 1996; Barstow
et al., 1997). Also, surrounding the LIC is the low den-
sity Local Bubble with dimensions of >200 pc. Soft X-ray
and OVI observations imply this region is largely filled with
a hot (T∼106K), extremely low-density (n∼10−2cm−3)
plasma. Frisch (1995) and Bruhweiler (1996) discuss possi-
ble origins of Loop I and the Local Bubble, but we need to
resolve two key questions. What is the geometry of the Lo-
cal Bubble? What are the distribution, physical conditions,
and kinematics of clouds in the Local Bubble? The answers
Springer

6Astrophys Space Sci (2006) 303:3–16
Fig. 2 View of the LISM within
400 pc of the Sun. The Sun is
located at the end of the dashed
line extending from the star β
CMa. The dashed line is
∼200 pc in length. The region of
the Local Void or Local Bubble
is denoted. Most of the volume
of the LISM is composed of
very low-density gas. The radio
loop structures are further
indicated. The stars of Sco-Cen
are interior to Loop I. Regions
of high dust extinction, regions
where dense clouds are found,
are also marked. From
Bruhweiler (1996)
to these questions also bear directly on whether interstellar
accretion is possible in specific white dwarfs.
2. The complex nature of the DA white dwarf
population
2.1. Composition and structure of the stellar
photospheres from UV spectroscopy
The emergence from the Asymptotic Giant Branch (AGB) of
the two H- and He-rich groups, is qualitatively understood
to arise from differences in precise evolutionary paths. The
He-rich group undergoes a late helium shell flash, which
causes convective mixing of the outer envelope. Hydrogen
is ingested and burnt or diluted by this process (e.g., Herwig
et al., 1999). The majority of young white dwarfs are of DA
type, out numbering the He-rich objects by a ratio of ∼7:1
(Fleming et al., 1986; Liebert et al., 2005). The more recent,
deeper, Sloan Digital Sky Survey appears to yield an even
higher value of 9:1 for the ratio of DAs to the total number
of DO and DB white dwarfs (Kleinman et al., 2004). How-
ever, the complex relationships between these groups and
a demonstrable temperature gap in the He-rich cooling se-
quence cannot yet be readily explained. For example, the ratio
of H-rich to He-rich progenitors (4:1) is lower than that of
the white dwarfs into which they evolve (Napiwotzki, 1999).
Furthermore, a gap in the He-rich sequence, between ∼45000
and 30000 K, separating the DO and DB white dwarfs (We-
semael et al., 1985; Liebert et al., 1986; Dreizler and Werner,
1996) implies that He-rich white dwarfs must temporarily be
seen as DA stars due to some physical process.
It is well established that the hottest (T∼50,000 K)
white dwarfs possess significant abundances of elements
heavier than He in their atmospheres. Typical ranges are
shown in Table I for the most important elements. Mass frac-
tions are given, with respect to H for DAs and He for DOs,
since they are essentially independent of nuclear burning,
if the element does not participate in the process. Observed
abundances may vary considerably with temperature, as dis-
cussed later. Hence, we list typical values and comment on
the possible ranges in each white dwarf pathway. Studying
how photospheric abundance patterns change as these stars
cool can delineate the evolutionary pathways followed by
white dwarfs. For example, abundance enhancements in C,
N and O may be a sign of near exposure of the C O core
during the AGB phase or of a late He-shell burning episode.
Springer

Astrophys Space Sci (2006) 303:3–16 7
Table 1 Abundances of
important heavy elements
observed in DA and DO white
dwarfs
DA DO
Element Typical value Comment Typical value Comment
C10
−6to 10−5Zero in coolest 3 ×10−3
N10
−6From zero to 10−25×10−5In a few cases
at extremes
O10
−4Zero in coolest 10−5
Si 10−4Seen at most temps 10−4
down to 10−8
Fe 2–5 ×10−4Zero in coolest – Not detected in any DOs
Ni 1–2 ×10−5Zero in coolest 10−4
Note that the values are much
more uncertain for the less
well-studied DO stars
Fig. 3 1230 ˚
A to 1280 ˚
A region
of the STIS spectrum of
REJ0558-373, showing
photospheric absorption lines of
N V (1238.821/1242.804 ˚
A) and
large numbers of Ni lines. The
best-fit synthetic spectrum is
shown offset for clarity. The
strong line near 1260 ˚
A, present
in the observation but not in the
model, is interstellar Si II (from
Barstow et al., 2003b)
Focused spectroscopic studies of the hottest white dwarfs
have begun to establish the general pattern of abundance with
respect to evolutionary status. Ultraviolet observations have
played an essential role by providing access to absorption
lines of elements heavier than H or He. Such features are
not present in optical or IR data except where photospheric
abundances are unusually high. Indeed, He is also hard to de-
tect, requiring abundances in excess of a few times 10−3in
the visible band. Typical detection limits in the UV are two
orders of magnitude lower. Therefore, the most important
and useful transitions, particularly many resonance lines of
elements heavier than H and He, lie in the far-ultraviolet (far-
UV, 1000–2000 ˚
A). However, since the lines are expected
to be weak and narrow, they are only normally visible at
high resolution (R>20,000). An example is a small sec-
tion of the high-resolution spectrum of the DA white dwarf
REJ0558-373, recorded with the Space Telescope Imaging
Spectrograph (STIS) onboard HST (Fig. 3), which shows the
interstellar 1260.4 ˚
A line of Si II together with photospheric
NV.
A survey of 25 hot DA white dwarfs, based on IUE and
HST data (Barstow et al., 2003b), shows that, while the
presence or absence of heavy elements largely reflects what
would be expected if radiative levitation were the support-
ing mechanism, the measured abundances do not match pre-
dicted values very well. These and earlier results are forcing
us to confront complexities in the real physical structure of
the stars. For example, it has become clear that the shape and
strength of the Balmer line profiles, from which Teff and log g
(and, indirectly, mass) are measured, are dependent on the
stellar photospheric abundances, requiring a self-consistent
analysis of each individual star based on data acquired at all
wavelengths (Barstow et al., 1998). Furthermore, we now
have direct observational evidence (Barstow et al., 1999;
Dreizler and Wolff, 1999) that photospheric heavy elements
are not necessarily homogeneously distributed (by depth) and
that more complex stratified structures must be considered.
While almost all stars hotter than ∼50,000 K contain heavy
elements, as expected, there is an unexplained dichotomy at
lower temperatures, with some stars having apparently pure
H atmospheres and others detectable quantities of heavy el-
ements (Barstow et al., 2003b, e.g., Fig. 4). In many of these
objects the photospheric opacity does not reveal itself in EUV
photometric or spectroscopic observations, implying that the
observed material resides in a thin layer in the uppermost re-
gion of the photosphere (see Holberg et al., 1999; Barstow et
al., 2003b). The effect of this stratification can be observed di-
rectly with high resolution UV spectroscopy. Figure 5 shows
Springer

8Astrophys Space Sci (2006) 303:3–16
Fig. 4 Measured abundances of
nitrogen (number ratio with
respect to hydrogen) as a
function of Teff for a sample of
25 DA white dwarfs (from
Barstow et al., 2003b)
Fig. 5 The STIS E140M
spectrum of REJ1032+532 in
the region of the NV resonance
doublet (histogram), compared
to the predicted line profiles
from a homogeneous model
atmosphere calculation (from
Holberg et al., 1999)
the HST/STIS spectrum of the white dwarf REJ1032+532,
compared to that predicted by a homogeneous model stel-
lar atmosphere. The N V line profiles in the model have
significantly broader wings than are observed, while a strat-
ified model (not shown) gives a much better match to the
data.
Recently, important progress has been made in incorporat-
ing radiative levitation and diffusion self-consistently into the
atmosphere calculations (Dreizler and Wolff, 1999; Schuh et
al., 2002). This work reconciles the overall spectral distri-
bution across the soft X-ray, EUV and far-UV bands with
the models (a problem with homogeneously distributed el-
ements) and explains the level of stratification inferred for
various elements. However, the abundance predictions do
not match observations for the known gravity of each star
observed, and agreement requires a higher surface grav-
ity than allowed by the optical data. In particular, we can-
not account for the large observed compositional differ-
ences between stars with identical temperature and surface
gravity.
In almost all the hot DA white dwarfs observed, the high
ionization resonance lines arising from photospheric heavy
elements have blue-shifted components, indicating that there
is some circumstellar gas present (Bannister et al., 2003).
Whether or not this material is a remnant of the planetary
nebula or due to ongoing mass-loss is unresolved. This has
important consequences for our basic understanding of stel-
lar composition. The effects of mass-loss, in the form of
weak winds ejecting material into the local ISM, and direct
accretion of material from the ISM are likely to be of great
importance in providing a plausible framework that can ex-
plain measured abundances.
Springer

Astrophys Space Sci (2006) 303:3–16 9
Hot He-rich DO white dwarfs are the progeny of stars
which constitute an interesting spectral class. Much of the
detailed physics involved in studying their atmospheres is
similar to that of the DA white dwarfs. However, the levi-
tation of heavy elements in these objects is far from under-
stood. The DO progenitors are the so-called PG1159 stars,
hot hydrogen-deficient objects, some of which represent the
hottest white dwarfs known, while others are still burning
helium in a shell.
Optical and UV spectral analyses have shown that the ef-
fective temperatures of PG1159 stars range between 75,000
and 200,000 K and the derived surface gravities are between
log g=5.5 and 8 (Werner, 2001). They are probably the out-
come of a late helium-shell flash, a phenomenon that drives
the fast evolutionary rates of three well-known objects (FG
Sge, Sakurai’s object, V605 Aql). Flash-induced envelope
mixing produces a H-deficient stellar surface (Herwig et al.,
1999). The He-shell flash transforms the star back to an AGB
star (born-again AGB star) and the subsequent, second post-
AGB evolution explains the existence of Wolf–Rayet cen-
tral stars of planetary nebulae (spectral type [WC]) and their
successors, the PG1159 stars. The photospheric composition
then essentially reflects that of the region between the H-
and He-burning shells in the precursor AGB star. It is dom-
inated by He, C, and O. Typical values are He =33%, C =
50%, O =17% (by mass), however, a considerable spread
of abundance patterns is observed, pointing to complicated
processes in the stellar interiors. Few stars show traces of
nitrogen (1%) or considerable amounts of residual hydrogen
(about 25%).
PG1159 stars provide the unique possibility of studying
the chemistry in the intershell region between the H- and
He-burning shells that is created after complicated and still
poorly understood burning and mixing processes during the
AGB phase. Usually the intershell material remains hidden
within the stellar interior. During the third dredge-up on the
AGB, however, intershell material can get mixed into the
convective surface layer and appears on the stellar surface,
though in rather diluted abundances. Nevertheless, this pro-
cess defines the role of AGB stars as contributors of nucle-
arly processed matter to the Galaxy. The motivation to study
PG1159 stars is based on the fact that these objects directly
display their intershell matter. It can be expected that grav-
itational settling is not affecting the composition, because
of ongoing mass-loss and convective motions. However, the
quantitative interpretation of the abundance analyses is still
premature because evolutionary calculations through a final
He-shell flash including a full nuclear network are not yet
available.
High-resolution UV spectroscopy was crucial in making
surprising discoveries which provide essential constraints to
calibrate theoretical modeling of stellar evolution. Generally,
UV spectra of PG1159 stars show only few photospheric
(absorption) lines, mainly from He II, C IV, O VI, and Ne
VII. Some of them display shallow N V lines and in many of
them we see sulfur. The S VI 933/944 ˚
A doublet in K1-16, for
example, suggests a solar abundance, which is in line with
the expectation that S is not affected by nuclear processes.
Silicon was also identified in some objects (Reiff et al., 2005),
but detailed abundance analyses remain to be done.
A very surprising result of UV spectroscopy was the
detection of a significant iron deficiency (1–2 dex) in the
three best studied PG1159 stars (Miksa et al., 2002). Obvi-
ously, iron was transformed to heavier elements in the in-
tershell region of the AGB star by n-captures from the neu-
tron source 13C(α,n)16 O (Herwig et al., 2003). Subsequently,
several other studies have also revealed an iron-deficiency
in [WC]-type central stars, which matches our picture that
these stars are immediate PG1159 star progenitors. Another
important result was accomplished by the identification of
one of the strongest absorption lines seen in FUSE spec-
tra of most PG1159 stars, located at 973.3 ˚
A. It is a Ne
VII line (Fig. 6) that allowed us to assess the neon abun-
dance in a large sample of objects (Werner et al., 2004).
Fig. 6 Discovery of a neon line (left panels) and a fluorine line
(right panels) in the hydrogen-deficient PG1159-type central star
PG1520+525 (top panels) and in the hydrogen-normal central star of
NGC 1360 (bottom panels). The neon and fluorine abundances in the
PG1159 star (given as mass fractions in the panels) are strongly en-
hanced, namely 20 times and 250 times solar, respectively, whereas
they are solar in NGC 1360. Note the strong Fe VII line (not included
in the models) at 1141.4 ˚
A in NGC 1360, which indicates a solar iron
abundance (Hoffmann et al., 2005). It is not detectable in the PG1159
star, probably due to a subsolar Fe abundance (Werner et al., 2004)
Springer

10 Astrophys Space Sci (2006) 303:3–16
It turns out that neon is strongly overabundant, (2%, i.e.,
20 times solar). This result clearly confirms the idea that
PG1159 stars indeed exhibit intershell matter. Neon is pro-
duced in the He-burning environment by two α-captures of
nitrogen, which itself resulted from previous CNO burning:
14N(α,γ)18 F(e+ν)18O(α,γ)22 Ne.
There are still many photospheric lines in UV spectra of
PG1159 stars which remain unidentified. Some of them may
stem from yet unknown Ne VII lines, or even of elements
which have not been detected in these stars at all. The latest
identification is that of a feature at 1139.5 ˚
A, which appears
rather strong in some objects. It is a line from highly ionized
fluorine (Fig. 6) and large overabundances (up to 250 times
solar) were derived for a number of PG1159 stars. This line
was also identified in “normal” hydrogen-dominated cen-
tral stars and, in contrast, solar fluorine abundances were
found (Werner et al., 2005). This again is a clear proof that
we see intershell matter on PG1159 stars. According to re-
cent calculations by Lugaro et al. (2004), their stellar models
show an effective fluorine production and storage in the in-
tershell, leading to abundances that are comparable to the
observed PG1159 abundances of fluorine. The general prob-
lem for fluorine production is that 19F, the only stable F
isotope, is rather fragile and readily destroyed in hot stel-
lar interiors by hydrogen via 19F(p,α)16 O and helium via
19F(α,p)22 Ne. The nucleosynthesis path for F production in
He-burning environments of AGB and Wolf–Rayet stars is
14N(α,γ)18 F(β+)18O(p,α)15 N(α,γ)19F.
All this underlines that AGB stars which dredge up ma-
terial from the intershell are contributing to the Galactic flu-
orine content (together with Wolf–Rayet stars and type II
SNe). This is completely in line with the detected fluorine
overabundances (up to 30 times solar) found from IR spectra
in AGB stars (Jorissen et al., 1992). To what extent PG1159
stars themselves return F to the ISM remains to be estimated.
The life time of a born-again AGB star is short in compari-
son to a usual AGB star, however, the fluorine fraction in the
mass lost by a wind of the former is much higher.
2.2. White dwarf masses and radii and the role of UV
imaging observations
Two of the most important physical parameters that can
be measured for any star are the mass and radius. They
determine the surface gravity by the relation g=GM/R2.
Hence, if log gis measured the mass can be calculated pro-
vided the stellar radius is known. One outcome of Chan-
drasekhar’s original work on the structure of white dwarfs
was the relationship between mass and radius, arising from
the physical properties of degenerate matter. Further theo-
retical work yielded the Hamada–Salpeter zero-temperature
mass-radius relation (Hamada and Salpeter 1961). How-
ever, white dwarfs do not have zero temperature, indeed
many are very hot. Hence, the Hamada–Salpeter relation
is only a limiting case and the effects of finite tempera-
ture need to be taken into account. Evolutionary calcula-
tions, where the radius of a white dwarf of given mass de-
creases as the star cools, have been carried out by Wood
(1992, 1995), Bl¨ocker (1995), Bl¨ocker et al. (1997) and
others.
Measurements of the surface gravity of samples of white
dwarfs show that the distribution of log gvalues and, there-
fore, of mass is very narrow (e.g., Bergeron et al., 1992;
Napiwotzki et al., 1999), with a peak mass of ∼0.6 M. This
is a direct consequence of the evolution of single stars, with
masses from 1 Mup to ∼8M. While the details of the re-
lationship between the initial mass of the progenitor star and
the final white dwarf mass are not particularly well under-
stood, it is clear that the small dispersion in the white dwarf
masses is related to a similarly small range of stellar core
masses and the fact that most of the outer stellar envelope
is expelled through several phases of mass loss along the
AGB. Importantly, any white dwarf with a mass outside the
approximate range 0.4–1.0 Mcannot arise from single star
evolution and must have an origin in a binary, where mass
exchange has taken place.
The basic model of the white dwarf mass-radius relation
is often used to derive masses from the spectroscopic mea-
surements of effective temperature and surface gravity (e.g.,
Bergeron et al., 1992; Napiwotzki et al., 1999). While this
is not in serious doubt, opportunities for direct observational
tests of the work are rare. This is particularly true of models
that take into account the finite stellar temperature and details
of the core/envelope structure, discussed above. Varying the
assumed input parameters in these models can lead to quite
subtle, but important differences in the model predictions.
To test these requires independent measurements of white
dwarf mass, which can be compared with the spectroscopic
results. Such information can be obtained spectroscopically
by measuring the gravitational redshift of absorption lines in
the white dwarf atmosphere (Vgr[km s−1]=0.636 M/R,M
and Rin Solar units), but this is only possible if the systemic
radial velocity is known as a reference point. Generally, this
is only the case if the white dwarf is part of a binary system
and then there is also the possibility of obtaining independent
dynamical information on the white dwarf mass from the sys-
tem orbital parameters. An additional important constraint is
knowledge of the stellar distance.
The four best white dwarf mass determinations, where we
have the most complete and accurate information, are for 40
Eri B, Procyon B, V471 Tauri B and Sirius B, where we can
combine the assembled data with the Hipparcos parallax to
test the mass radius relation (Fig. 7). While there is good
agreement between the observation and theory, there nev-
ertheless remains a high degree of uncertainty in the mass
determinations. As a result, for example, it is not possible
Springer

Astrophys Space Sci (2006) 303:3–16 11
Fig. 7 Comparison of mass
estimates for 40 Eri B, Procyon
B, V471 Tau B and Sirius B
with the evolutionary models of
Wood (1995), displayed at
various temperatures and with
“thin” and “thick” H envelopes.
The solid limiting curve
represents the Hamada–Salpeter
zero temperature relation for a
carbon core (figure produced by
Jay Holberg)
Fig. 8 Wide Field Planetary
Camera image of the binary 56
Per, where each component (A
& B) is itself resolved into a pair
(left). (right) Successive images
of the Aa/Ab pair taken ∼18
months apart clearly show the
orbital motion of the system
to distinguish between different models, such as those with
“thin” or “thick” H envelopes.
We have such complete data (dynamical masses, gravita-
tional redshifts and accurate parallaxes) for only a very few
white dwarfs. Therefore, it is important to extend the sample
of white dwarfs for which we have to more objects and, pos-
sibly, explore a wider range of masses and temperatures. The
role of direct imaging in the UV as means of discovering
new systems is particularly important. For example, a ma-
jor result of the EUV sky surveys was the discovery of many
unresolved binary systems containing white dwarfs and com-
panion spectral types ranging from A to K (e.g., Barstow et
al., 1994; Burleigh et al., 1997; Vennes et al., 1998). In visi-
ble light the presence of a companion of spectral type earlier
than mid-K will swamp the signature of the white dwarf
making it undetectable. In the EUV, where the companion
flux is generally negligible, the white dwarf stands out very
clearly. The UV wavelength range is an even more efficient
way of searching for these binaries, as the interstellar opacity
is much lower than in the EUV, and the GALEX sky survey
is finding many examples.
The large difference in the visual magnitudes makes these
systems generally impossible to resolve with ground-based
observations. However, in the UV where the contrast is far
better it is possible to measure their separations, or at least
provide improved constraints. HST Wide Field Planetary
Camera 2 images of 18 binary systems have resolved 9 ob-
jects (Barstow et al., 2001a). Figure 8 shows one of the most
interesting examples, 56 Per, a known binary in which each
component has been resolved into a pair, making it a quadru-
ple star system. The white dwarf is a companion to 56 Per
A and is labeled 56 Per Ab in the image. At a distance of
42 pc, the measured 0.39 arcsec separation indicates a binary
period of ∼50 years for the Aa/Ab system. Therefore, the or-
bital motion of the two stars should be readily apparent with
repeated exposures on timescales ∼1–2 years, from which
a dynamical white dwarf mass can ultimately be obtained.
This is clearly demonstrated in Fig. 8 which shows a zoomed
view of the Aa/Ab pair from the main image and, on a similar
scale, a second image obtained ∼18 months after the first.
Continued monitoring of systems like 56 Per will eventually
yield the orbital parameters and dynamical determinations
Springer

12 Astrophys Space Sci (2006) 303:3–16
of the component masses. This requires continued access to
UV imaging.
3. Problems of white dwarf evolution
Clearly significant progress has been made in the study of
white dwarf stars through ultraviolet observations. However,
these have not yet given us access to a detailed understand-
ing of the important physics because they have been limited
in scope. First the total number of objects studied with the
necessary spectral resolution and signal-to-noise is small.
Therefore, we do not know how representative of the general
population individual objects or small groups of objects may
be. This is exacerbated by the fact that there are strong selec-
tion effects present within the existing samples. For example,
most hot DA white dwarfs have been observed because of ex-
pectations that significant quantities of heavy elements were
present in their photospheres, and, as a result are mostly stars
with temperatures above 50,000 K. Therefore, a number of
significant questions regarding the evolution of white dwarfs
remain to be solved:
rWhat is the origin of the DO–DB gap and the relationship
between the H- and He-rich white dwarf branches? For
example, do DO white dwarfs appear as DAs through float
up of residual H?
rWhat mechanisms determine the compositions of the DA
white dwarfs as they cool? Radiative levitation and gravi-
tational diffusion are clearly important, but why do some
stars of the same apparent temperature and gravity have
widely differing compositions? Is accretion (from the ISM
or a companion) an important mechanism? Do the abun-
dance differences reflect differences in progenitor compo-
sition/prior evolution of the progenitor?
rWhat are the metal abundances in the He-rich white
dwarfs? Are the PG1159 stars and DOs part of a single
sequence or do they represent the separate progenitor evo-
lutionary paths. Do all hot He-rich white dwarfs eventually
become DBs?
rWhat is the 3-D structure of the local ISM? How do white
dwarfs interact with it and exchange material. Does mass-
loss continue beyond the planetary nebula phase? Is appar-
ently circumstellar material detected in some of the hottest
white dwarfs evidence of such mass-loss or merely the
residual signature of PN material?
rWhat is the initial-final mass relation for both H- and He-
rich white dwarfs and what is the upper limit on the possible
progenitor mass? Does the theoretical mass-radius relation
for white dwarfs stand up to close scrutiny? Do the most
massive white dwarfs have exotic core compositions be-
yond the C/O product of helium burning?
All white dwarfs that have ever been studied in the UV re-
side within our own galaxy and must have emerged from
stellar populations with different ages and environments. To
solve the outstanding problem and make significant further
progress in the study of white dwarfs requires a substan-
tial enlargement of the sample, to properly examine the full
range of temperatures, gravities and possible environmental
conditions.
rExpand the number of galactic white dwarfs by a factor 10
for which high resolution/high signal-to-noise UV spectra
are available.
rIncrease by a factor 10 the number of binary systems with
white dwarf components for which astrometric masses can
be obtained.
rBe able to study uniform, co-eval populations of white
dwarfs in globular clusters, the Magellanic Clouds and
nearby galaxies.
4. Future white dwarf research in the far ultraviolet
4.1. White dwarfs in the galaxy
A large-scale survey of the hot white dwarfs in the galaxy will
provide observations which can simultaneously address two
broad areas of astrophysics: the local interstellar medium,
its composition, ionization and structure; and the degener-
ate stars, their origin and evolutionary history as well as the
detailed modeling of critical physical processes in their pho-
tospheres. The primary broad scientific objectives would be
following:
rDefine the evolutionary history of the hot white dwarf stars
through detailed modeling of their photospheric composi-
tion and structure.
rStudy the occurrence of circumstellar material surrounding
the white dwarfs and their interaction with the ISM.
rMap out the 3-D structure of the local interstellar medium
(LISM) and determine its composition and ionization state.
rUse the morphology of the LISM and the estimated el-
emental diffusion (gravitational settling) times for white
dwarf photospheres to provide a crucial test of the abil-
ity of interstellar accretion processes to explain abundance
patterns in cooler white dwarfs.
rIdentify and characterise new non-interacting white dwarf
binary systems.
rCarry out a long-term astrometric programme to determine
white dwarf masses.
4.1.1. White dwarf composition
Although there is a qualitative understanding of how
the abundance patterns vary across the hot white dwarf
Springer

Astrophys Space Sci (2006) 303:3–16 13
population, the detailed picture is quite confused and some
very important questions need to be answered. The observed
abundances will reflect the balance of several processes, in-
cluding mass loss, radiative levitation, gravitational settling,
and accretion from the LISM. DA white dwarfs have been
selected for follow-up UV studies mainly on the basis of
their EUV fluxes, low values indicating the presence of pho-
tospheric metals. Thus, the existing observational sample is
strongly biased toward such stars. Few stars having appar-
ently pure-H atmospheres have been observed. In addition,
the existing data are highly non-uniform in wavelength cov-
erage, signal-to-noise and spectral resolution, which yield
non-uniform detection criteria for spectral features. There-
fore, the apparent absence of metals may be as much a func-
tion of the weak detection limits and too few observations of
appropriate stars than a real lack of metals. Hence, we have
no idea whether the small group of heavy element-rich stars
(see Fig. 4) are typical of the cooler group below 50,000 K,
or whether they are truly unusual. Are there really two dis-
tinct groups of stars with and without metals? Or, is this an
artifact of the small sample and in reality compositions range
between the two extremes? It is hard to explain why any DA
would have no heavy elements at all. Thus, establishing the
frequency of pure-H envelopes is an important component of
providing an answer to this problem. In particular, we need
to properly sample the lower temperature white dwarfs, es-
pecially within the 20,000–35,000 K region, which are not
well represented in earlier studies and existing data.
PG1159 stars are rare objects, about 40 are known. Only
a few of them have been studied in detail in the UV. High-
resolution UV observations are essential, because most di-
agnostic metal lines observed in these extremely hot stars
are located in this wavelength region. The wide spread in
element abundances, as well as the observed iron-deficiency
and neon- and fluorine-overabundances show that PG1159
stars have a large, and unique, potential to study mixing and
fusion processes whose consequences are usually unobserv-
able in other stars. As a consequence of a late He-shell flash,
PG1159 stars exhibit intershell matter that normally remains
hidden in the stellar interior. In contrast to DA and DO white
dwarfs, the observed element abundances in PG1159 stars
are not affected by gravitational settling, hence, abundance
patterns still do reflect the history of these stars.
4.1.2. Circumstellar material
What appears to be circumstellar gas has been detected in
most of the white dwarfs observed in high resolution HST
spectra. What is the nature of this material? Is it present in all
white dwarfs with photospheric metals or is there a temper-
ature cut-off? A lower temperature (greater age) limit would
imply that we are looking at a nebular remnant, which dis-
perses with time. Apportioning unresolved lines to circum-
stellar and photospheric components for the whole sample
of stars is essential for correct atmospheric abundance mea-
surements. For example, when observed by HST, the photo-
spheric C abundance of G191-B2B was really found to be a
factor 5 lower than the value obtained by IUE (see Barstow
et al., 2003b).
4.1.3. 3-D structure of the ISM
High-resolution spectra of LISM absorption lines from abun-
dant ionic species (C II, C II∗, N I, O I, Si II-III, S II, Fe II, Zn
II, Cr II, Mg I-II) will provide several quantitative measure-
ments of nearby interstellar gas. We will be able to measure
the line-of-sight densities and composition of the LISM, de-
rive velocities, to probe the gas kinematics, and determine
the ionization of the gas. It is important that the chosen white
dwarf sample includes enough different lines-of-sight to pro-
vide a true 3-D picture. The high resolution of the echelle data
for these white dwarfs will be instrumental in resolving the
inherent complex velocity structure seen even in the very
local gas within 15 pc (e.g., Lallement et al., 1986; Sahu
et al., 2000a,b), and provide the means to obtain reliable
ionic column densities for the individual velocity compo-
nents in the LISM. Does a single bulk velocity vector fit all
the lines of sight through a particular cloud, or is the gas
fragmented into filamentary structures more characteristic
of low velocity shocks? An adequate sample of white dwarfs
will provide extensive sampling for distances out to 50–100
parsecs.
The problem of the variability of the D/Hratio in the
LISM appears to be on the way to resolution, but there are
many details that still need to be addressed (Moos et al.,
2002; Sahu et al., 2000a,b). Specifically, obtaining reliable
D/Hratios is not easy. Extreme care must be taken in data
reduction and calibration and accounting for multiple veloc-
ity components along the line of sight. One must look at the
heavy ions to determine what velocities are present, because
they can be easily masked in the intrinsically broader profiles
of light HI and DI.
Column densities, measured from high resolution far-UV
spectra, will be useful for determining chemical abundances
in the LISM clouds. Comparing the LISM abundances with
those of Savage and Sembach (1996) would determine if
the cloud has abundances and depletions similar to warm
partially ionized gas observed in the more distant ISM.
Examination of the white dwarf data has revealed an un-
recognized problem, namely the frequent occurrence of in-
terstellar Si III λ1206 in many lines-of-sight. This is diffi-
cult to reconcile, since this ion has a high charge-exchange
rate with neutral hydrogen. Its presence suggests substan-
tial amounts of warm ionized gas have been unaccounted
for in the LISM. It is intriguing that low ionization species
seen in G191-B2B (Sahu et al., 2000a,b) are close to the
Springer

14 Astrophys Space Sci (2006) 303:3–16
velocity of interstellar Si III. An examination of S II/H I
and C II/H I ratios compared to those found for the LIC
at 19.3 km/s, suggests the H ionization fraction is near 0.7
for the 8.6 km/s gas. Another problem is that the short-
ward “circumstellar” components of C IV in G191-B2B
also have a close velocity coincidence with the Si III and
the 8.6 km/s low ionization species. Photoionization calcu-
lations find no way to have all of these ions in the same
gas. High resolution (R∼50,000–100,000) data for other
sightlines should be able to determine the origin of the
Si III.
4.1.4. Interstellar accretion
For the cooler DAZ stars (6000 <Teff <12,000 K) the ex-
istence of heavy elements such as Ca, Mg and Fe has tradi-
tionally been attributed to interstellar accretion (e.g., Dupuis
et al., 1993). For DA stars at intermediate temperatures
(20,000K<Teff <25,000 K) the presence of heavy ele-
ments poses a dilemma.
In DAZ stars the retention times for high Z elements in
the observable photospheres, while substantial (i.e., ∼104
yr), remains short compared to the thermal (cooling) time
scale of the photospheres. It is possible to imagine that DAZ
white dwarfs represent stars passing through, or recently
passed through, a diffuse interstellar cloud and having ac-
creted heavy elements. As long as the fraction of DAZ stars
remained small it was possible to entertain such views since
the number of these stars were the result of infrequent en-
counters between diffuse clouds and white dwarfs. However,
Zuckerman et al. (2003) have shown that the occurrence of
cool DAZ stars approaches 25%, which is well in excess of
the fraction that might be attributed to ISM accretion. Like-
wise, other possible explanations involving intrinsically rare
events such as comet impacts are equally untenable. Yet,
Zuckerman et al., were able to demonstrate a correlation be-
tween the presence of low mass companions and the DAZ
phenomena.
Thus, some form of ongoing circumstellar accretion ap-
pears necessary to explain the bulk of the DAZ stars. Typi-
cally the DAZ stars are too faint and too cool to search for the
UV presence of heavy elements. However, it is possible to in-
vestigate the DAZ phenomena in a hotter class of DA stars at
UV wavelengths. In earlier work, there has been little evalua-
tion of the actual conditions of the interstellar medium along
the lines of sight to the known DAZ white dwarfs. Neverthe-
less, knowledge of the morphology of the LISM is necessary
to evaluate and critically test the accretion model, since the
local distribution of interstellar clouds directly determines
the efficiency of interstellar accretion. In general, interstel-
lar features can be distinguished from stellar features on the
basis of velocity and the presence of excited fine-structure
lines. We can use the several density sensitive indicators to
probe the gas phase density of any accreting medium in the
vicinity of the star. For example, an important diagnostic tool
is the presence of the collisionally exited C II∗λ1335 line,
from which estimates of the ambient electron density can be
determined (Holberg et al., 1999). The presence of ground
state and exited C I lines indicate the presence of a medium,
which is sufficiently dense to be effectively self-shielding
with respect to UV radiation shortward of the C I ionization
limit. Even if excited lines are not detected, the mere pres-
ence of significant column density, as evidenced by strong
IS absorption, is important.
4.1.5. White dwarfs in binaries
Many new white dwarf binaries with main sequence com-
panions are being discovered by the GALEX UV sky sur-
vey. In these systems the presence of the companion ob-
scures the white dwarf at long wavelengths. Hence, for
these objects, far-UV spectroscopy is essential to deter-
mine the basic physical parameters of the white dwarfs.
For example, temperature and surface gravity can be de-
termined from model atmosphere analyses of the hydrogen
Lyman series lines (e.g., Barstow et al., 2003a), while pho-
tospheric composition can be determined from any heavy
element lines present, as discussed above. Since many
of the brighter companions will be members of the Hip-
parcos catalogue, and therefore have well-determined dis-
tances, the sample of binary white dwarfs can be used to
study the mass-radius relation and the initial-final mass
relation. The latter is extremely uncertain and is impor-
tant in validating potential models for type Ia supernovae
progenitors.
Only about half a dozen of the known binary systems
have sufficiently short orbital periods for astrometric infor-
mation to be obtained on sensible timescale. A subset of
any newly discovered binaries will also fall into this group.
Thus far astrometric orbits and directly determined masses
are only available for three white dwarfs, Sirius B, 40 Eri B
and Procyon B. All other WD masses are based on gravita-
tional redshifts or spectroscopic log gdeterminations, which
require theory-dependent assumptions, or on generally un-
certain measurements of interacting close binaries. Knowl-
edge of the masses is, in turn, vital in testing the theory of
WD structure (e.g., the mass-radius-core composition rela-
tion), understanding the history of star formation in the solar
neighbourhood, and setting limits on the age of the Galaxy.
4.2. White dwarfs outside the galactic disk
Imaging surveys of white dwarfs have been carried out
in globular clusters, but most of the individual stars are
only characterized by broad band photometry. This pro-
vides almost no information on the white dwarf masses and
Springer

Astrophys Space Sci (2006) 303:3–16 15
weak temperature constraints. Knowledge of the white dwarf
masses is essential for determining the cooling age of indi-
vidual stars and interpretation of the observed luminosity
functions. Study of white dwarfs in globular clusters yields a
number of advantages compared to the galactic population.
rAll the stars lie at the same, known distance.
rThe stars are co-eval.
rThe stars are all descendents of a uniform population.
Knowledge of white dwarf temperature and gravity (and,
therefore mass) in globular clusters will provide a direct cali-
bration of the initial-final mass relation. In particular, the up-
per limit of the progenitor mass will be reliably established
for the first time, which has important implications for mod-
els of SNIa systems. In the galaxy most of the white dwarf
progenitors for stars in the disk will probably have had popu-
lation I metallicities. With all progenitors in a globular cluster
being population II, the metallicity and, as result, their prior
evolution will be well determined. Since post-main sequence
evolution is affected by stellar metallicity, in particular in es-
tablishing core He burning, we would expect the resulting
white dwarf population to have different characteristics to
those in the galaxy.
5. The future need for far UV missions
In Section 4 we have outlined in detail the scientific goals
for future studies of white dwarfs. This wealth of science is
only possible through a programme of observations in the far
ultraviolet waveband. The principal need is for high resolu-
tion spectroscopy, but diffraction limited imaging is also of
importance. During the past 15 years, these joint capabilities
have been provided by the Hubble Space Telescope, follow-
ing on from 18 years of operations with IUE. Although of
great importance, the relatively small aperture of IUE limited
high resolution studies of white dwarfs to a handful of the
brightest examples. Using a variety of instruments HST has
provided us with a flow of high signal-to-noise and resolu-
tion (R∼50,000–100,000) spectra of white dwarfs and the
first diffraction limited imaging of white dwarfs in binary
systems. However, HST time has had to be shared across
a wider range of wavelengths including the visible and IR
bands. Since 1999, this has been complemented by the avail-
ability of the FUSE mission, working down to shorter wave-
lengths than HST (912 ˚
A cf. 1050 ˚
A), but with more modest
spectral resolution (R∼20,000). Sadly, the STIS instrument
on HST failed in August 2004, ending the UV spectroscopic
capability for the foreseeable future. Also, FUSE is prob-
ably nearing the end of its lifetime. While operations have
been maintained through the heroic efforts of the FUSE team,
continued degradation of its attitude control system (through
gyroscope and reaction wheel failures) will eventually lead
to its termination. Within current mission plans access to far
UV spectroscopy is likely to end soon with no prospect of any
replacement. It is astounding that, at a time when GALEX is
opening up the discovery space in UV astronomy in a major
way, that we will have no way of adequately following up its
surveys.
The current situation regarding continued access to the
far UV is full of complex programmatic and political is-
sues which are making it difficult to plan ahead. For ex-
ample, a shuttle mission to HST could carry out the instal-
lation of the Cosmic Origins Spectrograph (COS), which
will operate in the far UV with R∼20,000, and (possibly)
repair STIS. However, this is currently ruled out on safety
grounds. A robotic alternative is being studied as an exten-
sion to the attachment of the de-orbit module (which must
be carried out to control HST re-entry). If successful, any
HST repair/servicing would resurrect the far UV capability
and extend the mission lifetime to ∼2012. Alternatively, a
new mission called the Hubble Origins Probe (HOP) is be-
ing studied as a way of placing COS into orbit (with the Wide
Field Camera 3) using a new telescope and spacecraft.
Whatever is finally decided regarding HST and COS, the
result will not be ideal. HST operations will still have a lim-
ited life and COS does not operate at the spectral resolu-
tion needed for the work proposed here. Furthermore, with
its multi-waveband capability, HST has never provided as
much UV observing time as is really needed. Therefore,
there remains the problem of future provision for far UV
spectroscopy. This can be divided into two: the replacement
of 2-m class access to high resolution spectroscopy for the
∼2010–2015 period and the long-term provision of a larger
scale (4 to 6-m) facility beyond 2015.
There is an urgent need to provide a dedicated far UV
mission to follow HST. To achieve this in a relatively short
timescale requires the use of existing technology, but within
these constraints it should be possible to provide an instru-
ment with enhanced sensitivity through avoidance of com-
plex relay optics and improved (but still current) grating and
detector technology. A 2-m class telescope would be able
to address many of the science goals relating to observation
of white dwarfs in our own galaxy, provided the following
technical capabilities are achieved:
rGalactic white dwarf spectroscopic survey
◦λλ ∼912–3000 ˚
A, R∼50,000–100,000, Vlim ∼18
rAstrometric white dwarf masses
◦Diffraction limited imaging to V∼18
The World Space Observatory (WSO, see e.g., Barstow
et al., 2003c) is an example of a mission that could meet
these aims on the necessary timescale. Plans for WSO have
been developed over several years and phase A studies of the
Springer

16 Astrophys Space Sci (2006) 303:3–16
concept and key instruments have been carried out. There
is a good prospect that this mission will go ahead under the
leadership of the Russian Space Agency, with contributions
from many other countries.
In contrast, provision of a large (4 to 6-m) UV telescope is
not in the plans of any space agency. Some US studies have
been carried out on a large UV/optical space telescope but
no concrete plans have yet emerged. However, as agencies
begin to plan their programmes for the time-frame beyond
2015, it is absolutely essential that these should include a
large UV facility. It would preferable that such a telescope
should be UV only, but it may be inevitable that this would
need to be UV/optical mission on cost grounds. For white
dwarf research, the key requirements are for:
rGlobular cluster/Magellanic Cloud white dwarf surveys
◦Integral field spectroscopy λλ ∼912–1300 ˚
A, R∼1000,
Vlim ∼28
◦Wide field imaging (10 arcmin) to V∼35
References
Bannister, N.P., Barstow, M.A., Holberg, J.B., Bruhweiler, F.C.:
MNRAS 341, 477 (2003)
Barstow, M.A., Bond, H.E., Burleigh, M.R., Holberg, J.B.: MNRAS
322, 891 (2001)
Barstow, M.A., Good, S.A., Burleigh, M.R., Hubeny, I., Holberg, J.B.,
Levan, A.J.: MNRAS 344, 562 (2003a)
Barstow, M.A., Good, S.A., Holberg, J.B., Hubeny, I., Bannister, N.P.,
Bruhweiler, F.C., Burleigh, M.R., Napiwotzki, R.: MNRAS 341,
870 (2003b)
Barstow, M.A., Holberg, J.B., Cruise, A.M., Penny, A.J.: MNRAS 286,
58 (1997)
Barstow, M.A., Holberg, J.B., Fleming, T.A., Marsh, M.C., Koester, D.,
Wonnacott, D.: MNRAS 270, 499 (1994)
Barstow, M.A., Hubeny, I., Holberg, J.B.: 6C299, 520 (1998)
Barstow, M.A., Hubeny, I., Holberg, J.B.: 6C307, 884 (1999)
Barstow, M.A., Bannister, N.P., Cruddace, R.G., Kowalski, M.P., Wood,
K.S., Yentis, D.J., Gursky, H., Barbee, T.W., Jr., Goldstein, W.H.,
Kordas, J.F., Fritz, G.G., Culhane, J.L., Lapington, J.S.: Proc. SPIE
4864, 640 (2003c)
Bergeron, P., Saffer, R., Liebert, J.: ApJ 394, 228 (1992)
Bl¨ocker, T.: A&A 297, 727 (1995)
Bl¨ocker, T., Herwig, F., Driebe, T., Bramkamp, H., Sch¨onberner, D.: in:
J. Isern, M. Hernanz and E. Garcia-Berro (eds.), White Dwarfs,
Kluwer, Dordrecht, p. 57 (1997)
Bruhweiler, F.C.: Astrophysics in the Extreme Ultraviolet, C.S. Bowyer
and R.F. Malina (eds.), p. 261 (1996)
Burleigh, M.R., Barstow, M.A., Fleming, T.A.: MNRAS 6C287, 381
(1997)
Chandrasekhar, S.: ApJ 84, 81 (1931)
Chandrasekhar, S.: 6C95, 226 (1935)
Dreizler, S., Werner, K.: A&A 314, 217 (1996)
Dreizler, S., Wolff, B.: A&A 348, 189 (1999)
Fleming, T.A. et al.: ApJ 308, 176 (1986)
Fowler, R.H.: 6C87, 114 (1926)
Frisch, P.C.: Space Sci. Rev. 72, 499 (1995)
Frisch, P.C., Grodnicki, L., Welty, D.E.: ApJ 574, 834 (2002)
Hamada, T., Salpeter, E.E.: ApJ 134, 683 (1961)
Herwig, F., Bl¨ocker, T., Langer, N., Driebe, T.: A&A 349, L5 (1999)
Herwig, F., Lugaro, M., Werner, K.: in: S. Kwok, M. Dopita and R.
Sutherland (eds.), Planetary Nebulae, IAU Symp. 209, ASP, p. 85
(2003)
Hoffmann, A.I.D., Dreizler, S., Rauch, T., Werner, K., Kruk, J.W.: in: D.
Koester and S. Moehler (eds.), White Dwarfs, ASP Conf. Series,
vol. 334, 321 (2005)
Holberg, J.B., Bruhweiler, F.C., Barstow, M.A., Dobbie, P.D.: ApJ 517,
850 (1999)
Jorissen, A., Smith, V.V., Lambert, D.L.: A&A 261, 164 (1992)
Kleinman, S.J., Harris, H.C., Eisenstein, D.J., Liebert, J., Nitta, A.,
Krzesinski, J., Munn, J.A., Dahn, C.C., Hawley, S.L., Pier, J.R.,
Schmidt, G., Silvestri, N.M., Smith, J.A., Szkody, P., Strauss,
M.A., Knapp, G.R., Collinge, M.J., Mukadam, A.S., Koester, D.,
Uomoto, A., Schlegel, D.J., Anderson, S.F., Brinkmann, J., Lamb,
D.Q., Schneider, D.P., York, D.G.: ApJ 607, 426 (2004)
Lallement, R., Vidal-Madjar, A., Ferlet, R.: A&A 163, 204 (1986)
Liebert, J., Bergeron, P., Holberg, J.B.: ApJS 156, 47 (2005)
Liebert, J., Wesemael, F., Hansen, C.J., Fontaine, G., Shipman, H.L.,
Sion, E.M., Winget, D.E., Green, R.F.: ApJ 309, 241 (1986)
Lugaro, M., Ugalde, C., Karakas, A.I., et al.: ApJ 615, 934 (2004)
Lyu, Bruhweiler: ApJ 459, 216 (1996)
Margio, P.: A&A 370, 194 (2001)
Miksa, S., Deetjen, J.L., Dreizler, S., et al.: A&A 389, 953 (2002)
Moos, H.W., Sembach, K.R., Vidal-Madjar, A., York, D.G., Fried-
man, S.D., H´ebrard, G., Kruk, J.W., Lehner, N., Lemoine, M.,
Sonneborn, G., Wood, B.E., Ake, T.B., Andr´e, M., Blair, W.P.,
Chayer, P.: ApJS 140, 3 (2002)
Mueller, H.R., Frisch, P.C., Zank, G.P.: American Geophysical Union,
Fall Meeting (2003)
Napiwotzki, R.: A&A 350, 101 (1999)
Napiwotzki, R., Green, P.J., Saffer, R.A.: ApJ 517, 399 (1999)
Oppenheimer, B.R., Hambly, N.C., Digby,A.P.,Hodgkin, S.T., Saumon,
D.: Science 282, 698 (2001)
Pagel, B.E.J.: Chemical enrichment of the intracluster and intergalactic
medium, in: R. Fusco-Femiano & F. Metteucci (eds.), ASP Conf.
Series, 253, 489 (2002)
Redfield, S. and Linsky, J.: ApJ 534, 825 (2000)
Reiff, E., Rauch, T., Werner, K. and Kruk, J.W.: in: D. Koester and S.
Moehler (eds.), White Dwarfs, ASP Conf. Series, vol. 334, 225
(2005)
Sahu, M.S.: Proceedings of IAU 198, Light Element Abundances,Natal,
Brazil (2001a)
Sahu, M.S., Landsman, W., Bruhweiler, F.C., Holberg, J., Hubeny, I.,
Barstow, M., Linsky, J., Gull, T., Lindler, D., Lanz, T., Feggans,
K.: Bull AAS 199, 1101 (2001b)
Savage, B. and Sembach, K.: ApJ 470, 893 (1996)
Schatzmann, E.: White Dwarfs, North Holland Publishing, Amsterdam
(1958)
Schuh, S., Dreizler, S., Wolff, B.: A&A 382, 164 (2002)
Vennes, S., Christian, D., Thorstensen, J.R.: ApJ 502, 763 (1998)
Werner, K.: in: T. Bl¨ocker, L.B.F.M. Waters and A.A. Zijlstra (eds.),
Low mass Wolf–Rayet stars: Origin and evolution, Ap&SS, 275,
27 (2001)
Werner, K., Rauch, T., Reiff, E., Kruk, J.W. and Napiwotzki, R.: 2004,
A&A 427, 685.
Werner, K., Rauch and Kruk, J.W.: 2005, A&A 433, 641.
Wesemael, F., Green, R.F. and Liebert, J.: 1985, ApJS 58, 379.
Wood, M.A.: 1992, ApJ 386, 539.
Wood, M.A.: 1995, in: Koester D. and Werner K. (eds.), Lecture Notes
in Physics, White Dwarfs. Springer, Berlin, p. 41.
Zuckerman, B., et al.: 2003, ApJ 596, 477.
Springer

Astrophys Space Sci (2006) 303:17–31
DOI 10.1007/s10509-005-9018-3
ORIGINAL ARTICLE
Key Problems in Cool-Star Astrophysics
Isabella Pagano ·Thomas R. Ayres ·
Alessandro C. Lanzafame ·Jeffrey L. Linsky ·
Benjam´ın Montesinos ·Marcello Rodon`o†
Received: 18 April 2005 / Accepted: 10 October 2005
C
Springer Science +Business Media B.V. 2006
Abstract Selected key problems in cool-star astrophysics
are reviewed, with emphasis on the importance of new ultra-
violet missions to tackle the unresolved issues.
UV spectral signatures are an essential probe of critical
physical processes related to the production and transport of
magnetic energy in astrophysical plasmas ranging, for exam-
ple, from stellar coronae, to the magnetospheres of magne-
tars, and the accretion disks of protostars and Active Galactic
Nuclei. From an historical point of view, our comprehension
of such processes has been closely tied to our understand-
ing of solar/stellar magnetic activity, which has its origins in
a poorly understood convection-powered internal magnetic
dynamo. The evolution of the Sun’s dynamo, and associated
magnetic activity, affected the development of planetary at-
mospheres in the early solar system, and the conditions in
which life arose on the primitive Earth. The gradual fading
of magnetic activity as the Sun grows old likewise will have
profound consequences for the future heliospheric environ-
ment. Beyond the Sun, the magnetic activity of stars can
influence their close-in companions, and vice versa.
Cool star outer atmospheres thus represent an important
laboratory in which magnetic activity phenomena can be
†Deceased October 23, 2005
I. Pagano ()
INAF, Catania Astrophysical Observatory, Italy
T. R. Ayres
CASA, University of Colorado, Boulder CO, USA
A. C. Lanzafame ·M. Rodon`o
Department of Physics and Astronomy, Catania University, Italy
J. L. Linsky
JILA, University of Colorado & NIST, Boulder CO, USA
Benjam´ın Montesinos
IAA/CSIC, Granada and LAEFF/INTA, Madrid, Spain
studied under a wide variety of conditions, allowing us to
gain insight into the fundamental processes involved.The UV
range is especially useful for such studies because it contains
powerful diagnostics extending from warm (∼104K) chro-
mospheres out to hot (1–10 MK) coronae, and very high-
resolution spectroscopy in the UV has been demonstrated by
the GHRS and STIS instruments on HST but has not yet been
demonstrated in the higher energy EUV and X-ray bands. A
recent example is the use of the hydrogen Lyαresonance line
– at 110 000 resolution with HST STIS – study, for the first
time, coronal winds from cool stars through their interac-
tion with the interstellar gas. These winds cannot be detected
from the ground, for lack of suitable diagnostics; or in the
X-rays, because the outflowing gas is too thin.
A 2 m class UV space telescope with high resolution spec-
troscopy and monitoring capabilities would enable important
new discoveries in cool-star astronomy among the stars of the
solar neighborhood out to about 150 pc. A larger aperture fa-
cility (4–6 m) would reach beyond the 150 pc horizon to
fainter objects including young brown dwarfs and pre-main
sequence stars in star-forming regions like Orion, and mag-
netic active stars in distant clusters beyond the Pleiades and
αPersei. This would be essential, as well, to characterize
the outer atmospheres of stars with planets, that will be dis-
covered by future space missions like COROT, Kepler, and
Darwin.
Keywords Late-type stars ·Magnetic activity ·
Chromospheres ·Coronae ·UV astronomy
1. Introduction
Magnetic activity signatures, analogous to well-known so-
lar phenomena, are widely observed in cool stars. Dark cool
Springer

18 Astrophys Space Sci (2006) 303:17–31
spots in the stellar photosphere produce modulations of op-
tical light curves as the star rotates, and the migration of
Doppler shifted features through the line profiles. Flares,
coronal mass ejections (CME’s), and stellar winds (of lumi-
nous cool stars) are commonly observed. Magnetic activity
also is responsible for prominent emission-line spectra in the
ultraviolet and far-ultraviolet (UV/FUV) regions, and for the
entire stellar X-ray and radio emission.
In RS CVn-type binary systems, dMe stars, and young
rapidly rotating dwarfs, magnetic activity is of paramount
interest because of its extreme characteristics: in such stars
more than 50% of the stellar photosphere can be spotted,
chromospheric and transition region (TR) fluxes are so high
that they have topped out at a saturation limit, and the X-ray
luminosity can even reach a remarkable 10−3of the stellar
bolometric luminosity, ten thousand times larger than the
equivalent solar ratio.
Thanks to long-term systematic monitoring programs of
highly convective cool stars initiated in the mid sixties, no-
tably at Crimea (Chugainov, 1966), Catania (Godoli, 1968)
and Mt. Wilson (Wilson, 1978) and successively pursued
worldwide at other Observatories (e.g., Vienna, Konkoly,
Armagh, Potsdam and Fairborn), about forty years later we
are now in a position to state with confidence that many
cool stars show periodic variations of their emissions from
almost all atmospheric layers with characteristics similar to
the 11-year solar cycle (Wilson, 1978; Baliunas et al., 1995;
Rodon`o et al., 1995, 2000, 2002; Strassmeier et al., 1997;
Lanza et al., 1998, 2002; Cutispoto et al., 2001, 2003; Olah
and Strassmeier, 2002; Messina et al., 2001; Messina and
Guinan, 2002, 2003; Favata et al., 2004). However, there are
also phenomena that do not have analogues on the Sun: high-
latitude (Rodon`o, 1986) and polar spots (Strassmeier, 1990,
Strassmeier et al., 1991), very hot (>10 MK) coronal com-
ponents (Linsky, 2003; Audard et al., 2004), etc..1As for the
Sun, however, the activity phenomena occurring in the dif-
ferent atmospheric layers of cool stars appear to be closely
correlated (Catalano et al., 2000; Messina et al., 2002).
The Sun’s contemporary magnetic activity affects the
Earth’s biosphere and human civilization through a variety of
phenomena lumped under the heading “space weather.” Fur-
thermore, the Sun was considerably more active in its youth
than today (e.g., Ribas et al., 2005), and had a correspond-
ingly larger impact on the solar system, especially primitive
planetary atmospheres and the environment in which life be-
gan on Earth. Given the importance of accurately forecasting
solarrelated influences – now, in the past, and in the future
1Detailed descriptions and references for all the phenomena mentioned
above and the associated theoretical scenarios can be found in the Pro-
ceedings of the Cambridge Workshops “Cool Stars, Stellar Systems and
the Sun”, published in the Astronomical Society of the Pacific Confer-
ence Series.
– it is essential to investigate how magnetic activity depends
upon stellar parameters (mass, radius, rotation rate, binarity,
evolutionary stage, etc.) if we ever want to develop a theory
for these high-energy phenomena that has predictive power.
Hot plasma in magnetically disturbed cool-star outer at-
mospheres, with temperatures from 10,000 K to several
millions, can be observed in a single ultraviolet spectrum,
providing simultaneous information on thermal structures
of a wide range of atmospheric components. As an exam-
ple, Figure 1 illustrates the spectrum of αCen A (G2 V)
obtained with HST/STIS in the range 1140–1670 ˚
A with a
resolution R∼114,000 (E140H). In this spectrum Pagano
et al. (2004) identify a total of 671 emission lines from 37 dif-
ferent ions, including low temperature chromospheric lines
(e.g., C I,OI), TR lines (e.g., C II–IV,NIV,OIII–V,SiII–IV),
the coronal line Fe XII 1242 ˚
A, a number of intersystem lines
(e.g.,[O IV]) that are useful for measuring electron densities,
and two molecules (CO and H2).
In what follows, we emphasize the importance of the ul-
traviolet for addressing unresolved issues in the study of cool
stars and related objects.
2. A brief history of UV cool-star astronomy
Ultraviolet astronomy has been crucial to the study of cool
stars. Although the photospheric spectral energy distribution
of this type of star peaks at red wavelengths, the presence of
hot plasma in chromospheres, transition regions, and coro-
nae reveals itself most prominently through emission at the
shorter ultraviolet wavelengths, where conveniently the pho-
tospheric contribution typically is faint.
A comprehensive account of the development of UV
astronomy from the pioneering rocket experiments of the
1960’s until the era of the prolific International Ultraviolet
Explorer (IU E ) can be found in the review by Boggess and
Wilson (1987). In summary, the first ultraviolet stellar spectra
with sufficient resolution to detect emission lines were ob-
tained during a rocket flight in 1965. In the nine years from
1964 to 1972, NASA’s Orbiting Astronomical Observatories
(OAO’s) constituted the first programme of space facilities
designed specifically for ultraviolet astronomy. In particular,
the final one, OAO–C – launched in 1972 and designated
Copernicus–detected for the first time FUV emission lines
in two UV-bright cool stars, namely Procyon (αCMi) and
Capella (αAur). In parallel, the first European satellite car-
rying several UV experiments was TD–1, launched in 1972.
The technological efforts and the scientific needs of the
UV astronomical community crystallized in the joint NASA,
ESA and British SERC IUE project, launched in January
1978 and finally terminated in September 1996, after an ex-
traordinarily long and productive mission. The broad IUE
coverage (1100–3200 ˚
A), high spectral resolution (up to
Springer

Astrophys Space Sci (2006) 303:17–31 19
Fig. 1 The spectrum of αCen
A, a twin of the Sun, as observed
by HST/STIS. Only the strongest
transitions are labelled. This
spectrum contains a wealth of
emission lines that probe the
stellar atmosphere from the
chromosphere to the coronae
(from Pagano et al., 2004)
R=λ/λ =10000 in its echelle modes), and higher sen-
sitivity than previous UV missions provided numerous key
discoveries in many different areas of cool-star science, from
the atmospheres of T Tauri stars (Imhoff and Appenzeller,
1987), accretion processes in pre-main sequence stars, winds
in Herbig Ae/Be stars, chromospheres and transition regions
in all kinds of late-type stars (Jordan and Linsky, 1987), and
Doppler imaging of the components (and atmospheres) of ac-
tive binary systems, and stellar winds (Dupree and Reimers,
1987). Detailed summaries of the achievements in each of
these fields at the end of the IUE mission can be found in
the Proceedings of the Conference “Ultraviolet Astrophysics
beyond the IUE Final Archive” (Wamsteker and Gonz´alez-
Riestra, 1998).
Following IUE, three missions have made further and sub-
stantial contributions to UV astronomy of cool stars, namely
the Hubble Space Telescope (HST), the Extreme Ultraviolet
Explorer (EUVE) and the Far Ultraviolet Spectroscopic Ex-
plorer (FUSE).
Three UV instruments on board HST have made pivotal
discoveries in cool-star science. The Goddard High Res-
olution Spectrograph (GHRS) covered the interval 1150–
3200 ˚
A, with resolutions between 2000 and 80000. The com-
panion Faint Object Spectrograph (FOS) covered the broad
region from Lyαinto the visible, although at relatively low
resolution and with significant scattered light below 1900 ˚
A.
Nevertheless, FOS filled an important gap during the pe-
riod when GHRS was not able to use its own low-resolution
mode owing to an electrical failure (Ayres et al., 1996). The
second-generation Space Telescope Imaging Spectrograph
(STIS) covered the wavelength range 1150–10000 ˚
A, with
spectral resolving powers between 26 and 200 000. GHRS
and FOS were in operation from 1990 until 1997; STIS car-
ried on from then until a power supply failure in August
Springer

20 Astrophys Space Sci (2006) 303:17–31
2004. The HST UV instruments covered the same range as
IUE, but the much larger telescope aperture, higher spectral
resolution in some of the observing modes, more sensitive
digital cameras, and the ability to achieve high spatial res-
olution (in STIS long-slit modes) permitted observations of
cool stars many magnitudes fainter than the targets accessi-
ble in the pre-HSTera. A good example of the science added
by the HST instruments is a survey of FUV coronal for-
bidden lines carried out by Ayres et al. (2003a) in a sam-
ple of F-M dwarfs, giants and supergiants, detecting faint
lines from highly ionized species such as Fe XXI 1354 ˚
A,
formed at 107K. The UV coronal forbidden lines are di-
agnostically unique because they can be recorded at high
spectral resolution much more, easily than their permitted
counterparts in the X-ray region (where the current best
resolution is only about 1,500). Contributions in all other
UV fields, already opened by IUE were also remarkable and
valuable.
EUVE was launched in 1992 and operated until Jan-
uary 2001. It covered the extreme ultraviolet wavelengths
from 60 ˚
A up to the Lyman continuum (LyC) edge at 912 ˚
A
(although in practice few stars were observable longward of
∼400 ˚
A owing to interstellar extinction). Of the 734 sources
cataloged in EUVE’s all-sky survey, 55% were identified as
late-type stars. EUVE’s ability to resolve spectral lines from
a variety of high ionization stages was of tremendous value
for modeling the temperature structure of stellar coronae, and
many such studies of stars of different activity levels were
carried out. EUVE also enabled for the first time the study of
stellar coronal abundances and chemical fractionation phe-
nomena such as the so-called FlP-effect (elements with First
Ionization Potentials <10 eV are overabundant in the so-
lar corona compared with high-FIP counterparts, e.g. Osten
et al., 2003). Equally important, the long continuous stares
of days to weeks, motivated by operational considerations,
made EUVE an ideal platform to study coronal variability,
especially flare activity. An account of the achievements of
EUVE can be found in Bowyer et al. (2000).
FUSE, launched in 1999, records the 912–1187 ˚
A spec-
trum with four separate telescope/spectrograph channels
(further subdivided by redundant detector segments). The
FUSE range includes the strong Li-like O VI 1031,37 ˚
A reso-
nance doublet, formed at transition region (TR) temperatures
around 300 000 K, as well as the very strong C III 977 ˚
A reso-
nance line, formed near 60 000 K. Like STIS, FUSE provides
a link to coronal dynamics through the [Fe XVIII] 974 ˚
A and
[Fe XIX] 1118 ˚
A forbidden lines (cf. Redfield et al., 2003).
Molecular hydrogen absorption, and fluorescence of H2by
HLyα, add further diagnostics for determining plasma den-
sities, temperatures, and structural constraints in, for exam-
ple, circumstellar envelopes of Herbig Ae/Be and T Tauri
stars. Two important papers are the surveys of cool dwarfs
(Redfield et al., 2002) and giants (Dupree et al., 2005). Harper
(2004) and Ayres (2005) have reviewed recent FUSE results
on cool stars.
3. Open issues
In the remainder of the paper we focus in the remainder of
the paper on a few selected open issues in cool-star physics
that can be addressed by new observations in the ultraviolet
spectral range. These issues are, in our opinion, important
topics in stellar physics and key ingredients for progress in
the fields of “space weather,” “extrasolar planets,” and “life
beyond Earth.”
rStellar dynamos and the transport of magnetic energy in
plasmas;
rMagnetic activity of stars hosting planets;
rAstrospheres and solar-like stellar winds;
rActivity in young galactic clusters and star-forming re-
gions.
3.1. Stellar dynamos and the transport of
magnetic-energy in plasmas
In stars of spectral type early F and later, the coupling be-
tween differential rotation and turbulent convection in their
subphotospheric layers generates strong magnetic flux ropes,
by a mechanism known as “dynamo action”, first investi-
gated for the stellar case by Belvedere et al. (1980a,b,c). The
tubes buoyantly rise through the convection zone, penetrate
the stellar surface, become braided and twisted by surface
velocity fields, and ultimately reconnect releasing their mag-
netic free energy to power the nonclassical outer atmospheric
activity that is a focus of much solar/stellar research today.
Existing models of the dynamo are able to reproduce gross
features of the 11-year sunspot cycle, and the 22-year pe-
riod of the Sun’s global magnetic field reversals. However,
contemporary dynamo models cannot forecast the details of
solar activity on short time scales of months to years (cf.
review by Patern`o, 1998).
One of the challenges of dynamo theory today is to explain
the characteristic periods of the quasi-cyclic activity oscil-
lations of stars and, in particular, the problem of the large
observed ratio (>100) of cycle times and correlation times of
the turbulent eddies. While in an evolutionary scenario the
actual picture of the dynamo action ranges from an α2in fully
convective stars to α dynamo regime in solar-type stars, it
has been recognized that a necessary ingredient to explain
the observational features of magnetic activity in the Sun is
the inclusion of the meridional circulation together with a
small eddy diffusivity (Bonanno et al., 2005). In this case the
magnetic Reynolds number reaches 100–1000 and the dy-
namo action correctly reproduces all the observation(cycle
Springer

Astrophys Space Sci (2006) 303:17–31 21
times, butterfly diagram, sign of the current helicity, ratio of
toroidal/poloidal field strengths) with a differential rotation
profile as provided from helioseismoiogy. Another key issue
in understanding the “solar-stellar connection” in terms of
dynamo theory, and more generally the evolution between
α2regime and the α one, is the so-called “flip-flop” mech-
anism observed in some RS CVn’s and single young dwarfs
(see also Section 3.1.1).
Only in recent decades have we begun to appreciate the
effects of short-term solar variability on the Earth’s environ-
ment and human civilization. Major solar flares and CME’s
can cause electronic failures in commercial and scientific
satellites, disrupt long distance radio communications, and
induce ground currents than can overload electricity trans-
mission grids. Moreover, there is growing evidence that solar
activity can, to a certain extent, influence the Earth’s climate.
For example, historical records of sunspot counts show
that solar activity decreased for more than 50 years during
the 17th century when Northern Europe was experiencing the
“Little Ice Age.” On the other hand, a sustained increase of
activity, in a modern version of the “Grand Maximum” that
occurred during the 12th century, might cause climate warm-
ing and the increase in space storms and ultraviolet radiation,
which is harmful to life, particularly if protective ozone layers
continue to diminish as a consequence of anthropomorphic
emissions.
Butler (1994) and Pall´e Bag ´o and Butler (2001) have found
quantitative evidence that much of the warming of the past
century can be accounted for by the direct and indirect effects
of solar activity. However, at the moment we cannot fore-
cast long and short-term solar magnetic variability because
we do not have a comprehensive understanding of their root
causes: the dynamo mechanism and the physical processes
that shape plasma structure and control dynamics in the solar
outer atmosphere. One way to progress in this area is to step
back from the Sun, and instead evaluate how magnetic ac-
tivity phenomena depend on fundamental stellar parameters:
mass, rotation rate, chemical composition, binarity, and so
forth. The diverse stellar examples of activity might reveal
insights into the underlying physical processes that observa-
tions of the singular example of our Sun cannot, no matter
how detailed.
In what follows, potential new observational constraints
on the stellar dynamo and the mechanisms of coronal heating
are discussed.
3.1.1. Determining the locations and migration
patterns of active structures
Observational constraints on stellar dynamos can be obtained
by imaging the surface patterns of magnetic activity in large
samples of stars. We already know that on other stars the
mean locations of magnetic active regions and their migration
paths can be different from those observed on the Sun (Walter,
2003). In some cases, highly active stars show polar spots,
migration of active regions toward the poles, preferred active
longitudes, a “flip-flop” effect, i.e. sequential activation of
a pair of active longitudes (see, e.g., Rodon`o et al., 2000,
Berdyugina and Tuominen, 1998).
Much of the existing work on surface patterns of stellar
magnetic fields has involved optical observations of large
dark starspots, the sites of concentrated strong (kilogauss)
magnetic fields. However, there are important patterns of
smaller-scale fields on the Sun-plage and supergranulation –
that carry a large fraction of the global magnetic flux, and are
present even at cycle minimum when few or no spots are on
the disk. Just as the spots are associated with elevated chro-
mospheric and coronal emissions (in the surrounding active
region), so too are plage and the supergranulation network
areas, of enhanced chromospheric and transition region emis-
sions. For this reason, such areas show very large intensity
contrasts in the ultraviolet, even though they are not easily
distinguishable from the surrounding quiet photosphere at
visible wavelengths.
There are two basic approaches that one might utilize to
determine the migration patterns of active regions, plage, and
supergranulation in the chromospheric and transition-region
emissions of stars. The first is to obtain direct narrow-band
images (like TRACE does for the Sun) at very high spatial res-
olution; the second is to use high-resolution spectroscopy to
measure Doppler shifted signatures of active regions in time
series of emission line profiles (so-called Doppler Imaging).
HST, which has about 50 milliarcsecond resolution in the
mid-UV (Gilliland and Dupree, 1996), has successfully im-
aged in chromospheric light only star, αOri. Even though
the angular size of the chromosphere of a cool giant star can
significantly surpass that of the photosphere–the diameter of
Betelgeuse in Mg II 2800 ˚
A light is four times larger than its
optical size (Uitenbroeck et al., 1998) – the direct imaging of
stellar outer atmospheres is a daunting task. In fact, to obtain
a 100 ×100 pixel2UV map of the closest solar-like dwarf
stars (sufficient to record small active regions and crudely
image the supergranulation network) requires a spatial reso-
lution of about a tenth of a milliarcsecond, 500×better than
the best delivered by HST. Extending the sample to dwarfs
within 100 pc would require microarcsec imaging. Reaching
such an objective might seem to be far in the future. Neverthe-
less, one promising concept under study at the NASA GSFC
is the Stellar Imager mission, a kilometer scale interferom-
eter composed of ∼30 small telescopes formation-flying in
space (Carpenter et al., 2004).
Emission-Line Doppler Imaging, on the other hand, is a
powerful contemporary tool for probing high-contrast sur-
face structure or extended atmospheric zones in cool stars,
strongly complementing optical and X-ray imaging-which
preferentially maps the dark spotted areas, and the corona,
Springer

22 Astrophys Space Sci (2006) 303:17–31
respectively. The reader is referred to Lanza et al. (1998) and
Rice (2002) for the techniques to derive photospheric im-
ages, and to the reviews by (G¨udel, 2004) and (Favata and
Micela, 2003) for X-ray imaging techniques. Here we discuss
chromospheric and TR Doppler imaging methods. Discrete
features moving through the line profile, say, chromospherio
Mg II h and k: 2800 ˚
A or higher temperature C IV 1548 ˚
A, can
trace the locations of bright emission regions on the stellar
surface, and help us understand how the activity is organized
spatially and its relation to the dark starspots seen in optical
Doppler images. A few maps of the chromospheres of RS
CVn-type stars with rotation periods less than 3 days have
been obtained by monitoring the Mg II h and k lines with
IUE (Walter et al., 1987; Neff et al., 1989a; Bus`a et al., 1999;
Pagano et al., 2001). In this way, cool prominences have been
found at distances of 1-2 stellar radii, implying that multi-
temperature plasmas thread the circumstellar environments
of these systems. Spatially resolved surface fluxes in the chro-
mospheric plage maps suggested that these regions are sites
of “saturated heating” (Linsky, 1991) and that the relation-
ship between radiative and magnetic flux densities valid for
the Sun cannot be extrapolated to these extremely active stars
(Schrijver and Zwaan, 2000). A related phenomenon is that
of ”super-rotational” broadening seen in the TR lines of cer-
tain types of rapidly-rotating giant stars. The observed line
widths are up to twice that expected from the photospheric
vsin i, demonstrating that 100 000 K material is present at
levels up to 10 pressure scale heights above its equilibrium
altitude, raising questions as to how the gas got there and
how it can remain (Ayres et al., 1998).
From a practical point of view, Doppler imaging typically
requires spectral resolution ≥30 000 (<10 km s−1) to detect
discrete emission components migrating through the line pro-
file. An additional requirement is the ability to obtain a time
series of spectra with sufficient cadence to avoid smearing
the Doppler information, and long enough coverage to distin-
guish between persistent surface features and transient flare
activity. In fact, the best targets for emission-line Doppler
imaging are rapidly-rotating stars, and these by their nature
are highly active and flare frequently. For example, the single
attempt by HST to map a cool-star outer atmosphere, namely
HD 155555 (Dempsey et al., 2001), covered one stellar ro-
tation, and was only partially successful because flaring and
non phase-dependent variability could not be separated from
rotational modulation of unchanging active regions.
Another, more subtle, way to gain information on the small
scale geometrical organization of a stellar atmosphere is to
utilize fluorescent lines of molecules and atomic species. The
molecular features further are a guide to the presence of very
cool gas in the outer atmosphere, perhaps due to so-called
molecular cooling catastrophes (Ayres, 1981). Fluorescent
lines of molecules (e.g., H2and CO) and atomic species (e.g.,
Fe I&II,OI,ClI) are particularly strong in T Tauri stars
(owing to their disks) and in red giants (owing to the gen-
erally tenuous atmospheres and the presence of “clouds” of
very cool gas at high altitudes). For example, Herczeg et al.
(2002, 2004) identified 146 emission lines of H2pumped pri-
marily by Lyαin a STIS spectrum of the T Tauri star TW
Hya. The fluorescent processes rely on the excitation of spe-
cific low-excitation lines by wavelength-coincident radiation
fields produced in hotter, spatially separated regions of the
stellar (or disk) atmosphere. In the case of the red giants,
the dominant pumping lines are the resonance transitions
of atomic oxygen and hydrogen, and the hot chromospheric
layers in which these radiations form are too far removed
from the cooler photosphere to properly excite the observed
fluorescent transitions, at least if one considers traditional
1-D thermal models for such objects. The implication is that
hot and cold gas are much more intimately associated in the
outer atmospheres of these stars. It might be that magnetic
processes (which tend to produce filamentary hot structures
in an otherwise cool atmosphere) are more important in the
red giants, for example, than previously thought; and that,
in turn, might be an indication that the winds of such stars
could have a magnetic origin. In short, fluorescent processes
can be a guide to the geometrical organization of a stellar at-
mosphere (or accretion disk) on small physical length scales
not accessible to direct telescope observations.
In this connection, it should also be mentioned that studies
of the atmospheres of close-in exoplanets, through Doppler
shifted Lyαand other fluorescence emissions, could be fea-
sible with a sufficiently-sensitive next-generation UV spec-
trometer.
3.1.2. Velocity fields and plasma dynamics
The heart of understanding chromospheric and coronal heat-
ing – one of the major unsolved problems in solar and stel-
lar physics – is mechanical energy transport and dissipa-
tion. In the outer convection zone, immediately beneath the
photosphere, turbulence transforms gas kinetic energy into
sound waves and propagating electrodynamic disturbances.
Given the difficulty of transmitting sound waves through the
steep TR temperature gradient, electrodynamic processes are
thought to provide the bulk of coronal heating. However, the
shock dissipation of pure acoustic waves probably is an im-
portant heating source for nonmagnetic portions of the chro-
mosphere and the lower transition region. How sizeable this
contribution is with respect to the magnetic mechanisms is
not yet understood (c.f.,-Judge et al., 2004 and references
therein), and the quantitative details of the energy transport
and dissipation processes remain elusive. What is clear, how-
ever, is that plasma dynamics is an important byproduct of
the heating mechanisms, and thus a potential window into
their nature.
Springer

Astrophys Space Sci (2006) 303:17–31 23
Observations of the Sun (e.g., Teriaca et al., 1999 and ref-
erences therein) and late-type stars (e.g., Ayres et al., 1983,
1988, Wood et al., 1997, Pagano et al., 2004) have shown
that transition region emission lines are, on average, red-
shifted, and the that redshifts increase with increasing for-
mation temperature up to about 105K. This behaviour is not
anticipated by models of upward propagating acoustic waves,
for which both optically thin (e.g., Hansteen, 1993, Wikstøl
et al., 2000) and optically thick lines (Carlsson and Stein,
1997) are predicted to be blueshifted. On the other hand, sta-
tistically the observed solar redshifts are largest over active
regions (Brynildsen et al., 1996; Peter, 2000), compared with
quiet areas (Achour et al., 1995); and some TR models (Reale
et al., 1996) predict larger redshifts in regions permeated by
strong magnetic fields than in quiet regions.
The maximum redshift (∼15 km s−1) is reached at about
1−1.2×105K (active regions and quiet Sun, respectively).
At higher temperatures, the centroid velocities decrease,
crossing over to blueshifts at T ∼106K (reaching about
∼−10 km s−1). A similar behaviour is seen in αCen A
(Pagano et al., 2004), αCen B (Wood et al., 1997), Eri
(Jordan et al., 2001), and Procyon (Wood et al., 1996), al-
though the latter shows a maximum redshift at somewhat
higher temperatures (Wood et al., 1997). On the other hand,
the TR lines of the very active dM1e star AU Mic show hardly
any redshifts, and certainly show no conspicuous trend of
line shift versus formation temperature (Pagano et al., 2000;
Redfield et al., 2002).
In addition to redshifts, stellar transition region emission
lines also show a curious bimodal structure. AU Mic (dM1e)
was the first star for which this behaviour was noticed (Lin-
sky and Wood, 1994). The authors found that the Si IV and C
IV lines have broad wings superimposed on a narrower cen-
tral peak. The same bimodal structure of Si IV and C IV lines
subsequently was detected in TR spectra of other stars cover-
ing a range of spectral types, luminosity classes, and activity
levels, from RS CVn-type systems (e.g., HR 1099), to main
sequence dwarfs (e.g., AU Mic, Procyon, αCen A and B),
and giants (Capella, 31 Com, βCet, βDra, βGem, and AB
Dor) (Linsky et al., 1995; Pagano et al., 2000, 2004; Linsky
and Wood, 1994; Vilhu et al., 1998). The broad components
have widths comparable to, or wider than, the broadened C IV
profiles observed in solar transition-region explosive events
(see Dere et al., 1989), small bursts thought to be associated
with reconnections of emerging magnetic flux to overlying
pre-existing canopy fields. Wood et al. (1997) showed that
the narrow components can be produced by turbulent wave
dissipation or Alfv´en wave-heating mechanisms, while the
broad components – whose strength correlate with activity
indicators like the X-ray surface flux – can be interpreted
as a signature of “microflare” heating. Alternatively, for the
particular case of rapidly rotating AB Dor, Vilhu et al. (1998)
suggested that broad wings can arise from a ring of hot gas
at 2-3 stellar radii, like the “slingshot prominences” seen in
Hα.
Using SoHO/SUMER data, Peter(2001) showed that
broad components are a common feature in the thermal
regime 40 000K to 106K above the magnetically dominated
chromospheric network. The author presented evidence that
the narrow line core and broad wings are formed in radi-
cally different physical settings: small closed loops and coro-
nal funnels, respectively, the latter being the footpoints of
large coronal loops. Non-thermal widths of the broad com-
ponents follow a power-law distribution with respect to line-
formation temperature, a signature of upward propagating
magneto-acoustic waves (Peter, 2001). In stars other than
the Sun, the broad components have been observed over for-
mation temperatures of 40,000 K to 2 ×105K.
Usually, chromospheric emission lines do not show broad
wings in time-averaged spectra of normal stars, although very
optically thick, opacity-broadened features like the Mg II
h&k resonance lines would tend to mask such components if
present. However, in extremely active stars, or during large
stellar flares, chromospheric broad components occasionally
are seen. For example, transient broad redshifted components
of Mg II were recorded in the RS CVn binary AR Lac dur-
ing a large flare event early in the IUE mission by Neff et al.
(1989b). In another active binary, HR 1099, Bus `a et al. (1999)
suggested that persistent broad components observed in Mg
II h&k were associated with a large active region close to
the pole. Subsequently, Lanzafame et al. (2000) employed
a two-component NLTE model to synthesize the stellar Hα
and Mg II h&k lines, demonstrating that the broad compo-
nent could be explained as an effect of elevated densities
in the active region, while the narrow component would be
formed in the surrounding lower density quiet chromosphere.
This conceptually is a much different model than the dy-
namical origin proposed for the optically thin TR lines, but
still potentially is a valuable indicator of local plasma condi-
tions in active regions. Future observational and theoretical
studies are needed to determine which is the more realistic
model.
Extending the observational side of plasma dynamics to
higher, coronal temperatures is relatively straightforward on
the Sun, because many suitable strong coronal permitted
lines (e.g., Mg ×610,25 ˚
A) fall in the Lyman continuum
region immediately below 912 ˚
A, where high-resolution far-
UV spectroscopy still is practical. Unfortunately, these key
features are not accessible even in the nearest stars owing
to interstellar extinction. Observing spatially-resolved pro-
files of permitted coronal X-ray lines, say in the important
iron-L shell band at ∼1 keV, is not feasible at present, be-
cause contemporary high-energy missions like Chandra and
XMM-Newton have inadequate spectral resolution by an or-
der of magnitude, and there are no planned future missions
that will push that limit.
Springer

24 Astrophys Space Sci (2006) 303:17–31
Fig. 2 A very strong flare of the
RS CVn-type system HR 1099
observed in the Mg II h&k lines
by IUE. The dotted line is a
spectrum obtained during
quiescence. Arrows indicate the
Mg II 3d−3psubordinate lines
(from Bus`a et al., 1999)
However, a number of coronal forbidden lines in the UV
range long-ward of the Lyman continuum edge have been
observed in the Sun (see for example Doschek et al., 1975;
Feldman et al., 2000). On the stellar side, HSTs GHRS and
STIS instruments, and FUSE, have been able to detect highly
ionized iron forbidden lines in many cool stars (see for ex-
ample Maran et al., 1994, Pagano et al., 2000, Ayres et al.,
2003a, Redfield et al., 2003) with high spectral resolution
(up to R=40000).
The UV coronal forbidden lines detected in cool stars
for the most part show negligible Doppler shifts from the
photospheric radial velocities, suggesting that the emissions
arise mainly from confined structures, analogous to magnetic
loops on the Sun, rather than from a hot wind. Moreover, the
Fe XII and Fe XXI line widths generally are close to their
thermal values (FWHM 40-90 km s−1at T ∼106.2−107.0
K), except for the Hertzsprung-gap giants 31 Com (GO III)
and Capella (G1 III) and the K1 IV component of HR 1099,
all of which display excess broadening in Fe XXI. If the
additional broadening is rotational, it would imply that the
hot coronae of “X-ray deficient” 31 Com and Capella are
highly extended, compared to the compact structures sug-
gested by recent density estimates in a number of active
coronal sources. On the other hand, the more common case
of purely thermal line widths implies that supersonic turbu-
lent motions are absent in the coronal plasma, eliminating
shock waves as an important heating mechanism.
3.1.3. The physics of impulsive heating: Stellar flares
and microflares
Flares, lasting from a few minutes to several days, are the
most dramatic examples of transient energy release in so-
lar and late-type stellar atmospheres. A magnetic reconnec-
tion process is thought to power these events: magnetic free
energy is converted – in thin current sheets – into thermal
heating, Alfv´en waves, and the acceleration of relativistic
particles. Accordingly, the relaxation phenomenon is com-
plex and fast, producing radiation across the whole electro-
magnetic spectrum from nonthermal radio synchrotron, to
thermal emissions in the UV and X-ray ranges. In fact, the
flare phenomenon involves the whole atmosphere, from the
corona down to the lower chromosphere and photosphere,
which are blasted by hard radiations and particle beams from
the high-altitude flare kernel (Haisch et al., 1991). The in-
vestigation of detailed flare physics historically has relied on
multiwavelength simultaneous observations, involving both
spectroscopy and photometry.
Ultraviolet emission lines provide an important source of
plasma cooling, as well as lower-atmospheric heating, dur-
ing the gradual phase of stellar flares (e.g., Hawley et al.,
2003), and thus are valuable diagnostics of the flare evolu-
tion. Figure 2 shows Mg II h & k line profiles from a flare
on the RS CVn-type binary HR 1099 observed by IUE.At
the flare peak, the Mg II lines display very broad wings, and
the mid-UV continuum is enhanced. Bus`a et al. (1999) anal-
ysed this flare and concluded that material was ejected by the
secondary star in the direction of the primary.
Although large flares are a conspicuous contributor to
transient heating of stellar coronae, smaller scale events–
so-called micro- and nano-flares–might play a key role in the
“steady” heating of the outer layers of cool stars. Robinson
et al. (2001) studied the statistics of transient bursts in high
time resolution UV observations of AU Mic with HST STIS,
and concluded that the power-law slope of the occurrence
rate versus time-integrated flux was considerably steeper for
Springer

Astrophys Space Sci (2006) 303:17–31 25
low-energy flares than for the rarer high-energy ones; imply-
ing that microflares potentially can account for a significant
portion, if not all, of the coronal heating (see, e.g., Hudson,
1991 and G¨udel, 2003). However, this is such an important
issue that more conclusive evidence is needed: the investiga-
tion should be extended to a larger sample of flare stars with
sufficient time resolution and sensitivity to collect a signifi-
cant number of events for each target.
3.1.4. UV emissions from very late M dwarfs and brown
dwarfs
Magnetic activity decreases for spectral types later than ap-
proximately M7. However, flares have been observed in the
radio, optical, UV and X-rays from very late dM stars and
brown dwarfs (e.g., Linsky et al., 1995, Rutledge et al., 2000).
Several authors have suggested that hot gas in the atmo-
spheres of very low mass main-sequence stars and brown
dwarfs is present only during flares (Fleming et al., 2000,
Rutledge et al., 2000, Mohanty et al., 2002, Berger et al.,
2001, Berger, 2001). However, by recording UV spectra of a
sample of these stars, Hawley & Johns-Krull (2003) showed
that persistent quiescent chromospheres and transition re-
gions, similar to those observed in earlier type dMe’s, are
present at least through spectral type M9. The existence of
persistent magnetic activity in these fully convective stars
poses challenges for contemporary α-type dynamo mod-
els that require a shear layer between the convective outer
envelope and the radiative interior. The current thinking is
that a “distributed dynamo” might be in play, one that oper-
ates directly on convective turbulence and does not require
the catalyzing agency of differential rotation. Incidentally,
this same “α2” dynamo process probably also is operating
in the slowly rotating red giants, to account for the feable,
but nonetheless present, coronal activity of the inhabitants of
the so-called “coronal graveyard” (Ayres et al., 2003b), i.e.
the region of the HR diagram near K1 III where stars were
rarely detected in coronal proxy C IV (T ∼105K) lines and
in X-ray surveys.
Extending this work–transient vs. persistent hot plasma–
to the even lower mass range, and much cooler atmospheres,
of the brown dwarfs will require significantly more sensitive
UV spectroscopy than was available, for example, from HST
STIS.
3.1.5. Composition anomalies in stellar outer
atmospheres
Observations of solar and stellar coronae and of the solar
wind provide evidence for plasma chemical composition
anomalies with respect to the photosphere. The most famous
of these is the so-called First Ion-ization Potential (FIP) ef-
fect: elements with low FIP (I.P.<10 eV) are overabundant
with respect to high-FIP (>10 eV) species, relative to photo-
spheric ratios. This is the case, for instance, in closed mag-
netic flux regions on the Sun and in the slow solar wind (see
Feldman and Laming (2000) for a review). On stars, a variety
of behaviours are found, including the absence of any low-
FIP bias, normal low-FIP enhancements, and even low-FIP
abundance depletions (the so-called Metal Abundance De-
pletion [MAD] syndrome) seen in very active stars (Audard
et al., 2003; G¨udel, 2004).
The FIP chemical fractionation effect is thought to be a
byproduct of the processes that heat and structure the chro-
mosphere. For example, MHD waves passing through an am-
bient plasma with a temperature below about 7000 K, where
low-FIP elements are the only ones ionized, could selectively
sweep these ions into the corona, thereby boosting the local
abundance. An important goal for future work would be to
link stellar chromospheric spatial structure, as derived from
UV Doppler images, to the coronal abundance patterns de-
duced from rotational modulations of suitable coronal X-ray
lines, to constrain models of the FIP effect and its MAD
syndrome cousin.
Recently, Laming (2004) has proposed models for abun-
dance anomalies in the coronae of the sun and other late-
type stars following a scenario first introduced by Schwadron
et al. (1999). According to these models, the abundance
anomalies are produced by the ponder-motive (proportional
to the j×B) force on ions arising as Alfv´en waves propa-
gate through the chromosphere and depends sensitively on
the chromospheric wave energy density. The model can ex-
plain both the solar FIP effect and its variations as well as
the inverse FIP effect observed in some stars. A better under-
standing of the coronal abundance anomalies may therefore
offer a unique diagnostic of Alfv´en wave propagation be-
tween the chromosphere and corona.
3.2. Magnetic activity of stars hosting planets
More than 130 extrasolar planets have been detected so far,
mainly around solar-type stars. Most of these planets are
massive Jovian-types, observationally favored by the cur-
rently popular Doppler-reflex observing techniques. How-
ever, space missions like COROT (Baglin, 2003), Kepler
(Borucki et al., 2003), and Darwin (Fridlund, 2004) will
discover–and eventually allow us to characterize–Earth-sized
extra-solar planets.
How a planet might directly interact with its parent star is
a new field of research, motivated by the tight orbits of some
of the extreme “roasters.” Cuntz et al. (2000) predicted that a
giant planet orbiting close-by a star incites increased stellar
activity by means of tidal and magnetospheric interactions.
Experimental data supporting this prediction were reported
by Shkolnik et al. (2001, 2005), who found the strength of
the emission reversals in the Ca II H&K lines of a couple of
Springer

26 Astrophys Space Sci (2006) 303:17–31
stars hosting hot jupiters to be variable with the planet orbital
period. The effects of the nearby planet should be exaggerated
in the upper chromosphere, transition region, and corona;
thus UV, and EUV, spectroscopy will certainly contribute
importantly to exploring such planet-star Interactions.
The other side of the story is to understand how stellar
activity affects planets, especially habitable ones. The evolu-
tion of a planetary atmosphere under the joint erosive impacts
of coronal ionizing radiations and wind from the parent star
(e.g., Ayres, 1997), is an exciting scientific issue, with impor-
tant ramifications for understanding the evolution of Earth’s
paleoclimate and the birth of life (see Lammer et al., 2003).
UV spectroscopy of the parent stars of planetary systems
is feasible with a 2-m class telescope for most of the relatively
nearby stars in contemporary planet searches by the Doppler
reflex technique, as shown by HST. In fact, by using STIS
on HST, Vidal-Madjar et al. (2003) found ∼15 ±4% extra-
absorption in the Ly-αemission line of HD 209458 (V =7.6)
during the transit of its hot-jupiter planet, which can be un-
derstood in terms of escaping hydrogen atoms from the planet
atmosphere. Again looking at the UV stellar spectrum during
the planetary transit, Vidal-Madjar et al. (2004) detected C
and O in the atmosphere of HD 209458b. However, a more
sensitive 4–6-m class UV facility would be needed to reach
the next tier of fainter planet-stars that will be discovered in
future surveys by COROT and Kepler exploiting the transit
method.
In terms of detecting life on other worlds, we mention that
there are useful biomarkers in the UV: for example ozone,
O2,H
2,CO
+,CH
4.G´omez de Castro et al. in this volume
has a more extensive description of these biomarkers, so we
will not discuss them further.
3.3. Astrospheres and solar-like stellar winds
Winds of late-type stars are a byproduct of magnetic coronal
activity, certainly in dwarf stars like the Sun, but perhaps
also, to some extent, in red giants (e.g., Ayres et al., 2003b).
The winds of low-mass cool stars feed back on their coronal
evolution owing to the significant angular momentum carried
away by a fast, magnetized outflow. In the evolved red giants,
winds can potentially change the nuclear evolution of the
star by removing significant mass from the surface layers
(see e.g. Rauscher et al., 2002). Furthermore, as mentioned
previously, coronal winds can erode volatiles from primitive
planetary atmospheres by sweeping up charged molecules
and ions from exospheric regions ionized by the stellar UV
radiation field. Hence, exploring the winds of low-mass and
evolved stars is of fundamental importance from a number
of standpoints.
The velocity structure, mass loss rate, and ionization of
the outflowing gas in the low-temperature (∼104K) winds
of late-type giants and supergiants can be inferred from high
resolution spectra of optically thick UV resonance lines (see
e.g. Dupree et al., 2004, Young et al., 2005, Dupree et al.,
2005). The UV provides unique access to the dominant cir-
cumstellar wind absorption species in the red giants: H I,O
I,MgII, and C II. Unfortunately, the winds of main sequence
stars like the Sun are too hot and too thin to provide detectable
UV or X-ray absorption (or emission) signatures, yet under-
standing these coronal outflows is of paramount importance
to the wide range of issues described above. Fortunately,
however, in a few favorable cases of nearby stars, coronal
winds can be studied by subtle distortions of the H ILyαline
due to the “hydrogen walls” produced by the interaction of
the stellar outflow with inflowing interstellar gas (cf. Wood
et al., 2001, 2005a). In the stellar “astrosphere”, the stellar
analog of the solar “ heliosphere”, decelerated interstellar
hydrogen (in the stellar rest frame) produces an absorption
feature on the blue side of the interstellar absorption (seen in
the observer’s rest frame), as shown in Figure 3. By model-
ing these absorption features, Wood et al. (2002) performed
the first quantitative measurements of mass loss rates for G
and K dwarf stars. These authors found that the mass loss
rates increase with activity, as measured by the X-ray sur-
face flux, and thus with decreasing stellar age. By extending
their investigation on a slightly large sample of dwarfs, Wood
et al. (2005b) found evidence that winds suddenly weaken
at a certain activity threshold. These authors suggest that
in very active stars, winds may be inhibited by the strong
magnetic fields associated with stellar spots that have a large
filling factor on these stars. The sample of dwarfs for which
an astrosphere has been detected thanks to high resolution
profiles of their H ILyαlines consists of 14 stars only. While
it is necessary to consistently enlarge the number of studied
stars in order to sample with a great significance each ac-
tivity/age stage, the mass-loss/age relation inferred by Wood
et al. (2005b) gives us an empirical estimate of the history
of the solar wind. By implication, the solar wind might have
been 1000 times stronger when the Sun was very young, and
thereby likely played a major role in the evolution of plane-
tary atmospheres, particularly the stripping of volatiles from
primitive Mars.
3.4. Activity in young stellar clusters
The study of late-type stars in young (∼50–100 Myr) galactic
clusters, and even younger star-forming regions (1–10 Myr),
is a valuable window into not only the evolution of mag-
netic activity, but also its basic characteristics. In particular,
within a given cluster or other coeval group of objects, the
relationships between, for example, activity indicators and
rotation periods for stars of different spectral types will not
suffer a hidden age bias due to the common age and chemical
composition of the stars.
Springer

Astrophys Space Sci (2006) 303:17–31 27
Fig. 3 Comparison between the
Lyαspectra of αCen B
(grey-tone histogram) and
Proxima Cen (black-tone
histogram). The inferred ISM
absorption is shown as a
grey-tone dashed line. The α
Cen/Proxima Cen data agree
well on the red side of the H
I.absorption, but on the blue side
the Proxima Cen data do not
show the excess absorption seen
toward αCen (i.e., the
astrospheric absorption). The
blue-side excess Lyαabsorption
is fitted by a model of the αCen
astrosphere, corresponding to a
mass loss of ˙
M=4×10−14 M
y−1(adapted from Wood et al.,
2001)
It is generally accepted that late-type stars undergo
spin-down due to magnetic braking throughout their main-
sequence phase, especially near the beginning when their
angular momentum is highest thanks to spin-up by their na-
tal disks (see, e.g., Jianke and Collier Cameron, 1993; Collier
Cameron and Jianke, 1994 and references therein). Studies
of the “Sun in time” (see Ribas et al., 2005; Guinan et al.,
2003, Ayres, 1997, and references therein) have made use
of observations of stars with different rotation periods and
ages, but spectral types similar to the Sun to build a scenario
of what might have happened to the Sun at different stages
along its evolutionary path. These several efforts have used
observations of activity tracers in the optical, ultraviolet and
X-rays. In particular, in the ultraviolet domain IUE spectra
were intensively exploited to extract information on chro-
mospheric and transition region fluxes; while more recently
FUSE observations were used to study the outer atmosphere
(Guinan et al., 2003): emission features in the FUSE 920–
1180 ˚
A band probe hot plasma over three decades in temper-
ature: from ∼104K for the H ILyman series to ∼6×106K
for the coronal Fe XVIII 974 ˚
A line.
The general activity-rotation-age relationship in stars also
has been the subject of many studies. In one of the first efforts
(Simon et al., 1985) utilized IUE data of a moderate size
sample of solar-type field stars, compared with observations
of T Tauri stars, in order to determine whether the pattern of
main-sequence chromospheric decay shown by stars older
than about 100 Myr extended back to earlier times. They
showed that the activity-age relation for main sequence stars
older than about 100 Myr can be modeled by an exponential
law whose rate of decline depends on surface temperature.
Unfortunately, few subsequent extensive studies in this
direction have been carried out using the much more sen-
sitive contemporary instruments of HST or other missions.
Ayres (1999) obtained GHRS spectra (1150–1670 ˚
A) of three
solar-type dwarfs in the young clusters αPer (85 Myr) and
the Pleiades (125 Myr) to complement an earlier FOS study
of 10 cluster stars, including 5 in the Hyades (625 Myr);
with the aim to investigate the behavior of the stellar activity
and the dynamo at early ages (Ayres, 1997). Although the
study drew important conclusions on the activity-age rela-
tionship for TR lines like C IV, these referred specifically
to early G-type dwarfs, and generalization to other spectral
types and luminosity classes was not possible. The lack of
modern UV studies of the age-activity relation contrasts with
the very extensive surveys accomplished in the optical (see
e.g. Soderblom et al., 2001) and X-ray domains. In the latter,
for example, Pizzolato et al. (2003) investigated the relation-
ship between coronal X-ray emission and stellar rotation in
a sample of 259 dwarfs in the B–V range 0.5–2.0 observed
with ROSAT, including 110 field stars and 149 members of
the Pleiades, Hyades, αPer, IC 2602 (30 Myr) and IC 2391
(30 Myr) open clusters. The “missing” UV part of the puz-
zle is extremely important to problems such as the radiative
erosion of primitive planetary atmospheres by young hyper-
active parent stars because the dominant ionizing radiations
for abundant, atmospheric molecules like N2fall in the Ly-
man continuum (see Ayres, 1997); the crucial bright O v
629 ˚
A feature, for example, cannot be observed directly in
stars, but its strength can be inferred through measurements
of O IV,OV, and O VI UV lines longward of the 912 ˚
AHI
edge.
At ages less than about 107years, pre-main sequence T
Tauri stars are located in star-forming regions. The UV is ex-
tremely valuable in studies of the less-obscured of the clas-
sical and “naked” T Tauri stars. For example, Herczeg et al.
Springer

28 Astrophys Space Sci (2006) 303:17–31
Table 1 Instrumental requirements for research on cool stars in the UV domain
Telescope Spectroscopy Spectral Spectral Spatial Time
Issue class or Imaging resolution resolution monitoring resolution
Doppler imaging 2 m S ≥35,000 YES minutes
of field stars
Direct imaging of Interferometry 1 ∼50–100 µas YES minutes
chromospheres (e.g. 20 ×1m)
and TRs
Plasma dynamics in 2 m S ≥35000 YES minutes
chromosphere,
TR and corona
Stellar activity of 2 m S ≥35000 YES minutes
extrasolar planets’ 4–6 ma
parent stars
Astrospheres 2 m S ≥100000 YES minutes
solar like 4–6 ma
stellar winds
Flares 2 m S >10000 YES seconds
aTargets that will be explored by COROT,Kepler, and Darwin
(2004) has studied the fluorescent excitation of the Lyman
bands of H2lines by H ILyαto probe the physical conditions
in the accretion disk. UV signatures of the hot splashdown
point of the accretion stream can be exploited to estimate the
accretion rate and geometry (Johns-Krull et al., 2000; Calvet
et al., 2004). Furthermore, these very young objects, not sur-
prisingly, are hyperactive in terms of the usual UV and X-ray
indicators, and thus provide a laboratory for “magnetic ac-
tivity in extreme environments”. The dissection of the highly
complex TRs and coronae of these quite exotic objects is an
important challenge for future emission-line Doppler imag-
ing efforts.
4. Instrument requirements
Progress in exploring fundamental cool-star issues such as
the dynamo and coronal heating benefits tremendously from
high resolution UV spectroscopy. At a minimum, a resolu-
tion of R=30,000 is sufficient to resolve most of the narrow
chromospheric emission lines seen in dwarf stars, which typi-
cally have FWHM >10 km/s. The hotter TR lines are broader
owing to higher thermal velocities, but they also benefit from
good resolution for deblending purposes, dynamical studies,
and Doppler imaging. In addition to resolution, high sensi-
tivity is needed to reach the faint, interesting objects beyond
the solar neighborhood, out to at least 150 pc to include, for
example, the important young galactic clusters a Per and the
Pleiades, as well as the key TW Hya star-forming region.
For studies of cool winds and astrospheres, higher resolu-
tions are needed: R=100,000 has proved crucial in previous
GHRS and STIS work, and certain ISM-oriented problems
probably could benefit from R=200,000, or more. Here,
however, sensitivity is less of an issue, because most of the
key objects for wind studies are nearby bright red giants (or
bright hot stars for ISM work). For some sources the lack
of short-term variability (particulary in the ISM background
“light sources”) means that one can integrate for long periods
to build up sufficient signal-to-noise in the high-resolution
line profiles.
It also would be important to have a high sensitivity mode
at lower resolution, to reach the next tier of interesting ob-
jects beyond the horizon accessible to the R=30,000 mode
and to study flares and other types of variability. Histori-
cally (e.g., for GHRS or STIS) this would be a R=1000
resolution mode, but we would argue for at least 35000 (i.e.,
velocity resolution better than 100 km s−1) because then one
can have cleaner separation of close lines, some velocity dis-
crimination in short-period binary systems, and the possibil-
ity of measuring hypersonic dynamics in large flare events.
Furthermore, broad spectral coverage in the low-resolution
and mid-resolution modes is essential, from the viewpoint of
observing efficiency.
Together with point source spectroscopy, a long-slit imag-
ing capability, in at least the low resolution mode (R=
35000) and possibly also the mid-resolution mode (R=
35000), would be especially valuable in studies of close bi-
naries and the gaseous environments of PMS stars. In fact, a
simple direct imaging system with suitable narrow-band fil-
ters tuned to important TR lines like C IV could be exploited to
monitor flares and rotational modulations of active regions
in dozens of late-type stars at the same time, for example,
members of a compact galactic cluster or PMS candidates
in a star-forming region. A beam splitter could divert visible
light into a simple prism system that would recorde a low-res
spectral energy distribution as well as key emission features
Springer

Astrophys Space Sci (2006) 303:17–31 29
such as Ca II or Hα. In this regard, the “Optical Monitor”
on XMM-Newton has proven extremely valuable, although
in the cool-star application one would want an optical sys-
tem with a large dynamic range capable of recording bright
nearby stars as well as much fainter distant objects.
In summary, research on cool stars would benefit enor-
mously from a new 2 m class orbiting telescope feeding a
high sensitivity, high resolution UV spectrometer, with an
auxilliary strap-on optical/UV monitoring system. The tele-
scope should be placed far from the Earth in a drift-away,
L2, or Molniya-type high orbit suitable for long-duration
uninterrupted observations of targts to obtain high quality
Doppler maps, flare light curves and statistics, and for op-
erational flexibility to allow rapid response to targets of op-
portunity. High wavelength accuracy is required by many
of the topics we would address, hence a Pt-Ne calibration
system–missing from FUSE, but which has been incredibly
valuable for GHRS and STIS–is an essential component of
such a facility.
A larger aperture telescope, of 4–6 m class, would allow
the study of plasma dynamics and chromospheric/TR struc-
tures of fainter magnetic active stars, brown dwarfs, and cool
stars in more distant stellar clusters and star-forming clouds
beyond 150 pc: This more sensitive facility would also be
able to characterize the outer atmospheres of parent stars
of extrasolar planets that will be discovered by future space
missions like Corot, Kepler or Darwin.
A summary of instrumental requirements discussed above
is given in Table 1.
5. Conclusions
The term “stellar activity”, as applied to the Sun and other
late-type stars, includes a very broad range of phenomena,
such as starspots, plages, active regions, prominences, flares,
coronal loops, winds, mass ejections, and so forth. These
phenomena are observed directly on the Sun, and their ex-
istence on other stars is inferred by solar analogy from spe-
cific spectral properties or behaviors (following the so-called
”solar-stellar” connection).
The driving mechanism for stellar activity is the magnetic
field generated presumably by a dynamo process at the base
of or embedded in the convection zone. The activity that we
see in the outer atmospheres of late-type stars thus has its
roots deep within the star. This activity is important in its
own right, and also because of the influences of its ioniz-
ing radiations and wind on the heliosphere, or its extrasolar
equivalent; as a model of magnetodynamic processes oper-
ating broadly in the cosmos; and as a window into the deep
interior of the star.
Since in the near future, direct surface imaging of the mor-
phology connected with activity will not be feasible for other
stars, the key information must be extracted from remote-
sensing spectroscopy. The importance of the ultraviolet part
of the electromagnetic spectrum in this regard has been em-
phasized: cool-star magnetic activity produces hot material
above the stellar photosphere, from the chromosphere to the
million degree, or hotter, corona. There are very few lines
in the optical or infrared that are sensitive even to the cooler
end of this range. On the other hand, emission lines formed
at temperatures up to 107K are accessible in the UV long-
ward of the LyC edge, and achieving high spectral resolution
and large effective areas is much more feasible in this wave-
length domain than in soft X-rays. From a practical point of
view, cool stars usually have negligible photospheric emis-
sion below ∼1900 ˚
A, so the UV emissions from magnetically
disturbed hot gas can be recorded with high contrast.
We can confidently say that ultraviolet astronomy is abso-
lutely fundamental for understanding cool-star physics, and
further progress in this area absolutely requires a new gener-
ation of UV observatories in space.
Acknowledgements IP thanks EU for supporting the Network for Ul-
traviolet Astronomy (NUVA) within the OPTICON program funded in
the contest of its 6th Framwork Program. The work of BM has been
supported in part by the Spanish grant AYA2001-1124-C02, and TRA
by NASA grant NAG5-13058. JLL thanks NASA for support through
grant AR-09930 to the University of Colorado.
References
Achour, H., Brekke, P., Kjeldseth-Moe, O., Maltby, P.: ApJ 453, 945
(1995)
Audard, M., G¨udel, M., Sres, A., Rassen, A.J.J., Mewe, R.: A&A 398,
1137 (2003)
Audard, M., Telleschi, A., G¨udel, M., Skinner, S.L., Pallavicini, R.,
Mitra-Kraev, U.: ApJ 617, 531 (2004)
Ayres, T.R.: ApJ 244, 1064 (1981)
Ayres, T.R.: JGR (Planets) 102, 1641 (1997)
Ayres, T.R.: ApJ 525, 240 (1999)
Ayres, T.R.: Astrophysics in the far-ultraviolet: five years of discovery
with FUSE, in: Sonneborn, G., Moos, W., Andersson, B.-G. (eds.),
ASP Conference Series, in press (2005)
Ayres, T.R., Stencel, R.E., Linsky, J.L., Simon, T., Jordan, C., Brown,
A., Engvold, O.: ApJ 274, 801 (1983)
Ayres, T.R., Jensen, E., Engvold, O.: ApJS 66, 51 (1988)
Ayres, T.R, Simon, T., Stauffer, J.R., Stern, R.A., Pye, J.P., Brown, A.:
ApJ 473, 279 (1996)
Ayres, T.R, Simon, T., Stern, R.A., Drake, S.A., Wood, B.E., Brown,
A.: ApJ 496, 428 (1998)
Ayres, T.R, Brown, A., Harper, G.M., Osten, R.A., Linsky, J.L., Wood,
B.E., Redfield, S.: ApJ 583, 963 (2003a)
Ayres, T.R, Brown, A., Harper, G.M.: ApJ 598, 610 (2003b)
Baglin, A.: AdSpR 31, 345 (2003)
Baliunas, S.L., Donahue, R.A., Soon, W.H., Horne, J.H., Frazer, J., et
al.: ApJ 438, 269 (1995)
Berger, E. et al.: Nature 410, 338 (2001)
Boggess, A., Wilson, R.: Exploring the Universe with the IUE satellite,
in: Kondo, Y. (ed.), Kluwer Academic Publishers, Dordrecht, p. 3
(1987)
Springer

30 Astrophys Space Sci (2006) 303:17–31
Belvedere, G., Patern`o, L., Stix, M.: A&A 86, 40 (1980a)
Belvedere, G., Patern`o, L., Stix, M.: A&A 88, 240 (1980b)
Belvedere, G., Patern`o, L., Stix, M.: A&A 91, 328 (1980c)
Berdyugina, S.V., Tuominen, I.: A&A 336, L25 (1998)
Bonanno, A., Elstner, D., Belvedere, G., R¨udiger, G.: AN 326, 170
(2005)
Borucki, W.J. et al.: SPIE 4854, 129 (2003)
Bowyer, S., Drake, J.J., Vennes, S.: ARA&A 38, 231 (2000)
Brynildsen, N., Kjeldseth-Moe, O., Maltby, P.: ApJ 462, 534 (1996)
Bus`a, I., Pagano, I., Rodon`o, M., Neff, J.E., Lanzafame, A.C.: A&A
350, 571 (1999)
Butler, C.J.: Sol. Phys. 152, 35 (1994)
Calvet, N., Muzerolle, J., Brice˜no, C., Hern `andez, J., Hartmann, L.,
Saucedo, J.L.: AJ 128, 1294 (2004)
Carlsson, M., Stein, R.F.: ApJ 481, 500 (1997)
Carpenter, K.G., Schrijver, C.J., Allen, R.J. et al.: SPIE astronomical
telescopes and instrumentation, SPIE Paper #5491–28 (2004)
Catalano, S., Rodon`o, M., Cutispoto, G., Frasca, A., Marilli, E. et al.: in:
Ibanoglu, C. (ed.), Variable stars as essential astrophysical tools,
NATO ARW, 544, p. 687 (2000)
Chugainov, P.F.: Comm. 27 IAU, Inf. Bull. Var. Stars 122 (1966)
Collier Cameron, A., Jianke, L.: MNRAS 269, 1099 (1994)
Cuntz, M., Saar, S.H., Musielak, Z.E.: ApJ 533, L151 (2000)
Cutispoto, G., Messina, S., Rodon`o, M.: A&A 367, 910 (2001)
Cutispoto, G., Messina, S., Rodon`o, M.: A&A 400, 659 (2003)
Dempsey, R.C., Neff, J. E., Lim, J.: AJ 122, 332 (2001)
Dere, K.P., Bartoe, J.-D.F., Brueckner, G.E.: Sol. Phys. 123, 41 (1989)
Doschek, G.A. et al.: ApJ 196, L83 (1975)
Dupree, A.K., Reimers, D.: “Exploring the Universe with the IUE satel-
lite”, in: Kondo, Y. (ed.), Kluwer Academic Publishers, Dordrecht,
p. 321 (1987)
Dupree, A.K.: IAU Symp., 219, 623 (2004)
Dupree, A.K., Lobel, A., Young, P.R., Ake, T.B., Linsky, J.L., Redfield,
S.: ApJ 622, 629 (2005)
Favata, F., Micela, G.: SSRv. 108, 577 (2003)
Favata, F., Micela, G., Baliunas, S.L., Schmidtt, J.H.M.M., G ¨udel, M.
et al.: A&A 418, L13 (2004)
Feldman, U., Laming, J.M.: Phys. Scr. 61, 222 (2000)
Feldman, U., Curdt, W., Landi, E., Wilhelm, K.: ApJ 544, 508 (2000)
Fleming, T.A., Giampapa, M.S., Schmitt, J.H.M.M.: ApJ 533, 372
(2000)
Fridlund, C.V.M.: AdSpR 34, 613 (2004)
Gilliland, R.L., Dupree, A.K.: ApJ 463, L29 (1996)
Godoli, G.: in: Oehman, Y. (ed.), Mass motion in solar flares and related
phenomena, 9th Nobel Symp., p. 211 (1968)
G¨udel, M.: A&ARv 12, 71 (2004)
G¨udel, M., Audard, M., Kashyap, V.L., Drake, J.J., Guinan, E.F.: ApJ
582, 423 (2003)
Guinan, E.F., Ribas, I., Harper, G.M.: ApJ 594, 561 (2003)
Haisch, B., Strong, K.T., Rodon`o, M.: ARA & A 29, 275 (1991)
Hansteen, V.: ApJ 402, 741 (1993)
Harper, G.M.: in: Favata, F. et al. (eds.), Proc. of the 13th Cambridge
Workshop on Cool Stars, Stellar Systems and the Sun. ESA SP-
560, p. 51 (2004)
Hawley, S.L. et al.: ApJ 597, 535 (2003)
Hawley, S.L., Johns-Krull, C.M.: ApJ 588, L112 (2003)
Herczeg, G.J., Linsky, J.L., Valenti, J.A., Johns-Krull, C.M., Wood, B.
E.: ApJ 572, 310 (2002)
Herczeg, G.J., Wood, B.E., Linsky, J.L., Valenti, J.A., Johns-Krull,
C.M.: ApJ 607, 369 (2004)
Hudson, H.S.: Sol. Phys. 133, 357 (1991)
Imhoff, C.L., Appenzeller, I.: Kondo, Y. (ed.), Exploring the Universe
with the IUE Satellite, Kluwer Academic Publishers, Dordrecht,
p. 295 (1987)
Jianke, L., Collier Cameron, A.: MNRAS 261, 766 (1993)
Johns-Krull, C.M., Valenti, J.A., Linsky, J.L.: ApJ 539, 815 (2000)
Jordan, C., Linsky, J.L.: Exploring the Universe with the IUE Satellite,
Kondo, Y. (ed.), Kluwer Academic Publishers, Dordrecht, p. 259
(1987)
Jordan, C., Sim S.A., McMurry, A.D., Aruvel, M.: MNRAS 326, 303
(2001)
Judge, P.G., Saah, S.H., Carlsson, M., Ayres, T.R.: ApJ, 609, 392 (2004)
Laming, J.M.: ApJ 614, 1063 (2004)
Lammer, H., Selsis, F., Ribas, I., Guinan, E. F., Bauer, S.J., Weiss, W.W.:
ApJ, 598, L121 (2003)
Lanza, A.F., Catalano, S., Cutispoto, G., Pagano, I., Rodon`o, M.: A&A
332, 179 (1998)
Lanza, A.F., Catalano, S., Rodon`o, M., Ibanoglu, Tas, G. et al.: A&A
386, 583 (2002)
Lanzafame, A.C., Bus`a, I., Rodon `o, M.: A&A 362, 683 (2000)
Linsky,J.L.: Mechanisms of Chromospheric and Coronal Heating, Ulm-
schneider, P., Priest, E.R., Rosner, R. (eds.), (Berlin: Springer-
Verlag), 166 (1991)
Linsky, J.L.: Adv. Space Res. 32, 917 (2003)
Linsky, J.L., Wood, B.E., Judge P., Brown, A., Andrulis, C., Ayres, T.R.:
ApJ 442, 381 (1995)
Linsky, J.L., Wood: ApJ 430, 342 (1994)
Maran, S.P. et al.: ApJ 421, 800 (1994)
Messina, S., Guinan, E.F.: A&A 393, 225 (2002)
Messina, S., Guinan, E.F.: A&A 409, 1017 (2003)
Messina, S., Rodon`o, M., Guinan, E.F.: A&A 366, 215 (2001)
Messina, S., Pizzolato, N., Guinan, E.F., Rodon`o, M.: A&A 410, 671
(2002)
Mohanty, S., Basri, G., Shu, F., Allard, F., Chabrier, G.: ApJ 571, 469
(2002)
Neff, J.E., Walter, F.M., Rodon `o, M., Linsky, J.L., 1989a, Solar
and Stellar Magnetic Activity, Cambridge Astrophysics Series,
34.
Neff, J.E., Walter, F.M., Rodon `o, M., Linsky,J.L.: A&A 215, 79 (1989b)
Osten, R.A., Ayres, T.R., Brown, A., Linsky, J.L., Krishnamurthi, A.:
ApJ 582, 1073 (2003)
Olah, K., Strassmeier, K.G.: AN 323, 361 (2002)
Pagano, I., Rodon`o, M., Linsky, J.L., Neff, J.E., Walter, F.M., Kov`ari,
Zs., Matthews, L.D.: A&A 365, 128 (2001)
Pagano, I., Linsky, J.L., Carkner, C., Robinson, R.D., Woodgate, B.,
Timothy, G.: ApJ 432, 497 (2000)
Pagano, I., Linsky, J.L., Valenti, J., Duncan, D.K: A&A 415, 331 (2004)
Pall´e Bag ´o, E., Butler, C.J.: IAU Symp. 203, 602 (2001)
Patern`o, L.: Compt. Rend. Acad. Sci. Paris Ser. IIb 326, 393 (1998)
Peter, H.: A&A 360, 761 (2000)
Peter, H.: A&A 374, 1108 (2001)
Pizzolato, N., Maggio, A., Micela, G., Sciortino, S., Ventura, P.: A&A
397, 147 (2003)
Rauscher, T., Heger, A., Hoffman, R.D., Woosley, S.E.: ApJ 576, 323
(2002)
Reale, F., Peres, G., Serio, S.: A&A 316, 215 (1996)
Redfield, S., Linsky, J.L., Ake, T.B., Ayres, T.R., Dupree, A.K., Robin-
son, R.D., Wood, B.E., Young, P.R.: ApJ 581, 626 (2002)
Redfield, S., Ayres, T.R., Linsky, J.L., Ake, T.B., Dupree, A.K., Robin-
son, R.D., Young, P.R.: ApJ 585, 993 (2003)
Ribas, I., Guinan, E.F., G¨udel, M., Audard, M.: ApJ 622, 680 (2005)
Rice, J.B.: AN 323, 220 (2002)
Robinson, R.D., Linsky, J.L., Woodgate, B., Timothy G.: ApJ 554, 368
(2001)
Rodon`o, M.: in: M. Zeilik and D. Gibson (eds.), Proc. of the 4th Cam-
bridge Workshop on Cool Stars, Stellar Systems and the Sun. Lec-
ture Notes in Physics 254, 475 (1986)
Rodon`o, M., Lanz`a, A.F., Catalano, S.: A&A 301, 75 (1995)
Rodon`o, M., Messina, S., Lanza, A.F., Cutispoto, G., Teriaca, L.: A&A
358, 624 (2000)
Springer

Astrophys Space Sci (2006) 303:17–31 31
Rodon`o, M., Cutispoto, G., Lanza, A.F., Messina, S.: AN 322, 333
(2002)
Rutledge, R.E., Basri, G., Martn, E.L., Bildsten, L.: ApJ 538, L141
(2000)
Schrijver, C.J., Zwaan C.: ApJ 473, 470 (2000)
Schwadron, N.A., Fisk, L.A., Zurbuchen, T.H., 1999, ApJ 521, 859.
Simon, T., Herbig, G., Boesgaard, A.M.: ApJ 293, 551 (1985)
Shkolnik, E., Walker, G.A.H., Bohlender, D.A.: AAS 33, 1303 (2001)
Shkolnik, E., Walker, G.A.H., Bohlender, D.A., Gu, P.-G., K¨urster, M.:
ApJ 622, 1975 (2005)
Soderblom, D., Jones, B.F., Fisher, D.: ApJ 563, 334 (2001)
Strassmeier, K.G.: ApJ 348, 682 (1990)
Strassmeier, K.G., Rice, J.B., Welhau, W.H., Vogt, S.S., Hatzes, A.P.
et al.: A&A 247, 130 (1991)
Strassmeier, K.G., Bartus, J., Cutispoto, G., Rodon`o, M.: A&AS 125,
1 (1997)
Uitenbroek, H., Dupree, A.K., Gilliland, R.L.: AJ 116, 2501 (1998)
Teriaca, L., Banerjee, D., Doyle, J.G.: A&A 349, 636 (1999)
Vidal-Madjar, A., Lecavelier des Etangs, A., D´esert, J.-M., Ballester,
G.E., Ferlet, R., H´ebrard, G., Mayor, M.: Nature 422, 143
(2003)
Vidal-Madjar, A., D´esert, J.-M., Lecavelier des Etangs, A., Hebrard,
G., Ballester, G.E., Ehrenreich, D., Ferlet, R., McConnell, J.C.,
Mayor, M., Parkinson, C.D.: ApJ 604, 69 (2004)
Vilhu, O., Muhli, P., Huovelin, J., Hakala, P., Rucinski, S.M., Collier
Cameron, A. 1998, AJ 115, 1610 (1998)
Walter, F.M.: The Future of Cool-Star Astrophysics: in: Brown, A.,
Harper, G.M., Ayres, T.R. (eds.), Proc. of the 12th Cambridge
Workshopon Cool Stars, Stellar Systems, and the Sun, p. 14 (2003)
Walter, F.M., Neff, J.E., Gibson, D.M. et al., A&A 186, 241 (1987)
Wamsteker, W., Gonz `alez-Riestra, R. (eds.) Proc. of the Conference
“Ultraviolet Astrophysics beyond the IUE Final Archive”, ESA
SP-413
Wikstøl, Ø., Hansteen, V.H., Carlsson, M., Judge, P. G.: ApJ 531, 1150
(2000)
Wilson, O.C.: ApJ 26, 379 (1978)
Wood, B.E., Harper, G.M., Linsky, J.L., Dempsey, R.C.: ApJ 458, 761
(1996)
Wood, B.E., Linsky, J.L., Ayres, T.R.: ApJ 478, 745 (1997)
Wood, B.E., Linsky, J.L., Muller, H.-R., Zank, G.P.: ApJ 547, L49
(2001)
Wood,B.E., M ¨uller, H.-R, Zank, G.P., Linsky, J.L.: ApJ 574, 412 (2002)
Wood, B.E., Redfield, S., Linsky, J.L., Muller, H.-R., Zank, G.P: ApJS,
159, 118 (2005)
Wood, B.E., Muller, H.-R., Zank, G.P., Linsky, J.L., Redfield, S.: ApJL,
628, 143 (2005)
Young, P.R., Dupree, A.K., Espey, B.R., Kenyon, S.J., Ake, T.B.: ApJ,
618, 891 (2005)
Springer

Astrophys Space Sci (2006) 303:33–52
DOI 10.1007/s10509-006-8793-9
ORIGINAL ARTICLE
UV Capabilities to Probe the Formation of Planetary Systems:
From the ISM to Planets
Ana I. G´omez de Castro ·Alain Lecavelier ·
Miguel D’Avillez ·Jeffrey L. Linsky ·Jos´e Cernicharo
Received: 26 April 2005 / Accepted: 16 June 2005
C
Springer Science +Business Media B.V. 2006
Abstract Planetary systems are angular momentum reser-
voirs generated during star formation. Solutions to three of
the most important problems in contemporary astrophysics
are needed to understand the entire process of planetary sys-
tem formation:
The physics of the ISM. Stars form from dense molecu-
lar clouds that contain ∼30% of the total interstel-
lar medium (ISM) mass. The structure, properties and
lifetimes of molecular clouds are determined by the
overall dynamics and evolution of a very complex sys-
tem – the ISM. Understanding the physics of the ISM
is of prime importance not only for Galactic but also
for extragalactic and cosmological studies. Most of the
ISM volume (∼65%) is filled with diffuse gas at tem-
peratures between 3000 and 300 000 K, representing
about 50% of the ISM mass.
The physics of accretion and outflow. Powerful outflows are
known to regulate angular momentum transport dur-
A. I. G. de Castro
Instituto de Astronom´ıa y Geodesia (CSIC-UCM), Universidad
Complutense de Madrid, Madrid E-28040, Spain
A. Lecavelier
Institute d’Astrophysique de Paris, Paris, France
M. D’Avillez
Department of Mathematics, University of ´
Evora R. Rom˜ao
Ramalho 59, 7000 ´
Evora, Portugal; Institut f´
’ur Astronomie,
Universit¨at Wien, T¨urkenschanzstr. 17, A-1180 Wien, Austria
J. L. Linsky
JILA, University of Colorado and NIST, Boulder, CO
80309-0440, USA
J. Cernicharo
DAMIR-IEM-CSIC, C/. Serrano 113 & 121, 28006 Madrid, Spain
ing star formation, the so-called accretion–outflow en-
gine. Elementary physical considerations show that,
to be efficient, the acceleration region for the outflows
must be located close to the star (within 1 AU) where
the gravitational field is strong. According to recent
numerical simulations, this is also the region where
terrestrial planets could form after 1 Myr. One should
keep in mind that today the only evidence for life in the
Universe comes from a planet located in this inner disk
region (at 1 AU) from its parent star. The temperature
of the accretion–outflow engine is between 3000 and
107K. After 1 Myr, during the classical TTauri stage,
extinction is small and the engine becomes naked and
can be observed at ultraviolet wavelengths.
The physics of planet formation. Observations of volatiles
released by dust, planetesimals and comets provide an
extremely powerful tool for determining the relative
abundances of the vaporizing species and for studying
the photochemical and physical processes acting in the
inner parts of young planetary systems. This region is
illuminated by the strong UV radiation field produced
by the star and the accretion–outflow engine. Absorp-
tion spectroscopy provides the most sensitive tool for
determining the properties of the circumstellar gas as
well as the characteristics of the atmospheres of the
inner planets transiting the stellar disk. UV radiation
also pumps the electronic transitions of the most abun-
dant molecules (H2, CO, etc.) that are observed in the
UV.
Here we argue that access to the UV spectral range is
essential for making progress in this field, since the resonance
lines of the most abundant atoms and ions at temperatures
between 3000 and 300 000 K, together with the electronic
transitions of the most abundant molecules (H2, CO, OH, CS,
Springer

34 Astrophys Space Sci (2006) 303:33–52
S2,CO
+
2,C
2,O
2,O
3, etc.) are at UV wavelengths. A powerful
UV-optical instrument would provide an efficient mean for
measuring the abundance of ozone in the atmosphere of the
thousands of transiting planets expected to be detected by
the next space missions (GAIA, Corot, Kepler, etc.). Thus,
a follow-up UV mission would be optimal for identifying
Earth-like candidates.
Keywords UV astronomy ·ISM ·Pre-main sequence
stars ·Jets ·Winds ·Accretion disks ·Planets
1. Introduction
The formation of planetary systems covers a broad range
of physical and astrophysical processes ranging from the
physics of star formation (the interstellar medium (ISM),
molecular clouds, and initial mass function), to the physics of
accretion and outflow (accretion disk properties, winds gen-
eration, and disk instabilities) and finally, the formation of
planets (dust nucleation, planetesimal and planet formation,
planetary differentiation, planetary atmospheres and sustain-
able biological systems). The objective of this article is to
summarize the reasons why access to the UV range is an
essential requirement for making progress in these critically
important areas of astrophysics.
For this reason, the article has been split into three key sec-
tions: physics of the ISM, physics of accretion and outflow,
and planets and bio-markers. A summary has been added to
the end of this contribution with the required UV capabilities
to make progress in the field.
2. The physics of the ISM
Understanding the physics of the ISM is of prime importance
not only for Galactic but also for extragalactic and cosmolog-
ical studies. The ISM is everything observable in the Galaxy
except for stars, e.g., gas (ionised, atomic and molecular),
dust, high-energy particles (e.g., cosmic rays) and magnetic
fields. The ISM is a very complex, highly non-linear dynami-
cal system whose evolution controls star formation, gas mix-
ing, and, therefore, the chemical enrichment of the Universe.
The ISM is often classified into five components: Hot
Ionised Medium (HIM), Warm Ionised Medium (WIM),
Warm Neutral Medium (WNM), Cold Neutral Medium
(CNM) and dense Molecular Medium (MM) (see, e.g.,
Kulkarni and Heiles, 1988); the diffuse components (HIM,
WIM, WHM and CNM) appear to be in approximate pres-
sure equilibrium. The main properties of these components
are summarized in Table 1. X-ray observations are most sen-
sitive to the very hot gas with temperatures T≥106K. IR and
radio wavelength observations are the only tools for studying
the dense molecular gas where stars form. UV spectroscopy is
the most sensitive tool for measuring the properties (column
densities, temperatures, ionisation fractions, metallicity, de-
pletions, etc.) of the diffuse gas in the 3000–300 000 K tem-
perature range, the WNM and WIM.
In the last few years, old models, based on pressure equi-
librium between the various ISM phases (e.g., McKee and
Ostriker, 1977), have been replaced by detailed numerical
simulations that allow studying the ISM as it is, a dynamical
system. High-spatial resolution numerical simulations now
permit addressing a set of problems simultaneously, encom-
passing both large and small scales, provided that the appro-
priate grid size, resolution and numerical tools (e.g., adaptive
mesh refinement) are used. Among the most important of
these problems are: (i) global modelling to yield information
on the formation and lifetimes of molecular clouds, (ii) how
star-forming regions are influenced by large-scale flows in
the ISM, and (iii) the dynamic roles that SNe and superbub-
bles play in triggering local and global star formation (see
Heyer and Zweibel, 2004).
Key problems in ISM physics include the determination
of the relative contributions to the energy input from the var-
ious possible sources (SNe, massive star winds and radiation
fields, mass infall from the halo, galactic dynamics, cosmic
rays and magnetic fields) and the roles of MHD turbulence
and shocks in the energy cascade and structure formation.
During the last few years, a very efficient feedback loop has
been operating between radio observations and numerical
simulations to study the role of MHD turbulence in the energy
cascade within the densest regions of the ISM (H Iand molec-
ular clouds). A similar feedback loop needs to be established
with UV observations to understand the heating/cooling pro-
cesses and the overall ISM evolution, including the formation
of molecular clouds.
Numerical simulations predict that ∼65% of the ISM vol-
ume (within the disk: |z|<250 pc) is filled with gas at tem-
peratures between 103and 105.5K; in particular, the WIM
(104<T<105.5K) is expected to fill 25% of the disk vol-
ume, while the HIM filling factor is smaller (17%) because
it escapes to the halo (de Avillez and Breitschwerdt, 2004).
These predictions agree with recent observations. The Wis-
consin H-Alpha Mapper (WHAM) has observed O III emis-
sion extending to Galactic latitudes as high as |b|∼45◦, and
even He I(5876 ˚
A) emission has been detected. However,
X-ray and extreme UV observations now show that the fill-
ing factor of the HIM in the local ISM is small. Clearly, UV
instruments are required to make progress in our understand-
ing of the physics of the Galactic ISM. This physics can be
studied at two scales:
Large scale (>kpc scale): At this scale, the prime objectives
are understanding the overall star formation efficiency
and which parameters control the disk–halo circula-
tion (the Galactic fountain). The roles of supernova
Springer

Astrophys Space Sci (2006) 303:33–52 35
Table 1 Components of the ISM (reference properties)
Component n(cm−3)T(K) Ionisation fraction Spectral range
HIM Few 10−3(5–10) ×1061 X-ray, UV
WIMa,bFew 10−11041 UV, radio, optical
WNM 1–10 (3–8) ×1030.1 UV, optical, IR, radio
CNM 1–50 20–100 10−2UV, IR, radio
MM 103–10610 <10−4IR, radio
aIn the last few years, a new sub-classification has been introduced to distinguish between the WIM envelope
around molecular clouds, the so-called McKee & Ostricker WIM or MOWIM, and the other warm, ionised
components, the so-called Reynolds WIM.
bThis nominal temperature is often assigned because most of the observations come from Hαemission. However,
UV observations have pointed out the presence of significantly hotter gas traced by C IV or O VI
explosions, expanding H II regions and the overall
Galactic dynamics (shear, spiral arm shocks, high ve-
locity cloud shocks, etc.) is examined in detail. Some
attention is also devoted to understanding Galactic
magnetic fields and the Galactic dynamo.
Small scale (the Local Bubble scale): The Local Bubble rep-
resents the nearest region of the ISM and is thus an
ideal laboratory to test the details of the ISM physics:
dust depletion and abundances, non-equilibrium ioni-
sation, shocks, turbulence, etc.
2.1. Halo–disk interaction in the Milky Way
The Milky Way is surrounded by a large halo of hot gas
which must be replenished as the gas cools. The most di-
rect evidence comes from the detection of high-ionisation
UV resonance lines and of X-ray emitting gas surround-
ing the Galactic disk. The X-ray halo has a luminosity of
∼4×1039 erg s−1and the thickness of the emission is prob-
ably a few kpc. The temperature of the X-ray emitting gas
is (1–2) ×106K, well below the escape temperature, and
therefore it is gravitationally bound to the Galaxy. The high
temperatures of the gas in the halo must be explained by en-
ergetic processes, most likely occurring in the Galactic disk
(though external sources could contribute). The most likely
energy source for these processes are massive OB stars.
There is also evidence that hot halo gas has cooled, as pro-
vided by the presence of the high-ionisation UV resonance
lines. Some of the UV resonance lines might be produced by
photoionisation of gas near 104K, but there is good evidence
that the high-ionisation lines, such as Si IV,CIV and N Vare
due to gas that is cooling (Savage and Sembach, 1994). For
the Si IV and C IV absorption lines it was found that the ratio
of the column densities from these ions is almost a constant
value, implying that the ionisation state of the gas is nearly
identical in different parts of the halo. Benjamin and Shapiro
(1993) argued that in a galactic fountain, gas cooling from
106K would be opaque to its own radiation, causing it to
self-ionise, so that its ionisation state is determined by the
cooling process itself. They also showed that the absorption
strength of N Vis reproduced by this model. Martin and
Bowyer (1990) have shown that the emission from C IV and
OIII ions is consistent with being produced from a cooling
Galactic fountain.
The cooling of the fountain flow, is dominated by colli-
sional ionisation and radiative recombination as three-body
recombination (the inverse process of collisional ionisation)
is very unlikely to occur, because as the fountain gas, with
a temperature in excess of 106K, rises into the halo it ex-
pands and cools adiabatically. Hence, there is a reduction of
the plasma temperature and density (to be of the order of
10−2cm−3). Furthermore, the cooling timescales at consid-
erable heights above the disk can be much smaller than the
dominant microphysical processes and therefore, recombi-
Fig. 1 Ionisation stages of oxygen in a self-consistent calculation of
a galactic outflow (wind) (Breitschwerdt and Schmutzler, 1999). The
initial temperature of the flow is T=2.5×106K
Springer

36 Astrophys Space Sci (2006) 303:33–52
nation of highly ionised species lags behind (Breitschwerdt
and Schmutzler, 1999). An example is the O VI ion that is
abundant over a large range of temperatures from 4 ×105K
to temperatures as low as 104K (Fig. 1). Thus, care should
be taken while using a single ion like O5+as a diagnostic
element for plasma temperature (see also Schmutzler and
Tscharnuter, 1993).
Thus, measurements of the z-dependence of O VI,NV,
CIV and Si IV emission are of prime importance for deter-
mining whether the gas is in collisional equilibrium ionisa-
tion, in non-equilibrium ionisation, or whether other relevant
heating sources maybe present. These observations are crit-
ically needed to constrain numerical models of the ISM in
disk galaxies to provide further clues concerning how matter
and energy are transferred within the Galaxy. This is clearly
displayed in Fig. 2; adaptative mesh simulations of the dy-
namical evolution of the ISM (Avillez and Breitschwerdt,
2004, in press) show that matter in the disk is concentrated
in dense shells and filaments, while the halo acts as a pres-
sure release valve for the hot (T>105.5K) phase in the disk
thereby controlling its volume filling factor. The upper por-
tions of the thick ionised disk form the disk–halo interface
located about 2 kpc above and below the mid-plane. Here a
large-scale fountain is set up by hot ionised gas injected from
either the gas streaming out of the thick disk or directly from
superbubbles in the disk underneath. The gas then escapes in
a turbulent convective flow.
Radio observations can be used to map the clumpy dis-
tribution of matter in the disk, but the most important con-
straints will come from the study of the vertical distribution
of warm gas, which is best studied with UV spectroscopy.
This is also true for the high-velocity clouds (HVCs) detected
by their H I21 cm emission, which are surrounded by hot
ionised envelopes as pointed out by new observations from
the Far Ultraviolet Spectroscopic Explorer (FUSE) mission
and the Space Telescope Imaging Spectrograph onboard the
Hubble Space Telescope (HST/STIS). The detection of O VI,
CIV and Si IV absorption indicates that many HVCs have
a hot, collisionally-ionised component (Danly et al., 1992;
Tripp et al., 2003). UV absorption lines provide detailed in-
formation on the physical conditions and abundances of the
gas. Understanding the ionisation of such envelopes will per-
mit us to constrain the properties of the Galactic corona and
the Local Group medium. UV absorption lines are the most
sensitive probes for determining the abundances (and hence
their Galactic or extragalactic origin) of the HVCs (see, e.g.,
Richter et al., 2001). Note that the most robust specie for
constraining the metallicity of HVCs is O I, since oxygen is
only slightly depleted by dust grains (Moos et al., 2002) and
the ionisation potential of O Iis very similar to H I. Thus,
oxygen abundances based on the O I/H Iratio depend only
slightly on the ionisation of the gas for substantially ionised
plasmas.
Fig. 2 Slice through the 3D data set showing the vertical (perpendicular
to the mid-plane) distribution of the density at time 166 Myr. Red/blue
in the colour scale refers to lowest/highest density (or highest/lowest
temperature). The z-scale above 0.5 kpc and below −0.5 kpc is shrunk
(in order to fit the paper size) and thus, the distribution of the labels is
not uniform (from Avillez and Breitschwerdt, in press)
Springer

Astrophys Space Sci (2006) 303:33–52 37
2.2. The local ISM
The Sun lies inside the Local Bubble, a large ionised gas
bubble that extends outwards from the Sun to a neutral hy-
drogen column density log N(H I)≈19.2 (cm−2units), cor-
responding to a geometrical size of 100–200 pc depending
on the direction from the Sun. The Local Bubble’s morphol-
ogy has been identified by Na Iabsorption, which is formed
in the cold gas than surrounds the Local Bubble and deter-
mines its shape (Lallement et al., 2003). The Local Bubble is
thought to be an H II region formed by the explosions of su-
pernovae and the strong winds of young hot stars in the Lower
Centaurus Crux subgroup of the Scorpio–Centaurus Associ-
ation (Ma´ız-Apell´aniz, 2001; Bergh ¨ofer and Breitschwerdt,
2002). The temperature of the gas in the Local Bubble is
estimated to be about 106K if most of the soft X-ray back-
ground is due to thermal emission from the Local Bubble gas.
However, analysis of extreme ultraviolet emission obtained
with the CHIPS satellite has not yet led to an accurate tem-
perature or emission measure of the hot, low-density Local
Bubble gas, which may be far out of collisional ionisation
equilibrium.
Embedded in the Local Bubble are a number of warm gas
clouds. The Sun is located inside one of these clouds called
the Local Interstellar Cloud (LIC). The existence of the LIC
was first suggested by Vidal-Madjar et al. (1978). The LIC
was first identified by Lallement and Bertin (1992) on the
basis of measured Doppler shifts of interstellar absorption
lines in many directions that are consistent with a single
velocity vector, implying that all of the gas in this cloud is
moving with a common velocity away from the centre of the
Scorpio–Centaurus Association. Analysis of interstellar UV
absorption lines with the high-resolution echelle gratings in
the GHRS and STIS instruments on HST enabled Redfield
and Linsky (2000) to determine a temperature of the LIC gas
(7000 ±1000 K) and the morphology of the LIC. The LIC
centre is located in the anti-Galactic centre direction. The
maximum column density through the LIC is log N(H I)=
18.3, the LIC’s maximum dimension is about 6.8pc, and its
mass is about 0.32 M. The main evidence for the Sun being
located inside the LIC is that neutral helium flowing into
the heliosphere, which is not influenced by the solar wind,
has the same temperature and flow vector as the LIC (Bertin
et al., 1993; Witte, 2004).
Slavin and Frisch (2002) computed the ionisation of many
elements in the LIC taking into account UV and EUV ra-
diation from the most important ionising source, the star
CMa, hot white dwarfs and other stars, the diffuse UV
background, and the estimated radiation from the putative
conductive boundary between the warm clouds and the hot
gas of the Local Bubble. More sensitive UV observations are
required to study this boundary layer, if indeed it is present.
Their models assumed ionisation equilibrium and realistic
HIcolumn densities between the centre of the LIC and the
external sources of ionising radiation. One of their models
predicts the temperature, electron density, and ionisation of
many elements in good agreement with observations. Dust,
which is present in the LIC and other nearby warm clouds,
plays an important role in cooling the gas and in depleting
metals such as iron from the gas phase.
There are a number of other warm, partially ionised clouds
in the solar neighbourhood, which are also located inside the
Local Bubble. The so-called Gcloud identified by Lalle-
ment and Bertin (1992), which is situated in the Galactic
centre direction, is slightly cooler than the LIC and has a
somewhat different velocity vector. The closest star αCen
(1.3 pc), which is located in the Gcloud, shows no evidence
for absorption by gas at the LIC velocity even in very high
signal-to-noise GHRS echelle spectra. This places an upper
limit of 0.05 pc on the thickness of the LIC in the direction
of αCen and a time of <3000 years for the Sun to leave the
LIC and enter either the Gcloud or an unknown interface
region between the LIC and the Gcloud.
The broad UV spectral coverage, high spectral resolution,
and accurate wavelength scale of the STIS instrument al-
lowed Redfield and Linsky (2004) to measure absorption line
wavelengths, widths, and Doppler shifts for many atoms and
ions including H I,DI,CII,NI,OI,MgII,AlII,SiII and Fe II
along 29 different lines of sight through warm clouds in the
Local Bubble. They detected absorption at 50 different ve-
locities along these lines of sight, indicating about 12 clouds
with different velocity vectors. The observation of absorp-
tion lines from elements or ions with very different atomic
weights (2 for deuterium compared to 56 for iron) allowed
Redfield and Linsky (2004) to solve for the gas temperature
and non-thermal motions (turbulence) separately for each
velocity component. They found velocity components in the
local ISM with temperatures as high as about 12 000 K and a
few components with temperatures below 3000K. The mean
gas temperature is T=6680 ±1490 K, which is character-
istic of warm clouds, but there are some velocity components
inside the Local Bubble that could be cold clouds. In almost
all cases, the non-thermal motions are far smaller than the
thermal motions. The mean thermal pressure in the clouds,
PT/k=2280 ±520Kcm
−3. The magnetic fields in these
clouds have not yet been measured.
While STIS spectra of interstellar absorption lines formed
in the local ISM have begun to reveal its secrets, we have
sampled far too few lines of sight to identify the structure of
the local ISM in detail. In particular, we do not yet havea good
understanding of the amount of gas at different temperatures
in the local ISM, nor do we have a detailed understanding
of how the ionisation and temperature of the gas depends on
the radiation environment and past history. Understanding
the physics of the local ISM is required if we are to have
any confidence in understanding the physics of the Galactic
Springer

38 Astrophys Space Sci (2006) 303:33–52
disk, halo, and ISM in other galaxies. Since most interstellar
absorption lines are located in the UV and the absorption
lines are typically narrow with multiple velocity components,
a future sensitive high-resolution UV spectroscopic mission
is needed to extend the preliminary work provided by the
GHRS and STIS instruments on HST.
3. The physics of accretion and outflow
Understanding how stars form out of the contracting cores
of molecular gas is a major challenge for contemporary as-
trophysics. Angular momentum must be conserved during
gravitational contraction and magnetic flux is built up and
dissipated in the process, but the underlying mechanisms are
still under debate.
Solar-like protostars are an excellent laboratory for this
purpose, since their pre-main sequence (PMS) phases last
∼100 Myr. The collapse of low-mass protostars is sub-
alfv´enic, thus these protostars are expected to be magnetized.
The detection of kG fields in stars as young as a few million
years (Guenter et al., 1999; Johns-Krull et al., 1999) supports
this assumption. In the last few years, a new paradigm has
emerged to understand the basics of star formation. Protostars
are assumed to be magnetized and star growth is regulated
by the interaction between the stellar magnetic field and the
disk. The physics of this interaction is outlined in Fig. 3.
The disk–star interaction basically transforms the angular
momentum of the disk (differential rotation) into plasmoids
Fig. 3 The interaction between the stellar magnetic field and the disk
twists the stellar field lines due to the differential rotation. The toroidal
magnetic field generated out of the poloidal flux and the associated pres-
sure tends to push the field lines outwards, inflating them, and eventu-
ally braking the magnetic link between the star and the disk (boundary
between Regions I and II). Three basic regions can be defined: Re-
gion I dominated by the stellar wind, Region II dominated by the disk
wind and Region III dominated by stellar magnetospheric phenomena.
The dashed line traces the boundaries between this three regions. The
continuous lines indicate the topology of the field and the shadowed
areas represent regions where magnetic reconnection events are likely
to occur, producing high-energy radiation and particles (from G´omez
de Castro, 2004)
that are ejected from the system. There is a current sheet that
separates two distinct regions: an inner stellar outflow and
an external disk flow. Magnetic flux dissipation should occur
in the current layer producing the ejection of plasmoids, as
well as the generation of high-energy particles (cosmic rays),
X-rays, and ultraviolet radiation.
The phenomenon is non-stationary and is controlled by
two different temporal scales: the rotation period and the
magnetic field diffusion timescale. Stellar rotation is a
well-known parameter that controls the opening of the field
lines towards high latitudes. Plasmoid ejection, however, is
controlled by field diffusion which is poorly determined (see,
e.g., Priest and Forbes, 2000). All models can be fitted into
this basic configuration (Uzdensky, 2004).
During the last 5 years, numerical research on this interac-
tion region has gone into outburst (see, e.g., Goodson et al.,
1997, 1999). So far, most studies have analysed the interac-
tion between a dipolar stellar magnetic field and a Kepplerian
accretion disk. Numerical simulations show that the funda-
mental mechanism for jet formation is robust. The star–disk–
outflow system is self-regulating when various initial disk
densities, stellar dipolar field strengths, and primordial fields
associated with the disk are tested (Matt et al., 2002), al-
though strong stellar magnetic fields may disrupt the inner
parts of the accretion disk temporarily (Kueker et al., 2003).
Despite the numerical advances made so far, the real prop-
erties of the engine are poorly known because of the lack
of observations to constrain the modelling. Very important
open questions include the following:
1. How does the accretion flow proceed from the disk to the
star? Is there any preferred accretion geometry like, for
instance, funnel flows?
2. What roles do disk instabilities play in the whole accre-
tion/outflow process?
3. What are the dominant wind acceleration processes? What
are the relevant timescales for mass ejection?
4. How does this high-energy environment affect the chem-
ical properties of the disk and planetary building?
5. How important is this mechanism when radiation pressure
becomes significant as for Herbig Ae/Be stars?
Infrared and radio wavelengths cannot access this engine
because the spatial scales involved are tiny (<0.1 AU or
0.7 mas for the nearest star-forming regions compared with
ALMA’s resolution of 10 mas) and the temperatures are too
high (3000–300 000 K). High-resolution IR spectroscopy has
indeed confirmed the presence of warm molecular gas with
temperatures of 1500–3000 K in the innermost disk: both CO
(fundamental and overtone) and H2O emission have been de-
tected (see Najita et al., 2000 for a review or Carr, Tokunaga
and Najita, 2004 for more recent results). Fortunately, af-
ter 1 Myr, during the classical T Tauri Star (CTTS) phase,
the circumstellar extinction becomes small (AV<1 mag)
Springer

Astrophys Space Sci (2006) 303:33–52 39
and the engine described above can be properly tested at ul-
traviolet wavelengths. The UV spectral range is the richest
for diagnosing astrophysical plasmas in the 3000–300 000K
temperature range, since the resonance lines of the most
abundant species are located in the UV. In addition, as the UV
radiation field is strong in the circumstellar environment, flu-
orescence emission from the most abundant molecules (H2,
CO, OH, CS,S2,CO
+
2,C
2, and CS) is observed. As a result,
a single high-resolution spectrum in the 1200–1800 ˚
A range
provides information on the molecular content, the abun-
dance of very reactive species such as the O I, and the warm
and hot gases associated with the CTTSs. The potential of
UV spectroscopy for studying the physics of accretion during
PMS evolution is outlined in the following section.
3.1. UV observations of the jet engine in low-mass stars
The engine is a small structure (≤0.1 AU) with several differ-
ent constituents (the accretion flow, stellar magnetosphere,
winds, and inner part of the accretion disk) all radiating in the
ultraviolet. The UV spectrum of the T Tauri Stars (TTSs) has
a weak continuum and many strong emission lines. The con-
tinuum is significantly stronger than that observed in main
sequence stars of similar spectral types (G to M); the so-called
optical-veiling represents the low-energy tail of this excess
UV emission (Hartigan et al., 1990). The underlying photo-
sphere is barely detected, and only in warm (G-type) weak
line TTSs (WTTSs) is the photospheric absorption spectrum
observed. The UV continuum excess is significantly larger
in CTTSs than in WTTSs, as is well illustrated in the colour
(UV −V)–magnitude (V) diagram displayed in Fig. 4. Sim-
ple models of hydrogen free–free and free–bound emission
Fig. 4 The (UV −V, V) colour–magnitude diagram for the T Tauri
Stars observed with the IUE satellite in the Taurus region (a distance
of 140 pc to Taurus has been assumed). The crosses represent cool
TTSs (spectral types later than ∼K3) and the open circles represent
warm TTSs (spectral types earlier than ∼K3). The location of the main
sequence is marked by the spectral types. The stars closer to the main
sequence are the WTTSs (from G´omez de Castro, 1997)
added either to black bodies or to the spectra of standard
stars reproduce the UV continuum reasonably well (Cal-
vet et al., 1984; Bertout et al., 1988; Simon et al., 1990).
The fits yield chromospheric-like electron temperatures of
(1–5) ×104K. Three different mechanisms have been pro-
posed to generate this hot plasma and its UV continuum: (1)
a dense chromosphere (Calvet et al., 1984), (2) the release of
the gravitational binding energy from the infalling material
(Bertout et al., 1988; Simon et al., 1990; Gullbring et al.,
2000), and (3) an outflow (Ferro-Font´an and G´omez de Cas-
tro, 2003; G´omez de Castro and Ferro-Font´an, 2005). This
uncertainty in the formation of the UV emission points out
why high-resolution UV spectroscopy and monitoring are so
crucial for understanding and constraining the physics of the
engine.
3.1.1. Signatures of accretion
The most obvious signature of accretion is the detection of
narrow red-shifted absorption components on top of the emis-
sion profiles of singly ionised species such as Mg II or Fe II
with strong transitions in the UV at 2600 and 2800 ˚
A. It is
widely accepted that this absorption is produced in funnel
flows: magnetic tubes connecting the inner disk to the stellar
surface. However, there are no detailed maps of the funnel
flows except for some attempts made in the optical range
(Petrov et al., 2001; Bouvier et al., 2003). UV mapping is
crucial in determining the rigidity of the flux tubes and thus
the possible distortions induced by differential rotation and
the magnetic diffusivity of the disk.
Funnel flows are expected to radiate over a broad range of
temperature, from 3000 K at the disk end to some 100000 K
at the stellar surface. Since infalling material is nearly in free-
fall, its kinetic energy is finally released at the stellar surface
in an accretion shock that reaches temperatures of 106K. The
dominant output radiation is produced by the photoionised
pre-shock infalling gas (G´omez de Castro and Lamzin, 1999;
Gullbring et al., 2000). Thus, the full accretion column could
be tracked by monitoring CTTSs with a high spectral reso-
lution UV instrument, but this observation has not yet been
carried out! The only UV monitoring of CTTSs was by the
IUE satellite with low dispersion due to the small effective
area of its 40 cm telescope. Nevertheless, the results are very
promising as rotational modulation of the UV continuum and
line fluxes were detected in DI Cep and BP Tau (G´omez de
Castro and Fern´andez, 1996; G´omez de Castro and Fran-
queira, 1997a). This modulation is caused by the small size
of the accretion shock, which occupies only a small fraction
of the stellar surface.
An important result of these campaigns is that only ∼50%
of the UV continuum excess is rotationally modulated. Thus,
a significant fraction of the UV excess is not produced by the
accretion shock. Whether the wind or an extended magneto-
Springer

40 Astrophys Space Sci (2006) 303:33–52
sphere is responsible for the UV continuum excess remains
a matter of debate. In fact, the coexistence of several funnel
flows has been proposed to explain this fact (Muzerolle et al.,
2001).
3.1.2. Signatures of disks
High-resolution HST/STIS spectra have revealed, for the first
time, the rich UV molecular emission in CTTS. H2fluores-
cence emission has now been studied in detail in the nearest
CTTS, TW Hya, and the richness of the spectrum is over-
whelming: Herczeg et al. (2002) detected 146 Lyman-band
H2lines, representing 19 progressions (see Fig. 5)! The ob-
served emission is likely produced in the inner accretion disk,
as are the infrared CO and H2O lines. The excitation of H2
can be determined from the relative line strengths by measur-
ing self-absorption in lines originating in low-energy lower
levels, or by reconstructing the Lyαemission line profile in-
cident upon the warm H2using the total flux from each fluo-
rescing upper level and the opacity in the pumping transition.
Using this diagnostic technique, Herczeg et al. (2004) es-
timated that the warm disk surface has a column density
of NH2=3.2×1018 cm−2, temperature of T=2500 K, and
filling factor of H2as seen from the source of the Lyαemis-
sion of 0.25 ±0.08. The observed spectrum shows that some
ground electronic state H2levels with excitation energies as
large as 3.8 eV are pumped by Lyα. These highly excited
levels may be formed by dissociative recombination of H+
3,
which in turn may be formed by reactions involving X-rays
and UV photons from the star. A quick inspection of the
UV spectra in the IUE and HST archives shows that fluo-
rescent H2UV lines are observed in most of the TTSs (see
also G´omez de Castro and Franqueira, 1997b; Valenti et al.,
2000; Ardila et al., 2002).
The role of far-UV radiation fields and high-energy par-
ticles in the disk chemical equilibrium is now beginning to
be understood. Bergin et al. (2003) showed how strong Lyα
emission may contribute to the observed enhancement of
CN/HCN in the disk. The penetration of UV photons com-
ing from the engine in a dusty disk could produce an impor-
tant change in the chemical composition of the gas, allowing
the growth of large organic molecules. In this context, UV
photons photodissociating organic molecules at λ>1500 ˚
A
could play a key role in the chemistry of the inner regions of
the disk, while those photodissociating H2and CO will con-
trol the chemistry of the external layers of the disk directly
exposed to the radiation from the central engine (see, e.g.,
Cernicharo, 2004). Ultraviolet radiation also plays a very
important role in the evolution of the primary atmospheres
of planetary embryos (Watson et al., 1981; Lecavelier des
Etangs et al., 2004).
Strong continuum FUV emission (1300–1700 ˚
A) has been
detected recently from some stars with bright molecular disks
Fig. 5 A portion of the HST/STIS spectrum of the CTTS, TW Hya.
The narrow H2emission lines originate in the Belectronic state after
being pumped by the H ILyαline (from Herczeg et al., 2002)
including GM Aur, DM Tau, and LkCa 15, together with in-
ner disk gaps of few AUs (Bergin et al., 2004). This emis-
sion is likely due to energetic photoelectrons mixed into the
molecular layer that likely indicates the existence of a very
hot component in the disk. This very hot component is prob-
ably created by X-ray and high-energy particle irradiation
of the disk (Glassgold et al., 2004; G´omez de Castro and
Antonicci, 2005).
Spectroscopic observations of volatiles released by dust,
planetesimals and comets provide an extremely powerful tool
for determining the relative abundances of the vaporizing
species and for studying the photochemical and physical pro-
cesses acting in the inner parts of protoplanetary disks. The
UV studies of βPic-like systems illustrate the possibilities
of this spectral range for this purpose (Vidal-Madjar et al.,
1994, see also, Section 3.3). The relevance of UV observa-
tions to study comets is described in detail by Brosch et al.
(2005) in the Solar System chapter of this volume.
3.1.3. Signatures of winds
Large-scale outflows are observed as collimated jets or
Herbig-Haro objects in some TTSs (see Section 3.2). How-
ever, spectroscopic signatures of winds and outflows are de-
tected in all the TTSs. Three types of signatures have been
detected in the UV:
1. The emission profiles of the Mg II resonance lines (2796,
2803 ˚
A) show pronounced broad absorption in their blue
wing for the 16 TTSs observed with IUE or HST (see e.g.,
G´omez de Castro, 1997). Blue-wing absorption is also
observed in Lyα. Terminal velocities up to ∼300 km s−1
are observed. Unfortunately, the interpretation of these
profiles is very complex, since the Mg II and Lyαlines
are optically thick and there is no unambiguous method
Springer

Astrophys Space Sci (2006) 303:33–52 41
Fig. 6 Left: Variation of the
three velocity components:
rotation (Vt), radial expansion
from the axis (Vr) and axial
velocity (Vz) are represented.
Top panel:Vzand Vrare
represented with dashed and
solid lines, respectively. Bottom
panel: The ratio between Vtand
the Kepplerian velocity at
0.1 AU for a solar mass star
(155 km s−1) is plotted. Right:
Disk wind kinematics as shown
by the line profiles for an
edge-on system. Each profile
correspond to a ring of gas
perpendicular to the outflow
axis that is identified by its
distance (z) to the disk plane.
Bottom panel: Line profiles with
z<5 AU – the outflow passes
from being rotationally
dominated (inner broad or
double peaked profile) to radial
expansion dominated (double
peaked profiles with peak
velocity ∼120 km s−1.Middle
panel: Outflow passes from
being radial expansion
dominated to axial-acceleration
dominated. Top panel: Outflow
is dominated by axial
acceleration and the line
broadening is basically thermal
(from G´omez de Castro and
Ferro-Font´an, 2005)
for determining the underlying blue-shifted emission and
thus the true wind absorption.
2. Narrow and blue-shifted C III] 1909 ˚
A and Si III] 1892 ˚
A
emission has been detected at the same velocity as the op-
tical jets in some TTSs (G´omez de Castro and Verdugo,
2003a). This emission is produced by unresolved jets and
indicates that PMS star jets are hotter than previously in-
ferred from the optical observations in agreement with
the UV observations of protostellar jets and Herbig-Haro
objects (see Section 3.2).
3. Optically thin semi-forbidden lines tracing warm plasma
(C II] 2325 ˚
A, O II], and, most prominently, C III] and
Si III]) show long blueward-shifted tails with velocities up
to −300 km s−1as in RU Lup, together with slight shifts to
the blue of the line peak (G´omez de Castro and Verdugo,
2003a).
These data provide three key pieces of information: (i)
there is a broad range of temperatures in the outflows (3000–
30 000 K), (ii) outflows are not spherically symmetric, and
(iii) their kinematics produce line broadenings/asymmetries
similar to the jet velocity (or terminal velocity of the outflow)
in several sources.
The interpretation of the profiles requires a detailed com-
parison with theory, since the kinematics of MHD winds
from rotating structures is very complex. Three basic mo-
tions overlap: rotation, acceleration along the axis, and radial
expansion from the axis (see Fig. 6). Each kinematical com-
ponent dominates at different locations in the outflow. Rota-
tion is dominant close to the source of the outflow. Further
out, radial expansion is the most significant component up
to some height, z0, above the disk. For z>z0, the dominant
kinematical component is acceleration along the disk axis,
e.g., a collimated outflow or jet. For standard parameters, the
base of the wind is unresolved, z0=5 AU (see G´omez de
Castro and Ferro-Font´an, 2005). Thus, the only way to track
the velocity law of the wind is by a clever selection of spectral
indicators based on the thermal properties of the wind.
From a theoretical point of view, three possible types
of outflows can be fitted into this broad context: a stellar
wind, a disk wind, and an outflow driven from the inter-
face. Either centrifugal stresses or magnetic/thermal pressure
Springer

42 Astrophys Space Sci (2006) 303:33–52
Fig. 7 CIV,CIII] and C II] line
profiles for the unresolved
z<12 AU region. The profiles
are plotted for inclinations 0◦,
30◦,45
◦,60
◦and 90◦from
bottom to top
are involved in the acceleration of the outflows, but it re-
mains unclear which mechanism is dominant, whether it is
universal, and how this mechanism acts when radiation pres-
sure becomes significant as in Herbig Ae/Be stars. Numerical
simulations predict temperatures between ∼10 000 K for the
inner-disk winds up to ∼106–107K close to the magnetic
reconnection boundary (Goodson et al., 1997). The X-rays
(and the high-energy particles) produced in the reconnection
areas will be redistributed towards lower energies due to the
densities involved. Thus, the dominant radiative output from
TTS winds is expected in the UV range, as is observed. UV
line profiles calculated using a simplified warm disk wind
model are shown in Fig. 7. Notice that very broad or double
peaked profiles cantered at the rest wavelength of the line can
be produced in the wind (not only in the accretion disk) pro-
vided that the inclination is ∼90◦. These type of profiles have
been observed in some TTSs such as AK Sco or RW Aur. As
shown in Fig. 7, a clever selection of the UV spectral indi-
cators helps to dissect the kinematical structure of the wind.
The effect of internal wind extinction by dust lifted from the
disk mid-plane can also be traced through the flux ratios of
relevant lines.
Another important aspect of TTS winds is that a signifi-
cant fraction of the mass outflow is ejected in a non-stationary
manner. The timescales for these ejections range from a few
hours (Alencar, 2001; Bouvier et al., 2003; G´omez de Castro
and Verdugo, 2003b) to some 10 years (optical jets observa-
tions, see, e.g., L´opez-Mart´
in et al., 2003) or even to some
hundreds of years (molecular gas bullets, see, e.g., Bachiller,
1996). Recent observations have established a lower limit of
about 1 h, precluding the association of flares with the few
hours timescale variability in RW Aur, since the characteris-
tic decay time of flares in active stars is some few hundreds
of seconds (G´omez de Castro and Verdugo, 2003b). Several
ejection timescales typically coexist in the same object. For
instance, timescales of ∼1h, ∼5.5 days and ∼20 years are
observed in RW Aur. Despite the wealth of valuable infor-
mation that HST/STIS could have obtained to determine the
kinematics and properties of these outflows, the available
observations are few and not of long duration.
Springer

Astrophys Space Sci (2006) 303:33–52 43
3.2. UV observations of Herbig-Haro objects and jets
Observations of protostellar jets can provide important clues
concerning the collimation mechanism: in particular, the
role of episodic ejections or internal shocks in the jet, heat
dissipation, and major heating sources. Also, such obser-
vations provide important clues concerning the interaction
between the jet and the surrounding molecular gas, which
helps to discriminate between radiation-induced (photodis-
sociation) versus collision-induced (shocks) in the circumjet
chemistry.
Since early in the IUE project, it has been known that
protostellar jets and Herbig-Haro objects have a higher de-
gree of ionisation than previously inferred from optical data
(Bohm-Vitense, et al., 1982; Schwartz et al., 1985; see also
G´omez de Castro and Robles, 1999 for a compilation). High-
excitation objects like HH1 or HH2 produce strong emission
lines of C IV 1548, 1550 ˚
A, O III] 1664 ˚
A, Si III] 1892 ˚
A,
and C III] 1909 ˚
A (Ortolonai and D’Odorico, 1980). How-
ever, low-excitation objects like HH43 or HH47 are charac-
terized by the presence of the H2Lyman band emission lines
(Schwartz, 1983). H2emission lines, which are pumped by
the UV radiation generated in the internal shocks of the jet,
can measure the strength of the radiation field generated in
the shock and the Lyαline strength.
UV lines are variable and the variations of the
low-ionisation species are anticorrelated with the varia-
tions of the high-ionisation species and the short-wavelength
continuum. A detailed study of HH29 combining optical and
UV data led Liseau et al. (1996) to propose a two-phase model
with a warm component (T=104K and ne=103cm−3) and
a hot, dense component (T=105K and ne=106cm−3) with
a very small filling factor (0.1–1%).
Using IUE, Bohm et al. (1987) and Lee et al. (1988) de-
tected a variable and spatially extended short-wavelength
(1300–1500 ˚
A) UV continuum. At the end of the IUE mis-
sion, it was believed that the most likely mechanism for
its formation was continuum H2emission formed when H2
molecules are destroyed either by photodissociation by radi-
ation shortwards of 912 ˚
A or by collisions with low-energy
thermal particles. This assumption was based on the ab-
sence of the dominant Lyman band features (at 1258, 1272,
1431, 1446, 1505, 1547 and 1562 ˚
A), which are detected in
such low-excitation objects as HH43 or HH47. The high-
resolution spectra of HH2 obtained with the Hopkins Ultra-
violet Telescope (HUT) showed that the UV emission be-
low 1620 ˚
A is mostly produced by H2Lyman bands de-
tected below 1200 ˚
A and at 1510, 1580 and 1610 ˚
A (see
Raymond et al., 1997). Unfortunately, this is the only high-
resolution spectrum of a HH object obtained in the far
UV. HH47 was observed with HST/GHRS but the spec-
tral coverage was tiny 1262–1298 ˚
A (Curiel et al., 1995).
A low-dispersion (1000) spectrum of HH47 obtained with
HST/FOS/G270H shows no significant depletion of Fe in the
outflow.
The existing UV observations have left open many im-
portant questions that cannot be solved without further UV
observations. It is still unclear, how the kinetic energy of
the flow is damped into radiation. The non-detection of
OVI emission (Raymond et al., 1997) and the simultane-
ous detection of strong C IV and H2emission represent the
strongest and most promising arguments against radiative
cooling models. Another important question is how to un-
derstand the excitation mechanism of the H2line radiation,
since H2band structure is observed in high-excitation HH
objects. One proposed suggestion is collisional pumping of
the H2levels (Raymond et al., 1997).
3.3. Herbig Ae/Be stars
Herbig Ae/Be stars are intermediate-mass (2–10 M) PMS
stars. They are rather puzzling objects. Their larger masses
suggest that the gravitational collapse is superalfv´enic, so
magnetic fields are not expected to be strong. However,
Ae stars have a rich UV emission-line spectrum consistent
with the presence of a chromosphere above the photosphere
(Brown et al., 1996; Deleuil et al., 2005). Also, overionised
species (transition region or corona-like) are observed; a
marginal detection of magnetic fields has been reported for
HD 104237 (Donati et al., 1997). Thus, observations point out
that fields are present, at least, during the first ∼5×106years
of their PMS evolution.
UV-optical monitoring campaigns discovered azimuthal
structures in the wind of AB Aur (Praderie et al., 1986). Sub-
sequent optical monitoring campaigns, for example by the
MUlti-SIte COntinuous Spectroscopic (MUSICOS) consor-
tium, confirmed the presence of such azimutal structures in
the wind and in the chromosphere: the rotation period of
the chromospheric structures is 32 h (the stellar rotation pe-
riod), while the rotation period of the wind (traced by the
UV Mg II lines) is 45 h (Bohm et al., 1996; Catal´a et al.,
1999). Further UV observations detected clumps of very hot
gas, traced by N Vemission, in the wind of AB Aur (Bouret
et al., 1997). The generation of these azimuthal structures
and the very hot clumps is often interpreted by means of a
two-component wind model in some ways similar to the solar
wind consisting of the following:
A “slow”, dense outflow reaching terminal velocities of
∼300 km s−1, which produces the prominent P-Cygni pro-
files observed in the Ca II and Mg II[us1] lines and the
broad, blue-shifted absorption observed in C IV[uv1].
Mass-loss rates derived from semi-empirical models are a
few ×10−8Mper year (Bouret and Catal´a, 1998; Catal´a
and Kunasz, 1987).
Springer

44 Astrophys Space Sci (2006) 303:33–52
A “high” velocity component made by streamers of magnet-
ically confined gas.
Since Herbig Ae/Be stars are fast rotators, gas in the
streamers is forced to corotate up to the alfv´en point and
shocks are expected to occur between the “slow” and “fast”
components. As a result, dense azimutal structures are
formed in the corotating interaction regions (CIRs). The
existence of a magnetic collimator is further supported by
the detection of low-density Lyαjets in HD 163296 and
HD 104237 (Devine et al., 2000; Grady et al., 2004). Mag-
netic field dissipation also seems the most likely source for
the radiative losses in the chromosphere/wind that repre-
sent 4–8% of the stellar bolometric luminosity according to
Bouret and Catal´a (1998), although accretion flows may be
a non-negligible energy source (Blondel et al., 1983; Bouret
and Catal´a, 2000; Roberge et al., 2001). Accretion, how-
ever, may not be the driver of the outflow. Radiatively driven
winds are able to produce collimated outflows provided there
is a magnetic field (Sakurai, 1985; Rotstein and Gim´enez
de Castro, 1996). The ultimate source of the field remains
unidentified: turbulence and rotation could set up a dynamo
in the outer layers. Turbulence can be generated by stellar
pulsation; radial and non-radial modes have been detected
with periods from some tens of minutes to hours (see Catal´a,
2003 for a review). Also, the rotational braking produced by
the strong stellar wind could induce turbulent motions be-
low the stellar surface, forcing magnetic fields into the outer
stellar layers (Lignieres et al., 1996).
Future UV observations are required to characterize the
winds, to study the evolution of the hot clumps (presumably
shocks between the fast and slow components), and to study
the physical conditions of the disks. Herbig Ae/Be stars are
evolutionary precursors of the Vega-like stars, such as βPic-
toris (Vidal-Madjar et al., 1998), and thus are ideal laborato-
ries for studying planet formation. Radio and IR observations
are well suited to map the extended disk structure. Optical
and IR coronographic observations with the HST have pro-
vided high-quality images of the disk structure at large (e.g.,
Clampin et al., 2003). However, UV spectroscopy is the most
sensitive tool for determining the column density of hydrogen
and the fraction of hydrogen atoms in molecular form (see,
for instance, Bouret et al., 2003; Grady et al., 2005). This
high sensitivity is illustrated by the very low upper limits
provided by FUSE to the H2abundance of the circumstellar
disk surrounding βPictoris (Lecavelier des Etangs, 2001).
FUSE observations have allowed to detect H2in several disks
surrounding Herbig Ae/Be stars, providing estimates of tem-
perature, density and physical size of the emitting region (see,
e.g., Roberge et al., 2001; Lecavelier des Etangs et al., 2003).
Measurements of H2abundances can provide information on
the rate of H2formation on dust grains and the strength of
UV photoionising radiation.
3.4. In summary: The potentials of the UV
It is widely believed that infrared and radio wavelengths are
the most appropriate regimes for studying the formation of
stars. This perception is based on PMS stars typically be-
ing embedded in molecular clouds that strongly attenuate
UV and optical radiation. This perception is true for the very
early stages of star formation, but after about 1 Myr extinction
generally is low and TTSs are accessible to the very power-
ful UV diagnostic tools. To study the evolution of TTSs and
Herbig Ae/Be stars after 1 Myr is very important, because
planets are formed at this time and the inner disk can be ob-
served while the planets are forming. Also, the basic engine
that regulates the formation of stars, the accretion–outflow
engine, is naked at this time and can be properly observed.
Examples of some of the diagnostic capabilities of high-
resolution UV spectroscopy and monitoring are outlined in
Figs. 8 and 9.
UV line profiles can clearly disentangle accretion from out-
flow.
Figure 8 shows some UV lines in the spectrum of RY Tau.
The Fe II] 2506 ˚
A line shows a broad red-shifted pro-
file with a sharp edge at zero velocity. Since this line is
pumped by Lyαphotons, it should be formed in the ac-
cretion flow. The [O II] 2471 ˚
A line is blue-shifted and
traces the wind (G´omez de Castro and Verdugo, 2005).
Both lines are optically thin with no self-absorption. For
comparison, the strong Mg II 2796, 2802 ˚
A lines display
absorption components at the wind velocity and an ex-
tended red wing associated with the accretion flow. These
two physical components could not have been disentan-
gled from the analysis of the Mg II profile alone. In fact,
the long red-shifted tail would likely have been interpreted
as a signature of line saturation.
UV monitoring can be used to study the distribution of matter
in the circumstellar environment.
In the solar system, there are three very different types of
“flares”, which are sudden increases of the high-energy
radiation and high-energy particles flux: magnetic flares
(magnetic reconnection events), corotating interaction re-
gions or CIRs (shock fronts formed by the interaction be-
tween the slow and the fast component of the solar wind),
and coronal mass ejections. This classification also ap-
plies to TTSs and their circumstellar environments. High-
resolution UV spectroscopic monitoring is required to dis-
entangle the possible mechanisms for flares in protostellar
systems. This is feasible as shown in Fig. 9. AB Dor, a
nearby 30 Myr old star, is the only young star that has
been well monitored in the UV for flares. Nine events
were detected during the roughly 10 h of monitoring with
HST/GHRS! The C IV and Si IV UV line profiles pro-
duced by most of the events are narrow and red-shifted,
Springer

Astrophys Space Sci (2006) 303:33–52 45
Fig. 8 Profiles of some relevant
UV lines in the spectrum of RY
Tau. The C II] lines are a
multiplet with many transitions
producing this peculiar profile.
All of the lines, except for Mg II,
are forbidden or semi-forbidden
(from G´omez de Castro and
Verdugo, 2005)
Fig. 9 The C IV 1548 ˚
A profile
of AB Dor during a normal
stellar flare (left) and a transient
feature probably associated with
a CIR (right). Both events lasted
several kiloseconds. The left
profile is typical of three events
that occurred during the short
monitoring time, while the
profile on the right was observed
only once. Note the presence of
a narrow absorption and the very
broad line wings in the right
panel profile (see G´omez de
Castro, 2002 for more details)
indicating hot gas falling onto the star during the flare.
However, the strongest event produced a very broad pro-
file with narrow absorption slightly blue-shifted. This pro-
file lasted a few kiloseconds and thus the broad wings are
most likely tracing the front shock of a CIR (G´omez de
Castro, 2002).
In summary, IUE and HST (with its GHRS or STIS ul-
traviolet instruments) and FUSE have allowed us to begin to
grasp the enormous potential of the UV spectral range for the
study of the physics of accretion and outflow, including the
properties of the inner region of protoplanetary disks. Un-
fortunately, fewer than 10 TTSs were observed with spectral
resolution ∼50 000 (6 km s−1) during the lifetime of these
instruments, partly because HST/STIS was not sufficiently
sensitive. A UV instrument with sensitivity 50–100 times that
of HST/STIS would permit observations of about 100 TTSs
with magnitudes 10–13 located within 160 pc of the Sun. This
sensitivity limit would permit observations of much fainter
and more evolved WTTSs than was possible with HST/STIS.
Springer

46 Astrophys Space Sci (2006) 303:33–52
Fig. 10 Typical absorption
spectrum of an Earth-like planet
transiting in front of a solar-type
star from the UV to the near
infrared. We assumed the same
atmospheric structure as for the
Earth with, e.g., similar ozone
content. Thin solid, dashed and
dotted lines represent H2O, O2
and O3absorption. The vertical
scale represents the occultation
by the planet atmosphere of the
stellar flux during the transit
from Ehrenreich (2005)
4. Characterization of extrasolar planetary
atmospheres and the search for bio-markers
Since the mid-1990s, more than 100 extrasolar planets (here-
after called “exoplanets”) have been discovered. In the com-
ing decade, several observing programs will lead to the dis-
covery of an extremely large number of exoplanets. To ac-
quire a revealing picture of these new worlds, we need de-
tailed observations of a large sample of these exoplanets
to characterize planetary atmospheres well beyond the so-
lar neighbourhood with reasonable exposure times. The ob-
servation of UV and optical absorption when an exoplanet
transits its parent star is a very powerful diagnostic technique;
in fact, the most powerful technique for detecting Earth-like
planets because of the strong absorption of stellar UV pho-
tons by the ozone molecule in the planetary atmosphere (see
Fig. 10). Near future space missions including Corot, Kepler
or GAIA will lead to the discovery of a large number of ex-
oplanets transiting their parent stars. An adequate capability
for UV spectroscopic observations will be needed for de-
tailed follow-up observations to characterize the exoplanets,
their atmospheres, and their satellites.
4.1. The physical processes controlling the formation
and evolution of exoplanets
Since the unexpected discovery of the first hot-Jupiter by
Mayor and Queloz (1995), it is clear that exoplanets are an
extremely diverse group. With the discovery of more than
100 exoplanets, this diversity is clearly seen in their orbital
properties. We have “hot-Jupiters” with orbital periods as
short as 3 days, and several “very hot-Jupiters” with orbital
periods even shorter than 2 days. Less massive exoplanets
have recently been discovered (Santos et al., 2004; McArthur
et al., 2004; Butler et al., 2004), and the discussions on their
true nature show that a large variety is now expected and
certainly possible.
The same variety is also expected for the atmospheres of
these exoplanets. A quick look at the atmospheric content and
history of the solar system’s terrestrial inner planets shows
that with four planets, we find four very different possibili-
ties: Mercury has almost no atmosphere, Mars’ atmosphere
is tenuous with atmospheric pressure about one-hundredth
that of the Earth, and Venus is the extreme opposite with
more than 90 times the atmospheric pressure of the Earth.
Note that Titan, although much smaller than the Earth, also
has an atmosphere of 1.5 bar and is very different from other
giant planets satellites without atmospheres.
This diversity shows how difficult it is to predict what
should be the content of an exoplanet’s atmosphere. In the
solar system, the terrestrial atmosphere is unique with abun-
dant O2and O3produced by biological activity. Another
important characteristic of the terrestrial atmosphere is the
significant amount of water. The Earth and Titan both have
much N2in their atmospheres, but Titan has more methane
and no O2. Mars and Venus have similar atmospheric com-
position, but their total amounts are in a ratio of more than
104.
Springer

Astrophys Space Sci (2006) 303:33–52 47
Thus, there is no simple answer to the question of the
expected characteristics of planets and their atmospheres.
On the one hand, the solar system planets provide a first
hint of the expected diversity of the exoplanets and their
atmospheres. On the other hand, observations of exoplanets
and the detailed characterization of their atmospheres will
help us better understand the physical processes at work in
the building of a planet and its atmosphere.
It is clear that the detailed processes that created the so-
lar system planets is still a matter of debate and the im-
pact of many processes must still be clarified. In short, we
do not yet know the key physical parameters that govern
the formation, evolution and fate of a given planet and its
atmosphere.
How do the properties (temperature, stellar type, high-
energy particles, and metallicity) of the central star alter the
evolution of its planetary system? What effects do a planet’s
orbital parameters (orbital distance and eccentricity) have on
its size, mass and potential migration during the formation
process? Are there volatile-rich planets like the proposed
“Ocean-planets”? How do interactions with other planets and
planetesimals in their environment influence the evolution of
a planet? This last question is undoubtedly related to the
origin of water on the Earth. Are water-rich planets in the
“habitable zone” common, rare, or exceptions?
Several processes that are believed to play key roles in
building a planet can now be identified. To begin, we can look
at the best known planet, the Earth. Although still controver-
sial, it is generally accepted that the Earth’s original atmo-
sphere was accumulated simultaneously with the planet’s for-
mation. However, the heating of the atmosphere by the young
Sun’s UV and X-ray flux led to the hydrodynamical escape
of this primary atmosphere (as is observed on HD 209458b,
Vidal-Madjar et al., 2003, 2004). Tectonics, volcanism and
the planet out-gasing then formed the secondary atmosphere
in which we now live. Late bombardment by planetesimals
in the young planetary system contributed to a large fraction
of the terrestrial water but the fraction of water coming from
the Earth itself and the outside contribution is still a matter of
debate. Finally, life enriched the atmosphere in O2and ozone,
which are therefore considered as atmospheric bio-markers.
The observation of O2and ozone in the atmosphere of the
Earth or any exoplanet can lead to the conclusion that some-
thing very particular is happening there. Something which
could suggest the presence of life.
It therefore appears that we will soon discover many more
exoplanets, each one likely different from the others. As soon
as we will be able to characterize them in detail, there will
likely be many surprises. We cannot predict what will be
discovered, but this will be an unprecedented opportunity to
better understand the key processes at work in the shaping of
planets and, in particular, to better understand the origin of
our own Earth.
4.2. Ultraviolet observations of transiting planets
In the coming years, many exoplanets will be discovered
through transits, for example, by the Corot, Kepler and GAIA
missions. In particular, GAIA will likely identify thousands
of exoplanets transiting bright stars. These will be prime tar-
gets for detecting the atmospheric constituents through ab-
sorption line spectroscopy, thereby characterising the chem-
ical and physical properties of the atmosphere, including the
search for bio-markers.
Many molecules have strong electronic transitions in the
UV-optical domain. This wavelength range gives access to
the most important constituents of the atmospheres. In par-
ticular, bio-markers like ozone (O3) have very strong tran-
sitions in the ultraviolet (the absorption of UV radiation by
the Earth atmosphere is primarily due to O2and O3). The
Hartley bands of O3are the main absorbers at 2000–3000 ˚
A.
O2has strong absorptions in the range 1500–2000 ˚
A. CO has
strong bands below 1800 ˚
A, and weaker Cameron bands from
1800 to 2600 ˚
A. The CO+first negative bands are located in
the 2100–2800 ˚
A range. Finally, the presence of CO2can be
detected through the CO+
2Fox–Duffenback–Barker bands
from 3000 to 4500 ˚
A. We note also the important presence of
OIand C II lines at 1304 and 1335 ˚
A. Observation of these
species can be done easily, demonstrating that these atoms
and ions are present at very high altitude (several hundreds
of kilometres) and providing large absorption depths.
The electronic molecular transitions are several orders of
magnitude stronger than the vibrational or rotational transi-
tions observed in the infrared or radio range. These transitions
can be observed in absorption provided there is a sufficiently
strong UV background source. For this reason, the observa-
tion of UV and optical absorption when a planet transits its
parent star is intrinsically a more powerful diagnostic tech-
nique to characterize the atmospheres of the inner planets
than infrared observations of planetary emission. This is es-
pecially true for studying small Earth-like exoplanets. It is
far simpler to use the large number of photons in the stellar
continuum that are absorbed in spectral lines or molecular
bands by the planet’s atmosphere than to attempt to cancel
the huge stellar photon flux by a coronograph or interferom-
eter to search for the faint infrared emission (thermal and
scattered starlight) from the planet. In addition, the intrinsic
faintness of the target sources enhances potential difficulties
like confusion with exozodiacal light. Moreover, spectral ob-
servations of the atmospheres of satellites of giant planets,
suspected to be numerous, are even more difficult in emis-
sion.
The first observations of the atmosphere of an exoplanet
(HD 209458b, nicknamed Osiris) have been made through
UV-optical spectroscopy (Charbonneau et al., 2002; Vidal-
Madjar et al., 2003, 2004), demonstrating that it is an ideal
tool for probing the atmospheric content of transiting planets.
Springer

48 Astrophys Space Sci (2006) 303:33–52
It is noteworthy that we already have in hand four observa-
tions of the atmosphere of an exoplanet (including the de-
tection of oxygen). All of these present detections have been
performed: (1) in space, (2) in the UV-optical wavelength
range, and (3) in absorption during planetary transits. This
is not a coincidence but rather a consequence of this method
being the most powerful and its presenting the best trade-off
of scientific result versus technical feasibility.
4.2.1. Estimates of the expected detection rate for
Earth-like exoplanets
For a typical life-supporting Earth-like planet, the ozone layer
is optically thick to UV radiation incident at a grazing angle
up to an altitude of about 60 km. This ozone layer creates
an additional occultation depth of F/F∼2×10−6over
hundreds of Angstroms that can be compared to the 2 ×10−4
optical depth over 1 ˚
Adetected with HST on HD 209458.
We can estimate the minimum brightness of the parent star
(Fmin) relative to the brightness of HD 209458 (FHD 209458)
for a detection of the ozone absorption. We have:
Fmin =F/F
2×10−4−2λ
1˚
A−1S
SHST/STIS −1
FHD209458 .
With a telescope 50 times as sensitive as HST/STIS, ozone
can thus be detected in Earth-like exoplanets orbiting stars
brighter than V≈10 (easily identified by GAIA). This mag-
nitude corresponds to star at a distance d∼50 pc for the latest
type stars considered (K Vstars) and more than ∼500 pc for
the earliest stars (F Vstars).
With this estimate of the minimum stellar brightness
needed to detect a given species, we can evaluate the number
of potential targets. The number of targets with exoplanets for
which we can probe the atmospheric content, Npl , is simply
the total number of stars brighter than the limit N∗, multi-
plied by the fraction of stars having an identified transiting
exoplanet at a given orbital range, Ppl.
Npl =N∗×Ppl.
To evaluate the total number of stars, we must select the stel-
lar types to be included. The usual assumption is to limit
the estimates by counting only K-, G- and F-type main se-
quence stars. This is a conservative assumption based on the
bias of the present discoveries of exoplanets. We used the
conservative estimate of
N∗=48 000 ×100.6∗(V−10) .
The second term of the equation, Ppl, is more difficult to
quantify due to many unknowns: (i) there must be an exo-
planet orbiting the star, (ii) this exoplanet must be identified,
and (iii) it must be transiting the stellar disk. We made the as-
sumption that about 25% of stars will have identified orbiting
exoplanets in the near future. The probability of a transit for
a given exoplanet can be estimated to be Ptr ≈a/R∗, where
ais the orbital distance and R∗is the stellar radius. For an
exoplanet orbiting at 1 AU around a solar-type star, the prob-
ability is Ptr ∼0.5%. If we consider the habitable zone as the
most interesting orbital range, this probability increases for
the smaller and more numerous stars. Ptr ∼0.5% can thus
be considered as a conservative assumption for the habitable
zone around solar-type stars.
Finally, we have an estimate of the number of targets with
observable planets:
Npl ≈60 ×100.6∗(V−10).
A combination of this last equation with the minimum bright-
ness of the parent star needed to detect a given species, gives
the size of the exoplanet sample that a telescope can analyse
as a function of its sensitivity.
Similar calculations can be made for exoplanets very dif-
ferent from the Earth. The occultation depth is proportional to
the planetary radius (Rp) multiplied by the atmospheric scale
height (h), and the scale height is inversely proportional to
the planet’s gravity (g∝Mp/R2
p). These considerations de-
termine that the occultation depth is related to the planet’s
density (ρp)by
F/F∝R3
p/Mp∝ρ−1
p.
Hence, low-density exoplanets and planetary satellites will
give larger absorption depths than the high-density Earth (see
Fig. 11). In short, it is easier to probe the atmosphere through
transit spectroscopy in the case of Ocean-planets because
they are larger, or in the case of Titan-like satellites, because
they are less massive than the Earth and have larger atmo-
spheric scale heights. With an ozone layer similar to that in
the Earth’s atmosphere, the occultation depth of an Ocean-
planet, or a Titan-like satellite, will be F/F∼5×10−6.
The resulting number estimates of possible detections as a
function of the telescope sensitivity are given in Fig. 12.
Estimates of the minimum star brightness for detecting
atmospheric signatures can be translated into the number of
possible detections of such atmospheric signatures in tran-
siting exoplanets (nd). The number of possible detections is
related to the number of stars brighter than the minimum
brightness [n(F>Fmin)], the probability of finding an exo-
planet around these stars (Pp), and the transit probability at
a given orbital distance [Pt(d)]. Ppis unknown and we con-
sider Pp=0.25 as a reasonable number. We also estimate the
transit probability at 1 AU: Pt(1 AU) =0.5%. For the num-
ber of stars at a given brightness, we restrict the sample to
Springer

Astrophys Space Sci (2006) 303:33–52 49
Fig. 11 Typical absorption
spectrum of an Ocean-planet as
estimated by Leger et al. (2004)
when seen transiting in front of
a solar-type star. We assumed
the same atmospheric structure
as for the Earth but rescaled the
structure to the size and density
as presently expected for an
6 Earth mass and 2 Earth radius
volatile-rich exoplanet from
Ehrenreich (2005)
Fig. 12 Plot of the number of
expected detections of
atmospheric signatures as a
function of telescope sensitivity
for Earth-like and Ocean-like
planets
K-, G- and F-type main sequence stars. With these numbers,
we plot in Fig. 12 the number of possible detections of atmo-
spheric signatures as a function of telescope sensitivity. For a
UV telescope, the sensitivity depends on the mirror size and
the spectrograph efficiency. From now on, we quote the tele-
scope sensitivity (S) in units of HST/STIS sensitivity because
planetary atmospheres have already been detected and stud-
ied with this instrument and, henceforth, we are not limited
to a theoretical calculation that may ignore some potential
difficulties. We note that a 2-m class telescope including a
spectrograph with the efficiency of the COS instrument leads
to a sensitivity S∼20 HST/STIS.
We see that about 100 exoplanets are expected to transit
in front of K-, G- or F-type main sequence stars brighter than
V=10. In conclusion, using the transit probabilities in the
habitable zone, we find that the presence of bio-markers and
Springer

50 Astrophys Space Sci (2006) 303:33–52
other constituents in the atmospheres can be searched for in
more than about 100 Earth-like exoplanets orbiting K-, G-
and F-type main sequence stars. Further considerations to be
taken into account are given as follows:
The effect of stellar variety. The number estimates given ear-
lier is probably conservative. Indeed, we neglected the
stellar type in the estimates and considered the real ob-
servations of HD 209458b as a benchmark. However,
HD 209458 is a G-type star. A very large number of tar-
gets will be exoplanets orbiting K-type stars. For these
stars, the stellar radius is smaller and the absorption depth
due to the transiting planets will be larger (as observed
in the case of the recently discovered exoplanet TrES-1
transiting a K0 Vstar (Alonso et al., 2004)).
Moreover, late-type stars are expected to have habit-
able zones at orbital distances smaller than 1 AU assumed
in the earlier calculation. With smaller orbital distances,
the transit probability and the corresponding number of
targets increase. Since the previous calculation was done
for G-type stars, we expect a larger number of detections
for K-type stars. The final number of possible detection
of water bands should therefore be larger than is shown in
Fig. 12. Water bands are likely detectable in a reasonable
number of Earth-like exoplanets with a 50–100 HST/STIS
sensitivity telescope.
The spatial structure of the atmosphere can be studied by
time-tagged observations. Absorption spectroscopy of
transiting planets can also provide spatial information on
the physical and chemical properties of their atmospheres.
During partial phases when the planet partially covers the
observed stellar disk, time-tagged spectra provide a spa-
tial scan of the exoplanet’s atmosphere. The partial phase
lasts about 10 min for an Earth-size exoplanet orbiting at
1 AU from its parent star. For the closest stars (≤100 pc),
exposures of a few minutes will identify the atmospheric
diagnostics of the most important constituents. Detailed
time analysis of transit spectra can give information on the
spatial distribution of atmospheric characteristics along
the exoplanet’s surface, for example, the difference be-
tween poles and equator or the spatial inhomogeneity of
different chemical constituents.
5. Summary: The needed capabilities
The scientific program outlined in this article requires a broad
range of instrumentation from imaging to spectroscopic ca-
pabilities.
5.1. Imaging
Two imaging instruments are required: a large field-of-view
and a high-resolution imager.
The large field-of-view instrument will be used for two
main purposes: mapping of protostellar jets and tracking flare
timescales over large fields.
The high-resolution imager’s primary use is to resolve
the cooling structures of jets and to map protostellar disks.
Coronography is required. UV observations provide the
best contrast for detecting structures around young stars; for
instance, a Herbig Ae/Be star is a 100 times fainter at Lyα
than at Hα. Narrow filters cantered in the most prominent UV
lines like Lyα,CIV,CIII,CII,OI,HeII or O VI are required.
5.2. Spectroscopy
Most of the science program is oriented towards spec-
troscopy. Two basic modes are required: high-resolution
spectroscopy and medium-to-low long-slit spectroscopy.
High-resolution spectroscopy (R50 000) is required
for the Doppler mapping of circumstellar structures, flares,
winds and disks. It is also required for detailed studies of the
ISM. The spectral resolving power required to observe the
atmosphere of exoplanets is not a crucial capability. We have
seen that R=λ/λ =10 000 is more than enough. Even
lower resolving power, R1000, could be enough to detect
the broad-band signatures of many molecules.
Note, however, that in some cases higher resolving power
will resolve the thermal broadening of absorption lines in
planetary exospheres (Vidal-Madjar et al., 2003). In that case,
a high resolving power of R∼100 000 will provide impor-
tant constraints on the atmospheric structure.
Long-slit spectroscopy is required to map the spatially-
resolved jet emission, disks and circumstellar envelopes.
Spectral resolution as high as 10 000 is required.
Wavelength coverage. The target spectral range for the spec-
troscopic instruments goes from ∼1000 ˚
A (to include
the O VI lines and the H2bands) to ∼4000 ˚
Atohave
some overlap with optical telescopes and to cover most of
the molecular broad-band absorption expected from exo-
planet atmospheres. Extension to 10 000 ˚
A would provide
access to the strong water band, which is of prime interest
for the search, statistics and characterization of habitable
exoplanets and, consequently, for exobiology.
Sensitivity. An improvement by a factor of 20–100 over
HST/STIS capabilities will permit the study of the warm
ISM beyond the Local Bubble and observe gas high in the
halo towards the HVCs. It will also increase the sample
of TTSs observed in the UV from some 10 to about 200
including the WTTSs providing, for the first time, an un-
biased view of the accretion–outflow engine during PMS
evolution.
Springer

Astrophys Space Sci (2006) 303:33–52 51
The sensitivity of the spectrograph should be high around
prominent nebular lines like C III] 1909 ˚
A, Si III] 1892 ˚
A,
and C II] 2325 ˚
A.
Time-tagged observations. Accurate time information is es-
sential. The absolute accuracy of the timescale needs to
be precise to coordinate monitoring campaigns with other
instruments or to study exoplanet transits. The accuracy
and uniformity of the timing sets the spatial resolution for
Doppler mapping. Time-tagged observations can be con-
sidered as a proxy for spatial resolution at the level of the
exoplanet’s size.
5.3. Orbit
The orbit should permit efficient observations. A long-period
orbit will allow long uninterrupted observing with few Earth
occultations, little airglow pollution, and minimal geocoronal
emission. This will facilitate long-duration flare observations
and Doppler mapping on timescales of 1 day. An L2 orbit is
optimal for this purpose.
Acknowledgements This work has been supported by the European
Commissions 6th Framework programme under contract number RII3-
Ct-2004-001566 to the OPTical Infrared CO-ordination Network for
Astronomy (OPTICON). The authors are member of the Network for
UltraViolet Astrophysics (NUVA): this network is defined within the
OPTICON activities to structure the European astronomy. AIGdC ac-
knowledges support by the Ministry of Science and Technology of Spain
through grants AYA 2000-966, ESP2001-4637E and ESP2002-10799-
E. JLL acknowledges support by NASA through grant AR-09930. M.A.
would like to thank Dieter Breitschwerdt for enlightening discussions
on non-equilibrium ionisation.
References
Alencar, S.H.P., Johns-Krull, C.M., Basri, G.: Astrophys. J. 122, 3335
(2001)
Ardila, D.R., Basri, G., Walter, F.M., Valenti, J.A., Johns-Krull, C.M.:
Astrophys. J. 566, 1100 (2002)
Avillez, M., Breitschwerdt, D.: A&A 425, 899 (2004)
Avillez, M., Breitschwerdt, D.: A&A 436, 585 (2005)
Bachiller, R.: Annu. Rev. Astron. Astrophys. 34, 111 (1996)
Benjamin, R., Shapiro, P.: In: E. Silver and S. Kahn (eds.), Ultraviolet
and X-Ray Spectroscopy of Laboratory and Astrophysical Plas-
mas, Cambridge University Press, Cambridge, p. 280 (1993)
Bertout, C., Basri, G., Bouvier, J.: Astrophys. J. 330, 350 (1988)
Breitschwerdt D., Schmutzler T.: A&A 347, 650 (1999)
Bergh¨ofer, T.W., Breitschwerdt, D.: A&A 390, 299 (2002)
Bertin, P., Lallement, R., Ferlet, R., Vidal-Madjar, A.: J. Geophys. Res.
98(A9), 15193 (1993)
Blondel, P.F.C., Talavera, A., Djie, H.R.E.T.A.: A&A 268, 624 (1993)
Boehm, T.,Catala, C., Donati, J.-F., Welty, A., Baudrand, J., Butler, C.J.,
Carter, B., Collier-Cameron, A., Czarny, J., Foing, B., Ghosh, K.,
Hao, J., Houdebine, E., Huang, L., Jiang, S., Nefl, J.E., Rees, D.,
Semel, M., Simon, T., Talavera, A., Zhai, D., Zhao, F.: Astron.
Astrophys. Suppl. 120, 431 (1996)
Boehm, K.-H., Buehrke, Th., Raga, A.C., Brugel, E.W., Witt, A.N.,
Mundt, R.: Astrophys. J. 316, 349 (1987)
Boehm-Vitense, E., Cardelli, J.A., Nemec, J.M., Boehm, K.H.: Astro-
phys. J. 262, 224 (1982)
Bouret, J.-C., Catala, C.: A&A 340, 163 (1998)
Bouret, J.-C., Catala, C.: A&A 359, 1011 (2000)
Bouret, J.-C., Catala, C., Simon, T.: A&A 328, 606 (1997)
Bouret, J.-C., Martin, C., Deleuil, M., Simon, T., Catala, C.: A&A 410,
175 (2003)
Bouvier, J., Grankin, K.N., Alencar, S.H.P., Dougados, C., Fern´andez,
M., Basri, G., Batalha, C., Guenther, E., Ibrahimov, M.A., Ma-
gakian, T.Y., Melnikov, S.Y., Pelrov, P.P., Rud, M.V., Zapatero
Osorio, M.R.: A&A 409, 169 (2003)
Calvet, N., Basri, G., Kuhi, L.V.: Astrophys. J. 277, 725 (1984)
Catala, C.: A&SS 284, 53 (2003)
Catala, C., Kunasz, P.B.: A&A 174, 158 (1987)
Cernicharo, J.: Astrophys. J. 608, L41 (2004)
Charbonneau, D., Brown, T.M., Noyes, R.W., Gilliland, R.L.: Astro-
phys. J. 568, 377 (2002)
Clampin, M., Krist, J.E., Ardila, D.R., Golimowski, D.A., Hartig, G.F.,
Ford, H.C., Illingworth, G.D., Bartko, F., Ben´etez, N., Biakeslee,
J.P., Bouwens, R.J., Broadhurst, T.J., Brown, R.A., Burrows, C.J.,
Cheng, E.S., Cross, N.J.G., Feldman, P.D., Franx, M., Gronwall,
C., Infanle, L., Kimble, R.A., Lesser, M.P., Martel, A.R., Menan-
teau, F., Meurer, G.R., Miley, G.K., Postman, M., Rosati, P., Siri-
anni, M., Sparks, W.B., Tran, H.D., Tsvetanov, Z.I., White, R.L.,
Zheng, W.: Astron. J. 126, 385 (2003)
Curiel, S., Raymond, J.C., Wolfire,M., Hartigan, P., Morse, J., Schwartz,
R.D., Nisenson, P.: Astrophys. J. 453, 322 (1995)
Danly, L., Lockman, F.J., Meade, M.R., Savage, B.D.: ApJSS 81, 125
(1991)
Deleuil, M., Bouret, J.-C., Catala, C., Lecavelier des Etangs, A., Vidal-
Madjar, A., Roberge, A., Feldman, P.D., Martin, C., Ferlet, R.:
A&A 429, 247 (2005)
Devine, D., Grady, C.A., Kimble, R.A., Woodgate, B., Bruhweiler, F.C.,
Boggess, A., Linsky, J.L., Clampin, M.: Astrophys. J. 542, L115
(2000)
Donati, J.-F., Semel, M., Carter, B.D., Rees, D.E., Collier Cameron, A.:
Month. Not. R.A.S. 291, 658 (1997)
Ehrenreich, D.: PhD Thesis, Univ. de Paris VI (2005)
Ferro-Fon´an, C., G ´omez de Castro, A.I.: Month. Not. R.A.S. 342, 427
(2003)
Glassgold, A.E., Najita, J., Igea, J.: Astrophys. J. 615, 972 (2004)
Grady, C.A., Woodgate, B., Heap, S.R., Bowers, C., Nuth, J.A., III,
Herczeg, G.J., Hill, H.G.M.: Astrophys. J. 620, 470 (2005)
Grady, C.A., Woodgate, B., Torres, C.A.O., Henning, Th., Apai, D.,
Rodmann, J., Wang, H.S.B., Linz, H., Williger, G.M., Brown, A.,
Wilkinson, E., Harper, G.M., Herczeg, G.J., Danks, A., Vieira,
G.L., Malumuth, E., Collins, N.R., Hill, R.S.: Astrophys. J. 608,
809 (2004)
G´omez de Castro, A.I.: Month. Not. R.A.S. 332, 409 (2002)
G´omez de Castro, A.I.: A&SS 292, 561 (2004)
G´omez de Castro, A.I. Verdugo, E.: APJ, submitte (2005)
G´omez de Castro, A.I., Fern´andez, M.: Month. Not. R.A.S. 283,55
(1996)
G´omez de Castro, A.I., Franqueira, M.: IUE-ULDA Access Guide No.
8: T TauriStars, ESA Scientific Publication, ESA-SP 1205 (1997a)
G´omez de Castro, A.I., Franqueira, M.: Astrophys. J. 482, 465 (1997b)
G´omez de Castro, A.I., Ferro-Font´an, C.: MNRAS 362, 569 (2005)
G´omez de Castro, A.I., Lamzin, S.: Month. Not. R.A.S. 304, L41 (1999)
G´omez de Castro, A.I., Robles, A.: INES Access Guide No. 1: Herbig-
Haro Objects, ESA Scientific Publication, ESA-SP 1237 (1999)
G´omez de Castro, A.I., Verdugo, E.: Astrophys. J. 597, 443 (2003a)
Goodson, A.P., Boehm, K.-H., Winglee, R.M.: Astrophys. J. 524, 142
(1999)
Goodson, A.P., Winglee, R.M., Boehm, K.-H.: Astrophys. J. 489, 199
(1997)
Springer

52 Astrophys Space Sci (2006) 303:33–52
Guenther, E.W., Lehmann, H., Emerson, J.P., Staude, J.: A&A 341, 768
(1999)
Gullbring, E., Calvet, N., Muzerolle, J., Hartmann, L.: Astrophys. J.
544, 927 (2000)
Haffner, L.M., Reynolds, R.J., Tufte, S.L., Madsen, G.J., Jaehnig, K.P.,
Percival, J.W.: Astrophys. J. Suppl. Ser. 149, 405 (2003)
Hartigan, P., Hartmann, L., Kenyon, S.J., Strom, S.E., Skrutskie, M.F.:
Astrophys. J. 354, L25 (1990)
Heyer, M., Zweibel, E.: A&SS 292, 9 (2004)
Herczeg, G.J., Linsky, J.L., Valenti, J.A., Johns-Krull, C.M., Wood,
B.E.: Astrophys. J. 572, 310 (2002)
Herczeg, G.J., Wood, B.E., Linsky, J.L., Valenti, J.A., Johns-Krull,
C.M.: Astrophys. J. 607, 369 (2004)
Johns-Krull, C.M., Valenti, J.A., Koresko, C.: Astrophys. J. 516, 900
(1999)
Kueker, M., Henning, T., Ruediger, G.: Astrophys. J. 589, 397 (2003)
Kulkarni, S., Heiles, C.: Galactic and Extragalactic Radio Astronomy,
Springer-Verlag, Berlin and New York, p. 95 (1988)
Lallement, R., Bertin, P.: A&A 266, 479 (1992)
Lallement, R., Welsh, B.Y., Vergely, J.L., Crifo, F., Sfeir, D.: A&A 411,
447 (2003)
Lauroesch, J.T., Meyer, D.M., Blades, J.C.: Astrophys. J. 543, L43,
(2000)
Lecavelier des Etangs, A., Deleuil, M., Vidal-Madjar, A., Roberge, A.,
LePetit, F., et al.: A&A 407, 935 (2003)
Lecavelier des Etangs, A., Vidal-Madjar, A., McConnell, J.C., H´ebrard,
G.: A&A 418, L1 (2004)
Lecavelier des Etangs, A., Vidal-Madjar, A., Roberge, A., Feldman,
P.D., Deleuil, M., et al.: Nature 412, 706 (2001)
Lee, M.G., Bohm, K.H., Temple,S.D., Raga, A.C., Mateo, M.L., Brugel,
E.W., Mundt, R.: Astron. J. 96, 1690 (1988)
Lignieres, F., Catala, C., Mangeney, A.: A&A 314, 465 (1996)
Liseau, R., Huldtgren, M., Fridlund, C.V.M., Cameron, M.: A&A 306,
255 (1996)
L´opez-Martin, L., Cabrit, S., Dougados, C.: A&A 405, L1 (2003)
Ma´
iz-Apell´aniz, J.: Astrophys. J. 560, L83 (2001)
Martin, C., Bowyer, S.: Astrophys. J. 350, 242 (1990)
Matt, S., Goodson, A.P., Winglee, R.M., Boehm, K.-H.: Astrophys. J.
574, 232 (2002)
McKee, C.F., Ostricker, J.P.: Astrophys. J. 218, 148 (1977)
Moos, H.W., Sembach, K.R., Vidal-Madjar, A., York, D.G., Friedman,
S.D., H´ebrard, G., Kruk, J.W., Lehner, N., Lemoine, M., Son-
neborn, G., Wood, B.E., Ake, T.B., Andr´e, M., Blair, W.P., Chayer,
P., Gry, C., Dupree, A.K., Ferlet, R., Feldman, P.D., Green, J.C.,
Howk, J.C., Hutchings, J.B., Jenkins, E.B., Linsky, J.L., Murphy,
E.M., Oegerle, W.R., Oliveira, C., Roth, K., Sahnow, D.J., Sav-
age, B.D., Shull, J.M., Tripp, T.M., Weller, E.J., Welsh, B.Y.,
Wilkinson, E., Woodgate, B.E.: Astrophys. J. Suppl. Ser. 140,3
(2002)
Muzerolle, J., Calvet, N., Hartmann, L.: Astrophys. J. 550, 944 (2001)
Ortolani, S., D’Odorico, S.: A&A 83, L8 (1980)
Petrov, P.P., Gahm, G.F., Gameiro, J.F., Duemmler, R., Ilyin, I.V., et al.:
A&A 369, 993 (2001)
Praderie, F., Catala, C., Simon, T., Boesgaard, A.M.: Astrophys. J. 303,
311 (1986)
Priest, E., Forbes, T.: Magnetic reconnection: MHD Theory and Appli-
cations, Cambridge University Press, New York (2000)
Raymond, J.C., Blair, W.P., Long, K.S.: Astrophys. J. 489, 314 (1997)
Redfield, S., Linsky, J.L.: Astrophys. J. 534, 825 (2000)
Redfield, S., Linsky, J.L.: Astrophys. J. 613, 1004 (2004)
Reynolds, R.J., Chaudhary, V., Madsen, G.J., Haffner, L.M.: Astron. J.
129, 927 (2005)
Richter, P., Savage, B.D., Wakker, B.P., Sembach, K.R., Kalberla,
P.M.W.: Astrophys. J. 549, 281 (2001)
Roberge, A., Lecavelier des Etangs, A., Grady, C.A., Vidal-Madjar, A.,
Bouret, J.-C., et al.: Astrophys. J. 551, L97 (2001)
Rotstein, N,, Gimenez de Castro, C.G.: Astrophys. J. 464, 859 (1996)
Sakurai, T.: A&A 152, 121 (1985)
Savage, B.D., Sembach, K.R.: Astrophys. J. 434, 145 (1994)
Schmutzler, T., Tscharnuter, W.M.: A&A 273, 318 (1993)
Schwartz, R.D.: Annu. Rev. Astron. Astrophys. 21, 209 (1983)
Schwartz, R.D., Dopita, M.A., Cohen, M.: Astron. J. 90, 1820 (1985)
Shapiro, P.R., Moore, R.T.: Astrophys. J. 207, 460 (1976)
Simon, T., Vrba, F.J., Herbst, W.: Astron. J. 100, 1957 (1990)
Slavin, J.D., Frisch, P.C.: Astrophys. J. 565, 364 (2002)
Tripp, T.M., Wakker, B.P., Jenkins, E.B. Bowers, C.W., Danks, A.C.,
Green, R.F., Heap, S.R., Joseph, C.L., Kaiser, M.E., Linsky, J.L.,
Woodgate, B.E.: Astron. J. 125, 3122 (2003)
Valenti, J.A., Johns-Krull, C.M., Linsky, J.L.: Astrophys. J. Suppl. Ser.
129, 399 (2000)
Vidal-Madjar, A., Lagrange-Henri, A.-M., Feldman, P.D., Beust, H.,
Lissauer, J.J., Deleuil, M., Ferlet, R., Gry, C., Hobbs, L.M., Mc-
Grath, M.A., McPhate, J.B., Moos, H.W.: A&A 290, 245 (1994)
Vidal-Madjar, A., Laurent, C., Bruston, P., Audouze, J.: Astrophys. J.
223, 589 (1978)
Vidal-Madjar, A., Lecavelier des Etangs, A., D´esert, J.-M., Ballester,
G.E., Ferlet, R., H´ebrard, G., Mayor, M.: Nature 422, 143 (2003)
Vidal-Madjar, A., Lecavelier des Etangs, A., Ferlet, R.: Planet. Space
Sci. 46, 629 (1998)
Vidal-Madjar, A., D´esert, J.-M., Lecavelier des Etangs, A., H´ebrard,
G., Ballester, G.E., Ehrenreich, D., Ferlet, R., McConnell, J.C.,
Mayor, M., Parkinson, C.D.: Astrophys. J. 604, L69 (2004)
Watson, A.J., Donahue, T.M., Walker, J.C.G.: Icarus 48, 150 (1981)
Witte, M.: A&A 426, 835 (2004)
Springer

Astrophys Space Sci (2006) 303:53–68
DOI 10.1007/s10509-005-9016-5
ORIGINAL ARTICLE
Ultraviolet Studies of Interacting Binaries
Boris T. G¨aansicke ·Domitilla de Martino ·
Thomas R. Marsh ·Carole A. Haswell ·
Christian Knigge ·Knox S. Long ·Steven N. Shore
Received: 5 May 2005 / Accepted: 28 September 2005
C
Springer Science +Business Media B.V. 2006
Abstract Interacting Binaries consist of a variety of stellar
objects in different stages of evolution and those containing
accreting compact objects still represent a major challenge to
our understanding of not only close binary evolution but also
of the chemical evolution of the Galaxy. These end-points of
binary star evolution are ideal laboratories for the study of ac-
cretion and outflow processes, and provide insight on matter
under extreme physical conditions. One of the key-questions
of fundamental relevance is the nature of SNIa progenitors.
The study of accreting compact binary systems relies on ob-
servations over the entire electromagnetic spectrum and we
outline here those unresolved questions for which access to
the ultraviolet range is vital, as they cannot be addressed by
observations in any other spectral region.
B.T. G¨aansicke ()·T.R. Marsh
Department of Physics, University of Warwick, Coventry CV4
7AL, UK
e-mail: Boris.Gansicke@warwick.ac.uk
D. de Martino
INAF – Osservatorio di Capodimonte, Via Moiariello 16, 80131
Napoli, Italy
C.A. Haswell
Department of Physics and Astronomy, The Open University,
Milton Keynes MK7 6AA, UK
C. Knigge
School of Physics and Astronomy, University of Southampton,
Southampton SO17 1BJ, UK
K.S. Long
Space Telescope Science Institute, 3700 San Martin Drive,
Baltimore, MD 21218
S.N. Shore
Dipartimento di Fisica, Universit´a di Pisa, Largo Pontecorvo 2,
56127 Pisa, Italy
Keywords Close binaries ·Cataclysmic variables ·
Symbiotic stars ·X-ray binaries ·Evolution ·Accretion
discs ·Winds ·Magnetism
1. Scientific background and astrophysical context
The 20th century saw an impressive leap in the theory of stel-
lar evolution – leading from not even knowing what source
of energy powers the Sun to the extremely detailed models of
stellar structure and evolution available today. A number of
the present-day key research areas, e.g. galaxy evolution, are
deeply rooted in our understanding of stellar evolution. How-
ever, while we may feel comfortable about our understanding
of single stars, observational evidence collected throughout
the last few decades makes it increasingly clear that the ma-
jority of all stars in the sky are born in binaries, of which
many will interact at some point in their lives (Iben, 1991).
Virtually all of the most exotic objects in the Galaxy are de-
scended from such binary stars, including binary pulsars, all
the galactic black-hole candidates, low-mass X-ray binaries
(LMXB), millisecond pulsars, cataclysmic variables (CVs),
symbiotic stars, and many others. Binary stars are important
in many other contexts, too. Sub-dwarf B stars, which now
appear to be another product of binary evolution, dominate
the ultraviolet light of old galaxies. The Type Ia supernovae,
among the most important ‘standard candles’ in the deter-
mination of extragalactic distances on a cosmological scale,
are thought to arise from exploding white dwarfs driven over
their Chandrasekhar mass limit by accretion from a compan-
ion star. Even the class of short gamma-ray bursts, the most
powerful explosions in the Universe, may be related to the
merging of two neutron stars, again products of binary star
evolution.
Springer

54 Astrophys Space Sci (2006) 303:53–68
Interacting binary stars are showcases of the processes
of mass accretion and outflow, exhibiting a variety of phe-
nomena such as accretion discs, winds, collimated jets and
magnetically controlled accretion flows, thermal disc insta-
bilities, and both stable and explosive thermonuclear shell
burning. The plasma conditions in these accretion structures
span a huge range of physical conditions, including relativis-
tic environments and extreme magnetic field strengths. Con-
sequently, interacting binaries are also extremely versatile
plasma physic laboratories.
Despite their great importance for a vast range of astro-
physical questions, our understanding of close binary stars
and their evolution is still very fragmentary. The ultravio-
let (UV) is of outmost importance in the study of interact-
ing binaries, as a large part of their luminosity is radiated
away in this wavelength range, and, more importantly, as
the UV hosts a multitude of low and high excitation lines
of a large variety of chemical species. These transitions can
be used both as probes of the plasma conditions, as well as
tracers of individual components within the binaries through
time-resolved spectroscopy. Moreover, the physical status of
the binary components and in particular the accreting white
dwarf primaries in cataclysmic variables, symbiotic stars,
and double-degenerate binaries can be easily isolated and
studied in the UV range. Even though substantial scientific
progress has been achieved throughout the last three decades,
primarily using the International Ultraviolet Explorer (IUE),
the Hubble Space Telescope (HST), and the Far Ultraviolet
Spectroscopic Explorer (FUSE), these are still the early days
of UV astronomy of interacting binaries, and many key ques-
tions are yet without answer. Here we outline the enormous
potential that a major UV observatory has for our understand-
ing of interacting binaries, and how the expected findings re-
lated to much wider astrophysical contexts, including galaxy
evolution and cosmology.
2. Accreting white dwarfs
2.1. The complex interplay between stellar properties
and binary evolution
Compared to their isolated relatives, the evolution of white
dwarfs in interacting binaries is much more complex, and
closely related to the evolution of the binary as a whole,
and hence understanding close binary stellar evolution is im-
possible without detailed knowledge of the properties of the
white dwarf components in these stars. The most abundant
type of mass-transferring binaries containing a white dwarf
are the cataclysmic variables (CVs), which have mass trans-
fer rates in the range 10−11–10−9Myr−1. The accretion of
this material and its associated angular momentum affects
practically all fundamental properties of CV white dwarfs.
Compressional heating is depositing energy in the enve-
lope and the core of the white dwarf, effectively compensat-
ing the secular cooling, with the result that accreting white
dwarfs are substantially hotter than isolated white dwarfs
of comparable age and mass. Townsley and Bildsten (2002,
2003) have shown that the white dwarf temperature can in-
deed be used to establish a measure of the long term average
of the accretion rate sustained by the white dwarf. As the
secular average accretion rate is directly related to the rate at
which the binary is losing orbital angular momentum mea-
suring this parameter is of fundamental importance for any
theory of close binary evolution.
Accretion will increase the mass of the white dwarf. Even-
tually, if nothing else happens and the mass supply of the
companion star is sufficient, this will drive the white dwarf
over its Chandrasekhar mass limit, and it may turn in into a
supernova Type Ia. Accreting white dwarfs as possible SN Ia
progenitors are discussed in more detail in Section 2.2 below.
However, in most CVs the accreted hydrogen layer will ther-
monuclearly ignite once the density and temperature exceed
the critical condition. This hydrogen shell burning is typically
explosive, observationally designated as a classical nova, and
ejects a shell of material into space (see Section 2.3). As the
critical mass of the accreted hydrogen-rich layer is fairly low
(∼10−5–10−3M), a CV will undergo hundreds to thousands
of nova explosions. Currently, it is not clear what the mass
balance during the nova event is, i.e. whether the amount of
ejected material is equal to or even exceeds the mass of ac-
creted material, and, hence, the long-term evolution of the
white dwarf mass is not known.
Chemical abundances of white dwarf surfaces can be af-
fected by accretion and greatly modified by nova explosions.
While a roughly solar composition is expected for a freshly
accreted white dwarf atmosphere, many CVs were recently
found to possess an unexpected wide variety of departures
from (solar) abundances (Sion, 1999), opening new hori-
zons in the current understanding of binary evolution. In-
deed while in single white dwarfs metallic species and their
abundance reveal processes which oppose diffusion, those in
cataclysmic variables show a mix of chemical species and
abundances that cannot result from accretion from a normal
secondary star, thus pointing towards a thermonuclear activ-
ity in their past evolution. The hypothesis of CNO processing
as the source of the abundances has been further supported by
the detection of proton-capture material by Sion et al. (1997).
This has great implications for CV evolution and contribu-
tions to the heavy element content of the interstellar medium
(see also Section 2.2).
Rotation rates of non-magnetic white dwarfs in CVs were
unknown prior the HST era and its advent opened a new topic
in close binary evolution. Global rotational velocities are now
measured for a handful of dwarf novae systems (Sion, 1999)
and were found to be much larger (300–1200 km s−1) than
Springer

Astrophys Space Sci (2006) 303:53–68 55
the few tens of km s−1in isolated white dwarfs, implying
that accretion efficiently spins-up the primaries. However the
measured rates are much lower than expected on the basis
of the amount of angular momentum accreted during their
characteristic lifetimes (Livio and Pringle, 1998; King et al.,
1991) suggesting that part of the accreted angular momen-
tum is removed from the white dwarf during the expanded
envelope mass loss phase which follows a nova eruption.
This independently suggests that also dwarf nova experience
nova outbursts and then return to be dwarf novae again, as
suggested by the cyclic evolution scenario. Although this re-
sult has an enormous evolutionary implication, it is based
on only 5 CV white dwarfs for which reliable rotation rates
could be determined.
Whereas accretion alters the properties of white dwarfs in
interacting binaries, some of the white dwarf characteristics
will in turn deeply affect the accretion process – e.g. the mass
of the white dwarf defines the depth of the potential well, and,
thereby, the amount of energy released per accreted gram of
matter, the rotation rate of the white dwarf determines the
luminosity of the boundary layer, i.e. the interface between
the inner accretion disc rotating at Keplerian velocities and
the white dwarf itself, and finally the magnetic field of the
white dwarf determines the accretion geometry.
The observational study of accreting white dwarfs can
only be carried out in the UV, as the emission from the ac-
cretion flow dilutes or even completely outshines the white
dwarf at optical wavelengths. Because of the faintness of
most CV white dwarfs, the number of systems for which
medium-resolution (1–2 ˚
A) spectroscopic data, adequate
for temperature measurements, has been obtained is 35
(Sion, 1999; Szkody et al., 2002; Araujo-Betancor et al.,
2005) – out of a total of ∼1000 CVs known (Downes et al.,
2001). High-resolution (0.1 ˚
A) UV spectroscopy necessary
for accurate abundance and rotation rate determinations has
been obtained only for a handful of systems, most noticeably
for the nearest CV, WZ Sge (Fig. 1, see e.g. Sion et al., 2001,
2003; Long et al., 2004; Welsh et al., 2003), at the expense
of >20 HST orbits.
2.1.1. Future prospectives of UV astronomy
In order to fully assess the interrelation between the white
dwarf properties and the evolutionary state of CVs, a suffi-
ciently large number of systems has to be observed. Mapping
out the parameter space (Twd,Mwd , abundances and rotation
rate of the white dwarf, as well as the binary orbital period)
will eventually require adequate data for 100–200 systems.
Temperature measurements need a broad UV wavelength
coverage, optimally from the Lyman edge down to 3000 ˚
A,
at a low resolution (R1000–2000). Abundance/rotation
rate measurements rely on medium-resolution (R20000)
spectroscopy covering a sufficient number of transitions; the
traditional range 1150–1900 ˚
A is adequate even though sim-
ilar capabilities below Lyαwould be desirable. Throughput
is the crucial need for this science, as typical flux levels are
afew10−16 erg cm−2s−1˚
A−1.
2.2. Accreting white dwarfs as likely SN Ia progenitors
Whereas SN Ia are routinely used as beacons at cosmo-
logical distances (Filippenko, 2004), and generally associ-
ated with the thermonuclear disruption of a carbon-oxygen
white dwarf (Livio, 2001), the nature of their progenitors re-
mains elusive. Two different channels of SN Ia progenitors
are currently most favoured (Yungelson and Livio, 2000).
Fig. 1 High-quality UV
spectroscopy of accreting white
dwarfs in CVs is necessary to
determine their temperature,
mass, rotation rate, and
atmospheric abundances from
detailed model atmosphere fits.
Only a very limited number of
CVs has been bright enough to
be studied with the STIS high
resolution grating, such as e.g.
WZ Sge (from Long et al., 2004)
Springer

56 Astrophys Space Sci (2006) 303:53–68
In the double-degenerate channel two white dwarfs spiral
in under the effect of gravitational radiation until they fi-
nally merge, exceeding the Chandrasekhar mass limit. In-
tensive optical surveys have been carried out for this type of
SN Ia progenitors, most recently by (Napiwotzki et al., 2001),
identifying a few potential SN Ia progenitor candidates. In
the single-degenerate channel a white dwarf accretes from
a main-sequence companion. However, as outlined above,
most white dwarfs accreting hydrogen-rich material will go
through classical nova explosions and grow only little (or
even shrink) in mass. Only if the white dwarf is accreting
at a rate sufficiently high to sustain steady-state hydrogen
shell burning – the accreted hydrogen is thermonuclearly
processed at the rate it is accreted. These objects have been
predicted (Shara et al., 1977; Iben, 1982; Fujimoto, 1982) and
first found in the EINSTEIN X-ray survey of the Magellanic
clouds (Long et al., 1981), even though it took a fair amount
of time to identify their true nature (van den Heuvel et al.,
1992). Based on their observational hallmark – a very large
luminosity in soft X-rays, these objects are coined super-
soft sources, or more appropriately supersoft X-ray binaries
(G¨aansicke et al., 2000). The high accretion rates that are
necessary to fuel the steady-state shell burning in supersoft
X-ray binaries can be provided by a Roche-lobe filling main
sequence star if its mass is similar to or exceeds that of the
white dwarf. As mass is transfered from the more massive
to the less massive star, the binary period shrinks as a con-
sequence of angular momentum conservation, stabilising or
even enhancing the mass loss of the donor star. The mass
transfer ensues on a time scale which is too short for the
donor star to adjust its thermal structure, and in an evolu-
tionary jargon supersoft X-ray binaries are known as thermal
time scale mass transfer (TTSMT) CVs. In the absence of
nova eruptions, the white dwarfs in TTSMT CVs grow in
mass, and will, if the donor star provides a sufficient amount
of material, surpass the Chandrasekhar limit and potentially
explode in a SN Ia (Di Stefano, 1996; Starrfield et al., 2004).
If the donor star in a TTSMT runs out of fuel before the
white dwarf reaches the Chandrasekhar mass limit, the mass
ratio will eventually flip with the donor star being less mas-
sive than the white dwarf. Consequently, the mass transfer
rate decreases and the shell burning ceases. From this point
on, the system will evolve and look (at a first glance) like a
normal CV – with the dramatic difference that normal CVs
contain main-sequence donor stars, whereas post-TTSMT
CVs contain the CNO processed core of the previously more
massive star. Schenker et al. (2002) suggest that a significant
fraction (up to 1/3) of all present-day CVs may actually have
started out with a companion more massive than the white
dwarf, and underwent a phase of TTSMT. A recent HST/STIS
snapshot survey of 70 CVs showed that ∼10% of the systems
display a significantly enhanced N/C abundance ratio, which
suggests that these systems went through a phase of TTSMT
(G¨aansicke et al., 2003). So far, not a single progenitor of
supersoft X-ray binaries/TTSMT CVs has been identified.
2.2.1. Future prospectives of UV astronomy
Mapping out the population of failed SN Ia =post-
TTSMT CVs will require low (R1000–2000) res-
olution spectroscopy of several 100 CVs down to
10−16 erg cm−2s−1˚
A−1. Based this large sample, it will
be possible to determine the orbital period distribution of
post-TTSMT CVs with respect to the “normal” systems,
model their evolution, and finally extrapolate these popula-
tion models to the regime of true SN Ia progenitors. Follow-
up the brighter ones at high resolution (R20000) will
be necessary in order to determine their detailed proper-
ties. This will also help to answer the very important ques-
tion on whether the white dwarfs in these systems have
grown in mass, i.e. are they more massive than in normal
CVs?
Equally important is the search for true SN Ia progenitors.
However, as the TTSMT phase is very short the chance of
finding systems in this stage is small – in fact, in our Galaxy
(where absorption in the plane further decreases the prob-
ability of finding such systems) only two supersoft X-ray
binaries are known. Supersoft X-ray binaries can be located
in local group galaxies using high spatial resolution (1)
X-ray missions such as Chandra – however, X-ray data alone
is typically insufficient to determine the properties of the
objects. UV observations can substantially defeat crowding
problems (see Section 7), and are well-suited to obtain fun-
damental parameters such as orbital periods. This implies
large aperture high spatial resolution UV imaging capabili-
ties (supersoft X-ray binaries in M31 have V23).
A so far entirely unexplored potential is the search for
SN Ia pre-progenitors, i.e. detached white dwarf/main se-
quence binaries with Msec >1.6M(Langer et al., 2000;
Han, 2004). In the optical, these systems will be entirely dom-
inated by the main-sequence star, and follow-up UV studies
of main-sequence stars with UV excess (identified e.g. in the
GALEX survey) will be necessary to identify them.
2.3. The nova phenomenon
Novae are the most spectacular phenomenon encountered in
CVs and represent key objects to understand a wide variety
of physical conditions of accreting matter including super-
Eddington regimes and the interaction of ejecta in the in-
terstellar medium and its chemical evolution. They are fun-
damental standard candles up to the Local Group, having
hence important implications for cosmological distance cal-
ibrations.
Despite the enormous observational effort of the past
15 years, especially with multi-wavelength campaigns and
Springer

Astrophys Space Sci (2006) 303:53–68 57
datasets, there remain two fundamental uncertainties: what
are the masses and structures of the ejecta and what drives
the mass loss during the outburst? Radiative processes, which
might be the source of a stellar wind during the ejection phase
(e.g. Hauschildt et al., 1994) depends on the abundances (and
therefore the details of the spectral evolution during the initial
stages) and the bolometric luminosity. Explosions are pow-
ered directly by decays of radioactive isotopes generated dur-
ing the thermonuclear runaway following the initial envelope
expansion, but any subsequent mass loss must be driven by
the match of the flux distribution with the envelope opacities
(Shore, 2002). The UV, now inaccessible to observation, is
the driving spectral region for the phenomenological analysis
of novae. Only in this region it is possible to directly probe
the properties of the ejecta – abundances, structure, mass –
and determine the energetics of the thermonuclear runaway.
The reasons are simply that the photometric behaviour is
driven at all wavelengths longer than the UV by bolometric
flux redistribution from the evolving central remnant white
dwarf and that in the UV we can measure the resonance tran-
sitions off the dominant ions throughout the first few months
of outburst. To date, only novae in the Galaxy and the LMC
have been observed.
Novae have been used as distance calibrators for nearly a
century through the maximum magnitude – rate of decline
(MMRD) relation, but the origin of this relation has only
recently been understood. As the ejecta expand, the rapid and
enormous increase in the opacity from recombination-driven
strengthening of the line absorption redistributes flux into the
optical. But the correspondence between these two regions,
the completeness of the redistribution process, depends on
the details of the ejecta filling factors. If the ejecta initially
fragment and/or if they are not spherical in the earliest stages,
the process will be less efficient and the observed maximum
at longer wavelengths will be altered. Without the UV, it
is impossible to determine the bolometric luminosity and
therefore to constrain its constancy.
The two principal classes of classical novae are distin-
guished by their abundances, which reflect the composition
differences of the accreting white dwarf (CO and ONe). The
most extreme explosions may produce significantly altered
abundance patterns and there is an indication that the ejecta
for both of these types are also helium enriched. Without the
UV to provide access to resonance lines for the relevant ions,
abundance studies – and the determination of structure – are
limited by uncertainties in the equation of state for the ejecta.
Among recurrent novae, the two classes; those in com-
pact, cataclysmic-like systems and those with red giant com-
panions, appear to have very low mass ejecta (in agreement
with current models) but with abundance patterns that sug-
gest helium enrichment. The UV is the only way to study the
ejecta in the optically thick phases (which last only a matter
of days) to obtain unambiguously the abundances. It also is
not clear whether these systems show discs, or winds, during
quiescence. Finally, the interaction between the expanding
ejecta and the winds in the symbiotic-like systems (with red
giant companions) can only be studied effectively at high
spectral resolution in the UV where the resonance lines and
continuum of the white dwarf are accessible.
2.3.1. Future prospectives of UV astronomy
With increased aperture, especially in the 4–6 meter range,
and high spectral resolution (10 000 or higher), it would be
possible to study novae throughout the Local Group, espe-
cially M 31 in which the full range of novae appear to occur.
Novae are important contributors to several rare isotopes, es-
pecially 22Ne and possibly 22 Na, and also may be important
in ionising galactic halos (thus being important for under-
standing the halo ionisation and properties of LyαForest
systems formed therein). Since they are recurrent phenom-
ena, on many timescales, and remain hot for long periods they
may be important for understanding the UV upturn in ellip-
tical galaxies (they can mimic post-AGB stars, for example).
Finally, as bright, transient UV sources, they provide probes
of the interstellar medium throughout their host galaxy. Also,
the transition from the super-soft phase into the UV is essen-
tial but it has always been extremely difficult to determine.
Furthermore the determination of chemical abundances in
the ejecta is a fundamental parameter to test theories on the
processes that lead to the nova phenomenon and to under-
stand the state of the binary system. Indeed in no system
these abundances were found to be solar-like (Selvelli and
Gilmozzi, 1999) having important implications in the chem-
ical evolution of interstellar medium.
2.4. Symbiotic stars
The symbiotic systems are exotic and intriguing interact-
ing binaries. The nature of the accreting hot companion was
longly debated and proofs that the hot accreting compan-
ion is most likely a white dwarf and not a main sequence
star were provided by UV observations (e.g. Eriksson et al.,
2004). These systems however differ from the CVs because
of their wider orbits, with orbital periods from a few to a few
dozen years, and because the white dwarf accretes from the
stellar wind of a late-type giant rather than through Roche
lobe overflow from a main sequence star. The wind from the
cool star is ionised by the radiation from the white dwarf
resulting in the characteristic combination of sharp nebu-
lar emission lines and molecular absorption bands in their
UV and optical spectra (Birriel et al., 2000). An increasing
number of symbiotic stars are also found to show nova out-
bursts. In these systems, despite the much smaller outburst
amplitudes compared to those observed in novae, the total
energy associated with the outburst may significantly exceed
Springer

58 Astrophys Space Sci (2006) 303:53–68
that of a classical nova. Currently very little is known about
the line-emitting regions associated with the outburst of a
symbiotic nova because of the long timescales to reach the
maximum and the very much slower decays (Rudy et al.,
1999). Symbiotics also fall in the category of the supersoft
X-ray sources (Greiner, 1996) making them potential SN Ia
progenitors (Hachisu et al., 1999). Furthermore only a hand-
ful of symbiotics have been monitored trough their outbursts
in the UV, where the evolution of the hot accreting object can
be best followed and from which the energetics of the pro-
cess can be best studied and linked to the soft X-ray emission
(Gonz´alez-Riestra et al., 1999).
Among the different classes of interacting binaries dis-
cussed in this paper, symbiotic stars are by far the physi-
cally largest objects, and future UV/optical interferometric
missions with sub-milliarcsecond spatial resolutions will be
able to resolve the two stellar components, as well as the
wind / accretion flow from the companion star, and possi-
bly an accretion disc around the compact star. Carrying out
such studies in the different UV resonance lines will allow a
detailed mapping of the ionization structure in the accretion
flow.
2.4.1. Future prospectives of UV astronomy
To identify the hot accreting component and to determine
the UV luminosity and its evolution during outbursts the con-
struction of SED over a wide spectral range is necessary. This
requires low dispersion (∼2000) spectroscopy in the desir-
able range from the Lyman limit down to 3400 ˚
A. Our knowl-
edge of the population of symbiotic novae in our Galaxy and
Local group will enormously improve with UV imaging ca-
pabilities as described in Sections 2.2 and 7. Furthermore the
study of emission lines mapping a wide variety of physical
conditions of the accreted matter and outflow need moderate-
to-high dispersion spectroscopy (R∼10 000–20 000) in the
FUV range. UV imaging at at sub-milliarcsec resolution is
required to physically resolve the stellar components and ac-
cretion flow.
2.5. Accretion flows in magnetic systems
Accretion can be greatly influenced by the presence of mag-
netic fields of the primary star. In those CVs where the
white dwarf is strongly magnetised (B>105–108G) im-
portant modifications of the accretion flow occur already at
the distance of the donor star. The formation of an accre-
tion disc is prevented in the high field systems (B>10 MG)
or truncation of the accretion disc can occur in moderately
(B<5 MG) magnetised CVs. Hence, the wide range of ac-
cretion patterns encountered in mCVs allows to test different
physical conditions of accretion flow and X-ray irradiation.
In particular, previous observations of magnetic systems have
shown that the FUV continuum is dominated by the X-ray ir-
radiated white dwarf pole (G¨aansicke et al., 1995), while the
accretion funnel down to the post-shock regions contributes
in the NUV continuum and in the FUV emission lines of
CNO. The truncated disc is also a source of UV continuum
(Haswell et al., 1997; de Martino et al., 1999; Eisenbart et al.,
2002; Belle et al., 2003). However there are still important
open questions on the physical conditions (kinematics, tem-
perature and density) of the accretion flow:
(1) A strong potential to diagnose ionised gas is provided
by the resonance FUV emission lines (CNO), as their
FWZI ∼2000–3000 km s−1clearly indicates that they
map the accretion flow down to the white dwarf surface.
However, while past (low resolution) UV observations
have allowed significant progress in the understanding of
emission line formation ruling out collisional ionisation
and strongly favouring photoionisation models (Mauche
et al., 1997), there is still a great uncertainty in theory, as
they cannot simultaneously account for all the line flux
ratios observed in CVs. Among the magnetic systems
there seems to be a higher ionisation efficiency in the
hard X-ray intermediate polar systems with respect to
the soft X-ray polar systems (de Martino, 1999) suggest-
ing that the soft X-rays are efficiently absorbed likely due
to larger absorption column densities. Contribution to the
FUV lines can also arise from the X-ray irradiated hemi-
sphere of the secondary star and from material located in
different parts of the flow (G¨aansicke et al., 1998), where
plasma conditions can be very different from each other.
An important improvement can be achieved by obtaining
a systematic survey of phase-resolved (white dwarf spin
and orbital period) of the FUV lines in magnetic CVs to
perform Doppler tomogram analyses and to identify the
kinematical properties of the accretion flow as well as
to separate the different contributions of emission lines.
This can allow a proper test of line formation theory as
well as to understand the irradiation effects of the sec-
ondary star.
(2) Magnetic systems have the most complex accretion ge-
ometry which is very difficult to parametrise. Spectral
energy distributions (SEDs) from optical through the UV
to the X-rays are necessary to determine the energy bud-
get and to infer the mass accretion rates (Eisenbart et al.,
2002). In the case of truncated discs as in the moderately
magnetised systems, the UV SED is crucial to assess
temperature profile and extension of disc down to the
magnetospheric radius as the SED might show a turn-
of in the UV range. Also, the X-rays can be substan-
tially absorbed leading to accretion luminosities which
can be much lower than those determined with combined
UV and optical observations (Mukai et al., 1994). Up to
date only a handful of bright magnetic CVs have been
Springer

Astrophys Space Sci (2006) 303:53–68 59
observed in the UV so far, but a systematic UV study has
the potential to infer the relation between mass accretion
rate and system parameters such as inclination angle and
magnetic moments.
Furthermore, it is of fundamental importance to separate
the spectral contributions of the heated pole caps of white
dwarfs in polars from the unheated underlying white dwarf,
thereby determining both the effects of irradiation and the
white dwarf temperature. In this respect an important is-
sue is the tendency of white dwarfs in the magnetic CVs
to be cooler than those in non magnetic systems (Sion, 1999;
Araujo-Betancor et al., 2005) with significant differences at
all orbital periods. This trend might reflect a difference in the
mass transfer rate efficiency with respect to non-magnetic
CVs as the systems evolve. In particular, the white dwarf
magnetic field may reduce the secondary star magnetic brak-
ing efficiency (Wickramasinghe and Wu, 1994), a hypothe-
sis which observations seem to confirm. However, a statis-
tically significant sample of magnetic CVs is needed to be
observed especially at periods above the orbital 2–3 hr period
gap where only two systems have been covered so far with
HST. This range of the period distribution is essential as it
is dominated by angular momentum loss through magnetic
braking. It is therefore important to determine the tempera-
ture of the unheated white dwarf atmosphere by means of low
resolution UV spectroscopy over a wide wavelength range
either when these systems are in a low accretion state or
via phase-resolved observations which can allow to isolate
the heated atmospheric pole from the unheated white dwarf
atmosphere.
2.5.1. Future prospectives of UV astronomy
To map the accretion flow structure will need system-
atic phase-resolved UV spectroscopy at the orbital and
white dwarf rotational periods in high and low disper-
sion to study the phase dependence of the FUV emis-
sion lines and of the SED. In particular the determina-
tion of kinematical properties of the accretion flow and
the identification of the X-ray irradiated secondary star at-
mosphere require the coverage of the various FUV emis-
sion lines and hence a minimum range, of 1150–1800 ˚
A
(1000–1800 ˚
A desirable) with a spectral resolution of R∼
20 000. Only for a handful of bright magnetic systems high
resolution spectroscopy has been performed with HST and
FUSE. Furthermore, the construction of SEDs over the
widest spectral range from the Lyman edge down to 3400 ˚
A
is essential to determine simultaneously the different spec-
tral components (disc, white dwarf, accretion funnels). This
requires low resolution spectroscopy at R∼2000 and good
quality spectra at levels of a few 10−16 erg cm−2s−1˚
A−1.
Phase – resolved spectroscopy also demands large through-
put in order to achieve reasonable signal-to-noise ratio with
short exposure times. Timing capabilities of instrumentation
(e.g. photon – counting systems) allowing to explore differ-
ent types of variability (periodic, quasi-periodic and non pe-
riodic) on a wide range of timescales are essential. These can
allow the access to the mostly unexplored temporal domain
of UV emission in accreting magnetic systems.
3. Accretion discs
In order to form stars and galaxies, or to power active galac-
tic nuclei and gamma-ray bursts, matter must be compressed
by many orders of magnitude in size. This is possible while
gravity dominates over thermal, magnetic and rotational en-
ergy. This can require the radiation of substantial amounts of
thermal energy and the diffusion of magnetic field, but ulti-
mately rotation always puts a brake upon this process because
in a homologous collapse of a cloud of size Rthe rotation
energy scales as R−2while gravitational energy scales as
R−1. Nature’s solution to this problem is to re-distribute the
angular momentum in an accretion disc. In an accretion disc,
gas travels in near-circular orbits gaining angular momen-
tum from material at smaller radii, and losing it to matter at
larger radii. The transport of angular momentum is driven
by some form of viscosity, and only in recent years has a
plausible candidate for this been identified in the magneto-
rotational instability (Balbus and Hawley, 1998). Despite this
progress, our understanding of the viscosity of accretion discs
and precisely how energy is dissipated within them remain
the central unanswered questions in the field. A major ob-
stacle to making progress is that two important properties
of discs, their luminosity and temperature distribution, are
independent of viscosity in steady-state discs. Progress can
only be made through the study of phenomena that change
on the viscous timescale,tν∼R2/ν where Ris the size of
a disc and νis the kinematic viscosity or by examining the
vertical temperature structure of discs through their spectra.
The great advantage of close binary stars is their small scales
which lead to viscous timescales of only a few days or weeks,
making them amenable to direct observation.
The outbursts of dwarf novae are almost universally be-
lieved to be driven by changes in the viscosity of the material
in their accretion discs. In the standard disc instability model
developed in the 1980s, in the quiescent state, the viscosity
νis very low, and the viscous timescale, tν, is so long that
the disc cannot cope with the rate at which matter flows in
at its outer edge. Instead, mass piles up in the outer parts of
the disc until a critical point is reached and at some radius in
the disc the viscosity (and therefore viscous dissipation rate)
increases dramatically, by of order 100 times. This jump can
take only a few minutes, with the outburst following as a
heating wave propagates to all radii within the disc. A major
Springer

60 Astrophys Space Sci (2006) 303:53–68
goal of the study of accretion discs is to understand these out-
bursts: how they propagate, what triggers them, but above all,
why the viscosity ramps up so violently. This is thought to be
rooted in the ionisation of hydrogen (or in ‘ultra-compact’
binary stars, helium), but we have no detailed physical mech-
anism from which we can compute the viscosity for a given
composition, density and temperature. All current models,
which have been applied to accretion discs in a wide variety
of objects, are hence purely phenomenological.
The commonest, nearest and most easily studied accretion
discs are those of the cataclysmic variable stars which have
white dwarf accretors. Accretion discs around white dwarfs
vary in temperature from ∼6 000 K in their outermost parts
to over 100 000 K close to the white dwarf. The radius of the
outer disc is typically 10 to 50 times that of the white dwarf,
but it is from the hot, inner few white dwarf radii that most
of the energy is released. The UV is the key waveband for
seeing these regions. There are three other reasons for UV
observation of accretion discs in these binaries. First, one can
see absorption lines from the disc photospheres most easily
in the FUV (Fig. 2). These give a handle upon ionisation state
not available at optical wavelengths where emission from the
(barely understood) disc chromosphere is always dominant
and photospheric absorption lines are weak. Second, the UV
is where disc model atmospheres currently seem to fail most
severely, in general appearing too blue compared to obser-
vations (Orosz and Wade, 2003). The final unique feature of
the UV is its sensitivity to the geometry of the disc because
absorption of the inner by the outer disc is most easily seen
in the UV and was first established from UV observations
(Horne et al., 1994).
3.1. Modelling the spectra of accretion discs
Since the early days of black-body models, followed by
models based upon sums over standard stellar atmospheres
(Wade, 1984), accretion discs models based upon modern
model atmosphere codes have been developed (Wade and
Hubeny, 1998; Orosz and Wade, 2003). Such models are
however undermined by our ignorance of the mechanism of
viscosity and hence of the vertical temperature structure of
discs. In principle, spectra can be used in reverse to determine
vertical structure, but, so far, little progress has been made
in this area. A problem of long standing is that disc model
atmosphere spectra do not work very well, especially at FUV
wavelengths. This might be down to the vertical structure, or
it could be that the steady-state (radial) temperature distribu-
tions used so far are not accurate (Orosz and Wade, 2003),
even though it is hard to understand how this can be the case
in systems that hardly change on many viscous timescales.
The problem comes from the small size of accretion discs
in close binary stars, with a typical radius of a few 1010 cm.
While their sizes help keep viscous timescales small, so that
thermal instabilities are easily observed over the course of
days to months, it means that the discs are not spatially re-
solvable as they subtend at best a few 0.01 milli-arcseconds
for the closest systems. Direct imaging of the accretion discs
will be possible in the foreseeable future only in the much
larger symbiotic stars (Section 2.4), in which, however, the
viscous time scales are substantially longer, and the temporal
variability of symbiotics in terms of disc instabilities is much
less established compared to the situation in CVs. Thus the
spectra we see are the integrated spectra from all radii in the
Fig. 2 Model spectra of
accretion discs with a range of
radial temperature profiles,
Teff ∝R−γOrosz and Wade
(2003)
Springer

Astrophys Space Sci (2006) 303:53–68 61
disc. This makes it hard to know whether it is the radial or
vertical structure that is causing the problem. With integrated
spectra, one cannot be sure whether all radii are poorly mod-
elled or whether only specific effective temperatures are in-
volved, and thus it is not clear how to adapt models. We need
spatially resolved spectra, which can be achieved through a
technique known as eclipse mapping (Horne, 1985). Applied
at UV wavelengths, this technique has the capability to pro-
vide spatially resolved spectra of the inner parts of accretion
discs where most energy is released. We can then see where
it is in the discs that model atmospheres fail most severely.
The principle of eclipse mapping is as follows: in an eclips-
ing system, the light-curve of the disc as it is eclipsed depends
upon how concentrated the surface brightness is. For instance
a flat distribution of brightness leads to a shallow V-shaped
light curve, whereas a distribution which is strongly peaked
towards the centre of the disc has a deep U-shaped light curve
(Fig. 3)
Thus, in essence, the light curves can be used to deduce
the variation of surface brightness with radius. Eclipse map-
ping was developed by Horne (1985) for broad-band optical
light curves. A significant step in eclipse mapping came with
its extension to spectra (Rutten et al., 1993; Rutten et al.,
1994). The simple, beautiful, idea was to carry out eclipse
mapping on each pixel of low-resolution spectra to produce
spectra at every point of discs. This technique applied to UV
data has the capability of giving us spatially-resolved spectra
of the inner accretion discs of close binary stars. Only one
such analysis has been carried out in the UV with HST/FOS
(Baptista et al., 1998; Fig. 4) of the brightest of all high-state
systems, UX UMa.
Even on this, the brightest suitable system, the study was
limited by signal-to-noise ratio, especially at FUV wave-
lengths, precisely the most important part of the spectrum.
The signal-to-noise ratio in this region of the spectrum is a
modest 10% and yet the radial resolution is still only ∼4
white dwarf radii, which means that we are still seeing the
integrated light from a region which varies by a factor of
three in temperature from the inner to outer edge of the an-
nulus. In other words the problem of integrated spectra is
only partially solved in this study.
3.1.1. Future prospectives of UV astronomy
To substantially improve our ability to model the spectra of
accretion discs requires a low resolution, wide wavelength
coverage UV spectrograph of much greater sensitivity than
has been available to date. Low resolution (R∼300) be-
cause the spectral eclipse mapping technique cannot resolve
the ∼1000 km s−1motions within the disc. UV because, as
Fig. 3 Model eclipses in the
continuum from 1410 to 1530 ˚
A
for black-body (top) and model
atmosphere discs for a range of
radial temperature profiles,
Teff ∝R−γOrosz and Wade
(2003)
Springer

62 Astrophys Space Sci (2006) 303:53–68
Fig. 4 The spectra of the steady-state system UX UMa as a function
of radius deduced from HST/FOS observations of its eclipse (Baptista
et al., 1998). The spectra are those of annuli with central radius indicated
in units of the distance to the inner Lagrangian point. Note that the
spectra are plotted in fν; the UV is dominant energetically
said before, this is where most luminosity is radiated and
where the current disc atmospheres fail most severely. Wide
wavelength coverage (at least 1000 to 3000 ˚
A, and if pos-
sible extending to the Lyman edge) because it is the vari-
ation of continuum flux with wavelength which tells us
most directly about the vertical structure in stellar atmo-
spheres. Finally, and perhaps above all, in comparison to
any UV mission to date, high sensitivity is needed in or-
der to improve both the signal-to-noise ratio in the decon-
volved spectra and their radial resolution down to of order
a single white dwarf radius and so that the method can be
applied to systems fainter than UX UMa. Two other require-
ments needed for this work are an ability to take short ex-
posures (<2 seconds) which are accurately timed with abso-
lute times good to better than one-hundredth of the exposure
length.
Once progress in understanding gross properties of spectra
has been made, there will be a need for higher resolution
observations. Models can predict the changes in detailed line
profiles expected during eclipse (Orosz and Wade, 2003).
Again disc broadening means that moderate resolution is
sufficient (R∼5000), but high-sensitivity in order to allow
short exposures and therefore high spatial definition of the
disc are a must. The study of integrated spectra provides the
most stringent requirement for spectral resolution because
face-on discs (no eclipse) have significantly narrower line
profiles and suffer less from blending. For these R∼20 000
would be useful, covering from 900 to 1700 ˚
A.
3.2. Disc instabilities
It is the study of dwarf nova outbursts that lead to the disc
instability theory and the discovery of the strong dependence
of viscosity with the physical conditions within the disc. The
disc instability model has largely been used to explain the
gross features of outbursts, such as their duration and am-
plitude, but there have not been convincing detections of the
heating fronts which would allow us to confirm predictions
of the models in detail. Attempts have been made from opti-
cal observations of eclipses, but these lose resolution in the
inner disc because the outer disc dominates the light output
at optical wavelengths. The propagation of the heating fronts
into the inner disc is of particular interest because there is
evidence to suggest that the inner disc is strongly depleted
during quiescence (Schreiber et al., 2004). This is needed to
prevent outbursts triggering in the inner disc, which in some
systems would lead to too high an outburst rate. Propagation
of heating fronts can be measured from the development of
photospheric line profiles from the disc. As the front pro-
gresses inwards, the contribution from smaller radii in the
disc will contribute to broadening the line profiles because
Doppler broadening is largest in the inner disc. The pho-
tospheric lines are strongest by far in the UV and develop
dramatically during outburst (Fig. 5).
3.2.1. Future prospectives of UV astronomy
So far this sort of study has not been possible because of
limited sensitivity at FUV wavelengths. It also requires a
much higher duty cycle than possible with HST as outbursts
take of order a few hours to a day to start. Observations such as
these could also show the development of winds through the
resonance lines which are only visible at UV wavelengths.
A spectrograph covering 900 to 1700 ˚
A with a resolution
R>5000 is needed for this work.
4. Hydrogen-deficient systems
A unique aspect of close binary stars is that owing to the evo-
lution of their mass donor stars, some systems can show very
unusual abundances, adding an extra dimension to the devel-
opment of atmospheric models. The AM CVn systems for
example have helium white dwarf donors and accretion discs
which are >95% helium. Ultra-compact neutron star/white
Springer

Astrophys Space Sci (2006) 303:53–68 63
Fig. 5 The spectra of the dwarf
nova VW Hyi caught at the start
of an outburst with HST/STIS
(Sion et al., 2004). At first the
spectra are dominated by the
white dwarf but strong
photospheric absorption lines
develop as the heating front
reaches the inner disc
dwarf binaries can have carbon-oxygen and oxygen-neon-
magnesium donor stars. The element abundances are crucial
to understanding the evolution that leads to such stars. For in-
stance, an evolutionary path from cataclysmic variable stars
to AM CVn stars typically leaves a small amount of hydrogen
(Podsiadlowski et al., 2003), whereas a route via double white
dwarf mergers does not. Similarly, the ratios of the CNO el-
ements in such stars depends upon the initial mass of the
donor stars. Such information is invaluable in pinning down
evolutionary pathways, and therefore to predicting the num-
bers of systems. These binaries are so compact that they can
fit comfortably inside the Sun. At the same time their short
orbital periods means that gravitational radiation is strong
(such systems will be significant sources for LISA) and mass
transfer rates can be high. As a result, they emit mostly at UV
wavelengths and the UV is where photospheric lines from the
disc are strongest.
4.1. Future prospectives of UV astronomy
Over the next few years many more AMCVn systems are
likely to be discovered, but as they are relatively rare, the
majority will be faint. As for the hydrogen-rich systems, a
R∼20 000 spectrograph covering 900 to 1700 ˚
A is needed,
with high sensitivity the key feature.
5. Accretion winds
Mass loss is an ubiquitous feature of astrophysical systems
and the evidence of mass loss in disc-dominated cataclysmic
variables is unambiguous. In the wavelength range acces-
sible to IUE (1150–3200 ˚
A), the existence of outflows
in systems observed a lower inclination is indicated by P-
Cygni-like and/or blue shifted absorption profiles in reso-
nance transitions of N V,SiIV, and most commonly C IV.Ve-
locity widths of 3000–5000 km s−1, comparable to the escape
velocity from the primary, are regularly seen, especially in
CIV. The features are understood to result largely from scat-
tering of disc photons by the outflow. At low inclinations, the
process removes photons along the line of sight to the disc;
emission wings arise from photons scattered into the line of
sight of the observer, just as in the stellar winds of massive
stars. At higher inclinations, less direct light is observed from
the disc, and the resonance lines generally appear as broad
emission features. Indeed, after analysing 850–1850 ˚
A spec-
tra of Z Cam obtained with the Hopkins Ultraviolet Telescope
Knigge et al. (1997) suggested that virtually all of the lines
in the UV spectrum of a typical high-state CV are formed in
the outflow, either in the supersonic portion of the wind or in
a lower velocity portion of the wind near the interface with
the disc photosphere.
Although the strong 1s–2p transitions of Li-like or Na-like
ions dominate the line spectra of disc-CVs observed with IUE
and HST, the FUV spectra obtained with FUSE often show
narrower features from intermediate ionisation state transi-
tions of abundant ions such as N III,CIII,SiIII, and Si IV. These
intermediate level ionisation state lines often show orbital
phase dependent effects, even in systems of intermediate in-
clination such as Z Cam (Hartley et al., 2005) that suggest the
effects of the accretion stream must be included to complete
the picture of extra-planar gas in disc-dominated systems.
In RWSex and V592 Cas, enigmatic orbital variations in the
Springer

64 Astrophys Space Sci (2006) 303:53–68
blue edges of broad C III profiles indicate departures from
bi-conical symmetry in the high-velocity wind (Prinja et al.,
2003, 2004). Whether these are associated with disc tilts or
the accretion stream or some other mechanism is not under-
stood. Unfortunately the number of systems in which appro-
priate studies have been undertaken is small, and generally
speaking not intensive or lengthy enough to fully characterise
the phenomenology of the effects.
Originally, the possibility that the wind was a radial
wind was considered, but observations of eclipsing systems
showed changes in profiles shapes that are most straight-
forwardly interpreted as an indication of rotation, thereby
indicating that the wind emanates from the inner disc (Drew,
1987). Consequently, our basic picture of the high velocity
wind first observed with IUE is of a bi-conical flow emanat-
ing from the inner portion of the disc and/or rapidly rotating
boundary layer. Vitello and Shlosman (1993) were the first
to attempt to actually model the profile shapes of wind lines
as observed in high state CVs in terms of kinematic prescrip-
tion for a bi-conical wind. They found that the IUE-derived
(R=200) C IV profiles of three systems – RW Sex, RW Tri,
and V Sge – could be reproduced with moderately collimated
winds with the local mass loss rates of order 10% of the disc
accretion rate and terminal velocities of 1–3 times the escape
velocity at the footprint of each streamline. Subsequently,
Knigge and Drew (1997) succeeded, using a somewhat dif-
ferent kinematic parameterisation for a bi-conical flow, in
reproducing the C IV profile of UX UMa through an eclipse.
This analysis was important, not only because it was the first
attempt to model changes in the profile through eclipse, but
also because it suggested, at least in UX UMa, the existence
of a relatively dense, high column density, slowly outflowing
transition region between the disc photosphere and the fast
moving wind. Both the Vitello and Shlosman and Knigge
and Drew analyses suggested that the characteristic accel-
eration length for the high velocity winds observed in disc
dominated CVs is quite long, or order 100 Rwd. Most of the
analyses of the spectra of disc winds were limited to sin-
gle lines, but more recently Long and Knigge (2002) have
developed Monte Carlo radiative transfer codes which in a
few cases (see Fig. 6) are able to qualitative reproduce the
full UV spectrum of a disc dominated CV. Hydrodynamical
simulations of radiatively-driven CV winds are also been un-
dertaken, and when combined with a radiative transfer code,
these are also beginning to be compared to observed spectra
with mixed results (see, e.g., Proga, 2003).
Although, modelling of CV winds has progressed, funda-
mental basic questions about the winds still remain. We are
unable to measure basic parameters like the mass-loss rate,
and although the wind is assumed to be radiatively driven,
the observational and theoretical evidence for this is at best
murky. For example, on the observational side, if the wind
is radiatively driven, one might expect that the observational
signatures of wind lines would be strongest when systems
are brightest. But Hartley et al. (2002) found there was no
correlation between the strength of wind features and con-
tinuum brightness in the spectra of three observations each
of the two nova-like variables IX Vel and V3885 Sgr with
HST. (Unfortunately, the number of high state systems that
have been observed enough to begin to characterise their be-
haviour with time is quite limited.) And Drew and Proga
(2000) have argued that the luminosity of discs is at best
marginally enough to accelerate a high velocity wind. Thus,
alternatives to the emission or additions to radiation pressure
must be considered instead. These include viscous heating of
the upper portion of the disc atmosphere (Czerny and King,
1989b), and irradiation (Czerny and King, 1989a) as well
as magneto-centrifugal forces producing constant angular
Fig. 6 HUT spectrum of the
wind dominated spectrum of
IX Vel modelled compared to
one of the models of Long and
Knigge (2002)
Springer

Astrophys Space Sci (2006) 303:53–68 65
velocity out to the Alven surface (Cannizzo and Pudritz,
1988).
5.1. Future prospectives of UV astronomy
As observations of high mass transfer discs and winds in CVs
have improved, so has the complexity of phenomenological
descriptions of the wind structures emanating from the disc,
especially as the wavelength coverage has extended in the the
UV. But very few systems have been studied in enough detail
to isolate common from uncommon behaviour. Furthermore,
it is quite clear that the appearance of disc dominated sys-
tems is strongly modified by inclination, and as a result one
needs to observe a number of similar systems to the same
level of detail to be able to go beyond the general variations
that were observed with IUE. To carry out a study of this
type higher sensitivity is required so that the pool of targets
that can be studied is substantial and so that the observations
can be made at the resolution needed to resolve the nar-
rower lines that exist particularly in the FUV short-ward of
1200 ˚
A.
Higher sensitivity observations are also required to obtain
a better short term characterisation of the wind flow. Some
systems seem to have little or no short term temporal vari-
ability, whereas others, e.g BZCam (Prinja et al., 2000) are
highly variable. We do not know whether this is due to some
fundamental difference in the systems – a magnetic white
dwarf for example – or is it due simply to differences in the
accretion rate. What is the role of outer disc in the creation
of disc wind? Some hydrodynamical simulations show fast
steady flows emanating from the inner disc, but complex time
variable flows in the outer discs.
More systems need to be measured with high signal-to-
noise ratio. The number of systems actually observed with
HST were far fewer than observed with IUE, a fact that
was partially a result of the way IUE was scheduled com-
pared to HST and the fact that HST was never designed to
be a dedicated UV observatory, but a multi-purpose / multi-
wavelength facility. Instead the observations with HST have
focused on a few key systems. Therefore it has been quite
difficult to determine whether many of the phenomenolog-
ical models proposed to explain the wind features of disc
dominated features are founded on general characteristics
of winds in disc dominated CVs and how many are due to
individual systems.
To maximally constrain models of the wind, the wave-
length coverage of a new mission should extend to the region
containing O VI, and possible to the Lyman limit. Including
OVI (along with N IV,SiIV,CIV, and He II) is important not
only because O VI represents the next step up in the temper-
ature space ladder, but also because the FUV below 1150 ˚
A
is rich in lines of intermediate ionisation states. These lines
establish stringent constraints on physical conditions in the
region near the disc plane.
6. Black-hole binary stars
Most of the dynamically-confirmed black hole binaries are
transients, which spend most of their time in a low-luminosity
state. Recently there has been much debate surrounding com-
parison of these quiescent black holes with their neutron star
analogues in the attempt to detect “direct” evidence of event
horizons in the former systems. Neutron stars are brighter
in X-rays, as might be expected if the black holes advect
accretion energy through the event horizon. The theoretical
models for low-luminosity accretion flows onto black holes,
however, include variants where the flow is unbound so that
much of the accretion energy may be carried away as kinetic
energy of an outflow. Hence it is crucial to identify the correct
theoretical model before claiming event horizons have been
detected. The UV is a vital window for achieving this: almost
any model can reproduce the X-ray data alone by varying the
fit parameters, but simultaneously fitting the UV spectra is
much more exacting while optical and infrared wavelengths
are hopelessly contaminated by the donor star and outer disc.
The problem is that these systems are faint in quiescence and
only three quiescent UV spectra exist to date: of the black
holes A 0620-00 and XTE J1118+480 and of the neutron star
Cen X-4. The black hole spectra resemble each other and dif-
fer markedly from Cen X-4’s. This suggests a real physical
difference, but clearly insufficient to decide definitively be-
tween models. We need to observe more systems, and obtain
simultaneous X-ray and UV data.
In outburst, transients brighten by factors of ∼1000 across
the optical-UV-Xray spectra regions. This is attributed to the
same disc instability that drives the outbursts of cataclysmic
variable stars, but the black-hole binaries are complicated by
(i) irradiation of the optical-UV emitting disc by the central
X-ray source which changes the effective temperature dis-
tribution and causes warping of the disc, (ii) by large discs
which have no global stable high-state configuration below
the Eddington limit, (iii) for reasons which are not yet fully
understood, the inner accretion disc is often missing, being
replaced within a transition radius, Rtr, by an optically thin,
inefficiently-radiating advective flow. These factors substan-
tially alter the character of the sources: luminosity genera-
tion, outflows, duty cycle, and the mass accumulation by the
black hole are all changed. To understand these complica-
tions, UV observations are essential: the optical is dominated
by the outermost disc (which behaves more like CV discs)
and by the donor stars. The UV is required to see unam-
biguously the signatures of irradiation in the SED, to detect
self-occultation by warping, and to measure the transition
radii.
Springer

66 Astrophys Space Sci (2006) 303:53–68
6.1. Future prospectives of UV astronomy
Only a handful of black hole X-ray transient outbursts have
had their SEDs monitored throughout the outburst, and their
behaviour has been diverse. Transients outburst on time-
scales of decades and there are many that we have yet to
detect. To understand the outbursts in a systematic way, more
SED monitoring, including the UV, is required. Broad wave-
length coverage at low resolution and high throughput are
essential for this kind of studies.
7. Star clusters as laboratories for close binary
evolution
It has been known since the mid-1970s that there is a 100-fold
or so overabundance of bright LMXBs in globular clusters
(GCs), relative to the galactic field (e.g. Katz, 1975). This
quickly led to the realization that the high stellar densities
in the cores of GCs might open up entirely new dynamical
channels for the formation of interacting close binaries. The
most famous of these is tidal capture, a 2-body process result-
ing from a close encounter between a compact object (white
dwarf or neutron star) and an “ordinary” cluster members
(main sequence star or giant). During such an encounter, the
latter star experiences tidal distortions. This dissipates orbital
energy and can therefore lead to capture and binary forma-
tion (Fabian et al., 1975). However, interacting binaries can
also be formed via 3- and 4-body interactions, i.e. processes
involving existing binaries. For example, in a close encounter
between a low-mass (e.g. MS/MS) binary system and a high
mass (e.g. NS) single star, the most likely outcome is ejec-
tion of the lowest mass participant and formation of a NS/MS
binary system (Sigurdsson and Phinney, 1993).
Interacting binaries in GCs deserve careful study for two
basic reasons. First, they can in principle provide us with
large, uniformly-selected samples of systems at known dis-
tances. This is precisely what is needed to test theoretical
binary evolution scenarios. Second, close binaries are actu-
ally key players in controlling the late dynamical evolution
of GC themselves. Thus interacting binaries can actually be
used as tracers of the dynamically-formed close binary pop-
ulation in observational studies of GC evolution. In practice,
the inevitable feedback between binary and cluster evolution
will complicate things, but there is no doubt that interacting
binaries in GCs can provide us with unique insights into both
types of evolution.
UV astronomy has a key role to play in this area. Accreting
binaries tend to have much bluer spectral energy distributions
than the late-type main sequence stars that make up the bulk
of stellar clusters and galaxies. This immediately implies that
FUV observations should be an excellent way to find and
study these populations, even in optically crowded fields,
such as GC cores. This expectation is strikingly confirmed
in Fig. 7, which shows FUV and U-band images of the same
central regions of the GC 47 Tuc. The difference in crowding
is obvious, and several CVs and new CV candidates pop up
nicely in the FUV image. This image represents the deepest
FUV survey of any GC carried out to date, and utilises ob-
servations obtained with STIS onboard HST (Knigge et al.,
2002). Earlier generations of FUV/NUV detectors on HST
have also been used to search for and study interacting bina-
ries in GCs (e.g. Paresce et al., 1992; de Marchi et al., 1993;
Ferraro and Paresce, 1993; de Marchi and Paresce, 1994,
1996; Paresce and de Marchi, 1994; Cool et al., 1995; Sosin
and Cool, 1995). In the case of 47 Tuc, the lack of crowd-
ing in the FUV even makes it possible to carry out slitless,
multi-object spectroscopy in the cluster core (Knigge et al.,
2003; Knigge, 2004).
Fig. 7 Left Panel: A deep HST/STIS FUV image of the core of 47 Tuc.
The image is approximately 25 ×25 in size and includes the clus-
ter centre (marked as a white cross). For comparison, 47 Tuc’s core
radius is 23. The positions of previously known blue objects (green
squares), Chandra X-ray sources (large yellow circles) and CV candi-
dates (small blue circles) are marked. The four confirmed CVs within
the field of view are labelled with their most common designations. The
image is displayed on a logarithmic intensity scale and with limited dy-
namic range so as to bring out some of the fainter FUV sources. Right
Panel: The co-added HST/WFPC2/F336W (roughly U-band) image of
the same field. This image, too, is shown with a logarithmic intensity
scale and limited dynamic range. Figure reproduced from Knigge et al.
(2002) ( c
2002 The American Astronomical Society)
Springer

Astrophys Space Sci (2006) 303:53–68 67
In principle, open clusters and local group galaxies could
also be used as binary evolution laboratories. However, open
clusters contain fewer stars than GCs and are characterised
by lower central densities. Thus interacting binaries are not
as abundant in open clusters as in GCs, and the construction
of a statistically interesting sample would probably have to
involve studies of many such clusters. Local group galax-
ies obviously harbour large interacting binary population as
well. However, even with a 4 m class space telescope, UV ob-
servations reaching the depths required to study the quiescent
interacting binary populations will be extremely challenging
for the Magellanic Clouds and probably impossible for all
other local group galaxies.
7.1. Future prospectives of UV astronomy
Several additional galactic GCs have recently been imaged
in the UV with HST, so the UV picture of their interacting
binary populations will become clearer as soon as these new
data sets have been analysed. However, all of these studies
are seriously constrained by the small field of view of both
the STIS and ACS UV detectors (roughly 30 ×30); this
often makes it impossible to obtain a complete census of
the interacting binary population. For example, the deep UV
image of 47 Tuc in Fig. 7 covers only about 1/3 of the cluster
core. GALEX will be of some use in this regard (e.g. to
find sources in GC outskirts and open clusters), although the
benefit of its larger field of view is partially offset by its poorer
spatial resolution and lower sensitivity (relative to HST).
The optimal future UV imaging instrument would con-
sist of a large (≥4 m) mirror feeding a large-format detec-
tor producing images with diffraction-limited spatial resolu-
tion. However, the ability to obtain spectral information is
also crucial to allow secure classifications of the detected
UV sources. Single-slit/single-object spectroscopy is an ex-
tremely inefficient way of obtaining this information in a
cluster setting. As noted above, slitless spectroscopy may be
used in special cases, but what is really needed is a more gen-
erally applicable way to carry out multi-object spectroscopy
(MOS) in the UV. MOS using optical fibres is probably not
an option, since fibre losses rise steeply towards short wave-
lengths (at least in the current generation of fibres). Con-
figurable slit masks are probably also impractical in a space-
based observatory, since their use would require a large num-
ber of delicate moving parts. A simple, low-tech solution
is to provide a reasonably large selection of narrow-band
filters. An intriguing high-tech solution might involve su-
perconducting tunnel junction (STJ) detectors (e.g. Cropper
et al., 2003; Verhoeve, 2002, see also Romani et al., 1999).
These are able to provide an energy estimate for every photon
detected, so imaging and spectroscopy could, in principle, be
done in a single observation.
8. Requirements on future UV instrumentation
Here we summarise the instrumental requirements defined
by the scientific goals above.
1. Low-resolution spectroscopy (R1000–2000), with a
wavelength coverage as large as possible. Optimum would
be simultaneous data from Lyman edge down into the blue
optical (5500 ˚
A). The first priority is the broadband cov-
erage, and highest throughput. Continuum signal-to-noise
ratio of 10 at flux levels of a few 10−16 erg cm−2s−1˚
A−1
should be achieved short exposures (10–30min).
2. Medium-resolution spectroscopy (R∼20 000). The
“standard” wavelength range 1150–1800 ˚
A would be ade-
quate, covering the entire far UV down to the Lyman limit
would be preferable.
3. Detectors. Both low and medium spectrographs should
have photon counting detectors with absolute times accu-
rate down to fractions of a second.
4. Large field-of-view UV imager (10 arcmin) with high spa-
tial resolution (diffraction limited). Broad-band UV fil-
ters. Photon counting with accurate timing information.
UV/optical interferometrie providing sub-milliarcsec spa-
tial resolution.
References
Araujo-Betancor, S., G¨aansicke, B.T., Long, K.S., Beuermann, K., de
Martino, D., Sion, E.M., Szkody, P.: ApJ 622, 589 (2005)
Balbus, S.A., Hawley, J.F.: Reviews of Modern Physics 70, 1 (1998)
Baptista, R., Horne, K., Wade, R.A., Hubeny, I., Long, K.S., Rutten,
R.G.M.: MNRAS 298, 1079 (1998)
Belle, K.E., Howell, S.B., Sion, E.M., Long, K.S., Szkody, P.: ApJ 587,
373 (2003)
Birriel, J.J., Espey, B.R., Schulte-Ladbeck, R.E.: ApJ 545, 1020 (2000)
Cannizzo, J.K., Pudritz, R.E.: ApJ 327, 840 (1988)
Cool, A.M., Grindlay, J.E., Cohn, H.N., Lugger, P.M., Slavin, S.D.: ApJ
439, 695 (1995)
Cropper, M., Barlow, M., Perryman, M.A.C., Horne, K., Bingham,
R., Page, M., Guttridge, P., Smith, A., Peacock, A., Walker, D.,
Charles, P.: MNRAS 344, 33 (2003)
Czerny, M., King, A.R.: MNRAS 236, 843 (1989a)
Czerny, M., King, A.R.: MNRAS 241, 839 (1989b)
de Marchi, G., Paresce, F.: , A&A 281, L13 (1994)
de Marchi, G., Paresce, F.: ApJ 467, 658 (1996)
de Marchi, G., Paresce, F., Ferraro, F.R.: ApJS 85, 293 (1993)
de Martino, D.: Memorie della Societa Astronomica Italiana 70, 547
(1999)
de Martino, D., Silvotti, R., Mouchet, M., Mukai, K., Buckley, D.A.H.,
G¨aansicke, B.T. and Rosen, S.: in: Hellier, C., Mukai, K. (eds.),
Annapolis workshop on magnetic cataclysmic variables, pp. 41–
46, ASP Conf. Ser. 157 (1999)
Di Stefano, R.: in Greiner, 1996, p. 193, Springer (1996)
Downes, R.A., Webbink, R.F., Shara, M.M., Ritter, H., Kolb, U., Duer-
beck, H.W.: PASP 113, 764 (2001)
Drew, J.E.: MNRAS 224, 595 (1987)
Drew, J.E., Proga, D.: New Astronomy Review 44, 21 (2000)
Springer

68 Astrophys Space Sci (2006) 303:53–68
Eisenbart, S., Beuermann, K., Reinsch, K., G¨aansicke, B.T.: A&A 382,
984 (2002)
Eriksson, M., Johansson, S., Wahlgren, G.M.: A&A 422, 987 (2004)
Fabian, A.C., Pringle, J.E., Rees, M.J.: MNRAS 172, 15P (1975)
Ferraro, F.R., Paresce, F.: AJ 106, 154 (1993)
Filippenko, A.V.: in Measuring and Modeling the Universe, p. 270
(2004)
Fujimoto, M.Y.: ApJ 257, 752 (1982)
G¨aansicke, B.T., Beuermann, K., de Martino, D.: A&A 303, 127 (1995)
G¨aansicke, B.T., Hoard, D.W., Beuermann, K., Sion, E.M., Szkody, P.:
A&A 338, 933 (1998)
G¨aansicke, B.T., Szkody, P., de Martino, D., Beuermann, K., Long,
K.S., Sion, E.M., Knigge, C., Marsh, T., Hubeny, I.: ApJ 594, 443
(2003)
G¨aansicke, B.T., van Teeseling, A., Beuermann, K., Reinsch, K.: New
Astronomy Review 44, 143 (2000)
Gonz´alez-Riestra, R., Viotti, R., Iijima, T., Greiner, J.: A&A 347, 478
(1999)
Greiner, J. (ed.): No. 472 in LNP. Heidelberg: Springer (1996)
Hachisu, I., Kato, M., Nomoto, K.: ApJ 522, 487 (1999)
Han, Z., Podsiadlowski, P.: MNRAS 350, 1301 (2004)
Hartley, L.E., Drew, J.E., Long, K.S., Knigge, C., Proga, D.: MNRAS
332, 127 (2002)
Hartley, L.E., Long, K.S., Froning, C.S., Drew, J.E.: ApJ 623, 425
(2005)
Haswell, C.A., Patterson, J., Thorstensen, J.R., Hellier, C., Skillman,
D.R.: ApJ 476, 847 (1997)
Hauschildt, P.H., Starrfield, S., Shore, S.N., Gonzalez-Riestra, R., Son-
neborn, G., Allard, F.: AJ 108, 1008 (1994)
Horne, K.: MNRAS 213, 129 (1985)
Horne, K., Marsh, T.R., Cheng, F.H., Hubeny, I., Lanz, T.: ApJ 426,
294 (1994)
Iben, I.J.: ApJS 76, 55 (1991)
Iben, I.J.: ApJ 259, 244 (1982)
Katz, J.I.: Nat 253, 698 (1975)
King, A.R., Wynn, G.A., Regev, O.: MNRAS 251, 30P (1991)
Knigge, C.: Modern Physics Letters A 19, 2013 (2004)
Knigge, C., Drew, J.E.: ApJ 486, 445 (1997)
Knigge, C., Long, K.S., Blair, W.P., Wade, R.A.: ApJ 476, 291 (1997)
Knigge, C., Zurek, D.R., Shara, M.M., Long, K.S.: ApJ 579, 752 (2002)
Knigge, C., Zurek, D.R., Shara, M.M., Long, K.S., Gilliland, R.L.: ApJ
599, 1320 (2003)
Langer, N., Deutschmann, A., Wellstein,S., H ¨oflich, P.: A&A 362, 1046
(2000)
Livio, M.: in Supernovae and Gamma-Ray Bursts: The Greatest Explo-
sions Since the Big Bang. p. 334 (2001)
Livio, M., Pringle, J.E.: ApJ 505, 339 (1998)
Long, K.S., Helfand, D.J., Grabelsky, D.A.: ApJ 248, 925 (1981)
Long, K.S., Knigge, C.: ApJ 579, 725 (2002)
Long, K.S., Sion, E.M., G¨aansicke, B.T., Szkody, P.: ApJ 602, 948
(2004)
Mauche, C.W., Lee, Y.P., Kallman, T.R.: ApJ 477, 832 (1997)
Mukai, K., Ishida, M., Osborne, J.P.: pasj 46, L87 (1994)
Napiwotzki, R., Christlieb, N., Drechsel, H., Hagen, H.-J., Heber, U.,
Homeier, D., Karl, C., Koester, D., Leibundgut, B., Marsh, T.R.,
Moehler, S., Nelemans, G., Pauli, E.-M., Reimers, D., Renzini, A.
and Yungelson, L.: Astronomische Nachrichten 322, 411 (2001)
Orosz, J.A., Wade, R.A.: ApJ 593, 1032 (2003)
Paresce, F., de Marchi, G.: ApJ Lett. 427, L33 (1994)
Paresce, F., de Marchi, G., Ferraro, F.R.: Nat 360, 46 (1992)
Podsiadlowski, P., Han, Z., Rappaport, S.: MNRAS 340, 1214 (2003)
Prinja, R.K., Knigge, C., Witherick, D.K., Long, K.S., Brammer, G.:
MNRAS 355, 137 (2004)
Prinja, R.K., Long, K.S., Froning, C.S., Knigge, C., Witherick, D.K.,
Clark, J.S., Ringwald, F.A.: MNRAS 340, 551 (2003)
Prinja, R.K., Ringwald, F.A., Wade, R.A., Knigge, C.: MNRAS 312,
316 (2000)
Proga, D.: ApJ Lett. 592, L9 (2003)
Romani, R.W., Miller, A.J., Cabrera, B., Figueroa-Feliciano, E., Nam,
S.W.: ApJ Lett. 521, L153 (1999)
Rudy, R.J., Meier, S.R., Rossano, G.S., Lynch, D.K., Puetter, R.C.,
Erwin, P.: apjs 121, 533 (1999)
Rutten, R.G.M., Dhillon, V.S., Horne, K., Kuulkers, E.: A&A 283, 441
(1994)
Rutten, R.G.M., Dhillon, V.S., Horne, K., Kuulkers, E., van Paradijs,
J.: nat 362, 518 (1993)
Schenker, K., King, A.R., Kolb, U., Wynn, G.A., Zhang, Z.: MNRAS
337, 1105 (2002)
Schreiber, M.R., Hameury, J.-M., Lasota, J.-P.: A&A 427, 621 (2004)
Selvelli, P., Gilmozzi, R.: Memorie della Societa Astronomica Italiana
70, 565 (1999)
Shara, M.M., Prialnik, D., Shaviv, G.: A&A 61, 363 (1977)
Shore, S.N.: in Hernanz, M., Jose, J. (eds.), Classical Nova Explosions.
pp. 175–187 (2002)
Sigurdsson, S., Phinney, E.S.: ApJ 415, 631 (1993)
Sion, E.M.: PASP 111, 532 (1999)
Sion, E.M., Cheng, F.H., G¨aansicke, B.T., Szkody, P.: ApJ Lett. 614,
L61 (2004)
Sion, E.M., Cheng, F.H.,Sparks, W.M., Szkody, P., Huang, M., Hubeny,
I.: ApJ Lett. 480, L17 (1997)
Sion, E.M., G¨aansicke, B.T., Long, K.S., Szkody, P., Cheng, F., Howell,
S.B., Godon, P., Welsh, W.F., Starrfield, S., Knigge, C., Sparks,
W.M.: ApJ 592, 1137 (2003)
Sion, E.M., Szkody, P., G ¨aansicke, B., Cheng, F.H.,LaDous, C., Hassall,
B.: ApJ 555, 834 (2001)
Sosin, C., Cool, A.M.: ApJ Lett. 452, L29+(1995)
Starrfield, S., Timmes, F.X., Hix, W.R., Sion, E.M., Sparks, W.M.,
Dwyer, S.J.: ApJ Lett. 612, L53 (2004)
Szkody, P., Sion, E., G ¨aansicke, B.T., Howell, S.B.: in G¨aansicke, B.T.,
Beuermann, K., Reinsch, K. (eds.), The Physics of cataclysmic
variables and related objects. pp. 21–30, ASP Conf. Ser.261 (2002)
Townsley, D.M., Bildsten, L.: ApJ Lett. 565, L35 (2002)
Townsley, D.M., Bildsten, L.: ApJ Lett. 596, L227 (2003)
van den Heuvel, E.P.J., Bhattacharya, D., Nomoto, K., Rappaport, S.A.:
A&A 262, 97 (1992)
Verhoeve, P.: AIP Conf. Proc. 605: Low Temperature Detectors 605,
559 (2002)
Vitello, P., Shlosman, I.: ApJ 410, 815 (1993)
Wade, R.A.: MNRAS 208, 381 (1984)
Wade, R.A., Hubeny, I.: ApJ 509, 350 (1998)
Welsh, W.F., Sion, E.M., Godon, P., G¨aansicke, B.T., Knigge, C., Long,
K.S., Szkody, P.: ApJ 599, 509 (2003)
Wickramasinghe, D.T., Wu, K.: MNRAS 266, L1 (1994)
Yungelson, L.R., Livio, M.: ApJ 528, 108 (2000)
Springer

Astrophys Space Sci (2006) 303:69–84
DOI 10.1007/s10509-005-9012-9
ORIGINAL ARTICLE
The Need for Ultraviolet to Understand the Chemical Evolution
of the Universe, and Cosmology
Willem Wamsteker ·Jason X. Prochaska ·
Luciana Bianchi ·Dieter Reimers ·Nino Panagia ·
Andrew C. Fabian ·Claes Fransson ·
Boris M. Shustov ·Patrick Petitjean ·Phillipp Richter ·
Eduardo Battaner
Received: 25 May 2005 / Accepted: 31 August 2005
C
Springer Science +Business Media B.V. 2006
Abstract We identify an important set of key areas where
an advanced observational Ultraviolet capability would have
major impact on studies of cosmology and Galaxy forma-
tion in the young Universe. Most of these are associated
with the Universe at z<3–4. We address the issues asso-
ciated with Dark matter evidence in the local Universe and
the impact of the Warm-Hot Intergalactic Medium WHIM on
the local Baryon count. The motivations to make ultraviolet
W. Wamsteker†
INTA-LAEFF, Madrid, Spain
J.X. Prochaska ()
University of California, Santa Cruz, California, USA
L. Bianchi
GALEX, Space Research Lab., JHU, Baltimore, USA
D. Reimers
Remeis Sternwarte, Hamburg, Germany
N. Panagia
STScI, ESA-RSSD, Baltimore, USA
A.C. Fabian
IOA, Cambridge University, Cambridge, UK
C. Fransson
Stockholm University, Stockholm, Sweden
B.M. Shustov
INASAN, Russian Academy of Sciences, Moscow, Rusland
P. Petitjean
Inst. D’Astrophys., Paris, France
P. Richter
Sternwarte, University of Bonn, Bonn, Germany
E. Battaner
Institute of Physics, University Granada, Granada Spain
(UV) studies of supernovae (SNe) are reviewed and dis-
cussed in the light of the results obtained so far by means
of IUE and HST observations. It appears that UV studies
of SNe can, and do lead to fundamental results not only
for our understanding of the SN phenomenon, such as the
kinematics and the metallicity of the ejecta, but also for ex-
citing new findings in Cosmology, such as the tantalizing
evidence for “dark energy” that seems to pervade the Uni-
verse and to dominate its energetics. The need for additional
and more detailed UV observations is also considered and
discussed.
Finally we show the enormous importance of the UV for
abundance evolution in the Intergalactic Medium (IGM), and
the importance of the He II studies to identify re-ionization
epochs, which can only be done in the UV.
Keywords Ultraviolet astronomy ·Chemical evolution ·
Cosmology ·Galaxy formation ·Supernovae ·Intergalactic
Medium
1. Introduction
Dramatic progress has been made during the past decade in
the acquisition of observational evidence of the contents of
our Universe at high redshift. Special mention can be made
here of the coordinated efforts through the Hubble Deep
Fields (HDF-N and HDF-S; Williams et al., 1996) and the as-
sociated Great Observatory Origins Deep Survey, GOODS
(Giavalisco et al., 2004) multi-wavelength data collection
effort. In addition to these, the results obtained from two ma-
jor mapping efforts: the 2dF survey at the Anglo-Australian
Telescope (Hawkins et al., 2003) and the Sloan Digital Sky
Survey (SDSS; York et al., 2000), have supplied new in-
sights in the stellar and galaxy content of the Universe, in
Springer

70 Astrophys Space Sci (2006) 303:69–84
a redshift range extending from 0 to z>6 and higher. The
results on the structure of the Cosmic Background (CMB) by
the Wilkinson Microwave Anisotropy Probe (WMAP) mis-
sion (Bennet et al., 2003) have, at the same time, allowed a
much more detailed evaluation of the formation of structure
after the first inflationary phases of Big Bang Cosmologies
(Tegmark et al., 2004). However, all these efforts did not
succeed to give a definite answer to the structure formation
epoch, scale and evolution. These questions can not be an-
swered from the high zside alone, because the consequences
for the present state of the Universe are rather different for
different cosmologies. Tegmark et al. (2004) have shown that
the recent WMAP CMB studies appear to be best compatible
with a CDM cosmology (i.e. “vanilla type models with 6
parameters). Further new data on the small scale fluctuations
and the polarization characteristics of the CMB are expected
to be found by ESA’s Planck mission and from the various
studies made from Antarctica.
A complete new window on the high redshift Universe
can be expected to be opened up through the James Webb
Space Telescope mission (JWST; NASA/ESA foreseen to
be launched in 2012) and its precursor mission the Wide-
Field Infrared Survey Explorer (WISE; NASA; launch in
2008), which will perform a new high sensitivity IR sky sur-
vey. Some of the infrared veils appear to be lifted by the
Spitzer Space Telescope Observatory from NASA. Quite
interesting results have been obtained already (e.g. Eyles
et al., 2005). Further progress in this area is expected from
the Herschel mission of ESA expected to be launched in
2007.
All these surveys and future missions characterize the his-
tory of star formation, the evolution of galaxy morphology
and distribution in space, and in combination with the CMB
data, give information on large-scale structure. These results
will have a major effect on our understanding of the part of
the baryonic material which has passed through the process
of Galaxy and star formation in the Universe at all redshifts.
There remain however some fundamental issues which can
not be addressed through any of the observatories foreseen
for the future either on the ground, or in space.
One needs to evaluate how the Universe evolved over
its lifetime, after the first ionization phase. This requires a
proper understanding of the variation in time of a number
of different observables, such as abundance evolution, star
formation rate changes with time, identification of possible
re-ionization phases etc. All of these can only be addressed by
observations of baryonic matter out to redshifts of z≈3–4.
The main questions can be summarized as follows:
1. All luminous material in the Universe has been formed
from the gaseous matter in the Interstellar Matter and the
Intergalactic matter, and little data are available to describe
how this gas is cycled to feed star and galaxy formation.
2. Any self-consistent theory of Star Formation and Galaxy
Formation will need an understanding of the Star Forma-
tion Rate (SFR) including UV data. This will allow us to
incorporate the most massive and most rapidly evolving
stars. They play a critical role in the recycling of matter
in the Universe and are an important factor in the energy
cycling in galaxies. An understanding of these processes
is very important to clarify the nature and extent of “dark
matter”. Similarly, the possible existence and effects of a
major pressure associated with “dark energy” can be ex-
pected to be addressed through adequate observations in
the UV. Due to cosmological redshift we can observe ob-
jects emitting or absorbing in FUV (at z<∼2) only with
a space telescope.
3. The important task is to establish the connection between
the nearby (z<∼2) Universe, covering 80% of the cos-
mic time and containing most of the baryonic matter, and
the early Universe which is being studied in great detail at
the redshifted UV wavelengths with the new generation
of ground based telescopes.
From the results of observations related to the early phases
(CMB and high z quasars and ultraluminous galaxies at
z>3) the following constituents of the Universe have been
derived (=ρ/ρcrit with ρcrit =3H0/8pG =h2×1.88 ×
10−29 g/cm3and H0=h×100 km/s Mpc−1):
Total density of matter-energy =1.02 ±0.02
Dark energy density =0.70 ±0.03
Dark matter density m=0.27 ±0.07
Baryonic matter density b=0.044 ±0.01
Hubble constant h=0.72 ±0.05.
Most of these parameters are referred to, and obtained from
observations related to the first 3 Gyr of the Universe and
therefore are based on strong assumptions and priors. For
example, a major assumption underlying the quoted errors
above, is the adoption of the errors associated with each prior.
In particular, primordial gaussian adiabatic, scale-invariant
density fluctuations are adopted. If, for example, an admix-
ture of 30 per cent isocurvature fluctuations is included, con-
sistency with CMB data is still obtained, but the error bars
are up to an order of magnitude larger. Thus, many of these
assumptions have strong affects on the above indicated con-
stituent distribution in the Universe.
The very information on the (re)cycling of the IGM and
population evolution over90% of the lifetime of the baryonic
Universe is an essential requirement for the understanding of
the a physical transition from the early universe to the current
epoch (14 Gyr) in which we exist.
One of the most direct tests of the standard big bang nu-
cleosynthesis (SBBN) is the determination of the primordial
abundances of the different light elements, especially H, D,
Springer

Astrophys Space Sci (2006) 303:69–84 71
and 4He. The astration (destruction) of deuterium as gas is
cycled through stars will than give also a strong constraint on
the history of star formation (see Epstein et al., 1976). Var-
ious high redshift D/H ratios have been established through
ground-based and HST observations of Lyman limit systems
(and damped Ly-αsystems) of QSO’s with z>2.0. Many
attempts have been made in the recent years to establish the
primordial D/H ratio. Unfortunately the values derived for
D/H show a considerable spread for systems with zabs >2
(e.g. Pettini and Bowen, 2001). The prediction from SBBN
for the 4He abundance is Y=0.247 ±0.02 and a present
day baryon density of bh2=0.0193 ±0.0014 (Burles and
Tytler, 1998), implying a surprisingly high baryon-to-photon
ratio η=5.3±0.4×10−10.
The interesting problem remains that FUSE observations
have shown for the Local Bubble a relatively constant value
For D/H≈1.5×10−5while for somewhat more distant
stars values extending from 7 ppm to 25 ppm are derived for
D/H. It remains therefore an important challenge to find the
“exact” value of b, i.e. to obtain an independent estimate
amount of ordinary (baryonic) matter. The required D/H
values can only be established if enough objects are used
in the redshift range between 0 and 2, as that will allow to
establish the validity of the astration model in the context
of SBBN. A large amount of work is done in the theoretical
aspects of the outstanding cosmological questions, but with-
out good observational evidence to guide the theories for the
time interval between 0 <z<3 -which spans 80% of the
age of the Universe- no real choice can be made between the
different models. Independent of the different models for the
early Universe the major baryonic component of the Universe
at z<3 must be associated with the Inter Galactic Medium
(IGM).
In the following sections we will illustrate the critical role
of the UV domain to clarify the questions raised by this theo-
retically exciting, but observationally very unsatisfactory sit-
uation. We will evaluate the new observational capabilities
needed to address the physical properties of the Intergalactic
Medium (IGM) such as metallicity, dust content, ionization
state, temperature. These will lead to a much improved de-
termination of the total baryonic mass bar in the present
day Universe and its state evolution during the expansion
phase.
In this discussion we will not introduce different shades of
baryonic matter such as concept of “dark baryonic matter”
(Combes, 2003), but will confine ourselves to the distinc-
tion between baryonic matter and the generic term for non-
baryonic matter, i.e. dark matter. The possibility for a non-
zero cosmological constant is maintained conceptually as
“dark energy”, but a detailed discussion of the various mod-
els which can be invoked introducing non-standard physics,
is beyond the scope of this paper. As most accessible diag-
nostics for baryonic matter associated with the IGM lie in the
UV for redshifts z<2, new space missions will be required.
In Section 2 we will discuss the issues associated with
the baryonic mass fraction bar. In Section 3 we will discuss
the relevance of SN as probes of the Universe. In section 4
we will discuss the IGM cycling aspects and abundance is-
sues. In section 5 the ionization state and the re-ionization
epoch(s). Finally in section 6 the instrumental requirements
for future instruments will be outlined. All this should be
seen in the context of the results from the Galactic Evolution
Explorer (GALEX; launch April 2003) which makes the first
sensitive UV Sky Survey ever (Martin et al., 2005). The com-
plete results catalogue of GALEX is expected to be ready in
2007.
2. Baryonic content of the universe
(Weighing the universe)
2.1. What part of baryons do we observe?
The most direct way is to estimate relevant contributions
from observations of all baryonic components. However this
requires that we have identified all baryonic constituents.
Even for those baryonic matter constituents we know, we
can only observe part. From the mass-luminosity relation for
galaxies the luminous matter density is estimated as lumh=
0.002–0.006 i.e only up to 30% of b.
This implies that apart from the fact that a large fraction
of the Universe is made up of Dark Matter and Dark Energy
we do not even have a certainty about the baryons beyond
30%. The primary question to be addressed than is: where
are missing baryons? (Carr, 1994). The known phases of
baryonic matter are:
rCondensed: Stars and gas in or near galaxies; Easily ob-
served
rVery hot (107–8K): intracluster and intragroup gas: X-ray
observations
rDiffuse (Warm 104K) photoionized gas: Ly-alpha ab-
sorbers
rWarm-hot: intergalactic medium (WHIM) at 105–7K:
Difficult to observe – highly ionized low density gas.
(C)DM is required to trigger the formation of structure
after the inflationary phase in BB cosmologies. Support for
the existence of DM has been found in the consistency of
the Ly-αforest results with the predictions from CDM cos-
mologies. It must however be kept in mind that most these
results have been obtained from the ground and thus are re-
lated to the distribution at z>2 (i.e. where redshifted Ly-α
enters the optical passbands). Only very few sightlines have
been studied in the range from 0 <z<2. This has been
Springer

72 Astrophys Space Sci (2006) 303:69–84
mainly caused by a lack of dedicated observing capabilities,
but also by the lack of identified background UV sources. A
new UV mission would, with the expected 105QSO’s in the
GALEX catalogue, allow immediately sufficient statistics to
determine both the distribution and evolution of Ly-αforest
with redshift.
It has been well established that CDM is distributed in dif-
ferent ways throughout the Universe. While at high redshifts
the DM appears to supply the separation between the indi-
vidual Ly-αfilamentary structures, in the massive clusters of
galaxies strong gravitational lensing requires a distribution
closely following the luminous matter. This would naturally
lead to the idea that the DM distribution must show a strong
evolution over the time from z=3 to the current epoch. As
the evidence in the last 10 Gyr for DM is all associated with
observables in a semi-indirect way, it is clear that a firm
understanding of the dominant component of the baryonic
content of the Universe will have a direct and very important
influence on the verification of DM evolution, and supply
the, currently poorly understood, evolution of the Universe
after structure formation.
We will here only comment on the low redshift evidence
for non-baryonic DM and the damped Ly-αsystems (DLA).
As it is uncertain that observables can be defined to identify
the DM associated with Gravitational Lensing and the virial
masses of large clusters, we will not discuss those further
here.
2.2. Damped Ly-αsystems and their origin
We would like to comment on the conclusions derived
from the observations of the damped Ly-αsystems (DLA)
and the Lyman forests. Especially the first are strongly
influenced by the definitions of galaxy sizes, mainly on
the basis of the very nearby galaxies as studied at optical
and radio (21 cm) wavelengths. However, with the higher
sensitivity supplied by the Ly-αabsorption in the UV and
the information on the ionization conditions supplied by
the other strong UV lines (CIV, NIII, He II, NII etc.) a
completely new view can be expected to be derived from the
availability of a large number of background sources from
the GALEX survey. Already early GALEX results on M83
(NGC 5236) have shown that the galaxy extent can be much
larger than assumed in the normal DLA evaluations (Fig. 1).
Also star formation may take place in much more rarified
surroundings than was considered feasible till now (Thilker
et al., 2005). The fact that only a limited number of galaxies
can be expected to be discovered could limit the statistical
value of such studies. On the other hand even a few galaxies
will allow us to study the equivalent of many sightlines.
Through objects like M83, we will be able to study the
environment giving rise to the DLA systems, not only through
the absorption characteristics, but we will also be able to
evaluate the nature of the stars formed in such rarified media,
which is important for the interpretation of the DLA systems
Fig. 1 GALEX Color image of
M83 (NGC 5236) showing star
forming regions in the far outer
disk extending to R∼20 Kpc
from Thilker et al. (2005). This
is nearly 4 times the radius
where the majority of HII
regions are detected. The normal
size of galaxies is indicated by
the contour just touching the
outline of the galaxy body. The
deep 21 cm contours are from
Rogstadt et al. (1974) and
extend to a limit of 1021 nHI/cm2
Springer

Astrophys Space Sci (2006) 303:69–84 73
at higher redshifts. Finally it will be possible to extend the
empirical mass function of the Lyman forest to much lower
surface densities than is possible by any other means. This
as a consequence of the high sensitivity to relatively low
densities of the Ly-αline in the UV.
2.3. Baryons in the warm-hot intergalactic medium
While the diffuse photonionized ionized intergalactic
medium that gives rise to the Ly-αforest is expected to
account for ∼30 percent of the baryons at z=0, the so-
called Warm-Hot Intergalactic Medium (WHIM) at temper-
atures T=105–107K most likely contributes at a similar
level to the cosmological mass density of the baryons in
the local Universe, as predicted by cosmological simulations
(e.g., Cen and Ostriker, 1999). The WHIM is believed to
emerge from intergalactic gas that is shock-heated to high
temperatures as the medium is collapsing under the action of
gravity.
Directly observing this gas phase is a challenging task, as
the WHIM represents a low density (nH∼10−4–10−6cm−3),
high-temperature (T∼105–107K) plasma, primarily made
of protons and electrons together with traces of some highly
ionized heavy elements. The most promising approach to
study the WHIM is the search for absorption features from
the WHIM in the FUV and in the X-ray regime. Five-times
ionized oxygen (OVI) currently is the most important high
ion to trace the WHIM at temperatures of T∼3×105K
in the FUV regime. Recent measurements indeed imply that
intervening OVI absorbers contribute with bar (OVI)∼0.002
to the cosmological mass density at z=0 (e.g., Savage et al.,
2002). Next to high-ion absorption from oxygen and other
metals (e.g., NeVIII; Savage et al., 2005), observations with
STIS (Richter et al., 2004) suggest that WHIM filaments can
be detected in in FUV Ly a absorption of neutral hydrogen
(Fig. 2).
Although the vast majority of the hydrogen in the WHIM is
ionized, a tiny fraction (typically <10−6) of neutral hydrogen
should be present if the gas stays in collisional ionization
equilibrium. Depending on the total gas column density of a
WHIM absorber and its temperature, weak but broad HI Ly-
αabsorption at column densities 12.5<log N(HI) <14.0
may arise from WHIM filaments and can be used to trace
the ionized hydrogen component. Recent STIS observations
imply a mass density of the broad Ly-αabsorbers (BLAs)
of bar(BLA) >0.003. These absorbers therefore represent
a significant baryon reservoir in the low-redshift Universe.
The STIS FUV measurements of the WHIM are encourag-
ing, as they demonstrate that a large fraction of the baryons at
z=0 indeed is hidden in a highly-ionized, high-temperature
intergalactic medium. However, to more precisely pinpoint
the baryon budget of the WHIM and to explore its physical
state, a FUV instrument more sensitive than STIS is required.
Such a new FUV instrument would be of crucial importance
– to significantly improve the statistics of intervening OVI
and broad Ly-αabsorbers, and
– to achieve a higher S/N in QSO FUV absorption line data.
The latter point is particularly important to beat down the
detection limit for FUV WHIM absorbers and to provide
a more reliable estimate of the ionization conditions in the
WHIM.
These considerations will lead to a dramatically improved
determination of the baryonic content of the Universe in the
range of 0 <z<3, and will supply new and important con-
straints to which cosmological theories will have to match.
UV spectroscopic observations will be able to put strong
constraints on the interpretation of the cosmological obser-
vations made of the CMB. Therefore less reliance on priors
is needed for the analysis of structure formation, as derived
from the data on the cosmic background.
With an improved baryonic mass content, we can expect
to be able to make the connections needed for a Universe in
which the CMB and other data on the early Universe can be
understood in a coherent framework of physics in which we
can also exist.
3. Ultraviolet studies of supernovae
3.1. Introduction
Supernovae (SNe) are the explosive death of massive stars
and moderate mass stars in binary systems. They enrich the
interstellar medium of galaxies with most heavy elements
(only C and N can efficiently be produced and ejected into
Fig. 2 The broad Ly-αabsorber
(BLA) at z=0.18047 in the
STIS spectrum of H1821+643
is shown (Richter et al., 2005).
As clearly visible, the shape of
this BLA differs significantly
from the shape of the narrow
z=0.17924 Ly-αforest
absorption near 1433.5 ˚
A
Springer

74 Astrophys Space Sci (2006) 303:69–84
the ISM by red giants winds and by planetary nebulae, as
well as pre-SN massive star winds). The nuclear detonation
supernovae, i.e. Type Ia SNe (SNIa; see Section 3.2.1 below)
provide mostly Fe and iron-peak elements, while core col-
lapse supernovae, i.e., Type II (SNII) and Type Ib/c (SNIb/c),
mostly O and alpha-elements (see section 3.2.2 below).
Through the interstellar medium (ISM) enrichment they are
the primary drivers of the chemical evolution of the Universe.
Additionally, SN ejecta deposit approximately 1051 ergs in
the form of kinetic energy into the ISM of a galaxy. This com-
pares with the total mechanical energy of a Milky Way class
galaxy, which is approximately given by the product (galaxy
mass) ×(rotation velocity)2∼1043 ×4×1014 =4×1057
ergs, approximately equivalent to 4 million SN events. At
a SN rate of 4 SNe/century, these explosions double the
energy of a galaxy in about 100 Myrs (if there where no
losses and dissipation phenomena). It is quite obvious that
one can not ignore this energy input for the evolution of the
entire galaxy, both dynamically and, through cloud compres-
sion/energetics, for star-formation.
SNe are bright events that can be detected and studied up
to very large distances. Therefore they can be used for many
different approaches to trace the evolution of the Universe.
The general feasibility of these different techniques has been
amply and very successfully demonstrated (e.g. Blades et al.,
1988; Sonneborn et al., 1997; Gilmozzi et al., 1987) by the
extensive UV observations of SN1987A with the Interna-
tional Ultraviolet Explorer (IUE). Ultraviolet spectroscopy
is crucially important in order to:
(1) Study the metallicity of individual SNe
(2) Study the metallicity of the intervening ISM/IGM
(3) Study the kinematics of the fast moving (i.e. the outer-
most layers) of the ejecta through the analysis of strong
UV lines with P Cyg profiles
(4) Study the overall energetics of the SN explosion at early
phases (from shock breakout to optical maximum for
types of SNe, but most importantly for SNII).
(5) Study the strong emission lines produced in the interac-
tion of the ejecta with pre-SN circumstellar material, e.g.
NV1240 ˚
A and collisionally excited CIV1550 ˚
A, NIV]
1470 ˚
A, OIII] 1665 ˚
A, NIII] 1750 ˚
A, CIII] 1909 ˚
A, etc.
3.2. Types of supernovae
3.2.1. Nuclear Detonation Supernovae: Type Ia
Type Ia supernovae are characterized by a lack of hydrogen
in their spectra at all epochs and their optical spectra are char-
acterized by a number of broad, deep absorption bands, most
notably the Si II 6355 ˚
A feature (actually the blue-shifted ab-
sorption of the 6347-6371 ˚
A Si II doublet; see e.g. Filippenko,
1997), which dominate their spectra at early epochs. SNIa
are found in all types of galaxies, from giant ellipticals to
dwarf irregulars. However, the SNIa explosion rate, normal-
ized relative to the galaxy luminosity (H or K band) and,
thus relative to the galaxy mass, is much higher -up to a fac-
tor of 16 for the extreme cases of irregulars and ellipticals-
in late type galaxies than in early type galaxies (Panagia,
2000; Mannucci et al., 2005). This suggests that, contrary to
common belief, a considerable fraction of SNIa belong to a
relatively young (age 1 Gyr), moderately massive stellar
population of 3.5M<M(SNIa progenitor)<8M, and that
in present day ellipticals, SNIa are most likely the explo-
sion in stars resulting from the capture of dwarf galaxies by
massive ellipticals.
Classical Type Ia supernovae are important objects
throughout many fields of astrophysics. They are believed to
result from the explosion of an accreting white dwarf (WD)
in a binary system. They can be used to probe the physics of
thermonuclear burning in degenerate or partially degenerate
matter, under conditions not achievable in the laboratory. The
heavy elements they produce play a key role in the chemical
evolution of galaxies, and details of the burning front influ-
ence the elemental relative abundances. SNIa are especially
important because of their possible use as cosmological can-
dles: one uses their observed light curve shape and color to
standardize their luminosities.
3.2.2. Core Collapse Supernovae: Types II and Ib/c
Massive stars (M∗>8M) end their evolution by collaps-
ing onto their inner Fe core and producing an explosion by
a gigantic bounce that launches a shock wave which prop-
agates through the star and eventually erupts through the
photosphere. In this ejection several solar masses of mate-
rial are thrown into the surroundings at velocities of thou-
sands of km/s. The current view is that single stars explode
as type II supernovae, while the supernovae of types Ib and
Ic originate from massive stars in interacting binary systems.
Although the explosion mechanism is essentially the same
in both types, the spectrum and light curve evolution are
markedly different for each.
3.3. Ultraviolet observations
The launch of the International Ultraviolet Explorer (IUE)
satellite in early 1978 marked the beginning of a new era
for SN studies because of its capability to measure the ul-
traviolet emission of objects as faint as mB=15. Moreover,
just around that time, other powerful astronomical instru-
ments became available, such as the Einstein Observatory for
X-rays, the VLA in the radio domain, and a number of tele-
scopes with new and highly efficient IR instrumentation (e.g.
UKIRT, IRTF, AAT and ESO). As a result a wealth of new
multi-wavelength information became available in the early
Springer

Astrophys Space Sci (2006) 303:69–84 75
1980’s. The coordinating efforts of astronomers operating at
widely different wavelengths, have provided us with fresh
insights in the properties and the nature of supernovae of all
types. Eventually, the successful launch of the Hubble Space
Telescope (HST) could have opened new possibilities for the
study of considerably fainter supernovae, allowingus to study
SN spectra with a high accuracy and to reach beyond the local
supercluster .
From 1979 through 1996 all bright supernovae, and a num-
ber of fainter ones have been observed with IUE. A total of
25, out of which 8 are of Type II, 12 Type Ia and 5 Type Ib/c.
Of these 25 only 7 SNe (1979C, 1980K, 1981B, 1983N,
1987A, 1990N, and 1992A) were bright enough to obtain
fair quality ultraviolet spectra and/or to follow their time
evolution (Cappellaro, Turatto and Fernley, 1995). However
even after 18 years of IUE observations and 14 years of HST
observations, the number of SN events that have been stud-
ied in detail with UV spectroscopy remains quite small (no
more than two objects per SN type with high quality spec-
tra for more than three epochs). As a consequence, we still
know very little about the properties and the evolution of the
ultraviolet emission of SNe. On the other hand, the Ultravi-
olet observations of SN1987 (Pun et al., 1995) have shown
that it is just the UV spectrum of a SN, especially at early
epochs, that contains a wealth information that cannot be
obtained at other wavelengths. A full review on the Ultra-
violet observations of Supernovae can be found in Panagia
(2003).
3.3.1. What we Know and Can Learn from UV Spectra
of SNIa
The UV spectra of type Ia SNe decline rapidly with fre-
quency, making them hard to detect at short wavelengths
(λ<2500 ˚
A). This aspect is illustrated in Fig. 3, which dis-
plays the λ>2200 ˚
A spectra of a sample of 10 type Ia SNe
observed with IUE. In all cases, the observing epoch is within
three days of the optical maximum. The spectra do not have a
smooth continuum but rather consist of a number of “bands”
with somewhat different strengths. The most prominent fea-
ture is the apparent emission at λ∼2950 ˚
A with a half-power
width of ∼100 ˚
A, i.e. v ∼104km/s. This band is likely to
be the result an opacity minimum between strong absorptions
on both sides i.e. Mg II centered at ∼2800 ˚
A and Fe II ∼3060
˚
A, each with half-power widths corresponding to the expan-
sion velocity of ∼104km/s. Several other absorption features
can be recognized, which are present at all epochs of observa-
tion. Some of them are most likely associated with multiplets
of Fe I, Fe II and Mg II, but the majority of these absorptions
have not been unambiguously identified. The very fact that
the spectra are so similar for the first three SNe in Fig. 3 and
also at all epochs, is an important result. This supports the
concept of homogeneity in the properties of all type Ia SNe.
On the other hand, some clear deviations from “normal”
can be recognized in Fig. 3. While the UV spectra of most
SNIa shown in Fig. 3 are quite similar, and virtually indis-
tinguishable from the spectrum of SN1992A near maximum
Fig. 3 Ultraviolet spectra of a
sample of Type Ia supernovae
observed with IUE around
maximum light. For comparison
(see text) we show the spectrum
of SN1992A at maximum as a
dotted line
Springer

76 Astrophys Space Sci (2006) 303:69–84
light, one notices that both SN1983G and SN1986G display
excess flux around 2850 ˚
A, and a flux deficiency around 2950
˚
A. This suggests that the Mg II resonance line is consistently
weaker and may indicate a lower abundance of Mg in these
two SNIa. They were characterized by a fast-decline and
some under-luminosity. Contrary, SN1990N, SN1991T, and,
possibly, SN1989M show excess flux around ∼2750 ˚
A and
∼2950 ˚
A and a clear deficit around ∼3100 ˚
A, possibly due to
enhanced Mg II and Fe II features and showed a slow-decline
and over-luminosity.
The best studied SNIa is the “normal” type Ia supernova
SN1992A in the S0 galaxy NGC1380, that was observed both
with IUE and HST (Kirshner et al., 1993). The FOS spectra of
HST from 5 to 45 days past maximum light, are the best UV
spectra available for any SNIa and reveal, with good signal to
noise ratio also the spectral region at below λ≈2650 ˚
A. The
UV photometry taken with the FOC of HST in the F175W,
F275W, and F342W bands shows light curves that resemble
the SNIa template U-band light curve (Leibundgut, 1988).
Using data from SN1992A and SN1990N, Kirshner et al.
(1993) constructed a SNIa template light curve for the flux
region near 2750 ˚
A (Fig. 4) that is quite detailed from 14
days before maximum light to 77 days after maximum light.
This light curve resembles the template U-band light curve
although it drops off a bit faster.
It is thus clear that type Ia supernovae are consistently
weak UV emitters, and even at maximum light their UV
spectra fall well below a blackbody extrapolation of their
optical spectra. Broad features due to P Cygni absorption of
Mg II and Fe II are present in all SNIa spectra, with remark-
able similarity for normal SNIa and systematic deviations for
slow-decline, over-luminous SNIa (enhanced Mg II and Fe II
absorptions) and fast-decline, under-luminous SNIa (weaker
Mg II lines).
Despite the fact that SNIa are relatively weak UV emit-
ters, obtaining UV spectroscopy of them is of fundamental
importance for at least three reasons:
(a) Definition of the UV Luminosity/Light Curve
Shape/Color Relations for Cosmological Applica-
tions: At optical wavelengths SNIa are remarkably
uniform in luminosity (0:16 mag), once corrected for the
light curve shape and color. Recent work suggests the
existence of a general correlation between the U-band
light curve width and luminosity. Almost nothing is
known, however, about SNIa behavior at space-UV
wavelengths.
(b) The Nature of the Progenitors and Explosion Mecha-
nisms of SNIa: We have no idea how a white dwarf (WD)
reaches the Chandra limit (e.g mass accretion from a
main-sequence star, a subgiant, a red giant, or a merger
with another WD), or whether it even reaches the Chan-
dra limit, and how it explodes (off-center detonation, de-
flagration, or pulsed delayed detonation). Establishing
the SNIa progenitor systems and explosion mechanisms
are essential to a reliable use of SNIa as cosmological
probes, and will allow to determine evolutionary trends
from the progenitor age and its initial composition.
(c) Determination of the Metallicity and Other Effects in
SNIa: Given that the main dust extinction related issues
surrounding high-z SNIa appear to be largely resolved
possible evolutionary effects such as metallicity are now
the major unresolved aspect of using SNIa as cosmolog-
ical distance indicators.
Fig. 4 The light curve of
SN1992A in the near ultraviolet
Springer

Astrophys Space Sci (2006) 303:69–84 77
For cosmological studies the apparent brightness of high-
z SNIa must be compared to those of low-z SNIa in order to
measure accurate relative distances. Moreover, the physical
understanding of SNIa, as well as techniques for standardiz-
ing SNIa and correcting for host galaxy dust extinction come
from detailed observations of low-z SNIa.
Even more importantly, modern observations of high-
redshift SNIa have provided evidence for a recent (past sev-
eral billion years) acceleration of the expansion of the Uni-
verse, pushed by “dark energy”. If confirmed, this exciting
result may require new physics (Panagia, 2005 and references
therein; see also Section 1).
Thus, an efficient strategy to elucidate all relevant aspects
about SNIa is:
(a) To study the brightest, closest SNIa to make maximum
progress in advancing our understanding of the physics
of SNIa,
(b) to utilize Hubble flow SNIa, where accurate relative lu-
minosities can be determined, to search for, and constrain
subtle effects that can affect precision cosmology mea-
surements, and
(c) to compare high-z SNIa to local Universe SNIa to de-
termine cosmological parameters accurately and confi-
dently.
As the UV spectra of type Ia SNe decline rapidly with
frequency and time prompt, early UV observations are of
paramount importance
3.3.2. Type Ib/c Supernovae
Type Ib/c supernovae (SNIb/c) are similar to SNIa in not
displaying any hydrogen lines in their spectra. They are
also dominated by broad P Cygni-like metal absorptions,
but they lack the characteristic 6150A trough of SNIa. The
finer distinctions between SNIb and SNIc were introduced by
Wheeler and Harkness (1986) and are based on the strength of
He I absorption lines: the spectra of SNIb display strong He I
absorptions and those of SNIc do not. SNIb/c have been found
only in spiral galaxies associated with spiral arms and/or H
II regions and seem associated with the evolution of massive
stars in close binary systems.
The best observed SNIc is SN1994 both with IUE and
with HST- FOS. The high quality UV spectra were remark-
ably similar to those obtained for SN1983N and were taken
only at two epochs well past maximum light (10 days and
35 days). Synthetic spectra matching inferred a photospheric
velocity decrease from 17, 500 to 7, 000 km/s (Millard et
al., 1999). The kinetic energy carried by the ejected mass is
near the canonical supernova energy of 1051 erg. Such ve-
locities and kinetic energies for SN1994I are “normal” for
SNe and are much lower than those found for the peculiar
type Ic SN1997ef and SN1998bw (see, e.g. Branch, 2000)
which appear to have been hyper-energetic. Type Ib/c super-
novae are, like type Ia, weak UV emitters with the UV much
weaker than blackbody extrapolations of the optical and NIR
spectra. Their typical luminosity is about a factor of 4 lower
than that of SNIa, and thus the mass of 56Ni synthesized in a
typical SNIb/c is only ∼0.15 M.
3.3.3. Type II Supernovae
Type II supernovae display prominent hydrogen lines
(Balmer series in the optical) with a strong continuum and
broad P Cygni lines superimposed. SNII are considered to be
the result of a core collapse of massive stars exploding at the
end of their RSG phase. SN1987A was both a confirmation
and an exception to this model (see Arnett et al., 1989; Pana-
gia, 2003; Pun et al., 1995 for details on SN1987A). There are
two types of SNII, the so-called “linear” type (SNIIL), which
are characterized by an almost straight-line decay of the B
and V-band light curves, and the more common “plateau”
type (SNIIP) which display a flattening in their light curves
a few weeks after maximum light.
The SNII studied best in the UV, is SN1998S, a type II with
relatively narrow emission lines. SN1998S was discovered
several days before maximum. The UV spectral evolution of
SN1998S (Fig. 5) showed the spectrum to become gradually
steeper in the UV, from near maximum light on 16 March
1998 to about two weeks past maximum on 30 March, and
the blue absorptions weaken or disappear completely. About
two months after maximum (13 May 1998) the continuum
was much weaker, although its UV slope had not changed
appreciably, and it had developed broad emission lines, the
most noticeable being the Mg II doublet at about 2800 ˚
A.
This type of evolution is quite similar to that of SN1979C
(Panagia et al., 1980).
Type II plateau (SNIIP) supernovae account for a large
fraction of all SNII. However, the only SNIIP that has been
studied in some detail in the ultraviolet is SN1999em. An
analysis of the early optical and UV spectra (Baron et al.,
2000) indicates that, spectroscopically, this is a normal type
II. Very early spectra combined with sophisticated spectral
modeling can supply an independent estimate of the total
reddening of the supernova. When the spectrum is very blue,
dereddening leads to changes in the blue flux that cannot
be reproduced by altering the “temperature” of the emitted
radiation. Thus, detailed modeling of the early spectra allows
us to determine both the abundance and total extinction of
SNII.
Another sub-type of the SNII family is the so-called type
IIb SNe, which display strong Balmer lines early on, but
later the Balmer lines weaken significantly or disappear alto-
gether (e.g. Filippenko et al., 1997). At this point their spec-
tra become more similar to type Ib SNe. The prototype this
class is SN1993J. An HST-FOS UV spectrum of SN1993J
Springer

78 Astrophys Space Sci (2006) 303:69–84
Fig. 5 UV spectral evolution of
SN1998S (SINS project,
unpublished). Shown are spectra
obtained near maximum light
(March 16, 1998), about two
weeks past maximum (March
30, 1998), and about two months
after maximum (May 13, 1998)
was obtained about 18 days after explosion, and close to
maximum light. This spectrum (Jeffery et al., 1994) shows
that the region between ∼λλ1650–2900 ˚
A is smoother than
observed for SN1987A and SN1992A and lacks strong P
Cygni lines absorptions from iron peak element lines. The
UV spectrum of SN1993J is appreciably fainter than ob-
served in most SNII, thus revealing its “hybrid” nature and
some resemblance to a SNIb. Synthetic spectra calculated
using a parameterized LT procedure and a simple model at-
mosphere do not fit the UV observations. Radio observations
suggest that SN1993J is embedded in a thick circumstellar
medium envelope (Van Dyk et al., 1994). The UV spectra of
other supernovae that are believed to have thick circumstel-
lar envelopes also have the λλ1650–2900 ˚
A regions lacking
strong P Cygni absorptions. Interaction of supernova ejecta
with circumstellar matter may be the origin of the smooth
UV spectrum. UV observations of such supernovae provides
insight in the circumstellar environment of the supernova
progenitors.
Thus, despite their different characteristics in the details
of the UV spectra, all type II supernovae of the various sub-
types appear to provide clear evidence for the presence of
a dense circumstellar medium and enhanced nitrogen abun-
dance. They are important as background UV sources at early
phases with the strong UV excess relative to a blackbody ex-
trapolation of their optical spectra.
3.4. Cosmological applications
As mentioned before, SNIa are very good standard candles
(e.g. Macchetto and Panagia, 1999) to measure distances of
distant galaxies, currently up to redshift z≈1 and, consid-
erably more in the foreseeable future. HST observations of
Cepheids in the parent galaxies of SNIa have lead to very
accurate determinations of their distances and the absolute
magnitudes of normal SNIa at maximum light (e.g. Sandage
et al., 1996; Saha et al., 2001). With these calibrations it
is possible to determine the distances of much more distant
SNIa. The Hubble diagram of distant SNIa (30,000 km/s >
v>3,000 km/s) gives a Hubble constant of H0=59 ±6
km/s/Mpc (Saha et al., 2001) while the independent HST
calibration of SNIa absolute magnitudes at maximum light
from Freedman, Kennicutt, Mould and collaborators (Freed-
man et al., 2001) gave a Hubble constant of H0=71 ±8
km/s/Mpc.
Studying the more distant SNIa (i.e. z>0.1) it has been
possible to extend our knowledge to other cosmological pa-
rameters. These results (Perlmutter et al., 1998, 1999; Riess
et al., 1998; Knop et al., 2003; Tonry et al., 2003; Riess et al.,
2004) suggest a non-empty inflationary Universe, which is
characterized by M≈0.3 and ≈0.7. Correspondingly,
the age of the Universe can be bracketed within the interval
12.3–15.3 Gyrs to a 99.7% confidence level (Perlmutter et al.,
1999).
However systematic uncertainties are uncomfortably large
and observations of more high-z SNIa are absolutely needed.
This is a challenging proposition, both for technical reasons,
in that searching for SNe at high redshifts one has to make
observations in the near IR (because of redshift) of increas-
ingly faint objects (because of distance) and for more subtle
scientific reasons, i.e. one has to verify that the discovered
SNe are indeed SNIa and that these share the same proper-
ties as their local Universe relatives. One can only discern
Springer

Astrophys Space Sci (2006) 303:69–84 79
Fig. 6 Spectra of Type Ia SN1992A and Type IIn SN1998S near
maximum light, normalized so as to have the same average flux in
the rest-frame V band. Upper panel: Linear flux scale, original spec-
tral resolution R∼1500. Middle panel: Resolution degraded to R=50
(low resolution spectroscopy). Lower panel: Magnitude scale (mλ=
−2.5log(Fλ)+const), resolution R=3 (broad-band photometry)
Type I from Type II SNe on the basis of the overall proper-
ties of their UV spectral distributions (Panagia, 2003, 2005),
because Type II SNe are strong UV emitters, whereas all
Type I SNe, irrespective of whether they are Ia or Ib/c, have
spectra steeply declining at high frequencies, as illustrated
in Fig. 6 for SNIa 1992A and SNIIn 1998S. Figure 6 also
shows the same spectra that have been degraded to a resolu-
tion of R=50 (low resolution spectroscopy), and to R=3
(broad-band photometry, expressed in magnitude difference
relative to the V-band; bottom panel). We see that while the
characteristic spectral features are still easily recognized in
the R=50 spectra, the only property that in the R=3 spec-
tra distinguish a SNIa from a SNII is the UV slope. This
technique of recognizing SNIa from their steep UV spectral
slope was first suggested by Panagia (2003), and has been
successfully applied by Riess et al. (2004a,b).
4. The state, abundances and distribution of the
IGM at 0 <z<3
4.1. HI and metal-enrichment
Damped Ly-αsystems derive their name from the observed
quantum mechanical damping of the Ly-αtransition relating
to their very large HI column density N(HI). Because the Ly-
αprofile is dominated by this damping, a standard fit to the
observed profile has two free parameters: (i) the centroid or
zabs; and (ii) N(HI). Therefore, accurate measures of N(HI)
can be acquired with modest resolution and S/N data. and his
collaborators initiated surveys for these galaxies 20 years ago
(e.g., Wolfe et al., 1986; Storrie-Lombardi and Wolfe, 2000)
and the majority of research has been performed on 4 m-class
telescopes. The principal results from these HI surveys are:
the cosmic evolution of DLA, the universal mass density of
neutral gas in units of the critical density. The uncertainties
in the each of the data sets obtained with IUE and HST are
very large and there exists a stark disagreement between the
central values of the two surveys. The uncertainties empha-
size the current challenge of studying HI gas at z<2 with
existing UV spectrographs. While HST/COS would enable
a modest survey of DLA at z<0.6, the parameter space
z=0.6–1.7 will require a next generation space telescope
simply to survey Ly-α.
Aside from the HI content, the most basic measure of the
DLA is metallicity. Because of the large HI surface density of
DLA, ionization corrections are generally small (e.g. Vladilo
et al., 2001) and measurements of low-ions like Fe+,Si
+, and
Zn+yield accurate measures of the metallicity, i.e., [Zn/H] ∼
[Zn+/H0]. The only serious systematic error is dust depletion;
refractory elements like Fe and Si might be depleted from the
gas-phase such that Si+/H0and Fe+/H0are lower limits to the
true metallicity. In general, the depletion levels of the DLA
are small (Pettini et al., 1997; Prochaska and Wolfe, 2002)
and the basic picture is well revealed by any of these elements
at high z. Metallicity observations of a large sample of DLA
present two main results: (1) an N(HI)-weighted mean Z,
which is the cosmic mean metallicity of neutral gas; and (2)
metallicities for a set of galaxies which presumably span a
large range of mass, morphology, and luminosity.
Figure 7 presents over 50 metallicity measurements from
z∼2–4.5 (Prochaska and Wolfe, 2002). The principal re-
sults are: (1) the mean metallicity (weighted or unweighted)
is significantly sub-solar; (2) there is little evolution in the
mean metallicity over this redshift range with the possible
exception of a modest decrease at z>3.5; (3) no galaxy ex-
hibits a metallicity lower than 1/1000 solar. These optical
observations constrain models of chemical evolution at these
epochs (e.g., Pei et al., 1999) and give the first glimpse into
metal production in the early universe. It is crucial, however,
to press to lower redshift. The time encompassed by the red-
shift interval 2 <z<4.5 pales in comparison with z<2. Of
immediate concern is to determine how the mean metallicity
rises to the enrichment level observed today.
4.2. Relative abundances: Dust and nucleosynthesis
High resolution (R>30000), high S/N (>30 per resolution
element), observations of the damped Lyαsystems enable
detailed studies of nucleosynthetic enrichment and dust prop-
erties in the early universe. This level of data quality is crucial
Springer

80 Astrophys Space Sci (2006) 303:69–84
Fig. 7 The upper panel shows metallicity measurements versus redshift
for some 100 DLA. Overplotted are the HI-weighted and unweighted
means in several redshift bins. The lower panel presents the same mea-
surements against cosmic time. It is clear that the cosmic enrichment
history of the universe is severely undersampled over the past 10 Gyr
to achieving the better than 10% precision required by rela-
tive abundance studies. Currently, there is an entire ‘cottage
industry’ focused on this area (Lu et al., 1996; Prochaska and
Wolfe, 1999; Molaro et al., 2000; Pettini et al., 2000; Ledoux
et al., 2002). Figure 8 presents two of the principal results
from these efforts: (a) [Si/Fe] and (b) [N/α] measurements
against [Si/H] metallicity.
The super-solar Si/Fe ratios presented in panel (a) high-
light the greatest obstacle to interpreting relative abundance
measurements from gas-phase abundances: the competing
effects of nucleosynthetic enrichment and differential de-
pletion. In terms of dust depletion, one observes Si/Fe en-
hancements in depleted gas owing to the differential de-
pletion of these two refractory elements. Regarding nucle-
osynthesis, Si/Fe enhancements suggest Type II SN nucle-
osynthesis (e.g., Woosley and Weaver, 1995), whereas solar
ratios would imply Type Ia SN enrichment patterns. Cur-
rently, we interpret the plateau of Si/Fe values at low metal-
licity as the primary result of nucleosynthesis. The mean
enhancement matches the Galactic halo-star observations at
the same metallicity (e.g., McWilliam, 1997) and it would
be difficult to understand why differential depletion would
imply such a uniform enhancement. In contrast, the rise in
Si/Fe at [Si/H] >−1 is highly suggestive of differential deple-
tion. One expects a decrease in Si/Fe from nucleosynthesis
at higher metallicity due to the increasing contribution from
Type Ia SN. Furthermore, larger depletion levels are expected
at higher metallicity. Investigating evolution in abundance ra-
tios like these at z<2 would reveal the detailed enrichment
history of galaxies and the evolution of dust formation.
Overcoming this dust/nucleosynthesis degeneracy is
among the most active areas of DLA research. One avenue is
to focus on special pairs of elements which are largely non-
refractory. Panel (b) is an excellent example of this; plotted
are N/αpairs from recent compilations by Prochaska et al.
(2002), Pettini et al. (2002) and Centuri´on et al. (1998). For
N, S, and Si (the latter two are α-elements), depletion effects
are small and the results show the nucleosynthetic history of
N in the DLA. For comparison, we also plot [N/α], [α/H]
pairs for z∼0 HII regions and stars (see Henry et al., 2000).
The majority of DLA observations fall along the locus of
local measurements, in particular the plateau of N/αvalues
at [Si/H] <−1. In contrast, a sub-sample of low metallic-
ity DLA exhibit much lower N/αvalues which Prochaska
et al. (2002) interpret these in terms of a truncated or top-
heavy initial mass function (IMF). These observations have
important implications for the processes of star formation in
the early universe and measurements at z<2 would assess
the timescale of star formation in these galaxies and further
elucidate the nucleosynthesis of nitrogen.
Finally, we wish to emphasize that a subset of DLA
(<15%) exhibit very strong metal absorption (Prochaska et
al., 2003). These metal-strong DLA allow the measurement
of over 20 elements in a single galaxy including B, Fe, Ge, Pb,
and Sn At low redshift, the incidence of metal-strong DLA
is presumably higher as the mean metallicity approaches the
solar value. Furthermore, confusion with the Ly-αforest is
minimized and high precision measurements of transitions
with λrest <1200 ˚
A are possible. This analysis would re-
quire high S/N, high resolution observations, yet the impact
on studies of nucleosynthesis is extremely impressive.
5. HeII in the intergalactic medium
When the first generation of massive stars had been formed,
the up to then cold and dark universe was ionized and heated
(epoch of reionization). This happened at redshifts z>6. Af-
ter the reionization epoch, completely ionized intergalactic
Hydrogen gives rise to the so called Ly-αforest. Spectro-
scopic observations of the Ly-αforest confirm the general
theoretical picture of an intergalactic medium as a fluctu-
ating distribution of baryons organized by the cosmic web
of dark matter and photoionized by high-redshift starburst
galaxies and later also by QSOs.
Observations of the HeII 304 ˚
A Lyman αline in the line of
sight of a handful of bright QSOs have added to this picture
that HeII is reionized much later at redshift z=2.9 (Reimers
et al., 1997, 2005). This delayed reionization of HeII
Springer

Astrophys Space Sci (2006) 303:69–84 81
Fig. 8 Relative abundances of
(a) Si/Fe and (b) N/αversus
Si/H and α/H where αrefers to
either Si or S For the DLA. The
upper panes highlights the
competing effect of
nucleosynthesis and dust
depletion in interpreting the
gas-phase abundances. We
interpret the plateau of [Si/Fe] at
[Si/H] <−1.5 as the result of
nucleosythesis in Type II
Supernovae. The rise in Si/Fe at
[Si/H] >−1 is associated with
differential depletion (Prochaska
and Wolfe, 2002). UV
observations would allow one to
trace these two processes at
z<2. The lower panel
compares DLA (circles and
triangles; the latter are
upper/lower limits) against N/α
measurements for HII regions
and stars at z≈0 (Prochaska et
al., 2002; Henry et al., 2000).
Although the majority of the
DLA lie on a N/αplateau
associated with metal-poor HII
regions, a significant sub-sample
is identified at N/α < −1. This
sub-sample could present
evidence for a truncated or
top-heavy IMF (Prochaska et
al., 2002). Observations of N/α
at z<2 would be able to
confirm this result
compared to HI had been predicted by theoreticians as caused
by the later appearance of AGN (only their hard, nonthermal
radiation can ionize HeII) compared to stars and by the 5–6
times higher HeIII recombination rate. The apparent 1.4 Gyr
lag between HI and HeII reionization carries information on
the relative beginnings of galaxy (stars) and AGN formation.
Its quantitative understanding will pin down the first epoch
of star formation, independent of direct detection methods.
The HeII Ly-αforest has been observed and spectroscop-
ically resolved in the line of sight of two bright QSOs with
FUSE (Kriss et al., 2001; Reimers et al., 2004). The earlier
finding of a transition from continuous optically thick HeII
304 ˚
A absorption for z>2.9 (HeII Gunn-Peterson trough)
to a resolved HeII Ly αforest at z<2.9 has been confirmed,
i.e. the epoch of HeII reionization is in fact around z=3. For
z<2.9, the resolved HeII 304 ˚
A forest lines are observed
to be generally much stronger than the HI Ly αforest by
the factor η=N(HeII)(N(HI). The factor ηdepends mainly
on the ratio of the intensities of the ionizing radiation at the
Lyman edge to the HeII 228 ˚
A edge J(911 ˚
A)/J(228 ˚
A) and
is roughly 80. However, its value and correspondingly the
shape of the ionizing radiation field appears to vary strongly
between η∼1 and 400 on the scale of 1.3 Mpc h−170, much
smaller than the typical mean distances between AGN at
z=3 of roughly 30 Mpc (Shull et al., 2004). If these fluctu-
ations are real, this would have considerable implications for
elemental abundances in the diffuse IGM. Part of the fluc-
tuations can possibly be explained by a combination of the
Springer

82 Astrophys Space Sci (2006) 303:69–84
large range of EUV spectral shapes of AGN with additional
filtering (softening) by absorption through the IGM.
Future observations of the HeII 304 ˚
A forest at high spec-
tral resolution and S/N have the potential to map both the In-
tergalactic Matter and Radiation Field in very much detail. At
first, since the Ly-αforest “clouds” have kinetic temperatures
of ∼104K, the lines are largely thermally broadened (with
an additional turbulent/expansion component). This means
that from a comparison of line widths of HI with HeII (the
thermal widths of HeII lines are 2 times smaller) the kinetic
temperature of the Ly-αforest gas can be measured directly
with high spatial resolution. From theoretical modelling of
the IGM we know that after reionization, the IGM cools by
expansion, is reheated by the delayed HeII reionization at
z=3 and continues to cool with decreasing redshift z. Ob-
servations of the HeII Ly-αforest over the redshift range
2.1≤z≤2.9 will test this in the most direct way. Besides
observing the evolution of the mean IGM temperature, we
will observe its fluctuations and the latter’s relation to the
fluctuating radiation field. Both IGM temperature, density
and the ionizing background radiation field can be measured
at a spatial resolution of less than 1 Mpc (co-moving).
Having measured these quantities, the existing theoretical
models of the IGM can be tested with much more detail than
possible today, when basically only the statistical properties
of theoretical Ly-αforests are compared with observations.
Due to the possibility to observe simultaneously HI and
HeII spectroscopically, the redshift range 2.1≤z≤2.9cov-
ers the only cosmic epoch where the IGM and its evolution
can be studied in detail. As such, this is a test bed for the-
oretical modelling of the IGM. If successful, it can be ap-
plied to the epoch 1 ≤z≤2 where the major star formation
takes place. By-products will be improved determinations of
the cosmic baryon density (at z=3 more than 95% of the
baryons are not in stars). Studies of the distribution of galax-
ies close to the lines of sight of QSO’s with HeII forest ob-
servations will allow to study the influence of galactic winds,
SN explosions as well as UV radiation of starburst galaxies
and AGN on the temperature and density of the IGM.
For z<2.1 information on the shape of the ionizing
background will never be available from HeII/HI due to ab-
sorption by interstellar HI in our galaxy. This means that the
composition of most of the baryonic matter in the universe
for z<2.1 will be difficult to measure. On one hand, DLAs
where abundance studies are easy contain only of the order
of 5% of the baryonic component. On the other hand, the
determination of the composition of the remaining 95%, the
diffuse highly ionized component, requires knowledge of
the shape of the ionizing background. Therefore, the only
way to study heavy element abundances and the physical
state of this component Is to observe simultaneously several
ionization stages of abundant elements, e.g. OIII-OVI, CIII,
CIV, SIII-SVI, NeIII-NeVII, with the aim to model the ion-
izing UV background radiation field and the physical status
in absorption systems. Except a few single ions like SiIV,
CIV, and OVI observable from the ground and which are not
sufficient to determine the state of ionization, all relevant
lines are in the intrinsic EUV at rest wavelengths between
300 and 900 Angstroms. Consequently, quantitative infor-
mation on the bulk of baryons in the diffuse IGM for z<2is
only available by UV spectroscopy of QSO’s in the satellite
UV. In fact, most of the mentioned ions have been observed
in the UV in a few bright QSO’s like HS 1700+6416 with
the Hubble Space Telescope (Reimers et al., 1992).
6. Instrumentation roadmap for the cosmological
questions
We will here try to give first approach to the instrumental re-
quirements needed to address the questions which have been
raised in Sections 2–5. As the roadmap concept pursued un-
der NUVA is solely related to the UV, we will only address
globally the needs and requirements. As the most developed
technique for UV astrophysics are spectroscopy and quanta-
tive imaging (i.e. photometry), we will confine ourselves to
these two. And since the only way to make UV observations
is from space, The options are limited to those associated
with a telescope in space.
Independently of the technique – photometry or
spectroscopy- the requirements space can essentially be split
into two different groups distinguished by some very primary
technical constraints and a major difference in cost. The fun-
damental difference of the two approaches is in the size of
the associated telescopes. One can essentially separate this
in two major classes:
1. The 2-m class and
2. The 6-m class.
One can than try to establish the reachable goals for each
of these and evaluate the importance of each of these through
the net effect of being able to go to fainter flux levels. Al-
though the possibility exists with the larger telescopes to
obtain higher resolution this will require a major technology
development programme.
Examples for each of these two classes have been dis-
cussed on various occasions in the scientific literature and
at various workshops (e.g. Wamsteker and Shustov, 2004 for
the 2-m class and Shull, 2003 for the >4-m class). In practice
there are not very many differences in the type requirements
for the instrumentation. The main difference in capabilities
between the two classes is given by the about 10-fold in-
crease in sensitivity between group 1 and group 2. A second
also important difference is the cost. A rough estimate of
the cost ratio of the two classes is a factor of 5, in the sense
that the cost for a 6-m class mission is at least 5 times the
Springer

Astrophys Space Sci (2006) 303:69–84 83
cost of a 2-m class mission. This is mainly associated with
technology developments which are required before a high
quality UV telescope with a diameter larger then4mcan
be expected to be successfully launched. Therefore there is
a clear need for a roadmap in the UV, since otherwise no
coherent planning and technology development will be im-
plemented in the expectation that at some stage the results of
this are sufficient that a major (6-m class) UV mission can
be realistically considered.
Taking the position that a, in the UV, diffraction limited
6-m telescope will allow the range over which objects can
be studied to be some ten times more distant, the science
questions to be addressed with a 2-m class are very similar
at the current stage of knowledge.
Although the requirements for the various subject areas
considered in this paper are not completely identical the re-
straining properties of the instrumentation on a UV space
telescope can be summarized as follows:
– a resolution R∼1000 is recommended for spectroscopic
studies of SN ejecta.
– For ISM studies, a resolution R>50,000, and possibly as
high as R∼100,000 is recommendable.
Wavelength coverage. For the majority of QAL research,
coverage from Ly-α(1215 ˚
A) to 2000 ˚
A rest-frame is es-
sential. One of the problems associated with the wavelength
coverage in the UV is that in the FUV (i.e. λ<1200 ˚
A) com-
pletely different technology is required than for the longer
wavelengths. Many projects (e.g. D/H measurements, photo-
ionization assessment, H2observations) need coverage down
to 900 ˚
A rest-frame. Therefore we are not only considering
a single instrument but the full package of science described
here would need an instrument which covers both the far UV
and the near UV domain with associated increase in com-
plexity and cost.
Resolution to examine the DLA or perform QAL studies
in general is dependent on the specific research area. Never-
theless, we can lean on our extensive experience with high
z QAL studies with optical spectrographs from the ground.
For the majority of scientific applications R=30,000 is a
bare minimum. Only at this resolution can one confidently
distinguish Ly-αclouds from metal lines in the Ly-αforest,
obtain abundance measurements to greater than 0.1 dex pre-
cisions, resolve velocity fields, and investigate a multi-phase
medium.
S/N, a good lower limit is 30 per resolution element (i.e.
S/N=15 pix−1for 4 pixel sampling). At such level, one can
carefully address systematic effects like continuum place-
ment and analyze the absorption line diagnostics with a large
dynamic range (e.g. Si II1808, C IV 1550, O VI 1030). As
is obvious, higher S/N is desirable and many applications
would depend on higher sensitivity.
Observing power. To allow observations of a large enough
sample for QSO’s at Z<2, a UV telescope must achieve the
above resolution, S/N, and wavelength coverage for a QSO
at V≈18 in a reasonable exposure time (<10 h =36 Ksec).
This would provide enough targets to examine the physical
conditions at similar levels as achieved at z>2.
References
Arnett, W.D., Bahcall, J.N., Kirshner, R.P., Woosley, S.E.: ARA&Ap
27, 629 (1989)
Baron, E., et al.: ApJ 545, 444 (2000)
Bennet, C.L., et al.: ApJ. Suppl. Ser. 148, 1 (2003)
Blades, J.C., Wheatley, J.M., Panagia, N., Grewing, M., Pettini, M.,
Wamsteker, W.: ApJ 332, L75 (1988)
Branch, D.: in Livio, M., Panagia, N., Sahu, K. (eds.), The largest ex-
plosions since the big bang: Supernovae and gamma ray bursts,
Cambridge University Press, Cambridge, p. 96 (2000)
Burles, S., Tytler, D.: ApJ 499, 699 (1998)
Cappellaro, E., Turatto, M., Fernley, J.: ESA SP-1189, IUE-ULDA
Guide No. 6, Supernovae, Ed. W. Wamsteker (1995)
Carr, B.: ARA&A 32, 531 (1994)
Cen, R., Ostriker, J.: ApJ 514, 1 (1999)
Centuri´on, M., Bonifacio, P., Molaro, P., Vladilo, G.: ApJ 509, 620
(1998)
Combes, F.: New Astr. Review 46, 755 (2003)
Epstein, R.I., Lattimer, J.M., Schramm, D.N.: Nature 268, 198 (1976)
Eyles, L.P., et al.: MNRAS, Mon. Not. Roy. Astron. Soc. 364, 443–454
(2005)
Filippenko A.: ARA&A 35, 309 (1997)
Freedman, W.L., et al.: ApJ 553, 47 (2001)
Giavalisco, M., et al.: ApJ 600, L93 (2004)
Gilmozzi, R., et al.: Nature 328, 318 (1987)
Hawkins, E., et al.: MNRAS 346, 78 (2003)
Henry, R.B.C., Edmunds, M.G., K¨oppen, J.: ApJ 541, 660 (2000)
Jeffery, D.J., et al.: ApJ 421, L27 (1994)
Kirshner, R.P., et al.: ApJ 415, 589 (1993)
Knop, R.D., et al.: ApJ 598, 102 (2003)
Kriss, G.A., Shull, J.M., Oegerle, W. et al.: Science 293, 1112 (2001)
Ledoux, C., Bergeron, J., Petitjean, P.: A&A 396, 429 (2002)
Leibundgut, B.: PhD Thesis, University of Basel (1988)
Lu, L.-M., et al.: ApJ. Suppl.Ser. 107, 475 (1996)
Mannucci F. et al.: A&A, in press [astro-ph/0411450] (2005)
Martin, D., et al.: ApJ 619, L1 (2005)
Millard, J. et al.: ApJ 527, 746 (1999)
McWilliam, A.: ARAAp 35, 503 (1997)
Molaro, P., et al.: ApJ 541, 54 (2000)
Panagia, N.: in Weiler, K.W. (ed.), Supernovae and Gamma-Ray
Bursters, Springer-Verlag, Berlin, pp. 113–144 (2003)
Panagia, N.: in Calamai, G., Mazzoni, M., Stanga, R., Vetrano, F. (eds.),
Experimental Physics of Gravitational Waves, World Scientific,
Singapore, p. 107 (2000)
Panagia, N.: in Giovannelli, F., Mannocchi, G. (eds.), Frontier Objects
in Astrophysics and Particle Physics, It. Phys. Soc., in press [astro-
ph/0502247] (2005)
Panagia N. et al.: MNRAS 192, 861 (1980)
Pei, Y.C., Fall, S.M., Hauser, M.G.: ApJ 522, 604 (1999)
Perlmutter S. et al.: Nature 391, 51 (1998)
Perlmutter S. et al.: ApJ 517, 565 (1999)
Pettini M., Bowen, D.V.: ApJ 560, 41 (2001)
Pettini, M., Ellison, S.L., Bergeron, J., Petitjean, P.: A&A 391, 21 (2002)
Pettini, M., Ellison, S.L., Steidel, C.C., Shapely, A.L., Bowen, D.V.:
ApJ 532, 65 (2000)
Springer

84 Astrophys Space Sci (2006) 303:69–84
Prochaska, J.X., Wolfe, A.M.: ApJS 121, 369 (1999)
Prochaska, J.X., Wolfe, A.M.: ApJ 566, 68 (2002)
Prochaska, J.X., Henry, R.B.C., O’Meara, J.M., Tytler, D., Wolfe, A.M.,
Kirkman, D., Lubin, D., Suzuki, N.: PASP 114, 933 (2002)
Prochaska, J.X., Howk, J.C., Wolfe, A.M.: Nature 423, 57 (2003)
Pun, C.S.J., et al.: Ap.J. Supp.Ser. 99, 223 (1995)
Reimers, D., Fechner, C., Kriss, G.A., Shull, J.M. et al.: Proceed. FUSE
Conf. Victoria, [astro-ph0410588v1] (2004)
Reimers, D., Vogel, S., Hagen, H.-J. et al.: Nature 360, 561 (1992)
Reimers, D., Fechner, C., Hagen, H.-J., Jakobsen, P., Tytler, D. and
Kirkman, D.: A&A, in press [astro-ph/0507178] (2005)
Richter, P., Savage, B.D., Tripp, T.M., Sembach, K.R.: ApJS 153, 165
(2004)
Riess A.G. et al.: AJ 116, 1009 (1998)
Riess A.G. et al.: ApJ 600, L163 (2004a)
Riess A.G. et al.: ApJ 607, 665 (2004b)
Rogstad, D.H., Lockart, I.A., Wright, M.C.H.: ApJ 193, 309 (1974)
Saha, A., Sandage, A., Tammann, G.A., Dolphin, A.E., Christensen, J.,
Panagia, N., Macchetto, F.D.: ApJ 562, 314 (2001)
Sandage, A., Saha, G.A., Tammann, L., Labhardt, N., Panagia, F.D.,
Macchetto: ApJ 460, L15 (1996)
Savage, B.D., Sembach, K.R., Tripp, T.M., Richter, P.: ApJ 564, 631
(2002)
Savage, B.D., Lehner, N., Wakker, B.P., Sembach, K.R., Tripp, T.M.:
astro-ph 0503051 (2005)
Shull, J.M.: in Sembach, K.R., Blades, J.C., Illingworth, G.D.,
Kennicutt, R.C. (eds.), Hubble’s Science Legacy; Future Optical –
Ultraviolet Astronomy from Space, ASP Conf series Vol. 291, p. 17
(2003)
Sonneborn, G., et al.: ApJ 477, 848 (1997)
Storrie-Lombardi, L.J., Wolfe, A.M.: ApJ 543, 552 (2000)
Tegmark, M., et al.: ApJ 606, 702 (2004)
Thilker, et al.: ApJ 619, L79 (2005)
Tonry, J.L.: ApJ 594, 1 (2003)
Van Dyk, S., et al.: ApJ 432, L115 (1994)
Vladilo, G., Centuri´on, M., Bonifacio, P., Howk, J.C.: ApJ 557, 1007
(2001)
Wheeler, J.C., Harkness R.P.: in Galaxy Distances and Deviations from
Universal Expansion (1986)
Wamsteker, W., Shustov, B.M., in: Wamsteker, W., Albrecht, R.,
Haubold, H.J. (eds.), Developing Basic Space Science World-
Wide, Kluwer Acad. Publ., p. 373 (2004)
Williams, R.E., et al.: AJ 112, 1335 (1996)
Wolfe, A.M. et al.: ApJS 61, 249 (1986)
Woosley, S.E., Weaver, T.A.: ApJS 101, 181 (1995)
York, D., et al.: AJ 120, 1579 (2000)
Springer

Astrophys Space Sci (2006) 303:85–102
DOI 10.1007/s10509-005-9032-5
Starbursts at Space Ultraviolet Wavelengths
Rosa M. Gonz´alez Delgado
Received: 8 July 2005 / Accepted: 2 December 2005
C
Springer Science +Business Media B.V. 2006
Abstract Starbursts are systems with very high star forma-
tion rate per unit area. They are the preferred place where
massive stars form; the main source of thermal and mechani-
cal heating in the interstellar medium, and the factory where
the heavy elements form. Thus, starbursts play an important
role in the origin and evolution of galaxies. The similari-
ties between the physical properties of local starbursts and
high-zstar-forming galaxies, highlight the cosmological rel-
evance of starbursts. On the other hand, nearby starbursts are
laboratories where to study violent star formation processes
and their interaction with the interstellar and intergalactic
media, in detail and deeply. Starbursts are bright at ultravi-
olet (UV) wavelengths, as they are in the far-infrared, due
to the ‘picket-fence’ interstellar dust distribution. After the
pioneering IUE program, high spatial and spectral resolution
UV observations of local starburst galaxies, mainly taken
with HST and FUSE, have made relevant contributions to
the following issues:
–The determination of the initial mass function (IMF) in
violent star forming systems in low and high metallicity
environments, and in dense (e.g. in stellar clusters) and dif-
fuse environments: A Salpeter IMF with high-mass stars
constrains well the UV properties.
–The modes of star formation: Starburst clusters are an
important mode of star formation. Super-stellar clusters
have properties similar to globular clusters.
–The role of starbursts in AGN : Nuclear starbursts can dom-
inate the UV light in Seyfert 2 galaxies, having bolometric
R.M. Gonz´alez Delgado
Instituto de Astrof´ısica de Andaluc´ıa (CSIC), Apdo. 3004, 18080
Granada, Spain
luminosities similar to the estimated bolometric luminosi-
ties of the obscured AGN.
–The interaction between massive stars and the interstel-
lar and intergalactic media: Outflows in cold, warm and
coronal phases leave their imprints on the UV interstellar
lines. Outflows of a few hundred km s−1are ubiquitous
phenomena in starbursts. These metal-rich outflows and
the ionizing radiation can travel to the halo of galaxies and
reach the intergalactic medium.
–The contribution of starbursts to the reionization of the
universe: In the local universe, the fraction of ionizing
photons that escape from galaxies and reach the inter-
galactic medium is of a few percent. However, in high-z
star-forming galaxies, the results are more controversial.
Despite the very significant progress over the past two
decades in our understanding of the starburst phenomenon
through the study of the physical processes revealed at satel-
lite UV wavelengths, there are important problems that still
need to be solved. High-spatial resolution UV observations
of nearby starbursts are crucial to further progress in under-
standing the violent star formation processes in galaxies, the
interaction between the stellar clusters and the interstellar
medium, and the variation of the IMF. High-spatial resolu-
tion spectra are also needed to isolate the light from the center
to the disk in UV luminous galaxies at z=0.1–0.3 found by
GALEX. Thus, a new UV mission furnished with an interme-
diate spectral resolution long-slit spectrograph with high spa-
tial resolution and high UV sensitivity is required to further
progress in the study of starburst galaxies and their impact
on the evolution of galaxies.
Keywords UV astronomy ·Starburst galaxies ·Galaxy
evolution
Springer

86 Astrophys Space Sci (2006) 303:85–102
1. Introduction
1.1. Starburst galaxies: Definition and general
properties
Starbursts are a very significant component of the universe.
They are the preferred place for the formation of massive
stars, and hence they are a relevant energy source that drives
the cosmic evolution of galaxies. Heckman (1998) finds that
in the local universe, within 10 Mpc, the four most lumi-
nous starburst galaxies (M82, NGC 253, NGC 4945, M83)
account for about 25% of the recent star formation rate in
this volume. Massive stars have a significant impact on the
evolution of galaxies. They are responsible for the thermal
and mechanical heating of the interstellar medium. They are
the factory where most of the heavy elements form; which
are dispersed throughout the interstellar medium when mas-
sive stars explode as supernovae. From this enriched gas, new
stars will form.
Starburst galaxies are systems with a high star formation
rate. However, this rate can be sustained for much less than a
Hubble time, because the gas reservoir in a galaxy may only
last for a few 108yr (the gas consumption time). Starbursts
have a significant large population of massive stars that are
able to produce large numbers of Lyman continuum photons
to ionize the interstellar medium. When the gas cools down,
hydrogen Balmer and other recombination lines form with
intensities that can exceed 1039 erg s−1. A fraction of the
Lyman continuum photons are, however, absorbed by dust
grains and their energy is re-emitted at far-infrared wave-
lengths.
Starburst galaxies are mainly selected on the basis of their
strong continuum at ultraviolet (UV) wavelengths (defined
here simply as the range 900 ˚
A–3300 ˚
A), their nebular optical
emission lines, and/or strong far-infrared radiation. Due to
these selection criteria, starbursts constitute a mixed type of
star-forming systems, which include:
– Giant-extragalactic HII regions, such as 30 Doradus in the
Large Magellanic Cloud, which is considered by Walborn
(1991) as a Rosetta Stone. These star-forming regions are
regarded as ministarbursts.
– Starburst dwarf galaxies, such as IZw 18. They have blue
colors and an optical HII region spectrum, but with signs
of an underlying stellar population older (a few 108yr
to 1 Gyr) than the ionizing population. They include HII
galaxies (Terlevich et al., 1991), blue compact galaxies,
and blue irregular galaxies, such as NGC 1569.
– Nuclear starbursts, such as the prototype NGC 7714
(Weedman et al., 1981). They have a strong UV contin-
uum, and strong optical emission lines. Their hosts are
spiral galaxies. Balzano (1983) found that about 40% of
the Markarian galaxies can be classified as nuclear star-
bursts.
– Very luminous infrared galaxies. The IRAS satellite has
discovered many of these galaxies. They have far-infrared
luminosities larger than 1011 L. In most of these galax-
ies, the far-infrared flux is thermal emission by dust grains
heated by massive stars. A typical very luminous far-
infrared starburst is NGC 1640.
– Lyman break galaxies (LBG). They are star-forming
galaxies at cosmological distances (z≥2) (Steidel et al.,
1996; Williams et al., 1996) that provide a significant
fraction of the global star formation rate of the universe
(Madau et al., 1996). LBG show a strong rest-UV con-
tinuum with absorption lines very similar to those of
local UV-bright starburst galaxies (Meurer et al., 1997;
Gonz´alez Delgado et al., 1998a; Heckman et al., 2005).
The most famous LBG is MS1512-cB58 at z=2.7276,
known simply as cB58.
Terlevich (1997) proposed to distinguish between star-
burst galaxy and starburst region. The former is when the
galaxy luminosity is totally provided by the starburst, while
in a starburst region, the starburst luminosity is substan-
tial but smaller than the galaxy luminosity. So, starbursts
may simply be defined as compact (10–103pc) sites of
recent star formation (106–108yr), that often show dust
obscuration.
In the conference ‘Starbursts – From 30 Doradus to Lyman
break galaxies’ (de Grijs and Gonz´alez Delgado, 2005),
Heckman (2005) has argued against the gas consumption
time definition of starburst, proposing an alternative defini-
tion. The inverse of the consumption time, b, is related to the
birth-rate parameter; bis the ratio between the current and
the past average star formation. This parameter varies signif-
icantly and systematically with the properties of the galaxy,
and leads to a steep decline in the fraction of starbursts with
increasing galaxy mass, and a strong redshift dependence.
For this reason, Heckman proposes a more physically mean-
ingful definition, which is based on the star formation in-
tensity. Starbursts are defined as systems that have a star
formation rate per unit area which is much larger than that
in the disks of normal galaxies. Nearby starbursts have star
formation intensities ranging from 1 to 100 Myr−1kpc−2,
and similar values are found for LBG (Meurer et al., 1997).
The Galaxy Evolution Explorer (GALEX) satellite
(Martin et al., 2005) has already made a significant contri-
bution to establish the physical properties of starburst galax-
ies. Two categories of local (0.1≤z≤0.3) UV luminous
galaxies (UVLG) have been found (Heckman et al., 2005).
The main differences arise from the different UV luminosity
per unit area, i.e., the variation in the star formation inten-
sity. The large UVLG (IFUV ≤108Lkpc−2) are not star-
burst galaxies. They are massive, late-type disk galaxies that
Springer

Astrophys Space Sci (2006) 303:85–102 87
have star formation rates sufficient to build their stellar mass
in a Hubble time. In contrast, compact UVLG, which can
clearly be classified as local starbursts, are low-mass galax-
ies (Mstar ∼1010 M) with half-light radii less than a few
kpc. They have large enough star formation rates to build the
present galaxy in ∼1–2 Gyr.
1.2. The relevance of space UV observations of
starburst galaxies
UV observations of starbursts are relevant because:
– This range is very sensitive to the star formation history.
In fact, the UV energy distribution shows a strong evolu-
tion from very young to intermediate age (∼1 Gyr) stellar
populations.
– The UV light allows a direct detection of massive stars,
thus, it provides a direct measurement of the star formation
rate.
– UV wavelengths contain valuable tracers of the cold and
molecular phases of the interstellar medium, and are an
important probe of the ionized interstellar medium in star-
burst galaxies. Low and high-ionization absorption lines
allow us to study the interaction of the starburst with the
interstellar medium in an ample range of physical (density
and temperature) conditions.
We owe much of our understanding of the starburst phe-
nomenon to the International Ultraviolet Explorer (IUE)
which provided the first UV (1200–3300 ˚
A) spectra of star-
burst galaxies (Kinney et al., 1993); the Hubble Space Tele-
scope (HST), for its impact with UV high spatial resolu-
tion images and spectra of starbursts (Meurer’ et al., 1995;
Leitherer et al., 1996); the Hopkins Ultraviolet Telescope
(HUT) and FUSE for collecting spectra of starbursts below
Lyα, down to the Lyman limit (Leitherer et al., 1995b). The
impact of the GALEX mission has just started (see the first
results in ApJL, volume 619), but there is no doubt about its
important contribution to our understanding of the cosmic
evolution of galaxies and, in particular, of starbursts.
Along this paper, we discuss the relevance of high spatial
and spectral resolution observations of starbursts at space UV
wavelengths and their impact on the following issues: the
stellar content of starbursts, the interaction of massive stars
with the interstellar medium, the relation between starburst
clusters and the formation of globular clusters, the role of
starbursts in AGN, and the contribution of starbursts to the
reionization of the universe.
2. UV imaging morphology
HST has been the first telescope to provide UV high spa-
tial resolution images of nearby starbursts. FOC, WFPC2,
STIS and ACS/HR on board HST have been able to dis-
sect the anatomy of nearby starbursts with a spatial sam-
pling better than 0.025 arcsec/pixel (Fig. 1). Other instru-
ments, on board UIT, FOCA and now GALEX, are quite
useful to study the general UV morphology of galaxies
at intermediate resolution (e.g. Bianchi et al., 2005) and
Fig. 1 UV images of starburst
galaxies taken with HST+FOC
(NGC 1741), and HST+STIS/
MAMA (NGC 4214, NGC 3049
and Tololo 89) with a spatial
sampling of 0.014 and 0.0244
arcsec, respectively; 1 arcsec
corresponds to 260 pc (NGC
1741), 25 pc (NGC 4214), 80 pc
(NGC 3049) and 100 pc (NGC
3049). Clusters and diffuse
extended emission are detected.
Springer

88 Astrophys Space Sci (2006) 303:85–102
to map the extended outflows in starbursts (Hoopes et al.,
2005), but they have much less spatial resolution than
HST.
In a pioneer work, Meurer et al. (1995) obtained HST/FOC
images (at 2200 ˚
A with a spatial sampling of 0.014 arc-
sec/pixel) of a sample of 9 starbursts, selected for their high
UV flux in the IUE aperture (Kinney et al., 1993). All galaxies
show an irregular UV morphology, but they reveal two im-
portant structural characteristics: compact knots embedded
in a diffuse UV background.
Compact knots are marginally resolved stellar clusters and
provide about 25% of the total UV emission. These clusters
are distributed irregularly over the UV background, but the
brightest ones are located in the center of the starburst. Their
UV absolute magnitude ranges from −19 to −10, and their
sizes are less than 10 pc. Their masses, estimated from the
UV luminosity, range from 104to 107M. These knots have
ages of a few Myr to ∼100 Myr, and they may be formed in
bursts. The brightest knots are named super-stellar clusters
(SSCs). The luminosity function (LF) of the stellar clusters
follows a power law, dN/dL∼L−α, with index 1.5≤α≤2,
similar to that found in merger systems observed at optical
wavelengths (Whitmore et al., 1993). Meurer (1995) argues
that SSCs have properties similar to globular clusters if
fading and a spread of the star formation time is taken into
account.
The UV diffuse emission accounts for about 75% of the
total UV flux from the starburst. It extends about a few 100
pc. Several origins have been proposed: (a) Continuous star
formation (csf) lasting for a few 100 Myr; originally, the
stars form in clusters over the last few 100 Myr, but clusters
dissolve with age and disperse across the field. (b) The UV
radiation originates in dusty compact stellar clusters, but it
is scattered by dust to the field. (c) Individual massive stars,
unresolved even at the HST spatial resolution. Long-slit UV
spectra of starbursts taken with STIS point to the csf origin,
so that the UV field light is created via dissipation of aging
star clusters (Tremonti et al., 2001; Chandar et al., 2005).
3. UV spectral morphology
Starbursts are recognized at optical wavelengths by their neb-
ular emission line spectrum. In contrast, in the UV, starbursts
show a continuum filled with absorption lines (Figs. 2 and 3).
This spectral dichotomy is caused in part by the massive stars
that power the starburst. Massive stars emit photons with en-
ergies of several eV that are absorbed and re-emitted in their
stellar winds, producing ultraviolet resonance transitions.
However, stellar winds are optically thin to most high energy
(≥13.6 eV) ultraviolet photons, that can travel tens of parsecs
from the star before they are absorbed and photoionize the
surrounding interstellar medium. Subsequently, this ionized
gas cools down via an emission spectrum. Beside the stellar
wind origin, absorption lines can also form in the photosphere
of massive stars, and in the interstellar medium. The wind
and the interstellar absorption lines are resonant transitions.
Usually, low-ionization lines have an interstellar origin, but
high-excitation lines can be wind lines with some interstellar
contribution.
The most important characteristics of these lines are de-
scribed below, and labeled in Figs. 2 and 3.
–Photospheric: These lines form in layers that are in hy-
drostatic equilibrium. Mainly from C, N, O, S, Si and Fe
ions of low and high-excitation potential, they originate
from excited levels, so they are not resonant transitions,
and thus not contaminated by interstellar components. Al-
though much weaker, they are not too much affected by
stellar wind lines. These lines are useful to constrain the
age of the starburst, but also the metallicity (de Mello
Fig. 2 UV spectra of starburst
galaxies taken with
HST+STIS/MAMA (NGC
3049 and Tololo 89) and
HST+GHRS (NGC 7714).
Some of the most relevant
photospheric (full line), wind
(dotted) and interstellar lines
(dashed) are labeled. The
spectra correspond to the main
cluster only (see Fig. 1).
Springer

Astrophys Space Sci (2006) 303:85–102 89
Fig. 3 FUV spectra of M83
taken with FUSE (thin line) and
HUT (thick line). The main lines
are labeled. (Figure adapted
from Leitherer et al., 2002)
et al., 2000; Robert et al., 2003). In particular, blends of
these lines in the ranges 1360–1380 ˚
A and 1415–1435 ˚
A
show a strong dependence with metallicity (Leitherer et
al., 2001). Rix et al. (2004) find also that the FeIII (1935–
2020) index is a strong metallicity indicator. Some of the
most relevant of these lines in starburst are: SV λ1502,
CIII λ1426–1428 and CIII λ1176.
–Wind: Hot stars develop strong wind stellar lines due to
the radiation pressure in ultraviolet resonance lines. As a
result, all the strong ultraviolet lines show a blueshifted
absorption (about 2000–3000 km s−1) or a P Cygni profile.
The shape of the profile reflects the stellar mass-loss rate,
which is related to the stellar luminosity, and thus to the
stellar mass. Therefore, the shape of the line profiles in
the UV integrated light of a stellar population is related
to its content in massive stars. Thus, these features can
be used to constrain the properties (such as age, and the
slope and upper mass limit of the IMF) of the starburst.
But due to the dependence of the mass-loss rate with the
metallicity (Maeder & Conti 1994), wind lines are also
strongly affected by metallicity. The most relevant wind
lines in starburst are: NV λ1240, SiIV λ1400, CIV λ1550,
and HeII λ1640.
–Interstellar: Low ionization lines are very useful to study
the kinematics of the ionized gas because the interstellar
component usually dominates over the stellar contribution
in starbursts (Gonz´alez Delgado et al., 1998a). They are
also useful to derive the metallicity of the gas (e.g. Pettini
et al., 2000; Savaglio et al., 2004). The high ionization
interstellar lines are blended with the wind lines, and a
careful separation between both components is necessary.
However, when the starburst is young (2–8 Myr), wind
lines dominate over the interstellar ones.
In addition, the UV spectra of starbursts may show the
Lyman series in emission or absorption. In particular the two
strongest lines, Lyαand Lyβare very useful to study the in-
teraction of the starburst with the interstellar medium. These
lines may be in emission, if they are from the starburst HII
region; but they may be in absorption due to the interstellar
medium within the starburst. Photospheric components may
also contribute to these lines. Lyαand Lyβare weak in very
hot stars but they increase with decreasing effective temper-
ature (Valls-Gabaud, 1993; Gonz ´alez Delgado et al., 1997).
Section 6 below is devoted to explain the relevance of Lyα
in starburst galaxies.
4. Dust opacity in starbursts
The most direct way to measure the star formation rate in star-
bursts is through the UV luminosity. Unfortunately,starbursts
are often dusty, and the presence of dust affects significantly
the UV emission. The UV is more affected by extinction than
any other wavelength. So, objects that are optically thin at
visible wavelengths may be optically thick in the UV. Dust
absorbs a fraction of the UV photons and reradiates in the
far-infrared. Analyzing a sample of UV selected local galax-
ies observed with GALEX, Buat et al. (2005) have found that
only 33% of the UV emission escapes from the galaxies and
the remaining 66% is absorbed by dust and reradiated in the
far-infrared.
Using IUE spectra of a variety of starburst galaxies (blue
compact, starburst nuclei, also some luminous infrared galax-
ies, etc.), Calzetti et al. (1996) found that the UV spectral
energy distribution is well parametrized by a power-law,
F∝λβ, and the spectral slope of the continuum, β, correlates
Springer

90 Astrophys Space Sci (2006) 303:85–102
with the nebular optical extinction derived using the Balmer
decrement. On the other hand, from population synthesis
models, Leitherer and Heckman (1995a) found that starbursts
have an intrinsic spectral slope that changes very little with
the initial mass function (IMF) or the age of the starburst,
taking values around −2.3. Then, according with these re-
sults, any deviation of the spectral slope from the intrinsic
value can be interpreted as a reddening effect. This provides
a nice recipe to correct the UV observed emission for extinc-
tion, and to find an effective attenuation law to perform the
extinction correction (Calzetti 1997).
Gordon et al. (1997) built a model of the stars and dust dis-
tribution in a starburst, and concluded that the grey starburst
extinction law is compatible with a clumpy shell geometry,
with the UV radiation from the starburst viewed filtered
through the dusty gas clouds (see also Charlot and Fall 2000).
This distribution explains why the ionized gas extinction (de-
rived using the Balmer decrement) is a factor of two higher
than the stellar extinction (derived from the UV slope) in
starbursts, because the emission lines are seen through a
larger column of dust than the UV continuum (Fanelli et al.,
1988).
Meurer et al. (1997) show that βcorrelates with the ratio
of the far-infrared to UV fluxes, LIR/LUV . The straightfor-
ward interpretation of these results is that dusty starbursts
absorb a large fraction of the UV radiation, that is subse-
quently reradiated in the far-infrared. This is an energy bal-
ance relationship, that allows us to recover the UV radiation
without a detailed understanding of the dust grain properties
or the extinction law. Following in the same line, Heckman
et al. (1998) found that at low-metallicity, starbursts have blue
colors and a significant fraction of the UV radiation escapes
from the starburst. But at solar metallicity starbursts have
redder UV colors. They have LIR/LUV ≥10, indicating that
only less than 10% of the intrinsic UV luminosity escapes
from the starburst. These results imply that dustier starbursts
are more frequent in more metal-rich galaxies. Note also,
that the galaxies with higher star formation rate are, by the
simple principle of causality, the most massive ones. Due
to the mass-metallicity relation, they are the dustiest ones
(Heckman, 2005). So, the most massive galaxies host more
powerful, more metal-rich and dustier starbursts.
Is the LIR /LUV −βcorrelation found for UV selected star-
bursts applicable to other objects? Goldader et al. (2002)
obtained FUV and NUV STIS images for 9 ultraluminous
infrared galaxies (ULIRGs). They found that, after correct-
ing for dust reddening using the LIR/LUV −βcorrelation,
the UV luminosity is insufficient to account for the farin-
frared luminosity. On the contrary, GALEX data of several
samples of normal galaxies with star formation show that the
observed UV luminosity overestimates the far-ultraviolet at-
tenuation of these galaxies predicted from the relationship
(Buat et al., 2005; Seibert et al., 2005).
Several attenuation laws have been proposed in the 1200–
3000 ˚
A spectral range (Rosa and Benvenuti 1994; Mas-Hesse
and Kunth 1999; Calzetti 1997) to be applied to star forming
regions, but all of them are coincident in showing a 2175 ˚
A
bump weaker than in the galactic extinction law. Recently,
Leitherer et al. (2002) and Buat et al. (2002) have used HUT
and FUSE data to extend the starburst attenuation law to the
FUV.
5. Stellar content
Most starbursts are far enough that the stellar population is
unresolved in individual stars even with the high spatial res-
olution of HST. Thus, the stellar content of starbursts has
to be estimated through their integrated light. The first deep
integrated spectra of galaxies were taken by IUE with aper-
tures of 10 ×20 arcsec (see the spectral atlas by Kinney
et al. (1993). Sekiguchi and Anderson (1987) were the first
to build a stellar library of galactic O and B stars that could
be used to predict the equivalent width of CIV λ1550 and
SilV λ1400 of a non-evolving stellar population of massive
stars. Later, Mas-Hesse and Kunth (1991) extended this work
using evolutionary models to predict the starburst properties
as a function of age. The stellar content of starburst galaxies
(e.g. Mas-Hesse and Kunth (1999) and giant extragalactic
HII regions (e.g. Vacca et al., 1995; Gonz´alez Delgado and
Perez, 2000; Jamet et al., 2004) have been estimated using
these IUE spectra. High spatial resolution (sub-arcsec) HST
spectra have been obtained (Leitherer et al., 1996; Conti et al.,
1996; Gonz´alez Delgado et al., 1999, 2002; Johnson et al.,
2000; Chandar et al., 2003a) and the results attained are dis-
cussed in Section 5.2.
Starbursts have a strongly absorbing interstellar medium,
and a significant fraction of the equivalent width of the CIV
λ1550 and SiIV λ1400 may be due to the interstellar com-
ponent. This is particularly true if very massive stars do not
form or if the starburst is not very young (10 Myr or older). In
fact, the low spectral resolution of IUE (about 1000 km s−1)
did not allow us to separate the stellar from the interstellar
components in most of the starbursts observed. The analysis
of these spectra has provided an uncertain determination of
the age and stellar content of some starbursts. IUE had also
the capability of taking spectra at higher resolution; however,
no starburst was bright enough to be observed in this mode.
It has been later, in the HST era with higher spatial and spec-
tral resolution observations, that some galaxies previously
classified as very young starbursts have been recognized as
evolved starbursts with a strong interstellar medium; NGC
1705 is a good example (Heckman and Leitherer, 1997a).
FOS, GHRS and STIS on board HST were used to obtain
deep spectra at a resolution (∼200 km s−1) sufficient to re-
solve the interstellar from the stellar wind components. These
Springer

Astrophys Space Sci (2006) 303:85–102 91
observations allow detailed profile analysis of the wind lines
of nearby starbursts to investigate their stellar content. Evo-
lutionary synthesis models that predict the UV wind profiles
of a stellar population are used to estimate the age and the ini-
tial mass function (IMF) of the starburst. These models have
been developed mainly in two spectral ranges, 1200–2000 ˚
A
and at the FUV, 1000–1200 ˚
A. We first describe the mod-
els and then comment on the general properties of starbursts
from the UV line synthesis.
5.1. UV line synthesis
Stellar wind lines contain information on the stellar mass; this
is the basis of the UV line synthesis. Stellar winds are driven
by radiation pressure. In O stars, a fraction of the radiative
momentum (L/c) is converted to kinetic momentum ( ˙
Mv∞);
so
˙
Mv∞∝(L/c)
where ˙
M, v∞, L and c are the mass-loss rate, the wind termi-
nal velocity, the radiant luminosity of the star and the speed
of light. The profile of the wind line contains information
about the terminal velocity and, via the previous relation-
ship, about the stellar luminosity. So, the wind line profiles
of the integrated spectra carry information about the massive
stellar population of the starburst, hence about its IMF.
5.1.1. Range 1200–2000 ˚
A
Robert et al. (1993) and Leitherer et al. (1995c) have com-
puted an atlas of evolutionary synthesis models that predict
the line profile of NV λ1240, SiIV λ1400, CIV λ1550, HeII
λ1640 and NIV λ1720 as a function of age and IMF, for
an instantaneous burst and for continuous star formation.
Figures 4 and 5 show these lines for several ages and dif-
ferent IMF. The results indicate: (a) CIV always shows a P
Cygni profile when O stars with M≥50 Mare in the zero-
age main sequence; it is a good age diagnostic of the stellar
population. (b) SiIV shows a conspicuous wind profile when
O blue supergiants are present. A strong P Cygni profile ap-
pears between 3 to 5 Myr for a burst stellar population. It is
also strong, when there is a large fraction of blue supergiants
with respect to O main sequence stars, i.e., when the stellar
population forms with a top-heavy IMF. (c) NV has a similar
behavior to SiIV. (d) Hell and NIV appear as a strong broad
emissionfeature when a large fraction of Wolf-Rayet stars
are present, in the age range ∼3–4 Myr.
Photospheric lines, such as CIII λ1426, 1428, SV λ1502,
are also strong when the burst is only a few Myr old and the
UV light is dominated by O stars. The contribution of B stars
to the UV light has been predicted by de Mello et al., (2000).
Si lines, such as SiIII λ1295,1297,1299, SiIII λ1417, SiII
λ1485, are good age diagnostics for evolved starbursts (age
≥10 Myr).
Initially, the spectral library used as input to the mod-
els was built with hot stars of solar or slightly subsolar
metallicity. However, photospheric lines are much weaker
at lower metallicity; and the stellar-wind properties are af-
fected by line-blanketing, since the mass-loss rate scales with
(Z/Z)0.5(Kudritzki et al., 1999). Leitherer et al. (2001)
built a new stellar library with O and B stars in the LMC and
SMC observed with HST. They implemented it in Starburst99
(Leitherer et al., 1999), providing new evolutionary synthe-
sis models at 1/4 Z. However, the behavior of the stellar
wind lines is complex; while NV λ1240 and SiIV λ1400 do
not scale monotonically with metallicity, CIV λ1550 is sig-
nificantly affected, showing a weaker P Cygni profile. Thus,
while the wind NV and SiIV may be equally well predicted
using the solar stellar library, CIV and the photospheric lines
Fig. 4 UV synthetic spectra
generated with Starburst99 for
an instantaneous burst at
different ages that follows a
Salpeter IMF in a mass range of
1to100M. Note the change of
the SiIV and CIV profile with
the age of the burst
Springer

92 Astrophys Space Sci (2006) 303:85–102
Fig. 5 As in Fig. 4 for an
instantaneous burst 4 Myr old
with different assumptions of
the IMF. Note how weak
become SiIV and CIV when
very few massive stars form in
the starburst
are overpredicted in low metallicity starbursts, inducing a
wrong estimation of the age and of the IMF parameters.
5.1.2. Range 1000–1200 ˚
A
The first evolutionary synthesis models at intermediate spec-
tral resolution in the FUV were computed by Gonz´alez
Delgado et al. (1997). Using a stellar library built with hot
O and early B stars observed with Copernicus and HUT,
they predicted the wind line OVI λ1032,1038 and the pho-
tospheric component of Lyβ. OVI develops a P Cygni pro-
file when formed in stellar winds of the most massive stars.
When these stars are absent, no OVI is formed. In contrast,
Lyβis a very sensitive indicator of B stars. If these stars
dominate, as is the case in evolved starbursts (age ≥10 Myr),
Lyβis present as a strong absorption feature. Because of
the constant strength of OVI in O stars, OVI is not a good
discriminator between instantaneous versus continuous star
formation for ages when the starburst is in the nebular phase,
but the absence of OVI and the presence of stellar Lyβis
a good indicator of a short burst duration and of the galaxy
being in an evolved starburst phase (age ≥10 Myr). However,
careful attention to interstellar absorption of Lyβis needed
before estimating the ages and stellar content in starbursts
using these lines. Robert et al. (2003) have made an exten-
sion of these models predicting also the photospheric lines
in the 1000–1200 ˚
A range using O, B and Wolf-Rayet stars
in the Galaxy and in the LMC and SMC observed by FUSE.
Other wind lines observed in starbursts are NIV λ955, CIII
λ977, NIII λ991 and NII λ1083 (Keel et al., 2004).
In contrast to wavelengths above 1200 ˚
A, the FUV contin-
uum suffers from an age-reddening degeneracy. Because at
λ≤1200 ˚
A hot stars are outside the Rayleigh-Jeans regime,
the age effects are no longer negligible in the continuum slope
for starbursts in an instantaneous burst (Leitherer, 2005).
For starbursts in the continuous star formation regime, the
FUV continuum is less sensitive to the age than the near-
UV, because the rate of death and birth of stars is reached
earlier.
5.2. Results: Ages and IMF
5.2.1. Stellar Clusters
Intermediate (∼1 arcsec) and high (∼0.1 arcsec) spatial res-
olution spectra with HST have been obtained to constrain
the IMF and age in stellar knots detected in starbursts. The
main results derived from the UV light provided by stellar
clusters can be summarized as follows: the spectral range
1200–2000 ˚
A can be characterized by an instantaneous burst
a few Myr old, populated by a Salpeter IMF with stars more
massive than 50 ˙
M(e.g. Conti et al., 1996; Leitherer et al.,
1996; Gonz´alez Delgado et al., 1999; Chandar et al., 2003b).
When the integrated light is emitted by extended areas (∼100
pc), the UV spectra are equally well fitted by continuous star
formation lasting for a few Myr. These results indicate that
clusters form with a very small age spread. In fact, this is the
case for the starburst He2–10 (Johnson et al., 2000; Chandar
et al., 2003a) in which the clusters are chained along ∼100
pc with a mean separation ≤10 pc, and they are all 4–5 Myr
old. These clusters have masses of several 104to Several
105M, that are typical of proto-globular clusters (Ho and
Filippenko, 1996).
There are indications that the IMF and the global star
formation processes are the same in metal rich clusters as
they are in metal poor ones. A good example is the metal-
rich, barred starburst NGC 3049. HST observations done
with STIS/MAMA (FUV) indicate that most of the UV light
is emitted within the central arcsecond. The wind lines de-
tected in the spectrum indicate that the cluster(s) in the inner
Springer

Astrophys Space Sci (2006) 303:85–102 93
50 pc formed 3–4 Myr ago in an instantaneous burst. Even
though the metallicity of the stars is supersolar, stars more
massive than 50 Mform in the cluster(s) (Gonz ´alez Delgado
et al., 2002). This result has been confirmed by Chandar et al.
(2003b) for other metal-rich starbursts. HeII λ1640 has been
detected in these objects, indicating the presence Wolf-Rayet
stars in these starbursts. This finding provides an additional
evidence of the population of the upper part of the IMF in
high metallicity starbursts.
5.2.2. Diffuse UV Light
The main results in this topic come from high spatial obser-
vations taken with HST+STIS. The narrow slit (∼0.1–0.2
arcsec) capability of STIS is needed to isolate the stellar clus-
ters light from the diffuse component. Tremonti et al. (2001)
have obtained long slit spectra of several stellar clusters plus
the inter-cluster regions of diffuse light in the low-metallicity
galaxy NGC 5253. They find that the UV light of clusters and
that of the diffuse component have different spectral proper-
ties. The clusters are well fitted by an instantaneous burst with
ages of several Myr that follow a Salpeter IMF extending up
to 100 M. However, the field spectrum is better fitted by
continuous star formation models with either Mup =30 M
or an IMF slope steeper than Salpeter’s. Alternatively, the
field stellar population could be formed following a Salpeter
IMF but older than the clusters. Similar results have been
obtained by Chandar et al. (2005) for a sample of starbursts.
They propose that if the field is composed of older, dissolving
clusters, they have to dissolve on timescales 7–10 Myr to cre-
ate the field. If the field is composed of young clusters that are
unresolved in the STIS observations, they would consist only
of a few 100 Min order to be deficient in O stars. However,
sampling effects in the IMF (Cervi˜no et al., 2002) must be
taken into account before obtaining any realistic conclusion.
5.2.3. Lyman Break Galaxies
The UV-rest frame spectra of LBGs have been obtained from
the ground with ∼10m class telescopes. These spectra are
quite similar to local starburst galaxies in the sense that they
are dominated by absorption lines (e.g. Shapley et al., 2003;
Noll et al., 2004). They have strong high-and low-ionization
interstellar lines that are thick in their cores. Photospheric and
wind lines are also present. But, most of the high-ionization
lines are dominated by the interstellar contribution. Probably
due to the large spatial extension covered by these observa-
tions (1 arcsec =8 kpc at z=2.5, assuming a standard cos-
mology), the wind profiles of the integrated light look weaker
than in many nearby starbursts, suggesting that star forma-
tion proceeds continuously, they have older ages, and/or the
metallicity is lower.
Because of its gravitationally lensed nature, cB58 has the
highest signal-noise rest-frame UV spectrum obtained to date
for LBG (e.g. Pettini et al., 2002). Even so, the wind lines
can not constrain well the age. The PCygni profiles of CIV
and NV are compatible with csf during the last several 10
Myr, and a Salpeter IMF extended beyond 50 M.Noevi-
dence exists for a flatter IMF or an IMF deficient in massive
stars (Pettini et al., 2000; de Mello et al., 2000). Photospheric
lines were detected and they are compatible with a metallic-
ity below solar, 1/4 Z. Spectra for other LBGs have been
obtained with much lower signal-noise, and only a few of the
individual objects can be analyzed. Composite spectra with
the average of more than several dozen objects are more
suitable to be studied. The results obtained in this way for
LBGs at the ‘redshift desert’ (1.4≤z≤2.5) indicate that
these galaxies have stellar properties similar to cB58 (Steidel
et al., 2004). But the metallicity can be higher, closer to solar.
Mehlert et al., (2005) have found an increase of the average
metallicity of bright starbursts with cosmic time (decreas-
ing redshift). A metallicity higher than solar has also been
estimated using the photospheric 1425 ˚
A index in the K20
survey (Daddi et al., 2004).
6. Starbursts in AGN
HST ultraviolet observations of Seyferts 2 and LINERs have
contributed significantly to establish the role that starbursts
play in the active galactic nuclei (AGNs) phenomena. The
high spatial resolution provided by HST has been crucial
to detect starbursts formed by stellar clusters in the center
of galaxies with an AGN. This result implies a significant
advance to establish a connection between violent star for-
mation processes and nuclear activity, because in the IUE
era, it was assumed that all the UV light obtained in the
Seyfert spectra was produced by the AGN. An extended re-
view of the role of UV observations in establishing the nature
of AGNs is given elsewhere else in this book by Kollatschny
and Ting-Gui.
6.1. Starbursts in Seyfert 2 nuclei
According to the unified scheme of AGNs, the main compo-
nents of a Seyfert nucleus are: (1) A super-massive black hole
and its associated accretion disk. (2) A circumnuclear dusty
torus that collimate the AGN radiation through its polar axis.
So, a Seyfert 2 nucleus should be a Seyfert 1 that is viewed
close to the equatorial plane. This torus will facilitate the
detection of starbursts in the cirmunuclear region, blocking
away the continuum radiation from the AGN. (3) A mirror of
dust and warm electrons located along the polar axis of the
torus, that reflects and polarizes the AGN radiation. Seyfert
2 nuclei exhibit a featureless continuum (FC) that comprises
Springer

94 Astrophys Space Sci (2006) 303:85–102
much of the near-UV. It was long-thought that this FC was
light from the hidden Seyfert 1 nucleus. However, optical
spectropolarimetry (Tran, 1995) showed that this is not the
case. Cid Fernandes and Terlevich (1995) proposed that a
heavily-reddened starburst provides this FC.
Because of the high sensitivity of UV wavelengths to the
presence of massive stars, HST UV observations were done
to prove the role of starbursts in Seyfert 2 nuclei (Heckman
et al., 1997b; Gonz´alez Delgado et al., 1998b). HST high spa-
tial resolution (0.014 arcsec/pixel sampling) imaging shows
that the UV continuum source is spatially extended (∼100
pc) and it is resolved in knots with sizes of a few parcsecs
and properties similar to the stellar clusters detected in star-
burst galaxies (Fig. 6). GHRS spectra for four galaxies, corre-
sponding to the central ∼100 pc, were obtained. These galax-
ies were selected to have high UV surface brightness. The
data provided direct evidence of the existence of a nuclear
starburst. Absorption features formed in the photospheres
and in the stellar winds of massive stars are detected (Fig. 6).
Interstellar lines blueshifted by a few hundred km s−1with
respect to the systemic velocity are also detected, indicating
an outflow driven by the nuclear starburst (see Section 7).
Their UV colors indicate that the starburst is quite reddened.
Their bolometric luminosities are similar to the estimated
luminosities of the hidden Seyfert 1 nuclei.
Subsequently, near-UV and optical spectra of a large
sample of Seyfert 2 were obtained proving the unambigu-
ous identification of circumnuclear starburts in ∼40% of
nearby Seyfert 2 galaxies as well as their energetic signif-
icance (Gonz´alez Delgado et al., 2001; Cid Fernandes et al.,
2001).
6.2. Starbursts in LLAGNs
Low-luminosity active galactic nuclei (LLAGNs) constitute
a significant fraction of the nearby AGN population. These
include LINERs, and transition-type objects (TOs, also called
weak-[OI] LINERs) whose properties are in between classi-
cal LINERs and HII nuclei. LLAGNs comprise ∼30% of all
bright galaxies and are the most common type of AGN (Ho
et al., 1997). What powers them and how they fit in the global
picture of AGN has been at the forefront of AGN research for
over two decades. Are they all truly “dwarf Seyfert galaxies”
powered by accretion onto a nearly dormant super-massive
black hole, or can some of them be explained at least partly
in terms of stellar processes?
HST observations at UV wavelengths of a few LLAGNs
have also proven that at least weak-[OI] LINERs could be
powered by young massive stars (Maoz et al., 1998; Colina
et al., 2002; Gabel and Bruhweiler, 2002). Nuclear stellar
clusters are detected in these objects through high spatial
resolution UV imaging and spectra. NGC 4303 is probably
the best example of the few objects observed (Colina et al.,
2002). STIS imaging (0.027 arcsec/pixel sampling) shows
that the nuclear knot has a size of only a few parcsecs and
its spectrum (taken with a slit width of 0.2 arcsec) shows
characteristic broad P Cygni lines produced by the winds
of massive young stars. These features are quite similar to
those detected in the spectra of stellar clusters located in the
starburst ring (Fig. 7). The line profile analysis suggests that
the nuclear cluster formed in an instantaneous burst 4 Myr
ago with a mass of 105M.
HST UV monitoring observations of 17 LLAGNs have
detected variability with amplitudes from a few percent to
50% (Maoz et al., 2005). The variability is more frequently
detected in those LINERs that have a compact radio core, as
expected from bona fide AGNs.
Subsequent optical studies of a large sample of LLAGNs
have shown that the contribution of an intermediate age stellar
population is significant in TOs (Gonz´alez Delgado et al.,
2004). Unfortunately, the premature death of STIS has not
allowed us to find out the fraction of LLAGNs that have a
young nuclear stellar cluster like NGC 4303 and which are
“dwarf Seyfert galaxies” like those that show UV variability.
Fig. 6 UV image and nuclear
spectrum of the Seyfert 2 galaxy
NGC 7130 taken with
HST+GHRS and 1.74 ×1.74
arcsec aperture. The UV light is
dominated by the nuclear
starburst that has an effective
radius of ∼80 pc. See Gonz´alez
Delgado et al. (1998b) for
further explanations
Springer

Astrophys Space Sci (2006) 303:85–102 95
Fig. 7 UV image and spectra of
the LLAGN NGC 4303 taken
with HST +STIS/MAMA. The
nucleus has a compact cluster of
3 pc size. The spectrum of the
nuclear cluster is quite similar to
the stellar clusters in the ring.
See Colina et al. (1998b) for
further explanations.
7. Interstellar lines: Starburst outflows
The large-scale outflow of interstellar material is a generic
property of starbursts, since it is a consequence of the high
star formation activity in these galaxies. Outflows are driven
by the mechanical energy provided by the combined effect of
stellar winds and supernovae in the starburst. Only a few Myr
after the onset of the star formation in the starburst, the mas-
sive stars start to explode as supernovae. Hot gas bubbles (su-
perbubbles) form inside the starbursts due to the deposition
of mechanical energy. The hot gas expands, preferably along
the direction perpendicular to the galaxy disk, sweeping up
ambient material. When the superbubble reaches several ver-
tical scale heights of the galaxy, Rayleigh-Taylor instabilities
develop, and the wall of the bubble can dissipate. This allows
the interior of the hot gas to blow out into the galactic halo
in the form of a collimated bipolar outflow, called superwind
(e.g. MacLow et al., 1989; Tenorio-Tagle and Mu˜noz-Tu˜non,
1996). These outflows can accelerate the ambient halo gas,
producing the bipolar regions of emission lines detected in
the spectra of many starbursts at optical wavelengths, as well
as the blueshifted interstellar absorption lines at UV.
Thus, outflows are ubiquitous phenomena not only in
nearby starbursts, but also in LBGs. Note that, at some evo-
lutionary phases of the starburst (ages older than 10 Myr if
the starburst formed in a burst), the mechanical luminosity
injected into the interstellar medium can dominate over the
ionizing luminosity (Leitherer et al., 1992), becoming almost
the only heating source of the interstellar medium. Outflows
might also be the main source of chemical enrichment of
the intergalactic medium, because they can blow out and
escape from the gravitational potential of the galaxy, venting
the metals produced by massive stars into the intergalactic
medium. Outflows constitute an important energy source for
the evolution of galaxies through the heating and enrichment
of the interstellar and the intergalactic media.
UV is the perfect wavelength range to study the neutral,
cold and warm phases of outflows. FUV is also useful to
study the coronal phases, being the observations in this spec-
tral range complementary to the X-ray data. Lyαemitting
gas traces the neutral HI outflows (cf. Section 8 below); low
(e.g. CII, SiII) and high-ionization (CIV, SiIV, NV) inter-
stellar lines trace the cold and warm phases, respectively,
while interstellar OVI traces the coronal phase. GHRS+HST,
HUT and FUSE spectra of nearby starbursts have contributed
significantly to our understanding of the different outflow
phases.
Dusty outflows have been detected through near and FUV
images of the halo of the nearby starburst galaxies M82 and
NGC 253 taken with GALEX (Hoopes et al., 2005). The UV
luminosity in the halo is too high to be provided by continuum
and line emission from shockheated or photoionized gas.
They find that the UV halo light may be stellar continuum
of the starburst scattered into our line of sight by dust in the
outflow.
One critical point related with the outflows is to know
which fraction of the kinetic energy supplied by supernovae
is radiatively lost and which fraction is carried out into the
outflow. If radiative cooling is not negligible, outflows can
break out and are able to escape from the gravitational po-
tential of the galaxy injecting metals into the intergalactic
medium. This process is relevant for the evolution of star-
bursts and the intergalactic medium. A study of the different
phases of the outflows is needed to settle its relevance.
7.1. Cold phase
Most starbursts have low-ionization absorption lines with
equivalent widths of several ˚
A. These lines are optically
thick, and they are in the flat part of the curve of growth.
This means that the equivalent width of these lines is not
proportional to the column density of the gas, as would be
the case if they were in the linear part of the curve of growth.
In fact, these lines are saturated because the ratio of the equiv-
alent widths of two transitions of the same ion is not propor-
tional to the ratio of the fλ2of each transition, where λis
the rest-frame wavelength of the transition and fthe oscilla-
tor strength. Instead, the equivalent width is determined by
the velocity dispersion of the gas, and therefore an equiva-
lent width of 2–3 ˚
A implies a velocity dispersion larger than
Springer

96 Astrophys Space Sci (2006) 303:85–102
200–300 km s−1. This may indicate that several unresolved
velocities are observed. This is the first evidence suggesting
that the interstellar lines are not virialized but rather they are
related with large-scale motions.
Other evidence of the large-scale motions of the inter-
stellar gas in starbursts comes from the broadening of the
low-ionization interstellar lines. The line profiles are asym-
metric, and when observed at high spectral resolution, they
are resolved into several interstellar components (Gonz´alez
Delgado et al., 1998a).
An additional and the strongest evidence for outflows
comes from the radial velocity of the lines. Many starbursts
show the low-ionization lines, such as SiII λ1526, CII λ1335,
SiII λ1260, blueshifted by several hundred km s−1with re-
spect to the systemic velocity of the starburst determined
with the photospheric lines, or with respect to the HI sys-
temic velocity of the galaxy (e.g. Gonz´alez Delgado et al.,
1998a). Because these interstellar lines cover more than 50%
of the UV light, they cannot be produced by isolated clouds,
and they must be associated with a galactic-scale outflow.
The shift to blue wavelengths detected in these lines is an
unequivocal prove of the outflows in these galaxies.
7.2. Warm and coronal phases
High-ionization interstellar lines, such as NV λ1240, SiIV
λ1400, and CIV λ1550, can trace warm ionized gas outflows
which are at T≥104K. This gas is ionized by radiation
from the massive stars in the starburst as well as by colli-
sional processes associated with the outflow. However, mea-
suring the blueshift of these lines is more difficult than in the
low-ionization lines. This is due to the difficulty in isolating
the wind and the interstellar components of these lines in
intermediate spectral resolution observations. But when the
starburst is not in a wind phase, and/or the spectral resolution
is better than 100 km s−1, the blueshift of the line is also a
measure of the outflow speed.
The hot phase gas is traced by the OVI λ1032,1038 in-
terstellar component. OVI could arise from collisionally ion-
ized gas with T≥105K. OVI has been detected in a sample
of nearby starbursts observed by FUSE. The center of the
lines is blueshifted by several 100 km s−1. The lines are also
very broad, with maximum outflow speeds of ∼1000 km s−1
(Heckman, 2004).
These two outflow phases have been measured in the
nearby dwarf starburst galaxy NGC 1705. This galaxy hosts a
12 Myr old super star cluster (V´azquez et al., 2004). Outflows
in the warm phase were detected in the SiIV high-ionization
lines by Heckman and Leitherer (1997a). More recently,
FUSE observations have shown that the warm gas outflow
has a lower velocity (∼50 km s−1) than the coronal interstel-
lar gas (∼80 km s−1, Heckman et al., 2001a). The kinematics
of the warm gas is compatible with a model of an adiabatic
expansion of the superbubble driven by the kinetic energy
supplied by the supernova. However, the expansion speed of
the superbubble is too small to produce OVI behind its shock
front. Instead, the column density and velocity of the OVI is
compatible with a model in which the superbubble has begun
to blow out of the interstellar medium of NGC 1705. OVI
absorption is produced during the blowout phase, in which
the superbubble shell is accelerated and fragmented. The in-
teraction between the outflowing gas and the shell fragments
create a high temperature coronal gas that produces the OVI
absorption. Heckman et al. (2001a) found that the cooling
rate in this phase is much less than the supernova heating rate;
thus, they concluded that probably the outflow in NGC 1705
is able to blow out and to vent the metals into the intergalactic
medium.
8. The Lyαline: Outflows of neutral H gas
Lyαin emission can in principle be produced by recombina-
tion of hydrogen photoionized by the O and B stars. Massive
stars in the starbursts provide enough ionizing photons to pro-
duce a large Lyαflux. In fact, evolutionary stellar population
models predict Lyαto be the strongest emission feature in the
spectra of young starbursts. This is particularly true for pri-
mordial galaxies because in the absence of metals the cooling
is produced by Lyαand He recombination lines (Schaerer,
2002). Lyαhas been used to spectroscopically confirm galax-
ies at high-z. However, many observational programs in the
past have failed to find a significant population of primordial
galaxies based on the detection of Lyα. The reasons have to
be found in the complex structure of the Lyαline.
The complexity of the line was noted more than 20 years
ago with IUE. Lyαobservations of nearby starbursts show
that the line is weaker than the value expected from recom-
bination, and there was a tendency to smaller Lyα/Hβratios
with metallicity (Meier and Terlevich, 1981; Hartmann et al.,
1988; Terlevich et al., 1993). Several arguments were pro-
posed:
– Resonance scattering by neutral hydrogen: Lyαphotons
are attenuated by dust as a result of multiple resonant scat-
terings by hydrogen atoms that increase the path length of
the Lyαphotons and thus the probability that they will be
absorbed by dust.
– Extinction: The Lyαlight is affected by dust more than any
other Balmer recombination line since extinction curves
peak in the FUV. In fact, some starbursts have Lyα/Hβ
ratio consistent with simple recombination theory if the
ratio is corrected for reddening using the appropriate ex-
tinction law for the metallicity of the galaxy and the age
Springer

Astrophys Space Sci (2006) 303:85–102 97
Fig. 8 Lyαprofile of the starburst galaxy IRAS 0833 +6517 taken
with HST+STIS/MAMA. Note the PCygni profile of Lyαand a second
blueshifted Lyαemission that forms in the expanding shell. (Figure
adapted from Mas-Hesse et al., 2003.)
of the burst is taken into account (Calzetti and Kinney,
1992; Valls-Gabaud, 1993).
However, these arguments contrast with observations of
low metallicity starbursts that are very little extinguished. No
Lyαemission was detected in galaxies like IZw18 (Kunth
et al., 1994), even though the metallicity and dust content
are very low. HUT and HST observations of nearby galaxies
have brought new insight into the nature of Lyα(Lequeux
et al., 1995; Gonz´alez Delgado et al., 1998a; Kunth et al.,
1998, 2003; Mas-Hesse et al., 2003). GHRS and STIS spec-
tral resolution has been crucial to understand the role played
by outflows in the structure of the Lαemission.
– Neutral HI outflows: Many nearby starburst galaxies in
which Lyαis detected in emission show an asymmetric
profile, with the peak emission redshifted with respect
to the systemic velocity, and a deep Lyαabsorption is
detected blueshifted by several 100 km s−1with respect
to the emission (see Fig. 8). This shift is in agreement with
the blueshift observed in the interstellar lines of the same
galaxy. The natural explanation is that the neutral HI gas
producing the absorption is outflowing from the starburst.
Tenorio-Tagle et al. (1999) have developed a detailed
model to explain the different Lyαprofiles, that requires the
time evolution of an expanding shell created by the super-
nova explosions in the starburst. Some of the most relevant
phases are:
– An expanding supershell forms by the SN action. Lyα
photons will be absorbed by the HI galaxy disk. If the HI
column density is very high, Lyαwill show an absorption
profile centered at the systemic velocity of the galaxy. The
starburst in IZw18 is in this phase.
– Rayleigh-Taylor instabilities produce the shell fragmen-
tation, and the shell will blow out. Ionizing radiation es-
capes into the halo and the IGM, producing an extended
biconical emission line region. Lyαwill be detected in
emission at the systemic velocity of the galaxy. Recombi-
nation in the shell will produce an additional Lyαemission
blueshifted at the shell expanding velocity. Tol1214 could
be in this phase (Mas-Hesse et al., 2003).
– An HI trapped ionization front is formed at the external
side of the expanding shell. Lyαphotons are absorbed
there. Lyαwill show a PCygni profile. Backscattering and
emission from the receding part of the shell will produce
an extended red wing in the Lyαemission. IRAS 0833 +
6517 is in this phase (see Fig. 8).
– Finally the shell is slowed down in its expansion, and
it will be completely recombined. A damped Lyαcore
profile will be observed with only a small blueshift.
Thus, Lyαemission, like the interstellar lines, is driven by
the dynamical effects of the violent star formation processes
ongoing in the starburst, rather than by the gravitational po-
tential well of the galaxy.
9. Lyman continuum: The escape of ionizing
radiation
The origin of the diffuse ultraviolet background that reion-
ized the early universe is still unknown. Quasi-stellar ob-
jects (QSOs) are considered one of the main sources. How-
ever, QSOs alone cannot account for all the background ra-
diation that maintains the diffuse intergalactic medium and
the Lyαforest clouds highly ionized. Two other possible
sources are: highly obscured QSOs that cannot be observed
because of dust, and a large fraction of high-mass stars
formed in primordial galaxies (Miralda-Escud´e and Ostriker,
1990).
Observations of the Lyman continuum of starbursts have
been done to estimate the contribution of massive stars to
the reionization of the universe. However, this estimation
depends critically on the determination of the fraction of
ionizing photons ( fesc) that escape from the galaxies and
reach the intergalactic medium.
Evidence exists that the HI disks surrounding galaxies
may not be totally opaque to the ionizing photons. Bland-
Hawthorn and Putman (2001) estimate that 5–10% of the
ionizing photons escape from the Milky Way halo. Oth-
ers (e.g. Castellanos et al., 2002) have found that a sig-
nificant fraction of ionizing photons may locally escape
from the HII regions in nearby galaxies, but it is unknown
whether these photons will escape from the galaxies. Star-
bursts outflows may be an efficient mechanism to open chan-
nels in the HI halo disks of the galaxies through which
the ionizing photons can escape and reach the intergalactic
medium.
Springer

98 Astrophys Space Sci (2006) 303:85–102
HUT and FUSE have contributed significantly in deter-
mining the value of fesc in local starbursts. Leitherer et al.,
(1995b) find, in a small sample of four starburst galaxies,
that fesc ≤3%. Later, Hurwitz et al., (1997) reevaluate fesc
applying a detailed model of the absorption by interstellar gas
in our Galaxy. They are not able to detect the Lyman con-
tinuum flux, but they derive un upper limit fesc ∼10%. De-
harveng et al. (2001) obtained FUSE observations of Mrk54
and found that the Lyman continuum radiation is not de-
tected above the HI absorption edge in our Galaxy. Compar-
ing with the number of ionizing photons derived from the Hα
flux, they estimate that ∼6% of Lyman continuum photons
escape the galaxy without being absorbed by interstellar ma-
terial. Heckman et al. (2001b) also obtained FUSE data of
five of the UV-brightest local starburst galaxies. They found
that the interstellar line CII λ1036 is essentially black. Be-
cause the opacity of the neutral ISM below the Lyman edge
is larger than in the CII line, the residual UV intensity of the
line can be used to put a constrain on fesc . They found an
upper limit of 6%. Thus, local starburst galaxies seem to be
very opaque and they leak out only a few percent of their
ionizing radiation.
Observations of the Lyman continuum flux in LBGs have
provided more discrepant fesc results. Steidel et al., (2001)
built a composite spectrum of 29 LBGs at z=3.4. These
galaxies belong to the group of strong Lyαemitters found
by Shapley et al. (2003). They estimate fesc ≥50%. The
implication of this result is that LBGs contribute at least as
many ionizing photons as QSOs at z∼3. However, these
results have not been confirmed by others. Giallongo et al.
(2002) obtained VLT spectra of two of the LBGs from the
sample of Steidel and collaborators. They set an upper limit
of 16% to fesc . Lyman continuum has been estimated also
from deep HST images of the HDF (Fern´andez-Soto et al.,
2003) and LBGs at 1.1≤z≤1.4. Both works also found fesc
ofafew%(≤4%). High spectral resolution data are required
to better constrain the Lyman continuum flux in LBGs, and
to determine whether LBGs are opaque like local starburst
galaxies or they are leaking out most of their photons as
Steidel et al., estimate.
However, in most of the LBGs the absorption part of the
LyαPCygni profile is completely black. Therefore, the star-
forming regions seem to be completely covered by neutral
gas, and fesc should be close to zero, unless the covering is not
isotropic and the escape is produced along other directions.
Note, however, that many of the nearby starbursts for
which fesc has been estimated in a few % also have high-
velocity outflows of neutral gas. But these outflows are also
an ubiquitous phenomenon in LBGs. Thus, as it was pointed
out by Heckman et al. (2001a) these outflows could be a nec-
essary but not sufficient mechanism to open paths within the
interstellar medium through which the ionizing radiation can
escape.
10. Interstellar lines: Abundances
The large number of line transitions of many different ions
that occur at UV wavelengths, makes this spectral range quite
suitable to determine the chemical abundances and to study
the chemical evolution of galaxies. Because metal transitions
that form in the neutral gas phase are common in the UV, this
range is quite suitable to determine the chemical abundances
of the HI phase. Note, however, that most of our knowledge
about the chemical abundances in starbursts comes from the
collisional emission lines formed in the ionized gas associ-
ated to the HII regions that are observed at optical and infrared
wavelengths. Thus, these abundances correspond to the gas
ionized phase. This difference between the abundances de-
termined using optical or UV transitions (collisional vs. re-
combination lines, HII vs. HI) is especially important for the
study of the chemical evolution of galaxies and, in particular,
for starburst galaxies.
As we pointed out earlier, many of the strongest interstel-
lar lines in the spectra of starburst galaxies observed with
an intermediate spectral resolution are saturated. Thus, the
strength of these lines is related more with the kinematics
of the gas than with the metallicity. But, unsaturated absorp-
tion interstellar lines have a suitable information of the gas
metallicity. So, when a line is in the linear part of the curve-
of-growth, its equivalent width is proportional to the column
density of the corresponding species, and the metallicity of
the element can be estimated. On the contrary, when the line
is in the flat part of the curve-of-growth, the profile is quite
insensitive to the abundances. For example, as shown in Pet-
tini and Lipman (1995), the range of possible values of (O/H)
admitted by the profile of the saturated OI λ1302 absorption
line is very large, spanning a factor of ∼1000. Then, only
unsaturated, and presumably weak lines with a moderate or
low transition-strength, fλ, are useful to estimate the column
density of the ions. Therefore, only with good signal-to-noise
and high spectral resolution spectra is it possible to determine
the gas abundances.
Different approaches can be followed to determine the
column densities of the different ions in starbursts. They
are: (1) The curve-of-growth method, which relates the
equivalent widths of the lines with N fλ2, where N is the
column density. (2) The direct method, based on the fit
of the absorption profiles to all the lines in the spectrum
arising from transitions of the same ion. (3) The optical
depth method. The optical depth is deduced from the ob-
served intensity in the line at velocity v, and then, the col-
umn density that best fits the line profile is inferred (Sav-
age and Sembach, 1991). Metal abundances are then de-
termined through the column densities and assuming some
ionization correction fraction. A hypothesis about the dust
depletion has to be made to obtain the final values of the
abundances.
Springer

Astrophys Space Sci (2006) 303:85–102 99
Abundances for starbursting dwarf galaxies are easier
to obtain than for nuclear starbursts since the former have
lower metallicities. Starbursting dwarf galaxies are very
interesting systems from the cosmological point of view.
According to the hierarchical galaxy formation scenario,
dwarf galaxies could be the building blocks of larger and
massive galaxies that formed by merging. Thus, local star-
bursting dwarf galaxies could be considered the closest
analogue to primeval galaxies. Local starbursting dwarf
galaxies are gas rich and chemically relatively unevolved
objects, as their low abundances indicate. They have HII
region abundances between ∼1/50 to 1/3 Z, which are
certainly not primordial. But Kunth and Sargent (1986) sug-
gested that the ionized gas could be enriched with met-
als ejected by supernovae in a very short time-scale, the
time-scale of a burst of star formation. Then, the HII abun-
dances would not necessarily reflect the actual abundance
of the HI phase, being the former lower if self-pollution
is important. However, this hyphotesis is not supported by
Tenorio-Tagle (1996) that predicts a larger time-scale for
the mixing of supernova ejected metals with the interstellar
medium. This time-scale would be of the order of several
108yr.
FUV observations of nearby starbursting dwarf galaxies
taken with FUSE have contributed significantly to test if these
galaxies are primeval, experiencing their first burst of star
formation (Thuan et al., 2002; Lecavelier des Etangs et al.,
2004; Aloisi et al., 2003; Lebouteiller et al., 2004; Aloisi
et al., 2005; Cannon et al., 2005). These works agree in find-
ing lower αelement abundances in the neutral HI gas phase
than in the HII regions. However, because these abundances
are not really primordial, starbursting dwarf galaxies are not
experiencing their first burst of star formation, and they are
not primeval galaxies.
UV interstellar lines have also been used to estimate how
chemically evolved are high-zstar forming galaxies. In con-
trast to local starburst galaxies, HII region abundances are
not known for LBGs, since collisional lines at the rest-frame
optical wavelengths have not been observed yet for a signif-
icant population of high-zstar forming galaxies. Neutral gas
phase abundances have been determined for cB58 (Pettini
et al., 2002). They found that the interstellar medium of this
galaxy is highly enriched in elements released by type II su-
pernovae, with abundances of O, Mg, Si, P and S ∼2/5 Zodot.
But, N and Mn, Ni and Fe are underabundant by a factor ∼3.
Because these elements are produced by intermediate-mass
stars, the enrichment in cB58 has probably taken place within
the last 300 Myr, which is the lifetime of these stars and the
release time scale for N.
Savaglio et al., (2004) have estimated column densities of
Fe, Mg and Mn for a sample of 13 galaxies at redshifts 1.3≤
z≤2. These column densities are similar to those derived for
cB58. But they are considerably larger than typical values in
damped Ly αsystems. Making a rough estimation of the HI
column density and assuming a moderate Fe dust depletion,
they estimate an abundance of 1–0.2 Z. Then, these galaxies
are also metal-rich.
11. Summary and future prospects
IUE has made an important initial contribution to our knowl-
edge of starbursts providing the first high quality UV spectra;
however, most of our actual knowledge about UV in nearby
starbursts comes from HST and FUSE observations. Along
this paper, we have emphasized the contributions of high
spatial (≤0.025 arcsec/pixel) resolution imaging and inter-
mediate (∼100 km s−1) dispersion spectra taken with FOC,
GHRS and STIS in the UV, and the high-resolution spectra
with FUSE in the FUV (1200 ˚
A down to the Lyman break). A
significant progress has been made in the determination of the
stellar content of starburst galaxies and the physical, chemical
and dynamical properties of the interstellar medium in these
galaxies. In particular, it has been possible to advance in our
understanding of the role that stellar clusters play in starbursts
and AGNs only thanks to the high spatial resolution of the
imaging and spectral observations taken with the instruments
on board HST. Stellar clusters have sizes of a few pcs, so they
can be isolated from the background radiation in very nearby
starbursts if they are observed with resolution better than 0.1
arcsec.
During the next few years, the GALEX mission will cer-
tainly provide the means for an important progress in the
study of the cosmic evolution of starburst properties, as well
as providing a larger sample of local starbursts. But the lim-
ited capabilities of its spectroscopic mode (low spatial and
spectral resolution and sensitivity) will not help to progress in
understanding most of the physics that regulates the violent
star formation processes in galaxies. Spatial resolution below
0.1 arcsec is needed to isolate the UV light of the center from
the disk in the UV bright galaxies discovered by GALEX at
z=0.1–30.3. This resolution can provide in these galaxies a
spatial sampling better than 500 pc which is the typical size
of nuclear starbursts.
Thus, the lack of any actual (or scheduled) UV mission
with a high spatial (better than 0.1 arcsec) and intermediate
(better than 100 km s−1) spectral resolution long slit spec-
trograph and a high spatial resolution imager with high sen-
sitivity will slow down the progress in our knowledge of
starbursts, and of their impact on the origin and evolution of
galaxies.
There are still many open questions in starburst galaxies
that need to be answered through space UV observations.
Some of them, listed below, have been proposed by experts
in the field.
Springer

100 Astrophys Space Sci (2006) 303:85–102
11.1. Dr. Veronique Buat
– GALEX will observe thousands (and even millions) of
galaxies in its imaging mode and new classes of objects
will certainly be discovered, an example are the local
(0.1≤z≤0.3) UV luminous galaxies (Heckman et al.,
2005). The UV spectroscopic follow-up of these GALEX
sources detected in the broad NUV and FUV bands is of
prime importance. Indeed, the FUV-NUV color only gives
a very crude estimate of the shape of the UV continuum,
especially at high redshift (Burgarella et al., 2005). The
GALEX capabilities in the spectroscopic mode are only
limited to the brightest objects. The combined effects of
the star formation history, the IMF and the interstellar
dust will only be quantified with intermediate resolution
spectra to model the SED between 1000 and 3000 ˚
A.
– Our knowledge of the UV spectral distribution of star-
burst galaxies relies almost entirely on the IUE observa-
tions of bright nearby starbursting objects. But the IUE
aperture can cover only the central starburst of many of
the most nearby galaxies. The recent photometric obser-
vations of GALEX not only confirm that the UV charac-
teristics of the central starbursts are not valid for normal
star forming galaxies (cf. Section 4) but also show that
these central properties may well not be representative of
the entire starbursting galaxies: the interplay of the star
formation history and the dust attenuation are likely to
modify the UV spectrum in a rather complex way varying
from place to place even in starburst galaxies. Unfortu-
nately, the spectral capabilities of GALEX (slitless mode)
will not allow us to carry out a detailed analysis of the
UV spectral energy distributions in various media. The
ideal mode for such studies is integral field spectroscopy
on a large field with medium resolution (or at least slit
spectroscopy) in order to be also sensitive to the diffuse
emission.
11.2. Dr. Rosa M. Gonz´
Alez Delgado and Dr. Luis
Colina
– There are still many open questions related with the
starburst-AGN connection. In particular, the role that
young stellar clusters play in the energetics of AGNs; the
frequency of nuclear young stellar clusters in AGNs; their
properties (luminosities, masses, ages, IMF, metallicities,
etc.). High spatial resolution spectroscopy is needed to iso-
late stellar clusters (of a few pc size) in the region where
the dynamical influence of the black hole is significant,
within 10 pc of the Seyfert nuclei (Ferrarese et al., 2001).
Intermediate spectral resolution (better than 100 km s−1)
is needed to isolate the interstellar from the stellar com-
ponent of the high ionization UV lines.
11.3. Dr. Claus Leitherer and Prof. Timothy Heckman
– Is there direct evidence for macroscopic turbulence in the
ISM? Interstellar absorption lines are thought not to indi-
cate gravity but rather stirring by winds and supernovae.
Testing this hypothesis requires spectrographs with re-
solving power of tens of thousand and higher sensitivity
than, e.g., STIS. Such data would allow us to probe the
kinematic structure and morphology of the ISM and con-
struct a kinematic model for all its phases, including the
outflow.
– Do starburst galaxies enrich the IGM in metals? There is
paramount evidence for the existence of large-scale galac-
tic superwinds transporting the nucleosynthetic products
out of the star-forming regions into the galaxy halos.
The question of material actually escaping from starburst
galaxies is still unanswered. The next generation of spec-
trographs will need higher sensitivity to probe starburst
galaxy environments out to tens of kpc using background
quasars. This could be a decisive test of the metal escape
likelihood and IGM enrichment.
11.4. Dr. J. Miguel Mas-Hesse
– The emission of Lyαphotons is of paramount importance
to study star formation episodes at redshifts z≥2, when
the line becomes visible in the optical range, and Hαis
already shifted to the NIR. As discussed in Section 8 the
visibility of the line depends on several factors, including
the distribution and kinematics of the neutral gas, the cov-
ering by dust,... Understanding the process of emission
and absorption of Lyman alpha photons requires not only
spectroscopy, but also imaging, especially if combined
with Hαimaging of the same region. Kunth et al. (2003)
and Hayes et al. (2005, A&A, in press [astro-ph/0503320])
have shown that it is possible to obtain Lyαemission
maps of starburst galaxies, though the instrumental setup
of HST/ACS was not optimized for it.
The process of Lyαemission and absorption could be
better understood by performing imaging observations of
starburst galaxies at redshifts z=0 to 1, looking for cor-
relations between the visibility of the line and the mor-
phology, evolutionary state, size or metallicity of the dif-
ferent galaxies. This could be achieved with an UV imag-
ing camera with the adequate set of narrow/broad band
filters covering the 1200–2400 ˚
A range. Complementary
Hαobservations at this redshift range could be obtained
from ground.
Acknowledgements We are very grateful to Veronique Buat, Luis
Colina, Timothy Heckman, Claus Leitherer, and Miguel Mas-Hesse
for discussing the main open questions that need to be answered with
UV instrumentations. We also thank Enrique P´erez, Miguel Cervi˜no,
Springer

Astrophys Space Sci (2006) 303:85–102 101
Valentina Luridiana, Jorge Iglesias, Jes´us Maiz, Guillermo Tenorio-
Tagle and an anonymous referee for very useful comments that im-
proved the presentation of the paper. This work has been supported by
the Spanish Ministerio de Educai´on y Ciencia through the grant AYA-
2004-02703.
References
Aloisi, A., Savaglio, S., Heckman, T.M., Hoopes, C.G., Leitherer, C.,
Sembach, K.R.: Astrophysical Journal 595, 760 (2003)
Aloisi, A., Heckman, T.M., Hoopes, C.G., Leitherer, C., Savaglio, S.,
Sembach, K.R.: Starbursts near and far: from 30 Dor to Lyman
Break Galaxies. Astrophysics & Space Science Library 329,P2
(2005)
Balzano, V.A.: Astrophysical Journal 268, 602 (1983)
Bianchi, L., Thilker, D., Burgarella, D. et al.: Astrophysical Journal
619, L71 (2005)
Bland-Hawthorn, J., Putman, M.E.: in Hibbard, J.E., Rupen, M.P., van
Gorkom, J.H.: (eds), ASP Conf. Series 240, p. 369 (2005)
Buat, V., Burgarella, D., Deharveng, J.M., Kunth, D.: A&A 393,33
(2002)
Buat, V., Iglesias-P ´aramo, J., Seibert, M. et al.: Astrophysical Journal
619, L51 (2005)
Burgarella, D., Buat, V., Small, T. et al.: Astrophysical Journal 619, L63
(2005)
Calzetti, D., Kinney, A.L.: Astrophysical Journal 399, 39 (1992)
Calzetti, D., Kinney,A.L., Storchi-Bergmann, T.: Astrophysical Journal
467, 38 (1996)
Calzetti, D.: Astronomical Journal 113, 162 (1997)
Cannon, J.M., Skillman, E.D., Sembach, K.R., Bomans, D.J.: Astro-
physical Journal 618, 247 (2005)
Castellanos, M., D´ıaz, A.I., Tenorio-Tagle, G.: Astrophysical Journal
565, L79 (2002)
Cervi˜no, M., Mas-Hesse, J.M.: A&A 284, 749 (1994)
Chandar, R., Leitherer, C., Tremonti, C.A., Calzetti, D.: Astrophysical
Journal 586, 939 (2003a)
Chandar, R., Leitherer, C., Tremonti, C.A.: Astrophysical Journal 604,
153 (2003b)
Chandar, R., Leitherer, C., Tremonti, C.A., Calzetti, D., Aloisi, A.,
Meurer, G.R., de Mello, D.: Astrophysical Journal 628, in press,
astro-ph/0505024 (2005)
Charlot, S., Fall, M.: Astrophysical Journal 539, 718 (2000)
Cid Fernandes, R., Terlevich, R.: 272, 423 (1995)
Cid Fernandes, R., Heckman, T., Schmitt, H., Gonz´alez Delgado, R.M.,
Storchi-Bergmann, T.: Astrophysical Journal 558, 81 (2001)
Colina, L., Gonz´alez Delgado, R., Mas-Hesse, J.M., Leitherer, C.,
Jim´enez-Bail´on, E.: Astrophysical Journal 579, 545 (2002)
Conti, P., Leitherer, C., Vacca, W.: Astrophysical Journal 461, L87
(1996)
Daddi, E., Cimatti, A., Renzini, A., Fontana, A., Mignoli, M., Pozzetti,
L., Tozzi, P., Zamorani, G.: Astrophysical Journal 617, 746 (2004)
Deharveng, J.-M., Buat, V., Le Brun, V., Milliard, B., Kunth, D., Shull,
J.M., Gry, C.: A&A 375, 805 (2001)
de Grijs, R., Gonz´alez Delgado, R.M.: Astrophysics & Space Science
Library 329 (2005)
de Mello, D.F., Leitherer, C., Heckman, T.M.: Astrophysical Journal
530, 251 (2000)
Fanelli, M., M.N., O’Connel, R.W., Thuan, T.X.: Astrophysical Journal
334, 665 (1988)
Fern´andez-Soto, A., Lanzetta, K.M., Chen, H.-W.: MNRAS 342, 1215
(2003)
Ferrarese, L., Pogge, R.W., Peterson, B.M., Merritt, D., Wandel, A.,
Joseph, C.L.: ApJ 555, 79 (2001)
Gabel, J.R., Bruhweiler, F.C.: Astronomical Journal 124, 737 (2002)
Giallongo, E., Cristiani, S., D’Odorico, S., Fontana, A.: Astrophysical
Journal 568, L9 (2002)
Goldader, J.D., Meurer, G., Heckman, T.M., Seibert, M., Sanders, D.B.,
Calzetti, D., Steidel, C.C.: Astrophysical Journal 568, 651 (2002)
Gonz´alez Delgado, R.M., Leitherer, C., Heckman, T.: Astrophysical
Journal 489, 601 (1997)
Gonz´alez Delgado, R.M., Leitherer, C., Heckman, T., Lowenthal,
J.D., Ferguson, H.C., Robert, C.: Astrophysical Journal 495, 698
(1998a)
Gonz´alez Delgado, R.M., Heckman, T., Leitherer, C., Meurer, G.,
Krolik, J., Wilson, A. S., Kinney, A., Koratkar, A.: Astrophysi-
cal Journal 505, 174 (1998b)
Gonz´alez Delgado, R.M., Garc´ıa-Vargas, M.L., Goldader, J., Leitherer,
C., Pasquali, A.: Astrophysical Journal 513, 707 (1999)
Gonz´alez Delgado, R.M., P´erez, E.: MNRAS 317, 64 (2000)
Gonz´alez Delgado, R.M., Heckman, T., Leitherer, C.: Astrophysical
Journal 546, 845 (2001)
Gonz´alez Delgado, R.M., Leitherer, C., Stasinska, G., Heckman, T.M.:
Astrophysical Journal 580, 824 (2002)
Gonz´alez Delgado, R.M., Cid Fernandes, R., P´erez, E., Martins, L.P.,
Storchi-Bergmann, T., Schmitt, H., Heckman, T., Leitherer, C.:
Astrophysical Journal 605, 127 (2004)
Gordon, K.D., Calzetti, D., Witt, A.N.: Astrophysical Journal 487, 625
(1997)
Hartmann, L.W., Huchra, J.P., Geller, M.J., O’Brien, P., Wilson, R.:
Astrophysical Journal 326, 101 (1988)
Heckman, T.M., Leitherer, C.: Astronomical Journal 114, 69 (1997a)
Heckman, T.M., Gonz´alez-Delgado, R., Leitherer, C., Meurer, G.R.,
Krolik, J., Wilson, A.S., Koratkar, A., Kinney, A.: Astrophysical
Journal 482, 114 (1997b)
Heckman, T.M., Robert, C., Leitherer, C., Garnett, D.R., van der Rydt,
F.: Astrophysical Journal 503, 646 (1998)
Heckman, T.M.: ASPC 148, 127 (1998)
Heckman, T., Sembach, K.R., Meurer, G., Strickland, D.K., Martin,
C.L., Calzetti, D., Leitherer, C.: Astrophysical Journal 554, 1021
(2001a)
Heckman, T., Sembach, K.R., Meurer, G., Leitherer, C., Calzetti, D.,
Martin, C.L.: Astrophysical Journal 558, 56 (2001b)
Heckman, T.M.: Astrophysic in the Far-Ultraviolet, ASP Conf. Series.
astro-ph/0410383 (2004)
Heckman, T.M.: Starbursts: From 30 doradus to lyman break galaxies.
Astrophysics & Space Science Library 329, 3 (2005)
Heckman, T., Hoopes, C.G., Seibert, M. et al.: Astrophysical Journal
619, 35 (2005)
Ho, L., Filippenko, A.: Astrophysical Journal 472, 600 (1996)
Ho, L., Filippenko, A., Sargent, W.: Astrophysical Journal Supplement
112, 315 (1997)
Hoopes, C.G., Heckman, T.M., Strickland, D.K. et al.: Astrophysical
Journal 619, L99 (2005)
Hurwitz, M., Jelinsky, P., Van Dyke Dixon, W.: Astrophysical Journal
481, L31 (1997)
Jamet, L., P´erez, E., Cervi˜no, M., Stasinska, G., Gonz ´alez Delgado,
R.M., Vilchez, J.M.: A&A 426, 399 (2005)
Johnson, K.E., Leitherer, C., Vacca, W., Conti, P.S.: Astrophysical Jour-
nal 120, 1273 (2000)
Keel, W., Holberg, J.B., Treuthardt, P.M.: Astronomical Journal 128,
211 (2004)
Kinney, A.L., Bohlin, R.C., Calzetti, D., Panagia, N., Wyse, R.F.G.:
Astrophysical Journal Supplement Series 86, 5 (1993)
Kudritzki, R.-P., Puls, J., Lennon, D.J., Venn, K.A., Reetz, J., Najarro,
F., McCarthy, J.K., Herrero, A.: A&A 350, 970 (1999)
Kunth, D., Sargent, W.L.W.: Astrophysical Journal 300, 496 (1986)
Kunth, D., Lequeux, J., Sargent, W.L.W., Viallefond, F.: A&A 282, 709
(1994)
Springer

102 Astrophys Space Sci (2006) 303:85–102
Kunth, D., Mas-Hesse, J.M., Terlevich, E., Terlevich, R., Lequeux, J.,
Fall, S.M.: A&A 334, 11 (1998)
Kunth, D., Leitherer, C., Mas-Hesse, J.M., Ostlin, G., Petrosian, A.:
Astrophysical Journal 597, 263 (2003)
Lebouteiller, V., Kunth, D., Lequeux, J., Lecavelier des Etangs, A.,
D´esert, J.-M., H´ebrard, G., Vidal-Madjar, A.: A&A 415, 55 (2004)
Lecavelier des Etangs, A., D´esert, J.-M., Kunth, D., Callejo, G., Ferlet,
R., H´ebrard, G., Lebouteiller, V.: A&A 413, 131 (2004)
Leitherer, C.: Starbursts near and far: From 30 Dor to Lyman break
galaxies. Astrophysics & Space Science Library 329, 89 (2005)
Leitherer, C., Robert, C., Drissen, L.: Astrophysical Journal 401, 596
(1992)
Leitherer, C., Heckman, T.M.: Astrophysical Journal Supplement Series
96, 9 (1995a)
Leitherer, C., Ferguson, H.C., Heckman, T.M., Lowenthal, J.D.: Astro-
physical Journal 454, 19 (1995b)
Leitherer, C., Robert, C., Heckman, T.M.: Astrophysical Journal Sup-
plement Series 99, 173 (1995c)
Leitherer, C., Vacca, W.D., Conti, P.S., Filippenko, A.V., Robert, C.:
Astrophysical Journal 465, 717 (1996)
Leitherer, C. et al.: Astrophysical Journal Supplement Series 123,3
(1999)
Leitherer, C., Leao, J.R.S., Heckman, T.M., Pettini, M., Robert, C.:
Astrophysical Journal Supplement Series 550, 724 (2001)
Leitherer, C., Li, I.-H., Calzetti, D., Heckman, T.M.: Astrophysical Jour-
nal 140, 303 (2002)
Lequeux, J., Kunth, D., Mas-Hesse, J.M., Sargent, W.L.W.: A&A 301,
18 (1995)
MacLow, M.-M., McCray, R., Norman, M.: Astrophysical Journal 337,
141 (1989)
Madau, P., Ferguson, H.C., Dickinson, M.E., Giavalisco, M., Steidel,
C.C., Fruchter, A.: MNRAS 283, 1388 (1996)
Maeder, A., Conti, P.: ARA&A 32, 227 (1994)
Malkan, M., Webb, W., Konopacky, Q.: Astrophysical Journal 598, 878
(2003)
Maoz, D., Koratkar, A., Shields, J.C., Ho, L.C., Filippenko, A.V.,
Sternberg, A.: Astronomical Journal 116, 55 (1998)
Maoz, D., Nagar, N.M., Falcke, H., Wilson, A.S.: Astrophysical Journal
625, 699 (2005)
Martin, D. et al.: Astrophysical Journal 619, L1 (2005)
Mas-Hesse, J.M., Kunth, D.: A&AS 88, 399 (1991)
Mas-Hesse, J.M., Kunth, D.: A&A 349, 765 (1999)
Mas-Hesse, J.M., Kunth, D., Tenorio-Tagle, G., Leitherer, C., Terlevich,
R.J., Terlevich, E.: Astrophysical Journal 598, 858 (2003)
Mehlert, D., Tapken, C., Appenzeller, I., Noll, S., de Mello, D.,
Heckman, T.M.: Starbursts: From 30 Doradus to Lyman Break
Galaxies. Astrophysics & Space Science Library 329, 299 (2005)
Meier, D., Terlevich, R.: Astrophysical Journal 246, L109 (1981)
Meurer, G.: Nature 375, 742 (1995)
Meurer, G.R., Heckman, T.M., Leitherer, C., Kinney, A., Robert, C.,
Garnett, D.R.: Astronomical Journal 110, 2665 (1995)
Meurer, G.R., Heckman, T.M., Lehnert, M.D., Leitherer, C., Lowenthal,
J.: Astronomical Journal 114, 54 (1997)
Meurer, G.R., Heckman, T.M., Calzetti, D.: Astrophysical Journal 521,
64 (1999)
Miralda-Escud´e, J., Ostriker, J.P.: Astrophysical Journal 350, 1 (1990)
Noll, S., Mehlert, D., Appenzeller, I. et al.: A&A 418, 885 (2004)
Pettini, M., Lipman, K.: A&A 297, L63 (1995)
Pettini, M., Steidel, C.C., Adelberger, K.L., Dickinson, M., Giavalisco,
M.: Astrophysical Journal 528, 96 (2000)
Pettini, M., Shapley, A.E., Steidel, C.C. et al.: Astrophysical Journal
554, 981 (2001)
Pettini, M., Rix, S.A., Steidel, C.C., Adelberger, K.L., Hunt, M.P.,
Shapley, A.E.: Astrophysical Journal 569, 742 (2002)
Rosa, M.R., Benvenuti, P.: A&A 291, 1 (1994)
Rix, S.A., Pettini, M., Leitherer, C., Bresolin, F., Kudritzki, R.-P.,
Steidel, C.C.: Astrophysical Journal 615, 98 (2004)
Robert, C., Leitherer, C., Heckman, T.M.: Astrophysical Journal 418,
749 (1993)
Robert, C., Pellerin, A., Aloisi, A., Leitherer, C., Hoopes, C., Heckman,
T.M.: Astrophysical Journal Supplement Series 144, 21 (2003)
Savage, B.D., Sembach, K.R.: Astrophysical Journal 379, 245 (1991)
Savaglio, S., Glazebrook, K., Abraham, R.G. et al.: Astrophysical Jour-
nal 602, 51 (2004)
Schaerer, D.: A&A 382, 28 (2002)
Seibert, M., Martin, C.D., Heckman, T.M. et al.: Astrophysical Journal
619, L55 (2005)
Sekiguchi, K., Andersenon, K.S.: A&A 94, 644 (1987)
Shapley, A.E., Steidel, C.C., Pettini, M., Adelberger, K.L.: Astrophys-
ical Journal 588, 65 (2003)
Steidel, C., Giavalisco, M., Pettini, M., Dickinson, M., Adelberger,
K.L.: Astrophysical Journal 462, L17 (1996)
Steidel, C., Pettini, M., Adelberger, K.L.: Astrophysical Journal 546,
665 (2001)
Steidel, C.C., Shapley, A.E., Pettini, M., Adelberger, K.L., Erb,
D.K., Reddy, N.A., Hunt, M.P.: Astrophysical Journal 604, 534
(2004)
Tenorio-Tagle, G.: Astronomical Journal 111, 1641 (1996)
Tenorio-Tagle, G., Mu ˜noz-Tu˜non, C.: Astrophysical Journal 478, 134
(1996)
Tenorio-Tagle, G., Silich, S.A., Kunth, D., Terlevich, E., Terlevich, R.J.:
309, 332 (1999)
Terlevich, R.: RM×AC 6, 1 (1997)
Terlevich, R., M´elnick, J., Masegosa, J., Moles, M., Copetti, M.V.F.:
A&A 91, 285 (1991)
Terlevich, E., D´
iaz, A.I. Terlevich, R., Garc´
ia Vargas, M.L.: MNRAS
260, 3 (1993)
Thuan, T.X., Lecavelier des Etangs, A., Izotov, Y.I.: Astrophysical Jour-
nal 565, 941 (2002)
Tran, H.: Astrophysical Journal 440, 565 (1995)
Tremonti, C.A., Calzetti, D., Leitherer, C., Heckman, T.M.: Astrophys-
ical Journal 555, 322 (2001)
Vacca, W., Robert, C., Leitherer, C., Conti, P.S.: Astrophysical Journal
444, 647 (1995)
Valls-Gabaud, D.: Astrophysical Journal 419, 7 (1993)
V´azquez, G.A., Leitherer, C., Heckman, T.M., Lennon, D.J., de Mello,
D.F., Meurer, G.R., Martin, C.: Astrophysical Journal 1600, 162
(2004)
Walborn, N.R.: International Astronomical Union. Symposium The
stellar content of 30 doradus derived from spatially integrated
ultraviolet spectra: A test of spectral synthesis models 148, 145
(1991)
Weedman, et al.: Astrophysical Journal 248, 105 (1981)
Whitmore, B.C., Schweizer, F., Leitherer, C., Borne, K., Robert, C.:
Astronomical Journal 106, 1354 (1993)
Williams, R.E. et al.: Astronomical Journal 112, 1335 (1996)
Springer

Astrophys Space Sci (2005) 303:103–122
DOI 10.1007/s10509-005-9027-2
ORIGINAL ARTICLE
A View to the Future: Ultraviolet Studies
of the Solar System
Noah Brosch ·John Davies ·Michel C. Festou†·
Jean-Claude G´erard
Received: 9 August 2005 / Accepted: 21 November 2005
C
Springer Science +Business Media B.V. 2006
Abstract We discuss the status of ultraviolet knowledge of
Solar System objects. We begin with a short historical sur-
vey, followed by a review of knowledge gathered so far and
of existing observational assets. The survey indicates that
UV observations, along with data collected in other spectral
bands, are necessary and in some cases essential to under-
stand the nature of our neighbors in the Solar System. By
extension, similar observations are needed to explore the na-
ture of extrasolar planets, to support or reject astro-biology
arguments, and to compose and test scenarios for the forma-
tion and evolution of planetary systems.
We propose a set of observations, describing first the nec-
essary instrumental capabilitites to collect these and outlining
what would be the expected scientific return. We identify two
immediate programmatic requirements: the establishment of
†Deceased 11 May 2005
Dedication: Michel Festou, our co-author and a very important
contributor to this paper, passed away while this paper was being
completed. We dedicate it to his memory
N. Brosch
The School of Physics and Astronomy, Beverly and Raymond
Sackler Faculty of Exact Sciences, Tel Aviv University, Tel Aviv
69978, Israel
J. Davies
UKATC, Royal Observatory, Blackford Hill, Edinburgh EH9 3HJ,
UK
M. C. Festou
Observatoire Midi-Pyr´en ´ees 14, avenue E. Belin 31400 Toulouse,
France
J.-C. G´erard
Laboratoire de Physique Atmospherique ´et Planetaire, Universite
de Liege, allee du 7 aout,
4000 Liege, 94720 Belgium
a mineralogic database in the ultraviolet for the characteri-
zation of planetary, ring, satellite, and minor planet surfaces,
and the development and deployment of small orbital solar
radiation monitors. The first would extend the methods of
characterizing surfaces of atmosphere-less bodies by adding
the UV segment. The latter are needed to establish a baseline
against which contemporaneous UV observations of Solar
System objects must be compared.
We identify two types of UV missions, one appropriate
for a two-meter-class telescope using almost off-the-shelf
technology that could be launched in the next few years,
and another for a much larger (5–20 meter class) instrument
that would provide the logical follow-up after a decade of
utilizing the smaller facility.
Keywords Planets .Comets .Solar system .Ultraviolet
Introduction
The UV window extends from 20 to 400 nm. It covers the
spectral domain where atoms have most of their resonance
lines and where simple ions and molecules have their flu-
orescence transitions. This is also the region where most
molecules and atoms are photodissociated and photoionized.
Below 100 nm is the spectral segment where most rare gases
have their resonance lines. Solar system objects can be ob-
served in this spectral region either because they harbour hot
or violent environments or because the solar light that is be-
ing absorbed and then scattered by them contains thousands
of lines below 200 nm.
One of the basic questions in modern astrophysics is how
planets “work,” how planetary systems originated, and how
life emerged on Earth. By studying our Solar System we
are linking ‘local’ studies to the issue of the existence of
Springer

104 Astrophys Space Sci (2005) 303:103–122
Earth-like extrasolar planets and the conditions expected on
their surfaces. The discovery of biomarkers on Earth analogs
is the essence of the search for extra-terrestrial life addresses
the question “are we alone in the Universe?” These are the
basic questions being asked by NASA and ESA: “How did
our Earth and our Solar System evolve, where are we in
the Universe, where are we going, and where did life come
from?”
Planets and other minor bodies represent the end stage
of the Solar System’s formation and their present state is
the result of numerous accretion, coalescence, and evolution
processes. The various components are inter-related: plan-
ets were formed through planetesimal accretion, evidence
of which remains in the form of asteroids and other bod-
ies such as Trans-Neptunian Objects (TNOs). Some comets
are fragments of TNOs; others come from the Oort cloud,
both in very cold and protective environments of the Solar
System. Thus the minor bodies of the Solar System have
retained information that allows us to study the primordial
cloud that formed our Solar System some 4.5 Gyrs ago. The
traditional rocky asteroids are intermediate objects, between
planetesimals and fully-grown planets, and thus contain in-
formation on the stage of nebular matter condensation prior
to the planet accretion. The collisional history of the Solar
System can be investigated by studying the internal structures
and size distributions of asteroids, TNOs and Near Earth Ob-
jects (NEOs). Finally, the interplanetary dust is mostly the
result of the “grinding down” of Main Belt asteroids, proba-
bly TNO debris, and ejecta from cometary nuclei.
While many objectives of solar system research can be
achieved by optical and near-IR (nIR) imaging, surface min-
eralogic characterization requires a wide spectral range in-
cluding the UV. Observations of planetary aurorae from the
Earth are impossible, owing to the brightness of the sun-
lit planetary disks and the lack of contrast at visible and
nIR wavelengths. Planetary missions with UV capabilities
(such as the Mercury Atmospheric and Surface Composition
Spectrometer (MASCS) on the Mercury probe MESSEN-
GER, and the ALICE instrument on the Pluto/Kuiper Belt
New Horizons mission) are rare and far apart; so it is nec-
essary to consider Solar System studies in the context of
general-purpose astronomical payloads. Even though a UV
space telescope might not be fully dedicated to planetary as-
tronomy, a complement of valuable targets exists and Solar
System observations can be made without compromising the
astrophysical goals of such a mission.
Planetary studies require synoptic observations over peri-
ods of time ranging from a single revolution (hours to days),
to the length of a comet passage through the inner planetary
region and the seasons of telluric planets (months), to orbital
revolutions about the Sun and to seasons on giant planets
and their satellites (months to years). Since some planetary
phenomena, such as aurorae, are directly linked to the solar
activity, studies should span one or more 11-year solar cycles.
The variability of Solar System objects, sometimes on short
time scales, underlines the need for long-term studes to sep-
arate intrinsic from evolution-driven properties of planetary
bodies and comets.
For a given aperture size, UV astronomy from space can
achieve much higher spatial resolution than from the ground
because of the absence of the smearing effect of the Earth’s
atmosphere (“seeing”) and because of the smaller diffraction
limit of UV telescopes. Present-day 8-m and larger ground-
based optical telescopes, equipped with adaptive optics (AO)
and with fields of view restricted to the “coplanarity patch,”
offer the same angular resolution as a two-meter space UV
telescope but are restricted to the optical-near IR bands and
are hampered by the natural sky brightness. The sky back-
ground at UV wavelengths is darker by about five magnitudes
than at the best ground-based observatories (O’Connell 1987)
and allows observations of very faint objects, in particular
those with extended very low surface brightness features.
Why argue now for UV astronomy? Few instruments that
allow UV planetary observations are presently available and
none will be available in the near future as the few existing
missions reach their design and even extended lifetimes. The
heritage of past missions, and the expertise of scientists in
designing, performing, and analyzing UV planetary observa-
tions are dwindling as individuals reach the end of their active
careers. New scientists refrain from entering a field devoid
of a promise to access cutting-edge research instrumentation.
There is need to continue synoptic observations of variable
sources even after specific planetary probes complete their
missions, and it is necessary to maintain the know-how of
observing and working with specific data sets.
Review of achievements
The study of the UV emission from Solar System bodies,
which started as an exploratory task in the middle of the last
century, proved useful for the understanding of planetary at-
mospheres and of plasma phenomena. Results obtained about
distant bodies were applied to the understanding of our Earth.
Spectroscopy of comets in the UV revealed the presence of
new compounds and clarified the mechanisms accounting for
their presence.
Key missions of solar system UV exploration
Rocket flights of (very) short durations provided the first
exploratory data and flights of that nature still remain useful
for developmental and testing purposes. As late as 2004 a
Black Brant rocket lofted a telescope and spectrometer to
record the UV emission from Mercury. Modern payloads
Springer

Astrophys Space Sci (2005) 303:103–122 105
Table 1 Past space missions that performed UV observations of Solar System bodies
Dates Mission Agency Instrument Resolution Range (nm) Comment
1968 OAO-2 NASA Photometer +
spectrometer ∼100 100–400 H coma of Comet
Tago-Sato-Kosaka
1970 OGO-5 NASA Lyman alpha
photometer
Low 121.6 Interplanetary H. Comets
Bennett and P/Encke
1972 TD-1A ESRO S-59 +S2/68
Spectrometers ∼70 133–280 UV sky survey
1975 ASTP NASA EUV Telescope Imager 5–150 Manned space flight
1977–1989 Voyager 1/2 NASA UV Spectrometer ∼40 40–180 Planetary Mission
1972–1981 OAO-3 Copernicus NASA UV Telescope ∼200 75–300 Giant Planets and Comets
1983 Astron USSR/CNES Spectrometer ∼200 150–350 P/Halley
1993&96 ASTRO-SPAS NASA/DARA ORPHEUS ∼5000 Spectrometer 39–91 Shuttle Free Flyer
1993&96 ASTRO-SPAS NASA/DARA IMAPS ∼200,000 Spectrometer 95–115 Shuttle Free Flyer
1990&95 ASTRO 1/2 NASA HUT ∼100 45–185 Venus, Mars-Polarimetry
1990&95 ASTRO 1/2 NASA WUPPE ∼100 140–320 Moon, Mars, Io, Halley
1990&95 ASTRO 1/2 NASA UIT Imager 120–300
1978–96 IUE NASA/ESA UV Spectrometer ∼50 and ∼5000 115–320 Giant Planets, Aurora, 50+
Comets
1992–2000 EUVE NASA EUV Spectrometer
and Imager ∼400 8–75 Sky Survey & follow-up,
comets, Venus, Jupiter
1989–2003 HST NASA/ESA WFPC, GHRS Imaging, 2000–80000 >115 Superb images
1998–2004 HST NASA/ESA STIS 100–100000 115–310 All planetary objects
2000– FUSE NASA High Resolution
Spectrometer ∼27,000 90-120 H2on Mars, comets
Not included in this list are several small experiments carried on various missions such as Cosmos 51, 215, 262, Apollo 16, ANS, D2B-Aura,
Skylab and Soyuz/Salyut, as well as short-duration rocket flights.
Fig. 1 Time series observation of the UV spectrum of P/Halley obtained
by the ASTRON orbital observatory (see Table 1) showing the strong
OH feature at 308 nm.
[http://www.crao.crimea.ua/astron/astron.html]
can achieve more than a simple exploration but still lack
the temporal coverage characteristics of Solar System object
phenomena.
The Voyager spectrometers were non-imaging, with me-
chanical collimators defining their fields of view. Their spec-
tral coverage was from 40 to 180 nm, the throughput was
rather low and no spatial information was available. Even
so, their results regarding the properties of giant planet
atmospheres, obtained by occultation techniques or in di-
rect “airglow” mode, were unique because of the proximity
of the instruments to their targets.
NASA’s Copernicus satellite (Orbiting Astronomical
Observatory-3, launched in 1972 and operated till February
1981) performed Far-UV (FUV) investigations and allowed
the detection of important atomic and molecular species of
the local interstellar medium, among which molecular hy-
drogen and deuterium. OAO-3 offered a unique way to study
the upper atmospheres of planets and key components of
cometary atmospheres.
Copernicus was followed by ORFEUS (Orbiting and Re-
trievable Far and Extreme Ultraviolet Spectrometer) and
by HUT (Hopkins Ultraviolet Telescope), both Shuttle-
launched and retrievable payloads. HUT flew twice on the
ASTRO platform for missions of order 12 days, along with
WUPPE (Wisconsin Ultraviolet Photo-Polarimeter Experi-
ment), the only instrument that has provided polarization
measurements in the UV. WUPPE was used to study the
Moon, Mars, Io, and comet Halley. HUT observed Jupiter,
Venus and Mars. IMAPS (Interstellar Medium Absorption
Profile Spectrometer) operated, as ORFEUS did, on board
the AstroSPAS space shuttle-borne platform (1993).
Most missions mentioned above were limited in duration
or in the observing time allocated for Solar System studies.
Springer

106 Astrophys Space Sci (2005) 303:103–122
Their role was, by necessity, mainly exploratory. The excit-
ing results indicated the need for missions of much longer
duration.
The spectroscopic results obtained by the International Ul-
traviolet Explorer (IUE) spacecraft with its 45-cm telescope
surpass by far those obtained by any other spacecraft, perhaps
with the exception of the Hubble Space Telescope (HST).
IUE was launched in January 1978 and its three-year mission
was extended year after year until it was deliberately termi-
nated in September 1996. These mission extensions provided
much of the most valuable science return. IUE performed
comparative studies of auroral activity at Jupiter, Saturn, and
Uranus and demonstrated the impact of solar wind varia-
tions on the brightness of the Jovian aurora. The amount of
hydrocarbon absorption in auroral spectra was used to de-
termine the FUV color ratio, from which the energy of the
primary auroral electrons could be inferred. The spectra re-
vealed the existence of new sulphur-bearing compounds in
cometary spectra (S2in comet IRAS-Araki-Alcock; CS2in
numerous comets) and investigated the abundance of water
and carbon compounds such as CO and CO2, providing a
database on gas production rates in over 50 comets that is
only surpassed in size by OH 18 cm radio surveys and by the
ground-based survey of the UMD-Lowell group (A’Hearn
1995).
The field of Extreme UV (EUV) observations was cov-
ered by the Extreme Ultra Violet Explorer (EUVE) satellite,
launched in 1992 and operated until 2000. EUVE detected
emission from comets resulting from charge-exchange reac-
tions with highly charged Solar wind particles, observed the
dayglow of the atmosphere of Venus, showed the existence
of EUV emission from the Full Moon that indicated differ-
ences on the local albedo, and detected helium emission from
the atmosphere of Jupiter following the impact of km-sized
disruption fragments from D/Shoemaker-Levy-9 (SL9).
The workhorse of solar system studies in the field of high-
resolution imaging, or observations in spectral domains not
visible from the Earth’s atmosphere, has been the Hubble
Space Telescope (HST), a 2.4-meter telescope for the UV-
to-NIR domain launched in 1990. The HST has produced the
best imaging database of other celestial bodies obtained from
the Earth’s vicinity. The HST results range wide, from atmo-
spheric studies of the giant and the terrestrial-like planets, to
imaging the dynamics of Jupiter’s and Saturn’s aurora down
to 10-s variations and monitoring the energy of the impinging
auroral electrons in regions connected to different magneto-
spheric plasma sources, to single-pixel imaging of Sedna and
the coarse mapping of the surfaces of Pluto, 1 Ceres and 4
Vesta. HST tracked the disintegration of comets (SL9 and
C/1999 S4 (LINEAR)) and found ozone on Ganymede and
molecular oxygen in the atmosphere of Europa. With the
demise in 2004 of the Space Telescope Imaging Spectrome-
ter (STIS) and with a refurbishing mission that could repair
or replace STIS doubtful, the lack of a general-purpose UV
spectroscopic facility is becoming acute.
The Far Ultraviolet Spectroscopic Explorer (FUSE) satel-
lite was launched in 2000; it covers the spectral range from
90.5 to 119.5 nm with reasonably high resolution (R=
24,000–30,000) and is still operating when these lines are
written. FUSE has a 352 by 387 mm2aperture and a PSF
of 1.5 arcsec. The FUSE spectral range was chosen to con-
tain the most important interstellar and hot environment lines
of deuterium, H2, and lines arising from high level ioniza-
tion states of the most abundant atoms in the universe (O, C,
N, ...). FUSE has a relatively low sensitivity, with effective
areas of 20 cm2at 90 nm and 80 cm2at 120 nm, although
this is 10,000 times the sensitivity of Copernicus and the
resolution is much better than any of the space FUV instru-
mentation built before. Among the achievements of FUSE in
the field of Solar system studies we count the discovery of H2
on Mars (Krasnopolsky and Feldman 2001), presumably the
result of photo-dissociation of water and subsequent molecu-
lar formation at mid-altitudes of the Martian atmosphere, the
measurement of the D-to-H ratio from which the existence of
an old global ocean on Mars can be inferred, the detection of
numerous new lines in comet spectra and an unprecedented
coverage of the auroral phenomena in the giant planets. The
measured intensity distribution amongst H2lines affected by
self-absorption was used to infer the pressure level of the
Jovian aurora, which was shown to be quite different from
Saturn’s case.
The FUSE spectra of comet C/2001 A2 LINEAR provided
the first high resolution EUV spectrum of a comet. In addition
to the H2lines, many new lines have been discovered. Quite
a few of these features are still unidentified and require more
work to be understood, and some of them likely are due to
electron impact excitation.
Key results from past UV astronomy missions
Planetary atmospheres and magnetospheres
Results from comparative studies of planetary atmospheres
indicate that the Solar System family can be separated into
four groups:
1. Bodies with Nitrogen dominated atmospheres (Earth, Ti-
tan, Triton, Pluto)
2. Bodies with carbon dioxide dominated atmospheres
(Venus and Mars)
3. Hydrogen & Helium dominated giant planets (Jupiter, Sat-
urn, Uranus, and Neptune)
4. Bodies with thin atmospheres, separable further into three
subgroups:
a. Rocky surfaces (Mercury, the Moon)
b. Volcanic bodies (Io)
c. Icy surfaces (Europa, Ganymede, Callisto, Charon)
Springer

Astrophys Space Sci (2005) 303:103–122 107
Fig. 2 The EUV long wave
spectrum of comet Hyakutake
barely shows the emission lines
originating from
charge-exchange reactions with
solar wind particles (source
Krasnopolsky and Mumma
2001). The need for better
spectral resolution is evident
Studies of the atmospheres of Venus and Mars have helped
our understanding of the greenhouse effect and the impact of
the continuous release of anthropogenic gases such as CO2
into the Earth’s atmosphere. The CO2-dominated planets can
only be studied in the UV, as telluric CO and CO2features
severely limit what can be observed from the ground. The
cycle of water and water-dissociation products, in particular
the deuterated species, allows one to investigate the exchange
of water between the surface and the atmosphere and, ulti-
mately, the escape of atmospheric constituents. Recent FUSE
observations of the H2and D lines showed that in the past
Mars was probably covered by a thick water ocean (Yung
1998; Krasnopolsky 2003).
The NO ultraviolet night airglow on Venus was discov-
ered in 1978 by the Pioneer-Venus UV spectrometer; it was
followed in 1989 by the detection of the UV day glow. The
existence of a suitable orbital platform could have allowed
synoptic observations of these phenomena. On 24 June 1999
the Cassini spacecraft flew by Venus. Prominent features
detected by the UVIS instrument include the HI 121.6 nm
(Lyα) line, the OI 130.4 and 135.6 nm, the CI 156.1 and 165.7
nm multiplets, and the CO Fourth Positive bands. Weaker fea-
tures of NI, CI, CII, OI and CO were also present. The EUV
spectrum contained well-defined features at 58.4 nm (HeI),
83.3 nm (OII), 98.9 nm (OI), and 102.6 nm (HI Lyβ). Weaker
emissions included OI 104.1 nm, CO (C-X) 108.8 nm, NI
113.5 nm, and a blend of CO (B-X) and OI near 115 nm.
The compositon of the upper atmosphere of the giant
planets indicates how these planets and their atmospheres
originated and evolved. Studies of giant planets in our So-
lar System provide the knowledge base against which such
observations must be interpreted. As an example, let us men-
tion the spectacular Lyαabsorption produced by the exo-
planet HD209458b during its transits in front of its star disc
(Vidal-Madjar et al. 2003).
Images of Jupiter and Saturn auroral emissions with the
constantly improving sensitivity of the HST cameras have
opened a new era in the understanding of the interaction of
giant planets with their magnetic environment. They have re-
vealed the general morphology of the precipitation patterns
and their dynamical behaviour. Our knowledge of the com-
plex interaction between the solar wind, the giant planets’
magnetospheres, the current systems flowing through these
magnetised environments and their signature in the planet
ionospheres has thus dramatically improved. In particular,
Jupiter’s main auroral oval probably results from breakdown
of outward corotating and drifting plasma from Io, which
drives currents between Jupiter’s magnetosphere and iono-
sphere (Bunce and Cowley, 2001; Grodent et al., 2003). In
the region of upward current (downward-moving electrons),
field-aligned potentials accelerate electrons to auroral ener-
gies, producing the main auroral oval emissions.
Based on HST FUV images, a similar concept has re-
cently been proposed in the case of Saturn. It is now be-
lieved that the auroral oval at Saturn corresponds to a ring
of upward current bounding the region of open and closed
field lines. Following extensive study of Earth’s aurora, we
now have three cases of magnetosphere-ionosphere coupling
whose most dramatic aspect is the aurora. The discovery of
the FUV magnetic footprints of Io, Ganymede and Europa,
and of Io’s trailing tail (Clarke et al., 2002) generated new
theories on the acceleration of electrons in the flux tubes and
their time stability. Spectral observations of the FUV-EUV
H2line intensity distribution have indicated that the global
thermal structure of the giant planet upper atmospheres is
strongly controlled by the redistribution of the energy flux
occurring after the precipitation of magnetospheric plasma
(Grodent et al., 2002).
Cometary physics
The major result of the early years of exploration in the newly
opened UV window was the detection of water dissociation
products. The HI Lyman alpha (Lyα) emission was studied
Springer

108 Astrophys Space Sci (2005) 303:103–122
in detail by Bertaux et al. (1973) and Keller (1973a, 1973b)
who revealed the existence of hot hydrogen components
in the coma resulting from the photodissociation of water
molecules. This naturally led to the proposal by Blamont and
Festou (1974) and Keller and Lillie (1974), who measured
the scalelength and the production of OH in comet Kohoutek
(1973 XII) and Bennett (1970 II), respectively, that in comets
the water molecule was the parent of most observed hydrogen
atoms and OH radicals.
The use of UV and EUV space telescopes allowed the
detection of important secondary parent molecule emissions
such as those of CO, S2and CS2and the CO Cameron bands,
a proxy for the CO2molecule.Feldman et al. (2002) used
FUSE on C/2000 A2 (LINEAR) to detect new bands of CO at
98.9 and 198.8 nm and also the H2emission, as this molecule
is a natural breakdown product of water. The overall picture
of a comet that emerged around 1980 was that water was
indeed the most abundant nuclear species, as first speculated
by Whipple (1950).
The measurement of the elemental composition of the UV
coma, containing all the most abundant atoms, numerous
ions and simple molecules that are dissociation products of
nuclear species, yields the elemental composition of the nu-
cleus. As a consequence, as shown by Huebner and Benkhoff
(1999), it is possible to investigate the water/CO mixing ra-
tio variation with heliocentric distance. The measurement of
the major compound production rates, and their evolution
with time, delivers fundamental information on how parent
molecules are stored in the comet nuclei and released into
the comae upon heating by sunlight.
The observation of species evaporated from the dust in
Sun-grazing comets and the discovery of over 700 such
comets by SOHO, particularly by the visible-light corona-
graphs (e.g., Biesecker et al., 2002), demonstrates the value
of observations at low solar elongations. In the UV domain,
the results obtained by the SOHO/SWAN instrument (Maki-
nen et al., 2001) show the power of an all-sky Lyαinstrument
to discover and study comets with low water production rates.
The lack of complementary follow-ups, which would have re-
quired the use of other instruments capable of low-elongation
observations, is obvious. In such cases, spectroscopic obser-
vations would likely reveal the presence of many new lines
that are usually unseen because of the weakness of the excit-
ing flux.
Asteroids and planetary surfaces
The reflectivities of planet, asteroid and satellite surfaces
are important ingredients in modeling their thermal evolu-
tion (solar heating vs. cooling to space), in understanding
the weather in these planets, in devising scenarios for the
‘chemical’ weathering of their surfaces, etc.
Stern et al. (1991) obtained some interesting results re-
garding the UV reflectivity of Triton and of the binary
Pluto/Charon from HST observations, following the optical
band surface “mapping” of the two bodies through obser-
vations of the mutual events’ series in 1985–1990 (Young
et al., 2001). Triton’s reddish color in the optical does not
continue into the UV. Instead, a blue upturn appears short-
ward of 275 nm. Charon, on the other hand, is grey in the UV
as well as in the optical. The differences among these bod-
ies, which are at similar heliocentric distances and probably
evolved together, are not currently understood and require
more observations.
The UV spectra of Pluto and Charon obtained with the
HST (e.g., Krasnopolsky 2001) emphasize the difficulty of
obtaining observations of these distant bodies in the UV:
these observations demanded 24 and 16 orbits respectively
yet yielded only ambiguous results. The availability of an in-
strument with a larger collecting area, and of a space platform
with higher observing efficiency than the HST, is obvious.
The interplanetary medium
Cosmic dust covers the surfaces of atmosphereless bodies
and is an important component that rains down on Earth as
visible meteors. In principle, these could originate from the
grinding down of TNOs and Main Belt asteroids, or from
the disintegration of comets, but the relative ratios are un-
known. Each possible source could produce grains with dif-
ferent chemical compositions, porosity, and complexity and
this would result in different optical properties that would be
exhibited also in the UV.
The grains show evolution under the gravitational influ-
ence of the planets, solar radiation pressure, the Yarkovsky
and Poynting-Robertson effects, and solar wind drag. Recent
observations with the Wisconsin Hαmapper (Reynolds et al.,
2004) show that the zodiacal cloud has a definite prograde
signature. The broadened widths of the Hαprofiles, together
with the large amplitude variations in the centroid velocity
with elongation angles, indicate that a significant population
of the dust grains is in elliptical orbits. These grains could
be part of the debris trails of comets seen in IRAS and ISO
maps (by Davies et al., 1984; Davies et al., 1997; Sykes et al.,
1986; Sykes and Walker, 1992).
The reason to study the Zodiacal Light in detail in the UV
is that an accurate measurement of the diffuse UV light in
the Solar System is still lacking. Scattering off small dust
particles in the zodiacal cloud, at heliocentric distances up to
5–10 a.u., provides the UV light (Leinert et al., 1998). The
UV maps of the zodiacal light are much less complete than
in other spectral bands. Information about the UV emission
from zodiacal dust is necessary for the computation of the
thermal equilibrium of dust grains.
Springer

Astrophys Space Sci (2005) 303:103–122 109
Table 2 Science issues and
instrument requirements Scientific focus Instrumentation and mission
•Time-variable solar system phenomena •Large collecting aperture (>2m)
•Interaction of the solar wind with planets, their
satellites and rings •Diffraction limited high angular resolution
(better than 0.02)
•Size distribution and chemical composition of
comets and TNOs and diversity among comets •Pointing accuracy and guidance stability
•Ability to observe moving targets at full
performance
•Circulation and dynamics of planetary
atmospheres •Imaging and high resolution
spectroscopy
•Geochemical provinces on planetary surfaces,
volatile transport
processes
•Minimum solar elongation <20◦
•Mission lifetime >15–20 years, real-time
operations
•Comparative planetology: Mercury to the
TNOs and beyond, Solar System evolution •Data from solar photon and particle monitors
•Other planetary systems •High operational efficientcy
One of the basic parameters in understanding cosmology
lies in the measurement of the background light. While this
background light component has been well mapped in the IR
by COBE/DIRBE, in the optical the still-quoted measure-
ments date back to Pioneer 10 (Toller and Weinberg, 1985).
Worse still, in the UV, where the Local Universe is observed,
the background emission by the zodiacal light is not well
defined at all. The brightness distribution over the sky varies
much less than in the optical, where a factor of three be-
tween the ecliptic poles and ecliptic equator has been mea-
sured. This would imply that while the dust grains that scatter
the visible light are somewhat confined to the ecliptic, those
that produce the UV zodiacal light are more spherically dis-
tributed. The different spatial distribution may reflect a dif-
ference in particle size, with the UV scatterers being smaller,
on average, than the particles responsible for the optical emis-
sion, as well as a possible different origin. Henry and Murthy
(1998) also claim the appearance of a spectral feature at ∼280
nm that has no readily available explanation.
A future UV observatory for solar system studies
In this section we derive observational requirements for fu-
ture UV instruments. We start by listing scientific goals that
will drive the instrumental requirements. Note that Solar Sys-
tem phenomena are time-variable. The atmospheres of many
planets reveal structures and variation with longitude, lati-
tude, and season and everything changes with the solar cycle
phase. Such studies require the performance of synoptic ob-
servation for all types of objects.
The primary scientific objectives of a UV observatory are
a complete characterization of the properties of all types of
solar system bodies, from object sizes, internal structure, and
rotational properties, to surface properties and atmospheric
composition. From these observations, comparisons can be
made to determine how solar system objects form and evolve.
Table 2 lists a number of science questions in the left column
and instrumental requirements in the right column. These
requirements are derived from a number of concurrent con-
siderations; for example, large apertures are required not only
to compensate for the low solar flux in the UV, but also to
counterbalance the low albedos in this spectral region.
Future scientific results
We mentioned above that a particular advantage of observ-
ing in the UV is the possibility of achieving a high angular
resolution by designing diffraction-limited telescopes used
at as short as possible a wavelength. Because of basic op-
tics, the resolving power of the JWST at its main operational
band (4-µm) can be achieved by a half-meter aperture UV
telescope. The WSO/UV telescope, which is one option for
a next generation UV telescope facility, will have in imaging
mode at 100 nm a spatial resolution one order of magnitude
better than that of the JWST.
Note that the ELT (Extremely Large Telescope) as was dis-
cussed in 2005 for possible realization in about two decades,
with a filled 100-m aperture, will offer a field of view of
three arcmin that, with the adaptive optics option, will have
a 1.4 mili-arcsec (mas) resolution with pixels that are less
than one mas in size. This angular resolution is now only the
province of VLBI, but could be achieved in the UV without
the AO option with a ∼20-m spaceborne telescope. In the
sub-mm domain, when the Atacama Large Millimetre Array
(ALMA) will be operational, its highest angular resolution
will be only 5.8 mas (at 720 GHz).
It is possible to translate this angular resolution into spatial
resolution by considering a few examples where we give the
linear and angular size of bodies at their typical distances.
We ask the reader to consider the amount of information
Springer

110 Astrophys Space Sci (2005) 303:103–122
retrievable for a nominal angular resolution of one mas (∼one
km at one a.u.); note that this is the diffraction limit at 100
nm for a 20-m telescope while the same limit is 0.01 for a
2-m telescope.
Planetary atmospheres and magnetospheres
Some key research areas in the field of atmospheric chemical
composition and dynamics are:
rMars: CO2and ozone absorptions, transport of water and
CO2. Loss of ancient oceans. Dust storms. CO2recycling.
rThe role of cold traps and the transport of condensable
species on planets (Mercury, Moon, Mars, Jovian satellites,
Triton, Pluto, Charon and TNOs).
rGlobal circulation of the giant planets, especially the
poorly studied planets Uranus and Neptune. The longitudi-
nal distribution of aerosols and UV absorbers can be deter-
mined and studied over many successive planet rotations.
The comparison of the properties of Uranus (“rolling”
along the ecliptic) and of Neptune will be particularly en-
lightning.
rStudy of the evolution of local atmospheric phenomena in
giant planets by imaging the different kinds of spots and
measuring the evolution of their shape and their transport.
rDistribution of SO2on Venus and on Io, and its relation to
the internal and volcanic activity.
rThe solar Lyαline nearly coincides with a line of the fourth
positive (14,0) band of CO; fluorescent scattering in the
(14,v) band is therefore observable and can be used to
monitor this species’ abundance in the upper atmosphere
of Mars and Venus.
rThe emission lines of atomic oxygen at 130.2 and 135.6
nm in the dayglow spectra of Mars and Venus are very
strong and are tracers of the thermospheres of those planets.
These observations, coupled with those of CO, O3and
CO2(via their absorptions) are essential to understand the
aeronomy of the telluric planets, in particular atmospheric
species’ transport and production/destruction mechanism
as a function of solar cycle activity and input of energetic
solar wind particles.
rSearches for Ar, Ne and N in all atmospheres of solar
system objects to constrain the temperature distribution
in models of the presolar nebula.
rUse of emissions from H, C, O, S and N and their ions to
investigate the physics of the planetary ionospheres.
rInvestigate on a long term basis the role and sources of
neutrals in Saturn’s magnetosphere.
rSeach for activity on low-gravity bodies, in particular
TNOs, by detecting traces of cryovolcanism through ejec-
tion of dust particles.
Sunlight at long UV wavelengths is reflected from strato-
spheres of planets by a combination of Rayleigh (atomic and
molecular) and aerosol scattering. Signature absorptions at
specific wavelengths can be used to identify trace organic
species (e.g., hydrocarbons, nitriles, etc.), many of which
are produced by non-equilibrium processes such as auroral
chemistry and photochemistry. A global latitudinal study of
the abundance of the atmospheric contituents allows one to
separate these two sources of compounds. These tracers can
also be used to study the dynamics of the atmospheres. A
spatial resolution of order 100 km at Jupiter would permit
unprecendented synoptic observations of the atmosphere of
this planet.
The atmosphere–surface interactions at Mars, Io,
Ganymede, Mercury, and the Moon could be studied with
unprecedented details. The surface of these bodies feeds the
atmosphere with volatiles and solid particles and the surface
is more or less covered with condensable materials during the
planet/satellite seasons. The solid particles play a key role in
the thermal balance of some atmospheres (Mars, Titan and
possibly Pluto, if the presence of a haze layer in this planet’s
atmosphere is confirmed).
UV observations allow the characterization and monitor-
ing of the transport of condensable species on planetary sur-
faces such as those of Mercury, the Moon, Mars, Triton, Io,
Pluto and TNOs. The study of volatile escape from low-
gravity bodies, in particular TNOs, is greatly facilitated. The
transient atmosphere of most low-gravity bodies such as the
Moon, Mercury, asteroids, and giant-planet satellites is due
to impact of highly energetic particles. The mid-UV region
offers the possibility to study most of these atmospheric com-
pounds.
Observations of stellar occultations can be used to ob-
tain atmospheric density profiles. The high time reso-
lution of large instruments allows a very good resolu-
tion of the atmosphere height where the occultation takes
place. The occultation signature is stronger in the UV
because of the higher refraction coefficient. The atmo-
spheric absorptions that could possibly be studied are
those of H2,CO
2, and hydrocarbons. Note though that
the occulation rate will mostly be determined by the pop-
ulation of occulted stars, which must be rather strong
UV emitters, rather than by the spatial resolution of the
telescope.
Characterization of the Jovian magnetized environment,
and observations of auroral phenomena at Jupiter and Sat-
urn, which is traditionally done with probes in the vicinity of
these planets, can be routinely performed with a spaceborne
UV telescope. Combining the two options, an orbital space-
craft and a telescope near the Earth as done by e.g., Mauk
et al. (2002), would show the similarity of auroral processes
on Jupiter and on our Earth. Under this heading we could
include also charge-transfer reactions with exospheric gases
on low-magnetic field objects (Moon, Mercury, and Mars)
and comets.
Springer

Astrophys Space Sci (2005) 303:103–122 111
Table 3 Physical and Angular
size of solar system bodies Object D (km) D (arcsec) Type of observation
Mercury 4,878 4.6–12.5 Map of specific regions
Venus 12,102 10–60 Map of specific regions
Moon 3,476 1.865 Map of specific regions
Mars 6,787 14–24 Map of specific regions
Largest asteroids Up to 960 Up to 0.8 Complete maps
Asteroids D >50 km 50 >0.05 From coarse to detailed maps
Jupiter 143,000 47 Maps of regions of interest
Io 3,640 1.2 Complete map
Ganymede 5,280 1.7 Complete map
Saturn +rings 121,000 19.5 Maps of regions of interest
Titan 5,150 0.8 Complete map
Tethys & Rhea 1,500–1060 0.17–0.25 Complete map
Uranus 52,400 4.0 Complete map
Oberon & Titania 1,575 0.12 Complete map
Neptune 50,000 2.4 Complete map
Triton 3,200 0.15 Complete map
Pluto 2,200 0.1 Complete map
Charon 1,200 0.055 Coarse map
Sedna 1,600 0.037 Coarse map
P/Encke 5 0.007 A few points
P/Halley 10 0.015 A few points
Hale-Bopp 50 0.075 Coarse map
Charged particles moving along magnetic field lines pro-
duce auroral phenomena (Jupiter, Saturn). The diagnostic H2
emissions (self-absorbed below 110 nm) can be used to mon-
itor this interaction all over the giant planets with special em-
phasis on auroral regions. The H2emission is very sensitive
to the height of the emissions, and thus may be used to deter-
mine the mean energy of the precipitated electrons, as was
illustrated with FUSE spectra (Gustin et al., 2004). However,
future instruments should have improved spatial resolution
so that specific areas of the aurora may be selected.
Equally interesting would be a detailed study of the flux
tube footprints in the auroral regions of Jupiter, first identi-
fied in HST STIS spectra. The study of the multiple heads
observed at the Io magnetic footprint, in particular their
relative intensities and inter-separations, and relationship
with the location of Io in its plasma torus, requires system-
atic FUV surveys of Io and its environment at all orbital
phases.
Io provides a unique opportunity to study how volcanism
is related to the ionosphere of the satellite and to its accom-
panying torus, while the ensemble is also dependent on the
highly dynamic and variable Jovian magnetosphere. The en-
vironment of Io and of the Io torus is very hot and many of
its emissions (in particular those of the multiply-ionized O,
S and C ions) are in the EUV and FUV regions. A detailed
study of such an intricate system requires synoptic obser-
vations over large times scales made at spatial resolutions
commensurate with the size of the sources and sinks of the
neutral and ionized species. In particular, measurements of
FUV emissions from the major ionic species in the torus
can provide images of the warm plasma in the main torus
region with sufficient spatial resolution to detect detailed
structures, yielding a map of plasma conditions dependent on
radius, longitude, latitude, and local time. Several unsolved
questions concerning the stability and variability of the torus
with longitude, local time and intrinsic temporal variability
Fig. 3 The three pictures show
the evolution of spatial
resolution in observing the
Jupiter aurora. The STIS
MAMA image (at the right) is
the only one showing the trail of
the Io magnetic footprint below
the auroral oval. Source: Gerard
et al. (2003
http://lpap.astro.ulg.ac.be/jupiter)
Springer

112 Astrophys Space Sci (2005) 303:103–122
Fig. 4 FUV spectrum of
Jupiter’s aurora obtained with
the FUSE large aperture. The
observed intensity distribution
of the H2lines is very well
reproduced by the synthetic
spectrum for a foreground H2
column of 7 ×1020 cm−2,
corresponding to a pressure
level of 1 microbar at 800 K
(from Gustin et al., 2004)
(possibly linked to volcanic activity) can be addressed with
multispectral FUV imaging.
While auroral and airglow emissions have been detected
by a few UV instruments in the past, none studied in de-
tail the variation of these phenomena as a function of local
time, magnetic longitude and solar activity. How these ef-
fects control the observed emissions is not yet understood.
The magnetospheric activity seems to correlate with the so-
lar activity. A dramatic example of UV imaging capability,
emphasizing the need for narrow-band filters and excellent
time resolution by using STIS on HST, was given by Waite et
al. (2001). The auroral oval light is comprised of the Lyman
and Werner bands of molecular hydrogen and the Lyαline
of atomic hydrogen. UV observations of the aurora offer a
much stronger contrast of the emission against the disk of
the planet than does the optical imaging.
It was long assumed that Saturn’s magnetosphere and au-
rora are intermediate between the case of the Earth, where
the dominant processes are solar wind driven, and the case
of Jupiter, where processes are driven by a large source of
internal plasma, the Io torus. Recent HST UV images show
instead that Saturn’s auroras differ notably from those of
the Earth and of Jupiter. Saturn’s auroral emissions are only
in partial corotation, the auroral oval quickly moves toward
higher latitudes in response to solar wind enhancements, and
it often exhibits an unexpected “spiral” structure.
The study of the production and evolution of neutrals in
the highly variable and neutrals-dominated magnetosphere
of Saturn is imperative. Saturn’s magnetosphere presents a
unique feature: it contains an abnormaly large abundance of
neutrals, in particular H and OH, both dissociation products
of water vapour (O has recently been detected by the Cassini
spectrometer). The source of OH is poorly known, although
it is probably from the sputtering of icy satellite surfaces
by energetic ions and interplanetary particles. Saturn will be
studied by the UV spectrometer of the Cassini mission, al-
though not on a regular basis because of mission constraints.
Paradoxically, one constraint is that the viewing direction is
not easily directed towards this large object; remote sensing
techniques are actually better suited for such observations
The magnetospheres of Uranus and Neptune are poorly
known; the role of their icy and rocky planetary cores on the
planet dynamos has not yet been investigated. Why Uranus
does not emit as much energy as Neptune is still mysteri-
ous. The two planets have very similar interiors but very
different rotation axis orientations, hence very different in-
teractions with the solar wind particles. Auroral phenomena
are essential to unravel the properties of the magnetic fields
of these two distant planets. Neptune and Uranus change
their magnetic configuration on time scales shorter than 24
hours because of the relative inclinations of their magnetic
and spin axes. Uranus also shows secular changes, because
of the high spin-to-orbit inclination. Thus, frequent observa-
tions of their auroral activity, achievable only by planetary
probes or by a space UV facility, are necessary to investigate
these interactions.
Studying the influence of satellites on planetary plasmas
and magnetic fields, such as the relation between Io and
Jupiter, Triton and Neptune, or between Saturn’s satellites
and its rings, requires the availability of UV imagers. Al-
though quite spectacular, the results that the UVIS spec-
trometer on board Cassini (Esposito et al., 2003) will deliver
would span only a short time period, in comparison to the
time constant of the phenomena to be investigated. Moreover,
Springer

Astrophys Space Sci (2005) 303:103–122 113
since this instrument resides constantly within the magneto-
sphere of Saturn, its data would be complementary to what
an Earth-based instrument would deliver.
Planetary surfaces and rings
The repartition of ices and condensable materials due to sea-
sonal effects on Mars, Ganymede, Io, Triton, Pluto, Charon,
and the larger TNOs can be studied through albedo maps.
The transport of condensable gases on the surface of plan-
etary bodies is controlled by seasonal effects. We are cur-
rently witnessing the displacement of the nitrogen frost from
the surface of Pluto, and the surface of Triton may possibly
be a template for an unchanged Pluto surface since its he-
liocentric distance hardly varies. In 10–20 years from now,
the surface of Pluto should be very different from what it
is today and could reveal underlying layers of yet unknown
composition. In addition, the UV is an excellent location to
look for organic absorptions on these surfaces.
The global characterization and long-term variation of the
properties of the rings of Jupiter, Saturn, Uranus and Nep-
tune has not been done (Cassini will do this only during its
four-year mission). A space UV facility could monitor the
evolution of ring structures coupled to satellite motions by
seeing these almost continuously. These observations could
determine the lifetime of ring systems and derive the for-
mation times of the existing rings. In low resolution mode,
images of Saturn’s rings could reveal the amount of absorp-
tion by water ice. Note that one may also image the rings
in reflection and in absorption at Lyαto attempt a detection
of H, a water dissociation product. As the majority of the
absorption features lie below 150 nm, and in order to acquire
a significant signal-to-noise result, it would be necessary to
observe the ring absorption during occultations of O, B and
A stars.
The rings can be imaged in the visible or in the NIR (e.g.,
the 2-µm absorption band) but this will not be done at 0.02”
resolution. Also, UV observations will help the search for mi-
nor species that are diagnostics of the chemical composition
Fig. 6 Saturn’s rings imaged by the Cassini probe from 3.7
million km with a resolution of 22 km per pixel. Such im-
ages could be routinely obtained by a large orbital UV
telescope. [Source: JPL PR at http://saturn.jpl.nasa.gov/
cgi-bin/gs2.cgi?path=../multimedia/images/large-moons/images/
PIA06142.jpg&type=image]
of the ring particles. The ring properties can be fairly well
represented by assuming that their constituent particles are
mostly water ice. However, the red color of the rings is an
indication that ice contaminants, probably refractory organ-
ics, are present. Since the ring particles are leftovers of the
disruption of what was once an icy satellite (or a number
of such satellites), a major objective for ring particle studies
is the determination of the nature and abundance of these
organics. Photometric studies could also shed some light on
the abundance of small inorganic particles, as well as on their
production and destruction timescales.
Comets
One of the key goals of cometary studies is to reveal the com-
position of the solid nucleus, the source of gases and dust
particles that scatter sunlight in the coma and tails. Comets
have spent most of their life at vast distances from the Sun,
Fig. 5 Polar projections of
individual FUV STIS images.
The South pole is at the center,
noon is downward, dawn to the
left and dusk to the right. The
left image shows a clear “spiral
structure,” while the image to
the right exhibits a bright spot,
tentatively identified as a cusp
signature observed at higher
latitude, probably due to
high-latitude reconnection.
(G´erard et al., 2004)
Springer

114 Astrophys Space Sci (2005) 303:103–122
essentially preserved in “deep freeze” in either the TN belt
or the Oort cloud until they enter the inner Solar System. If
the composition of a cometary nucleus can be determined,
then the temperature and density of the material from which
these remnants of the primordial solar nebula formed, would
be probed. Nuclei cannot be directly studied by telescopic
observation except at large heliocentric distances where they
are only little shielded or fully unshielded by the gas and dust
coma, yet at such distances they are very faint. It is through
spectroscopic observations of the coma and tail that key infor-
mation on the nuclei is obtained. In situ observations concern
only a few individual objects and need to be complemented
by observations of representatives of the entire population.
Molecules are expelled from the nucleus as this latter is
heated by the Sun. The ejected molecules are dissociated by
solar photons into molecular fragments and may become ion-
ized. The main nuclear component of comets, water, appears
in the coma as atomic H and O, and as the OH radical. All
three species have very strong transitions in the UV region.
The UV is the only domain where the three emissions can
simultaneously be detected and imaged at high spatial reso-
lution. Some of the primary molecules, such as CO and CO2,
can also be studied in the UV.
The abundant CO2molecule can be probed using its ion
and dissociation products, all with their strongest transitions
in the UV region. The CO 4th positive bands and the CO
Cameron bands are used for this purpose. The first emis-
sion is due to fluorescence of the CO molecule, while the
second is produced by prompt emission upon dissociation of
the CO2molecule or by electron impact onto a CO molecule.
UV observations allow one to separate the production mech-
anisms and distinguish between the emission parent species.
Since observations of the strong CO2band at 4.7 µm are
difficult, thus rare, observing the CO Cameron bands is the
only way to probe the elusive but cosmogonically important
CO2molecule. The statistical basis for a comparison of the
abundances of CO and CO2is currently insufficient to derive
any firm conclusion, but the existing evidence leans towards
a fairly large and stable abundance of the latter molecule in
cometary nuclei. Aging processes could be less efficient for
CO2because of the way the molecule is stored in nuclear
cometary ices.
Spatial information is crucial to disentangle the various
production and excitation mechanisms at play in the coma.
The best example is that of CO, which can be produced ei-
ther directly from the nucleus, from the dissociation of large
molecules ejected into the coma, or indirectly by desorption
from ejected grains from the surface of the nucleus. The na-
ture of the extended gas sources in the coma needs further
investigation, and requires simultaneous UV observations of
the CO and C I lines, as well as of the dust grains. High
sensitivity observations of the reflectivity of grains near 220
nm will indicate whether the interstellar absorption bump,
observed in the ISM at 217.5 nm, is present in Solar Sys-
tem grains, and if some interstellar grains have survived the
collapse of the pre-solar nebula indicating a possible role of
these grains as a source of the observed carbon. Radio and
NIR observations of CO do not have the sensitivity, or the
spatial and temporal resolution of UV observations, leading
to coarser results.
The UV region contains emissions of other short-lived
nuclear species, namely S2and CS2(via the CS emission).
These molecules have abundances relative to water that vary
greatly from comet to comet, especially S2. The storage of
these species in the nucleus, and the relation of their out-
gassing to that of the water and other ices, is unknown. High
spatial resolution UV observations are required to reveal their
onsets of production and their spatial distributions, hence
their production and storage mechanisms. Radio and NIR
observations complement the UV in the detection and moni-
toring of the global production of most parent species, but the
spatial resolution achievable in the UV is still unsurpassed in
comparison with that in other wavelength domains. The short
lifetime of CS2makes CS an interesting probe for the even-
tual anisotropic distribution of gas sources on the surface of
the nucleus.
The sensitivity of the instrumentation must allow one to
completely map the key atomic species H, O, C, N and S
emissions. Production processes and coma production rates
could then be investigated and the global atomic budget of
the volatiles in the nucleus obtained. The sensitivity must
be sufficient to detect these emissions beyond the Solar Sys-
tem “snow line” at ∼5 a.u. This would allow the detection
of yet unseen species, in particular N or N2and rare gases.
The presence of an H emission beyond the snow line would
directly give the conribution of the non-water H-atom carri-
ers. Since they have low sublimation temperatures, Ar and
Ne with transitions at 104.8 and 106.6 nm and at 63 nm re-
spectively can only be trapped in amorphous ice cages, from
which they should have escaped rather rapidly if the thermal
history of the comets is characterized by high temperature
episodes. EUVE and FUSE observations indicate strong de-
pletions of these species by factors of 10-25 but very deep
searches still remain to be conducted. Given the fact that
CO has about the same volatility as Ar, one may speculate
that the polarity of CO could explain its high abundance in
some comets by providing an electrostatic force to retain it
in low-temperature ices.
The EUV region contains numerous ion emissions that
allow the detailed study of the solar wind interaction with
the comet ionosphere. Observations with FUSE revealed the
presence of the important H2molecule, a direct tracer of
water. The Lyman series of atomic hydrogen lines can also
be observed. In addition, the H2(6,v”) bands can be seen
because of the coincidence of the P1line of the (6,0) H2
Lyman system with Lyα(Bowen fluorescence mechanism).
Springer

Astrophys Space Sci (2005) 303:103–122 115
Fig. 7 IUE UV spectrum of
comet Hyakutake
A similar phenomenon explains the strength of the O I 130.2
nm triplet: the solar Lyβline at 102.572 nm resonates with
theOI3D-3P transition at 102.576 nm. The Ne line at 63 nm
could also be detected, thanks to a coincidence with a strong
O V line.
The FUV region contains the important deuterium line
at 121.53 nm, very close to Lyα, that was first detected in
2004 in comet NEAT (Weaver et al., 2004). Such observa-
tions can be used to determine the D/H ratio in comets as a
function of dynamical type, thus allowing an investigation
of the source regions of comets in the solar nebula. Based
on our knowledge of the D/H ratio in the giant planets, and
if comets formed mostly by nebular gas recondensation, one
does not expect those formed in the Jupiter region to be as
deuterium-rich as those formed at greater heliocentric dis-
tances. So far, four comets have yielded a very similar value
that could indicate that the D/H ratio in comets is mostly
fixed by the solid component of the presolar nebula that has
probably not been fully evaporated. This hypothesis requires
a larger observation sample than presently available to be
confirmed.
The interaction of the solar wind with the coma will be
studied in detail through:
(1) Observations of the O+(83.4 nm), N+(108.5 nm), S+
(125.0 – 125.9 nm), C+(133.5 nm), CO+(200 – 230
nm), and CO+
2(289.0 – 289.6 nm) emissions,
(2) Emissions produced ahead of the comet by charge-
exchange reactions with highly charged solar wind ions.
These emissions are spread all over the EUV and FUV
regions (some are likely to be present in the FUSE spectra
of comets).
The O+and CO+emissions may be the best tracers of
the solar wind interaction with the comet. O+will fill out a
larger coma volume than CO+and will oppose a weaker force
to the solar wind than CO+, a mostly nuclear species with
large densities near the nucleus. The slowing down of the
solar wind particles will be completely revealed by imaging
in these lines. The coma response to a solar wind disturbance
will also be documented for the first time, and at high spatial
resolution. Note that the CO+lines are hard to separate in the
optical from weak C2and C3lines, and they appear on top of
a sometimes strong continuum. The continuum contribution
does not exist, or is extremely weak, in the UV domain and
this offers very clean data.
The continuum spectrum of comets below 300 nm has
rarely been studied. The spectrum seems to be solar, with
a slope of order 10% per 100 nm near 290 nm. Since the
albedo of solar system surfaces exibits large changes in the
280–400 nm range, which are caused by the physical as well
as the chemical nature of the surface materials, this part of
the spectrum should be observed with great care in the hope
of measuring a size/chemical nature-dependent change of the
albedo curve of the targets with wavelength.
Detailed observations of the UV spectrum of a periodic
comet are still to be made. One expects, due to a very different
thermal history, that the spectrum of periodic comets would
be severely depleted in some highly volatile species, e.g., CO.
This can be tested by detecting and monitoring the emissions
of CO and CO+
2, or those of C and O when the former are
not detected.
Other small bodies in the solar system
Objectives of a UV observational program for small Solar
System bodies would be the imaging of the surface, the pro-
duction of albedo maps, and the characterization of the re-
golith properties for all objects larger than 50 km (0.04 arc sec
at 2 a.u.). The determination of shapes and rotational proper-
ties could be inferred, from which the collisional history for
these objects would result automatically.
The characterization of the surfaces of asteroids and satel-
lites is traditionally done with optical and NIR information.
The UV could add a new dimension to planetary mineralogy,
but this would require a baseline laboratory program to char-
acterize the likely minerals in this spectral region. Lunar UV
mapping was done for the first time using UIT on the Astro
platform (Henry et al., 1995). The Astro results indicate a
lunar albedo of 0.038(±10%) at 170 nm and no opposition
Springer

116 Astrophys Space Sci (2005) 303:103–122
effect was found. The reflectivity of the lunar material below
400 nm was also discussed by Zou et al. (2004).
The requirement for surface characterization is adequate
surface resolution: 0.01 arcsec is 25 km at 3 a.u., necessary to
resolve approximately 1,000 Main Belt asteroids or coarsely
map TNOs at 60 a.u. with 500 km resolution. Such angular
resolution capabilities, or better, are also necessary in order to
resolve binary asteroids to derive their densities, in particular
for TNOs. Typical asteroid rotation periods are a few hours; in
order to avoid smearing the data, the necessary information
must be obtained within ∼10 minutes (surface brightness
constraint).
Using the UV to search for comet-like activity in small
bodies is beneficial, because in many cases activity mani-
fests itself as grain ejections and these will only scatter sun-
light. This diffuse scattered light would be seen in the UV
better than in the visible, because of better contrast against
the darker sky. The detection of activity from very distant
bodies would confirm the predictions of models of thermal
evolution at large heliocentric distances, mainly caused by
the transport of gases at low temperatures.
The study of transitional objects, dormant and/or ex-
tinct comets, is particularly fascinating. The asteroid 3200
Phaeton, the source of the Geminid meteors, is presumably
an extinct comet despite the failure to detect a coma (Hsieh
and Jewitt, 2005). A claim that asteroid 2001 YB5is the
source of a new meteor shower visible on January 7.5 was
recently published by Meng et al. (2004). As this object is
a Potentially Hazardous Asteroid (PHA), and such objects
have been linked theoretically to extinct comets, it is possible
that this is another case of a transitional object. Note how-
ever a dissenting opinion by Lupishko and Lupishko (2001).
The use of UV is beneficial because of the ability to image
low surface brightness features; therefore even low levels of
activity that would result in a very faint tail or a dusty coma
ejection could produce measurable signals.
A search and study of Trojans of all giant planets will
help determine whether it is possible that the Trojans were
accreted to their present locations at the early stages of solar
system formation. Peale (1993) showed that the Lagrangian
locations could be stable at the planet accretion phase; the
planetesimals could have been captured there by drag in
the primeval nebula. The question whether the Trojans are
comet-like bodies or not could be solved, as in the case of
extinct comets among the NEOs, by searching for comet-like
outgasing activity.
A search for multiple asteroid systems in order to deter-
mine their density, from which their interior structure could
be inferred, can be conducted better with a UV space tele-
scope than with an optical one because of the superior angular
resolution that can be achieved. Note that even ESO’s Very
Large Telescope Interferometer (VLTI) will not achieve a
resolution of one mas. With high angular resolution, from
a space platform that is not affected by seeing, it would be
possible to perform companion searches at much closer dis-
tances from the main body of a binary system than if using
visible light.
Outer solar system objects
After the discovery of 90377 Sedna (2003 VB12), it appears
that bodies in the inner part of the Oort cloud or scattered out
of the TNO belt are observable. Sedna never approaches the
Sun closer than ∼76 a.u., it was at about 90 a.u. at discovery,
and will become slightly better located for observations in the
next decades as it approaches perihelion. This is an opportu-
nity to characterize the nature of a body which has resided in
the far outskirts of the Solar System since its formation and
can show best the influence of eons of cosmic ray irradiation.
Other similar bodies, some possibly larger than Pluto (e.g.,
2003 UB313), exist at tens of a.u. heliocentric distances and
should also be included in such studies.
Observations of Sedna and its cousins imply that not
only ground-based optical observations from the largest tele-
scopes are required, but also the use of adequate space assets.
In particular, the use of low-resolution spectroscopy may be
preferable to photometry through a few filters to obtain a
global reflectivity profile. Sedna is a 20.5 mag object in the
R-band; assuming its UV albedo is only 10% of its visible
one, it would look like a 23st mag UV star. Broadband pho-
tometry in the UV, or very low resolution spectroscopy, is
possible for such sources using a 2-m class spaceborne tele-
scope. There are tens of other bright TNOs that could be
characterized in the FUV region and a new taxonomy for
these objects will emerge when all wavelength ranges will
have been observed.
Recent models predict cometary activity of TNOs at rather
large (tens of a.u.) heliocentric distances. This arises from the
influx of heat input at perihelion passage building up the CO
release from within the icy body, combined with transport
and recondensation of volatiles within the TNO. The ther-
mal evolution model by Choi et al. (2005) for the scattered
TNO 1999 TD10 showed an outburst at ∼60 a.u. after 75 or-
bits! Cometary-like activity has been reported for a number
of TNOs, with candidates for activity including the above-
mentioned scattered TNO (Choi et al., 2003) although this
claim was rejected by Mueller et al. (2004) and by Rous-
selot et al. (2003), as well as the earlier observation of 1996
TO66 (Hainaut et al., 2000). It is possible that cryovolcanism
activity could take place, whereby the ejection of a dust/ice
grains plume is observed due to pressure build-up in the icy
body. Imaging in the UV could facilitate the detection of such
cometary activity through the detection of faint dust features
because of the low sky background in the UV.
The size distribution of TNOs should be determined to es-
tablish to what extent the TN belt is the source of short-period
Springer

Astrophys Space Sci (2005) 303:103–122 117
comets. Even though faint objects like TNOs will probably
not yield a strong signal below ∼180–190 nm, and the ice ab-
sorption at 175 nm may not be seen, the advantage in angular
resolution of UV observations could allow the estimation of
size for some of the objects. The UV segment is necessary
also to investigate a possible signature near 220 nm from
possible traces of interstellar compounds that would serve
for surface characterization.
Interplanetary material
A long-term goal could be the complete characterization of
the properties of the interplanetary H and He, and the detec-
tion of temporal variations in the emissions/gas temperatures.
This, like zodiacal light studies, could be done if a large FOV
is available. Apart from a better characterization of the zodi-
acal light, a large FOV integral field spectrometer would also
allow the measurement of the UV spectrum of interplanetary
particles and its comparison to that of comet dust and aster-
oidal surfaces in order to determine the relative contributions
of the sources.
The influx of interstellar material into the Solar System
may be an elemental ingredient of planetary evolution as
the influx of interstellar material could modify the reflec-
tion properties of outer Solar System bodies. The Solar
System moves through the local ISM at ∼20 km/sec with
the strongest contribution from ∼5 a.u. upstream and with
a local cavity extending to ∼20 a.u. downstream. The so-
lar wind interacts with the ISM and reaches equilibrium at
∼200 a.u., at the heliosphere boundary. Dust particles, de-
tected by the Ulysses spacecraft, have increased by a factor
of three from 1997 to 2000. Landgraf et al. (2003) argue
that this could increase by another such factor by the end
of the next solar cycle, in 2012–3. It is possible to detect
very nearby (to the solar system) gas by studying at high
resolution the absorption lines of very nearby stars. Frisch
(2004) discusses the interaction of the heliosphere with the
very nearby ISM. The incoming ISM, detected as gas by
Frisch and others, is probably also related to the particu-
late influx detected by probes in the vicinity of the giant
planets.
The Tools
The proposed Solar System research in the UV relies on
several observational techniques. We list these below, be-
fore describing the necessary facilities. The techniques, to-
gether with the scientific goals of the different projects,
set the instrumental constraints of any proposed space
facility.
Photometric imaging
This technique yields the position, location and structure of
sources. We assume that any instrument designed for imag-
ing will have a photometric capability, good calibration, and
long-term stability. In order to be efficient in this task, it
should be done in narrow spectral bands centered on strong
emisson or absorption features. However, as a database for
planetary mineralogy in the ultraviolet range is not yet es-
tablished, this requires preliminary laboratory work.
Few UV images of solar systems objects not acquired
by specific planetary missions exist, mostly obtained during
rocket flights or by HST for target acquisition. This is quite
surprising, since such data are very powerful and offer a
global view of extended objects that helps characterize global
variations of the object appearance on a scale of order 1/100
the size of the object and on timescales that vary from minutes
to days. Each specific filter requires at least one adjacent
“continuum” filter for background subtraction.
Table 4 lists examples of spectral features that could be
observed during a program of high-resolution UV imaging.
The list is limited to the stronger features shortward of 315
nm, is by no means exhaustive, and no attempt was made to
prioritize the spectral bands.
Spectroscopy
In high resolution mode, R∼100,000 (0.001 nm resolution
at 100 nm) is required to resolve plasma velocities in a num-
ber of environments such as cometary comae. This has to
be achieved at a spatial resolution of 0.05 arcsec or better. In
low-resolution mode the emphasis is on detecting and charac-
terizing faint objects, and these figures become 2000 to 3,000
and 0.1 arc sec. An integral field spectrograph is highly de-
sirable as this would reduce the total exposure time. Another
solution is a spectrograph like STIS with scanning possibil-
ities.
The efficiency of classical spectrometers is notoriously
low, partly, because of the low efficiency of reflection grat-
ings. This has been mitigated by the use of high-efficiency
transmission gratings and grisms but it is possible that
Volume-Phase Holographic Gratings (VPHG) would offer
an additional 50% gain in efficiency. However, it will be
necessary to extend to the UV the present effort, which con-
centrates on visible and near-IR light.
(Photo)Polarimetry
Light scattered by dust and magnetized media is often po-
larized so studies of comet comae and auroral regions could
benefit from polarization measurements. For comets, and if
the instrument sensitivity is sufficiently high, one could ex-
pect to detect the continuum emission down to about 200 nm
Springer

118 Astrophys Space Sci (2005) 303:103–122
Table 4 UV Spectral features
observable in Solar System
targets
Wavelength Targets (non-
(nm) Species comprehensive list) Note
58.4 HeI Interplanetary material,
comets
Can be observed in 2nd order
63 Ne Comets, Saturn rings Resonates with OV
83.4 OII Comets, Saturn rings
98.9 OI Comets, Saturn rings
102.6 HI LyβInterplanetary material,
comets
102.576 OI Comets, Saturn rings Resonates with Lyβ
104.1 OI Comets, Saturn rings
108.5 NII Comets
108.8 CO (C-X) Comets
112–115 H2Lyman &Werner
bands
Giant planets, comets Aurora phenomena, satellite
footprints
113.5 NI Comets, Triton, Pluto
115 CO (B-X)+OI Comets Blend of features
121.53 DI LyαComets, IPM, giant planets
121.6 HI LyαInterplanetary material,
comets, Mars
The interplanetary H I signal can be
absorbed by various H2O clouds
125.0-125.9 SII Comets
130.2 OI Comets, Saturn rings
133.5 CII Comets
135.6 OI Comets
156.1 CI Comets
165 H2O1
st continuum
band
Comets, Saturn rings, icy
satellites
Absorption feature
165.7 C I Comets
175 Water ice TNOs Absorption feature
198.8 CO Comets
198-220 SO2absorptions Venus Surface/atmosphere interaction
200–230 CO+Comets
217.5 (wide) ISM absorption Comets, TNOs, zodiacal
dust
250 Continuum Comets, zodiacal dust Dust size distribution
255.4 O3Hartley Mars, Venus, Callisto?
257.6 CS Comets
289.0–289.6 CO+
2Comets, Venus, Mars
300.4 O3Hartley Mars, Venus, Callisto?
302 Continuum Comets Differential dust size distribution
308 OH Comets, Saturn rings H2O dissociation product
and thus collect information on the very small and irregular
particles that should polarize these wavelengths more than
larger particles.
Asteroid and other atmosphere-less bodies’ surfaces will
polarize the light they scatter. The polarization properties of
asteroid and comets should differ if they are due to particles
of different nature and shape. This would also be useful in
establishing whether the bodies have been heavily fractured
(“rubble piles”) or are solid.
Astrometry
This is the accurate measurement of the position and motion
of the source. We assume that any imaging device will have
the required electronic and mechanical stability to ensure
astrometric capability.
Occultations
Occultations offer a unique and efficient way of probing the
atmosphere of faint bodies such as Pluto, Charon, TNOs, and
comets. To be of use the telescope must offer (a) a large col-
lecting area to reach faint targets of occultations, and (b) high
observing efficiency to allow the utilization of rare observing
opportunities.
Absorptions in comets will mostly be due to water vapour
in the first continuum band centered at 165 nm. The IPM Lyα
emission could also be absorbed by the water vapour cloud
Springer

Astrophys Space Sci (2005) 303:103–122 119
surrounding the nucleus once the structure of the H I comet
emission has been subtracted. This can be used to map of the
water cloud, hence to provide information on its emission
pattern at the nucleus.
In spectrographic mode, the absorption of a hot star spec-
trum must be easily seen in the second continuum absorp-
tion of water vapor in bright comets within about 100 km
from the nucleus, if opacities of order 1/100 could be mea-
sured. Note that absorptions by coma species other than wa-
ter are probably much harder to detect, because their column
densities times the absorption cross section are significantly
smaller. In imaging mode, the opacity of the dust comae can
be probed through the observation of O-B-A stars occulted
by a comet, from which a column density profile independent
of any model assumptions can be derived.
The atmosphere of Pluto, discovered in 1985 by Brosch
(1985, 1995) when observing a stellar occultation by Pluto
and confirmed by other occultations observed from the
ground (e.g., Elliot et al., 2003), and the possible existence of
an atmosphere around Charon, could be probed during stel-
lar occultations observed in the UV from a space platform.
Mink (1993) found 26 possible occultations of stars brighter
than 16th mag by Pluto and 25 by Charon from 1993 to 2010
as seen from Earth. The observations of stellar occultations
by Pluto during the years following the first atmosphere de-
tection revealed the expansion of the atmosphere while the
planet is receeding from perihelion (Elliot et al., 2003). A
series of occultation measurements over the years will reveal
further secular changes.
The presence of atmospheres around TNOs can be tested
for in a similar manner once the number of TNOs with ad-
equate orbits increases and a search is conducted for occul-
tation opportunities. The use of the UV allows the tailoring
of the observations to specific absorption bands for definite
gases and, for thin atmospheres, allows a higher sensitivity
due to the higher refraction index of gases in the UV. On the
other hand and as already mentioned above, this requires the
identification of an UV-bright star that becomes occulted by
the target; the rareness of suitable candidates emphasizes the
need for large collecting optics. The large number of TNOs
compensates partly for the low occultation rate.
What is currently available?
At present and in the near future UV capabilities are very
limited. With the demise of STIS on HST there is now no
capability for UV spectroscopy, but some UV imaging is still
viable through WFPC2 and, for small angular extent targets,
also with the ACS. Note that UV imaging using CCD detec-
tors is fraught with red leaks of UV filters which prevent the
derivation of proper photometry. If an HST refurbishing mis-
sion is performed, and if that includes COS, some form of UV
spectroscopic capability will return. Similarly, “wide-field”
imaging would return with WFC3. Note that long uninter-
rupted HST observations will always be hampered by the
availability of a suitable target in the Continuous Visibility
Zone; otherwise typical 30 min observations are the rule for
this and other missions in Low Earth Orbits. With FUSE, the
capability exists for FUV spectroscopy from 90 to 119 nm but
for most planetary targets, is very limited by the brightness
of the objects.
GALEX offers the capability of low-resolution (∼7 arc-
sec) UV imaging and photometry in two UV bands from 135
to 300 nm, and of low-resolution R∼200 “objective prism”
mode spectroscopy in the same spectral region. The latter
is useful for obtaining the spectral energy distribution in the
UV for relatively bright objects, in cases when the objects are
at high galactic latitudes (to avoid confusion). Given the use
of non-integrating, time-tagged photon detectors in GALEX,
this mission has the possibility of electronically “tracking”
and following a target for as long as it stays in the field of
view, in order to build up the S/N or to provide a time-resolved
photometric light curve. Similar imaging capabilities, with
slightly different collecting apertures, plate scales, and spec-
tral stretch as defined by the filter sets and detector response,
will be offered by TAUVEX on GSAT-4 (launch planned
for 2006), and by ISRO’s ASTROSAT (launch planned for
2007-8), though in both cases the effective photon-collecting
efficiency will be similar to that of GALEX. This implies sen-
sitivity sufficient to observe a 12th mag (monochromatic)
object in a few seconds with S/N∼5.
UV instruments on planetary probes produce science only
when they encounter the planet. Although the information
collected by these probes is invaluable, the results are not
useful for synoptic studies because of the spotty coverage.
The same is true of various small instruments for the Shut-
tle or the ISS. Similarly, the small imagers on major satel-
lite platforms (e.g., the XMM Optical/UV Monitor) provide
a certain very restricted UV capability. The restrictions, in
these cases, come from pointing limitations that are dictated
by the main satellite instrument, from the low throughput due
to the relatively small aperture, and from the tailoring of the
filter complement. In most cases, these auxiliary instruments
have no spectroscopic capability.
Mission requirements
In this section we derive the requirements for two future
UV telescopes that could satisfy the planetary community’s
needs until the mid-21st century. The view is guided through
the perspective of the HST achievements and aims to provide
significant advantages over that mission.
The objectives, which need not be realized by a single
mission, would be to:
Springer

120 Astrophys Space Sci (2005) 303:103–122
rExplore the EUV region simultaneously with the FUV re-
gion.
rOperate with a smaller Sun-avoidance angle than HST.
rHave an optical system much faster than the HST, even for
a similar size instrument.
rProvide off-axis or Gregorian optics for efficient scattered
photon rejection.
rHave solar-blind detectors with high quantum efficiency.
rHave higher observational efficiency (90% or more of each
orbital revolution).
rAvoid contamination by the Earth geocoronal emissions
and from the interplanetaryHILyαbackground.
rHave near real-time response capability for targets of op-
portunity.
rProvide higher angular resolution than HST.
rOffer a large number of imaging filters.
rAllow simultaneous use of a number of filters.
rRegain UV capabilities lost with the HST and provide a
superior future facility.
rProtect detectors from bright FUV or UV targets.
Spatial resolution
Aside from the Moon, Venus, Jupiter and Saturn and comets,
most Solar System objects have small spatial extensions and
require very high spatial resolutions. One hundredth of an
arcsec is required to adequately resolve of order 1,000 aster-
oids and a few hundred TNOs, or to map planetary surfaces
and atmospheric phenomena, and can be achieved with a
modest-sized telescope (see table above). A second gener-
ation telescope could benefit from the lessons of HST and
of the first generation instrument, as well as from tech-
niques developed for telescopes operating in other spec-
tral bands, to yield the milli-arcsec resolution argued for
above.
Because of detector limitations, there must be a trade off
between spatial resolution and FOV. Large solar system ob-
jects such as cometary comae, the zodiacal light, etc., re-
quire large FOVs, although key scientific programs need
not generally cover the targets entirely. A medium (10 by
10 arcsec at sub-arcsec resolution) and a high-resolution
mode (1–2 by 1–2 arcsec at 0.01 arcsec resolution) should be
available.
Apart from the planetary probes, the Earth orbiters are
also generally limited by their ability to point accurately at
details of the planetary surfaces (or atmospheres) and by the
capability for tracking features on planets and satellite. To
track the Great Red Spot of Jupiter and keep it (or part of
it) in the spectrometer slit, for example, requires a tracking
accuracy of ∼1 mas/sec (1 km/sec at 5 a.u., assuming zonal
winds of ∼one km/sec). Any jitter in the tracking would
smear the spectral resolution, which will dilute the feature
response by that of its neighborhood, or will require closing
the tracking loop using imaging in the telescope itself (this
task is deemed to be very difficult).
Spectral resolution
This refers to the capabilities of an instrument operating with
any of the two telescopes considered here. The spectral res-
olution must be matched to the detector’s capabilities and to
the spectral stretch considered, unless one wishes to consider
an echelle configuration.
Pointing range, accuracy and stability
Targets near the Sun must be observable which imposes a
requirement for strong rejection of sunlight. Mercury and
comets close to perihelion must be accessible. The minimum
elongation angle is of order 15◦. If technically not feasible,
the elongation limit would be set by Venus (about 40◦) and
by low-q comets, in which case the limit must be as low as
possible since comets inside the orbit of Venus are usually ex-
tremely bright (this requires protection against damage from
bright targets).
Moving target capabilities should allow observations of
targets at a rate of at least one arcsec per second of time. Note
that the reflex motion of a body at the distance of Neptune is
∼four arcsec/hour.
Spacecraft and spacecraft orbit
A high apogee orbit is required to allow nearly continuous
operations and minimum obscuration by the Earth disc. To
avoid penetrating for too long deep into the Earth geocorona,
a highly elliptical or at least a geosynchronous orbit is pre-
ferrable. It should be possible to use the Earth’s shadow to
perform observations that require special protection from
sunlight. The second Lagrangian point L2location would be a
bonus, but solar panel constraints would restrict the availabil-
ity of any Solar System target, unless fully orientable solar
panels and good stray light suppression are implemented.
The radiation background at L2would be similar to that en-
countered in any high Earth orbit.
The above-mentioned requirements cannot be realized
in a single payload. To achieve an angular resolution of
one mas that would be diffraction limited at 100 nm
requires a 20-m aperture. It makes sense to combine this
with other cutting-edge technologies that are in design stages
now but are likely to mature within one or two decades. Such
technologies might include in-space coating of the mirrors
with aluminum for best-possible throughput. Such a mission
is our long term objective. However, to achieve the goals of
quick realization and continuity of UV astronomical efforts,
we must also have a mission that would provide a limited
Springer

Astrophys Space Sci (2005) 303:103–122 121
Table 5 UV missions for Solar
System exploration Property Two-meter aperture Twenty-meter aperture
Straylight rejection 10−11 10−13
Minimal elongation for observation 60◦20◦
Angular resolution@100 nm 10 mas 1 mas
5σimaging detection@100 nm 10−16 10−18
in one sec [ergs−1cm−2A−1]
Orbit Geosynchronous or L2L2or deep space
Number of imaging filters 5? 50?
FOV for imaging Few arcsec Few arcsec
Spectrometer Single-object, R∼50000 Single-object, R∼200000
Long slit, R∼1000 Imaging spectrometer,
R∼1000
Stability 0.1/sec 1 mas/sec
Moving target capability 1/sec 1/sec
enhancement with respect to the HST capabilities in the very
near future. A longer term goal might include a larger instru-
ment with enhanced capabilitites. Therefore, we conclude
with a proposal for two UV space observatories, as detailed
in Table 5.
A solar monitor
Phenomena observed in Solar System objects are either trig-
gered by sunlight and impinging solar particles (resonance
fluorescence, excitation of line emissions) or are the conse-
quence of the input of solar energy on them (thermal spec-
trum, evaporation or sublimation processes). In order to prop-
erly interpret the observations of various Solar System bod-
ies, one requires a good knowledge of the solar flux impinging
onto these bodies.
The solar spectrum produced in the EUV/FUV regions is
due to physical phenomena occuring mosty in the chromo-
sphere and corona of the sun, while the photospheric part
dominates beyond 200 nm. Below 150 nm, the solar spec-
trum is mostly composed of individual lines. The variablity
of the solar flux is wavelength-dependent and variabilities of
order 10% are observed in the 150–200 nm regions and may
reach 50% or so near 120 nm. In the EUV region, flux varia-
tions can reach one order of magnitude. These line flux vari-
ations translate into highly variable excitation rates of comet
transitions on timescales of hours to years. The 27-day and
11-year cycles are particularly important. The knowledge of
the solar spectrum at a resolution of order 0.1 nm is required
to correctly interpret observations of Solar System objects.
Because of the slow rotation of the Sun about its axis and,
often, significant differences in the ecliptic longitudes of the
Earth and the target, observations from the Earth are often
not sufficient. Latitudinal solar flux variations and solar line
shapes have to be modelled, as they are not easily measurable
without complex instrumentation.
In principle, it would be possible to achieve this by ob-
serving with a low-resolution spectrometer a point-like con-
stant object that reflects solar radiation, but the requirements
for a long duration mission, excellent instrumental stability,
and high sampling frequency would make this approach ex-
tremely difficult. It is possible to design such an instrument
around a small-aperture payload locked onto the Sun. The
Solar Radiation Monitor will be a basic tool for planetolog-
ical research. Because of possible azimuthal differences in
the solar emission that may influence the results obtained for
other planets and bodies in the system, it is necessary to de-
ploy 2–3 such instruments in orbit around the Sun. As most
Solar System objects are confined close to the ecliptic, it is
probably not necessary for monitoring purposes to have such
instruments in solar polar orbits.
Conclusion
We argued above the necessity of a two-stage approach in
assuring the continued access to the space ultraviolet for the
planetary science community. A first stage should be an in-
strument that would provide the community with better UV
capabilities than HST but would be fully dedicated to obser-
vations in this spectral segment. Because of this conservative
approach, we estimate that this goal could be achieved rel-
atively cheaply and could be implemented almost immedi-
ately. If this first step were to be adopted, a suitable instrument
could be functioning within five years.
A second stage must represent a breakthrough in all
the characteristics of a space telescope and should pro-
vide a UV capability commensurate in angular resolu-
tion with that of the cutting-edge instruments of the mid-
21st century: ELT and ALMA. By adopting a 20-m aper-
ture, the throughput with respect to the HST or with the
first generation instrument would be increased by two or-
Springer

122 Astrophys Space Sci (2005) 303:103–122
ders of magnitude. With proper optical construction and
platform design, such an instrument could observe Mer-
cury in the ultraviolet and follow Sun-grazing comets rel-
atively soon after their perihelion passage. The angular
resolution and sensitivity limit will allow the mapping of
Sedna and of other inner Oort cloud objects when these are
discovered.
We identified the need for the construction and deploy-
ment of solar monitors, to establish the baseline signal that
activates atmospheric phenomena in planets, satellites, and
comets. We also identified the need for laboratory miner-
alogic studies to provide the baseline information for UV
planetology.
Acknowledgements NB is grateful for continued support of the UV
astronomy efforts from the Ministry of Science and Technology of the
Israel Government, the Israel Space Agency, and the Austrian Friends of
Tel Aviv University. JCG is supported by the Belgian Fund for Scientific
Research (FNRS).
NUVA is supported by OPTICON, a project funded by the European
Commission under contract RII3-CT-2004-001566
References
A’Hearn, M.: Icarus 118, 223 (1995)
Bertaux, J.L., Blamont, J.E., Festou, M.: A&A 25, 415 (1973)
Biesecker, D.A., Lamy, P., St. Cyr, O.C., Llebaria, A., Howard, R.A.:
Icarus 157, 323 (2002)
Blamont, J.E., Festou, M.: Icarus 23, 538 (1974)
Brosch, N.: IAUC 4117 (1985)
Brosch, N.: MNRAS 276, 571 (1995)
Bunce, E.J., Cowley, S.W.H.: P&SS 49, 261 (2001)
Choi, Y.-J., Brosch, N., Prialnik, D.: Icarus 165, 101 (2003)
Choi, Y.-J., Prialnik, D., Brosch, N.: DPS 37, 56.04 (2005)
Clarke, J.T. et al.: Nature 415, 997 (2003)
Davies, J.K. et al.: Nature 309, 315 (1984)
Davies, J.K. et al.: Icarus 127, 251 (1997)
Elliot, J.L. et al.: Nature 424, 165 (2003)
Esposito, L. et al.: Space Science Reviews, in press (2004)
Feldman, P.D., Weaver, H.A., Burgh, E.B.: ApJ 576, 91 (2002)
Frisch, P.C. Adv. Space Res. 34(1), 20 (2004)
G´erard, J.-C. et al.: (http://lpap.astro.ulg.ac.be/jupiter) (2003)
G´erard, J.-C. et al.: J. Geophys. Res. 109(A9), 207 (2004)
Grodent, D. et al.: J. Geophys. Res. 106(A7) (2001)
Grodent, D. et al.: J. Geophys. Res. 108(A7), 1366 (2003)
Gustin, J. et al.: Icarus 171, 330 (2004)
Hahn, J.M. et al.: Icarus 158, 360 (2002)
Hainaut, O.R. et al.: A&A 356, 1076 (2000)
Henry, R., Murthy, J.: AFRL report from MSX (1998)
Henry, R. et al.: ApJL 454, L69 (1995)
Huebner, W.F., Benkhoff, J.: Space Sci. Rev. 90, 117 (1999)
Hsieh, H., Jewitt, D.: ApJ 624, 1093 (2005)
Keller, H.U.: A&A 23, 269 (1973a)
Keller, H.U.: A&A 27, 51 (1973b)
Keller, H.U., Lillie, C.F.: A&A 34, 187 (1974)
Krasnopolsky, V.A.: Icarus 153, 277 (2001)
Krasnopolsky, V.A., Feldman, W.C.: Science 294, 1914 (2001)
Krasnopolsky, V.A., Mumma, M.J.: ApJ 549, 629 (2001)
Krasnopolsky, V.A. et al.: Science 280, 1576 (1998)
Krasnopolsky, V.A.: Sixth International Conference on Mars (2003)
Landgraf, M. et al.: J. Geophys. Res. 108(A10), LIS5 (2003)
Leinert, Ch. et al.: A&AS 127, 1 (1998)
Lupishko, D.F., Lupishko, T.A.: Solar System Res. 35, 227 (2001)
Makinen, J.T.T. et al.: A&A 368, 292 (2001)
Mauk, B.H. et al.: Nature 415, 1003 (2002)
Meng, H. et al.: Icarus 169, 385 (2004)
Mink, D.J.: At the Pluto/Charon conference in Flagstaff (1993)
Mueller, B.T.A. et al.: Icarus 171, 506 (2004)
O’Connell, R.C.: AJ 94, 876 (1987)
Peale, S.J.: Icarus 106, 308 (1993)
Reynolds, R.O., Madsen, G.J., Moseley, S.H. et al.: ApJ 612, 1206
(2004)
Rousselot, P. et al.: A&A 407, 1139 (2003)
Stern, A. et al.: Icarus 92, 332 (1991)
Sykes, M.V. et al.: Science 232, 115 (1986)
Sykes, M.V., Walker, R.G.: Icarus 95, 180 (1992)
Toller, G.N., Weinberg, J.L.: In: Giese and Lamy (eds.), Properties and
Interactions of Interplanetary Dust, Reidel, p. 21 (1985)
Vidal-Madjar, A. et al.: Nature 422, 143; (2003); ApJ 604, L69 (2004)
Young, E.F. et al.: AJ 121, 552 (2001)
Yung, C.: Science 280, 1545 (1998)
Zou, Y.-L. et al.: Chin. J. Astron. Astrophys. 4(1), 97 (2004)
Waite, J.H. et al.: Nature 410, 987 (2001)
Weaver, H.A., A’Hearn, M.F., Arpigny, C., Combi, M.R., Feldman,
P.D., Festou, M.C., Tozzi, G.-P.: 36th DPS, paper 23.01, November
(2004)
Whipple, F.L.: ApJ 111, 375 (1950)
Springer

Astrophys Space Sci (2006) 303:123–132
DOI 10.1007/s10509-005-9019-2
ORIGINAL ARTICLE
Active Galaxies in the UV
Wolfram Kollatschny ·Wang Ting-Gui
Received: 4 October 2005 / Accepted: 10 October 2005
C
Springer Science +Business Media B.V. 2006
Abstract In this article we present different aspects of AGN
studies demonstrating the importance of the UV spectral
range. Most important diagnostic lines for studying the gen-
eral physical conditions as well as the metalicities in the
central broad line region in AGN are emitted in the UV. The
UV/FUV continuum in AGN excites not only the emission
lines in the immediate surrounding but it is responsible for
the ionization of the intergalactic medium in the early stages
of the universe. Variability studies of the emission line pro-
files of AGN in the UV give us information on the structure
and kinematics of the immediate surrounding of the central
supermassive black hole as well as on its mass itself.
Keywords Ultraviolet: galaxies .Galaxies: active .
Galaxies: seyfert .Quasars .Emission lines .Quasars .
Absorption lines
1. Introduction
Active Galactic Nuclei (AGN) are the most luminous ob-
jects in the universe. Their luminosities, their spectral en-
ergy distribution from the radio to the γ-ray range, as well
as their emission line ratios cannot be generated by normal
stars. Galaxies containing an active nucleus are called active
galaxies. We divide the AGN in different subclasses such as
Quasars, Seyfert galaxies and Liners.
W. Kollatschny ()
Institut f¨ur Astrophysik, Universit¨at G ¨ottingen,
Friedrich-Hund-Platz 1, D-37077 G¨ottingen, Germany
e-mail: wkollat@astro.physik.uni-goettingen.de
W. Ting-Gui
Center for Astrophysics, University of Science and Technology of
China, Hefei, 230026, China
Many aspects of the generation of the energy in AGN are
still unknown. Accretion of gas onto a central supermassive
black hole (SMBH) is generally accepted to be the dominant
physical process generating the enormous energies we are
observing (Rees, 1984). The accretion flow is the source of
the non-thermal continuum emission in the UV, X-ray and
optical. The spectral energy distribution (SED) of the non-
thermal continuum emission in typical AGN has its maxi-
mum in the UV.
The central continuum source ionizes the circumnuclear
gas in the so called broad line region (BLR) and narrow
line region (NLR). The majority of the most important
emission lines are emitted in the UV spectral range. The
overall continuum distribution as well as the UV spectral
lines (narrow emission lines, broad emission lines, absorp-
tion lines) are tracers of the physical conditions of those
regions where these emission lines originate. The emis-
sion line region of the narrow lines is spatially resolved in
some nearby objects. They originate at distances of pc to
kpc from the central ionizing source. However, the broad
emission lines originate at distances of light days to light
months only from the central ionizing source. This BLR
is unresolved by orders of magnitudes even for the nearest
AGN.
Various excellent reviews about AGN have been pub-
lished over the past years. Different aspects of AGN spectra
were highlighted in those papers as (e.g., Netzer, 1990; Urry
and Padovani, 1995; Koratkar and Blaes, 1999; Hamann and
Ferland, 1999; Veron and Veron, 2000; Ho, 2004; Heckman,
2004; Peterson et al., 2004).
This article is devoted to the UV spectral range of AGN.
The UV spectral range is important for our understanding of
active galaxies because:
– The maximum flux of AGN is emitted in the UV.
Springer

124 Astrophys Space Sci (2006) 303:123–132
– The rest frame EUV continuum in highly redshifted
AGN is important for our understanding of the early
universe.
– The UV spectra of the class of low luminous AGN can
only be observed in the local universe because of their
faintness.
– The most important diagnostic emission and absorption
lines are emitted in the UV: they give information on the
physical conditions in the emission line region next to the
central ionizing source.
– For the study of the cosmological and chemical evolution
of AGN the UV spectra of ’nearby’ objects (Z=0–2) have
to be known.
– Important far UV diagnostic lines can only be observed in
the UV – even for high redshift objects.
– Variations of the emission lines give us information on the
structure and kinematics of the innermost AGN regions.
The most important lines next to the central black hole are
emitted in the UV/FUV.
2. The AGN Family
2.1. Seyfert galaxies and quasars
Quasars are the most luminous subclass of the AGN family
having nuclear magnitudes of MB<−21.5. Seyfert galaxies
are by definition those AGN with MB>−21.5. Besides a
strong non-thermal continuum their spectra are dominated by
broad permitted emission lines in the UV and optical. Typical
observed line widths (full width at half maximum (FWHM))
are 3000–6000 km s−1with maxima of up to 30,000 kms−1.
The line widths are interpreted as Doppler motion of the
BLR clouds where these lines are emitted. The non-thermal
ionizing source in AGN is surrounded by the central BLR
clouds at distances of less than 1 pc (1015 to about 1017 cm).
Typical electron densities in these emission line regions are
ne=109–1011 cm−3for temperatures of about T ∼20.000 K.
Most of the important diagnostic lines of this BLR are emitted
in the UV spectral range – except for the optical Balmer and
a few Helium lines.
In the spectra of Seyfert 2 galaxies only narrow (permit-
ted and forbidden) emission lines with typical line widths
(FWHM) of 300–500 km s−1are present in contrast to those
of Seyfert 1 galaxies and quasars. These narrow emission
lines originate at distances of about 100 to 1000 pc from the
center. Electron densities in the range from 102to 104cm−3
are derived for typical electron temperatures of 10.000–
25.000 K. Even if many of the narrow emission lines are
emitted in the optical wavelength range too – the most im-
portant ones are emitted in the UV.
2.2. Low luminosity AGN
Low Luminosity AGN (LLAGN) refers to those objects with
Hαluminosities less than 1039 erg s−1. They are the most
abundant type of AGN and reside in 40% of bright galaxies
in the local universe (Heckman, 1980; Ho et al., 1997). There
are evidence that LLAGN may consist of two different sub-
classes. The first subclass is accretion onto small black holes,
i.e., a scaled version of Seyfert galaxies (Filippenko and Ho,
2003; Barth et al., 2004). In the second subclass, it is the
very low accretion rate that leads to low nuclear luminosity
but otherwise with black holes of similar masses to those in
quasars and Seyfert galaxies (e.g., Di Matteo et al., 2003).
Both classes of objects have attracted much attentions in the
past decade because of their role in the history of black hole
growth in the universe and the accretion physics. The black
hole-host galaxy connection in the low mass end of black
hole, which is likely in their infants, is crucial to the origin of
such relations in the massive quiescent and active galaxies,
which were found in the last five years (e.g., Magorrian et al.,
1998; Gebhardt et al., 2000; Ferrarese et al., 2001), and clues
to the formation of seeded BH in the early universe. The state
of very lower accretion rate is the end point of the AGN evo-
lution and provides the test-bed for accretion process at very
low rate, which is in a very different form from those seen
in Seyfert galaxies and quasars. Very low radiative efficiency
and the lack of big blue bump is the major prediction of theo-
retical models for the latter type (e.g., Narayan et al., 1998).
Thus the ultraviolet observation is critical to discriminate the
two possibilities.
Owing to the weakness of the active nuclei, stellar light
usually dominates the continuum emission in the optical
band even at the resolution of Hubble Space Telescope. As
stellar spectrum drops rapidly towards ultraviolet in most
of LLAGN, UV observation is one of the most important
spectral regimes for exploring the continuum properties of
those objects. Reverberation mapping of broad line region
described in the next section can only be carried out in ul-
traviolet for this type of AGN since one has to measure pre-
cisely small variations in the continuum flux. In addition,
these AGN are so faint, only nearby objects can be studied
in detail. However, they are much less studied in the UV
than other type of AGN due to their intrinsic faintness (Maoz
et al., 1999).
The majority of these sources show characteristics of
Low Ionization Nuclear Emission Line Region (LINER),
which can be produced either through photo-ionization of
the AGN/young stellar clusters or shock process (Heckman,
1980). Some key issues that might be solved with future
UV observations include: (1) How much fraction of LIN-
ERs are powered by nuclear activity, how much by star
forming process and what is the role of shocks? Measur-
ing high excitation lines (such as CIII/CII) in UV is critical
Springer

Astrophys Space Sci (2006) 303:123–132 125
to distinguish photo-ionization process by the central con-
tinuum from opaque shock ionization models (Dopita and
Sutherland, 1996). The measurement of UV absorption lines
of young stellar component or the featureless AGN contin-
uum will allow to determine the contribution of the ionizing
source, directly. (2) What is the UV continuum spectrum of
these active nuclei, which is closely related to the truncate
radius of the geometrically thin and optically thick part of the
disk and coupling between electron and proton in the case
of low rate accretion onto large mass BH (Quataert et al.,
1999), or to the global energy output in the accretion onto
low mass AGN. (3) How does the BLR structure of LLAGN
fit into the whole picture of AGN? There is indirect evidence
that the size of BLR in LLAGN deviates systematically from
the relation extrapolated from the known one for Quasars
and Seyfert galaxies (Wang and Zhang 2003). But a direct
measurement of the size of BLR by reverberation mapping
is required.
3. Spectral Energy Distribution and Rest Frame
EUV Continuum in AGN
The mean broadband continuum spectral energy distribution
(SED) for radio-quiet and radio-loud AGN is shown in Fig.
1. The flux scale has been normalized at 1 µm. The AGN
continuum flux is relatively flat from the radio to the X-ray
range. The bulk of this flux is thought to arise from syn-
chrotron emission. Besides a bump in the infrared due to
thermal dust reemission the overall continuum flux peaks
additionally in-between the optical and soft X-ray spectral
range in the UV. This spectral feature is sometimes called
the big blue bump. More than half of the bolometric lumi-
nosity of an (un-obscured) AGN is emitted in this big blue
bump. The big blue bump is thought to arise from an accre-
tion disk surrounding the central black hole. Gravitational
energy from the central accretion flow is converted into the
observed UV radiation of the disk. The thermal emission in
the UV corresponds to typical temperatures of 105K (e.g.,
Koratkar and Blaes, 1999).
The study of the UV/EUV spectral range is very difficult
because of the absorption caused by our own galaxy, the in-
trinsic absorption in distant galaxies, as well as the absorption
in the intergalactic medium. Figure 2a shows the UV com-
posite spectrum derived from more than 2000 AGN spectra.
Before combining the spectra (Telfer et al., 2002) corrected
them for internal and external extinction as good as possible.
The dotted line shows accretion disk models of (Mathews
and Ferland, 1987). The dashed line corresponds to simple
power law models with a thermal cutoff corresponding to a
temperature of 5.4×105K. Figure 2b again shows a com-
posite optical–soft X-ray spectrum for radio-loud and radio-
quiet quasars. One can see the flux is peaking not as extreme
as model calculations of accretion disk models predict (e.g.,
Laor et al., 1997). The accretion disk models cannot repro-
duce in a simple way the observed spectral shape. There are
indications in the observed composite AGN spectra that the
spectral index brakes at ∼1000 ˚
A. Observational difficulties
are caused by dust obscuration and the contamination of the
host galaxy. Furthermore, the composite spectrum has been
derived from different classes of AGN. Far more observations
in the UV of all classes of AGN are needed to understand the
details of accretion disks surrounding the central black hole
in AGN.
The knowledge of the UV/FUV spectral shape of quasars
is of outmost importance for our understanding of the evo-
lution of the early universe. The UV continuum of quasars
ionizes the intergalactic medium at the end of the dark ages.
At z≥6 the neutral hydrogen has been re-ionized by the
ionizing radiation of quasars at very early stages of the uni-
verse. The epoch of the ionization of HeI and HeII is even
less clear. Figure 3 shows a spectrum of the high redshift
Fig. 1 Schematic representation
of the mean spectral energy
distributions (SED) for a sample
of radio-quiet (solid lines) and
radio-loud (dashed lines) QSOs
(from Elvis et al., 1994)
Springer

126 Astrophys Space Sci (2006) 303:123–132
Fig. 2 Left: Composite
optical–soft X-ray spectrum for
the RQQs and RLQs in our
sample (thick solid line). Three
X-ray–weak quasars, and PG
1114+445, which is affected by
a warm absorber, were excluded
from the composite. Right:
Observed energy distribution of
some quasars vs. two accretion
disk model spectra (Laor et al.,
1997)
Fig. 3 De-noised,
full-resolution spectrum of
SDSS J1030+0524 with
matched templates from the
LBQS and Telfer et al. (2002).
The template is a very good
match to the quasar redward of
the LyαIGM absorption (White
et al., 2003)
quasar SDSSJ1030 +0524 (z=6.28) with the UV spectral
template of Telfer et al. (2002). The Gunn-Peterson absorp-
tion troughs show no emission over a redshift interval of 0.2
starting at z =6.
4. UV Emission Line Diagnostics
A UV spectrum of the Seyfert 1 galaxy NGC 4151 is shown
in Fig. 4. Some emission lines as well as some absorption
features are indicated in Fig. 4. The spectrum has been taken
with the Hopkins Ultraviolet Telescope (HUT) (Kriss et al.,
1992). The most important AGN diagnostic lines between
950 and 2000 ˚
A are: CIII 977, NIII 991, Lyβ+OVI 1034,
Lyα, NV 1240, OI 1303, CII 1335, SiIV+OIV] 1394,1402,
NIV] 1486, CIV 1549, HeII 1640, OIII] 1663, NIII] 1750,
and CIII] 1909. These emission lines show a wide range of
ionization states. They originate at different distances from
the central ionizing source in clouds with densities from ne=
108–1012 cm−3.
Photoionization calculations predict line flux ratios we
can compare with the observations. Figure 5 shows a series
of calculations of emission line ratios for different slopes
of the ionizing continuum flux (from Hamann and Ferland
(1999)).
4.1. Metalicities
The determination of the heavy element abundances in AGN
is one further aspect of quasar emission line studies. This is
connected with the investigation of the chemical evolution of
the universe as quasars can be observed at extreme distances
and therefore at very large look-back times. Surprisingly,
the broad line spectra of nearby AGN resemble those of the
most distant quasars. Furthermore, there are indications in the
spectra of some distant luminous quasars that their metalicity
abundances are very high even at z ≥5 (e.g., Ferland et al.,
1996).
In early investigations of AGN spectra the collision-
ally excited inter-combination lines NIII]λ1750, NIV]λ1486,
OIII]λ16664, CIII]λ1909 have been used to derive the abun-
dance ratios of the elements nitrogen, oxygen and calcium
(e.g., Shields, 1976; Davidson, 1977; Baldwin and Netzer,
1978). But these diagnostic lines are weak in most spectra.
Furthermore, the densities in the BLR (ne=109–1011 cm−3)
are near the critical densities of these lines. Therefore these
lines have different degrees of collisional suppression.
Springer

Astrophys Space Sci (2006) 303:123–132 127
Fig. 4 The ultraviolet spectrum
of the Seyfert galaxy NGC 4151
obtained by the Hopkins
Ultraviolet Telescope (HUT).
Emission line features and
absorption features are marked.
They are due to various
ionization states of different
elements in the hot gas present
in the nucleus of this active
galaxy (from Kriss et al., 1992)
Fig. 5 Predicted line flux ratios,
gas temperatures and
dimensionless equivalent widths
in Lyαplotted for clouds
photo-ionized by different
power-law spectra. (from
Hamann and Ferland, 1999)
Permitted lines might be better candidates for deriving
the element abundances in AGN. Detailed calculations have
been carried out (e.g., Hamann et al., 2002) proving the
sensitivity of the UV broad emission lines with respect to
the metalicity in AGN spectra. The most important diag-
nostic lines are NIIIλ991, NVλ1240, CIIIλ977, CVλ1550,
CIV+OIVλ1034, HeIIλ1640.
All these diagnostic lines are emitted in the UV. It is possi-
ble to derive the metalicities only for distant (z≥2) as well as
luminous quasars when the diagnostic lines are shifted into
the optical range. Very few is known about nearby and/or low
luminous AGN. But it is necessary to have this information
for deriving the chemical evolution of the universe.
A few very interesting AGN show clear indications of
abundance anomalies as e.g. Q0353-383 (Osmer and Smith,
1980). But they are rare and nothing is known about their
evolution and their number in the present day universe. Very
recently (Bentz et al., 2004) checked the Sloan Digital Sky
Survey for all nitrogen-rich quasars. They investigated more
than 6000 quasars with appropriate redshifts that the impor-
tant UV diagnostic lines were shifted into the optical range.
Only four candidates show very strong nitrogen emission
lines comparable to those in the spectrum of Q0353-383
(see Fig. 6). This means that only about one in 1700 dis-
tant quasars (z≥2) has extreme nitrogen over-abundances.
Further spectra of nearby and distant, as well as of bright and
low luminous AGN are needed to understand these galax-
ies within the overall AGN population. There is the basic
question whether the nitrogen enrichment is a short phase in
an AGN lifetime only or whether only a certain percentage
of quasars reaches extremely high metalicities. We need UV
spectra to detect high or even very high metalicities in present
day AGN to answer this question.
4.2. Far UV diagnostic lines
Very few is known about line strengths of diagnostic emis-
sion lines in the extreme ultraviolet spectral range between
300 and 900 ˚
A. Composite far UV spectra have been con-
structed from the spectra of highly redshifted QSOs taken
with the Hubble Space Telescope (HST) and the Far Ultravio-
let Spectroscopic Explorer (FUSE). They show the HeIIλ304
and HeIλ584 lines as well as the high ionization NeVIII +
OIV lines at 772 ˚
A and OIII at 831 ˚
A (Telfer et al., 2002,
Springer

128 Astrophys Space Sci (2006) 303:123–132
Fig. 6 Rest-frame spectra of (a) Q0353–383 and (b) the SDSS com-
posite, composed of 2204 quasar spectra. Both spectra are plotted in
semi-log format to enhance fine details (Bentz et al., 2004)
Scott et al., 2004) (see Fig. 7, Telfer et al., 2002). Consider-
ably more UV spectra of intermediate and high redshift AGN
are needed to compile far UV spectra with better S/N ratio
and to investigate the spectral details of different classes of
AGN.
The UV and EUV diagnostic lines are of outmost impor-
tance to understand the AGN phenomenon. Their properties
reflect the highest energetic areas next to the central black
holes in AGN.
4.3. UV absorption lines
Broad blue shifted resonant absorption lines in the ultra-
violet have been detected in 10–20% optically selected
quasars (Weymann et al., 1991), while narrow intrin-
sic absorption lines are much more common (∼40% of
Seyfert galaxies and ∼20–30% in quasars (Hamann and
Sabra, 2004, and references therein)). The predominance
of blue-shift among absorption lines suggests that partially
ionized gas outflows from the active nucleus. Recent X-ray
observations with moderate spectral resolution have found
similar blue-shifted absorption lines in the X-ray bands (e.g.,
Collinge et al., 2001). There are suggestion that the mass loss
rate and kinetic energy associated with the outflow may be
large and can have significant impact on the structure of disk
itself if it is disk-wind and on the ISM of the host galaxies. But
evidence for this is still ambiguous for following reasons. Be-
cause strong UV absorption lines may be severely saturated
and partially covering, the column density and ionization
state of major UV absorbing ions are poorly determined (e.g.,
Arav et al., 2001). Although the total absorption column den-
sity can be better determined from photo-electronic absorp-
tion in X-rays, very little information about velocity structure
of X-ray absorption line/edge can be obtained from the cur-
rent data. Resonant line absorptions in X-ray can be a very
Fig. 7 Overall mean composite
QSO spectrum in 1 ˚
A bins with
some prominent emission lines
marked. The dotted line shows
the best-fit broken power-law
continuum, excluding the region
below 500 ˚
A. The lines at the
bottom indicate the continuum
windows used in the fit (Telfer
et al., 2002)
Springer

Astrophys Space Sci (2006) 303:123–132 129
powerful diagnostics of properties of ions at different level
of ionizations, but spectral resolution comparable to those in
optical and UV band will not be available within next decade.
Therefore, it is necessary to observe the weak absorption lines
of the same elements that produce strong absorption lines in
order to derive both the covering factor and column density
as a function of velocity. Most these lines fall in the spectral
domain of far to extreme ultraviolet. Figures 8 and 9 shows
absorption lines in the UV spectral range of 3C191 taken with
Keck (Hamann et al., 2001) and of Mrk509 taken with FUSE
(Kriss et al., 2003). UV observations, simultaneously in soft
X-rays with future more sensitive X-ray missions, may im-
prove our understanding of the problem in several aspects: (1)
Simultaneous observations of UV and soft X-ray absorption
of low red-shift AGN would allow to determine the total col-
umn densities of material, especially those at the ionization
level similar to that of UV absorbing material, and ionization
states of the X-ray absorbing material (Wang et al., 2000).
At the same time we get velocity structure information from
UV absorption lines. This will permit a detailed modeling
of the physical state of outflows. (2) By studying absorp-
tion lines in UV bright z∼2 BAL QSOs, we will obtain the
kinematical properties of absorption lines of highly ionized
species at far UV. Comparison of those with low ionization
species will allow to study the changes in the kinematics
with ionization state, thus to bridge the gap between that
with X-ray absorbing material in these objects. Observing
bright z=2 BAL QSOs will also allow to better determine
Fig. 8 High resolution Keck
spectrum of 3C191 showing the
strong associated absorption
lines in the UV (Hamann et al.,
2001)
Fig. 9 FUSE spectrum of NGC
7469 in the Lyβ/O VI region
(thin black line) (Kriss et al.,
2003)
Springer

130 Astrophys Space Sci (2006) 303:123–132
the shape of the ionizing continuum, one uncertainty in the
modeling of the ionization structure of absorbing gas. (3)
Variations of intrinsic UV absorption lines can put strong
constraints on the density of the absorbing material, and thus
give an upper limit on the distance to the continuum source.
If these observations are carried out simultaneously in soft
X-rays for low-zAGN, one might distinguish the variations
caused by changes in the flow and ionization effect (e.g.,
Gebel et al., 2002). (4) Comparison of abundances derived
from absorption lines with those from emission lines will
give an independent check of those derived from emission
lines.
5. Structure and Kinematics of the Central Region
in AGN
The innermost line emitting region in AGN – the broad emis-
sion line region (BLR) – surrounds the central supermassive
black hole at distances of about 1015 to 1017cm. This corre-
sponds to radii of light days to light months. The motions
of the line emitting clouds give us information on the mass
of the central black hole (e.g., Kaspi et al., 2000; Peterson
et al., 2004). The broad-line region is spatially unresolved
even in the nearest AGN. But we can derive the structure and
kinematics with indirect methods by studying their line and
continuum variability (e.g., Kollatschny, 2003; Horne et al.,
2004).
5.1. Reverberation mapping
In a first step one has to correlate observed light-curves of
integrated broad emission line intensities with the ionizing
continuum light curve. It is of great advantage to observe the
ionizing flux in the UV since the optical continuum flux is far
more contaminated by the stellar continuum flux of the host
galaxy. Figure 10 shows the results of an optical/UV variabil-
ity campaign (including HST observations) of the prototype
Seyfert galaxy NGC5548 (Peterson and Wandel, 1999). Plot-
ted is the time lag of the emission lines with respect to contin-
uum variations as a function of their linewidth (FWHM) in the
rms profiles. The time lag corresponds to the mean distance
of the line emitting region from the central ionizing source.
One can see a clear trend: the higher ionized lines originate
closer to the central source. UV lines originate about ten times
closer to the center than optical emission lines. The most suc-
cessful monitoring campaign of the integrated UV lines of
an AGN has been carried out for NGC5548 so far (Clavel
et al., 1991; Korista et al., 1995). Variability campaigns of
e.g. 3C390.3 (O’Brien et al., 1998) and Akn 564 (Collier
et al., 2001) demonstrated the power of UV reverberation
studies but the S/N ratio and/or the fractional variability am-
Fig. 10 Time lags (cross-correlation function centroids τcent) in days
(1 lt-day =2.6×1015 cm) for various lines in NGC 5548 are plotted
as a function of the FWHM of the feature (in the rest frame of NGC
5548) in the rms spectrum. The filled circles refer to data from 1989,
and the open circles refer to data from 1993. The dotted line indicates
a fixed virial mass M=6.8×107M(Peterson and Wandel, 1999)
plitudes of the continuum variations were not strong enough
for detailed line profile studies.
Future monitoring campaigns of many galaxies includ-
ing the UV spectral range of the highly ionized OVI lines
(λλ1032,1038) e.g. will uncover the innermost broad line
region in AGN. The clear trend that higher ionized emis-
sion lines originate closer to the center has been seen
in optical variability campaigns of e.g. Mrk 110 too (see
Fig. 11) (Kollatschny, 2003). But the most important lines
for reverberation studies are: CVλ1550, SiIV+OIV]λ1400,
HeIIλ1640, NVλ1240, CIV+OIVλ1034 (see Fig. 10). These
lines give us information about the immediate surrounding
Fig. 11 The distance of the Balmer and Helium emitting line regions
from the central ionizing source in Mrk 110 as a function of the FWHM
in their rms line profiles. The dotted and dashed lines are the results
from model calculations for central masses of 0.8, 1.5, 1.8, 2.2, and 2.9
·107M(from bottom to top) (Kollatschny, 2003)
Springer

Astrophys Space Sci (2006) 303:123–132 131
of the central black hole one order of magnitude closer than
we can do it with optical lines.
The line profile variations of UV lines should be studied
in a second step. They gives us information on the kinemat-
ics in the broad line region. Detailed profile variations have
been studied in the optical lines of Mrk 110 (Kollatschny
and Bischoff, 2002; Kollatschny, 2003) only so far. Dif-
ferent delays of emission line segments (the velocity-delay
maps) measure the geometry and flow of the line emitting
gas when we compare observed two-dimensional velocity-
delay maps with model calculations (e.g., Welsh and Horne,
1991). Figure 12 shows the correlation of Hβand HeIIλ4686
line profile segments with continuum variations. The data are
from the variability campaign of Mrk 110 taken with the 10 m
Hobby Eberly Telescope at McDonald Observatory. Only
Keplerian disk BLR models can reproduce the observed fast
and symmetric response of the outer line wings. The Hβline
center originates at distances of 25 light-days while the HeII
line center originates at distances of 4 light-days only.
5.2. Central black hole mass in AGN
It is possible to calculate the central black hole mass in AGN.
One has to know the distances of the line emitting clouds
as well as the velocity dispersion of these clouds (e.g., Pe-
terson et al., 2004). We derived a central black hole mass
of 1.4∗108M◦in Mrk110. In that case we used additional
information about the projected angle of the accretion disk
where the broad emission lines originate (Kollatschny, 2003).
A Schwarzschild radius rsof 4 ×1013 cm corresponds to this
Fig. 12 The 2-D CCFs(τ,v)
show the correlation of the
Balmer and Helium line segment
light curves with the continuum
light curve as a function of
velocity and time delay (grey
scale) in Mrk110. Contours of
the correlation coefficient are
over-plotted at levels between
.800 and .925 (solid lines). The
dashed curves show computed
escape velocities for central
masses of 0.5, 1, 2 ×107M
(from bottom to top)
(Kollatschny and Bischoff,
2002; Kollatschny, 2003)
Springer

132 Astrophys Space Sci (2006) 303:123–132
Fig. 13 Schematic model of the innermost region in the Seyfert galaxy
Mrk110 derived from 2D-reverberation mapping (Kollatschny, 2003)
black hole mass. Figure 13 shows the inner broad line re-
gion structure of Mrk 110 derived from 2D-reverberation
mapping. The HeII line originates at a distance of 230
Schwarzschild radii only from the central black hole. The
monitoring of highly ionized UV lines in AGN enables us
to study the physics of the immediate environment of black
holes even more closer to the center. This helps us to de-
rive the central black hole mass more precisely. Finally, we
will achieve a clear progress in our knowledge of black
hole physics by monitoring different types of AGN in the
UV.
References
Arav, N., et al.: ApJ 561, 118 (2001)
Baldwin, J.A., Netzer, H.: ApJ 226, 1 (1978)
Barth, A.J., Ho, L.C., Rutledge, R.S., Sargent, W.L.C.: ApJ 607,90
(2004)
Bentz, M.C., Hall, P.B., Osmer, P.S.: AJ 128, 561 (2004)
Clavel, J. et al.: ApJ 366, 64 (1991)
Collinge, M.J., et al.: ApJ 557, 2 (2001)
Collier, S., Crenshaw, D.M., Peterson, B.M., Brandt W.N., et al.: ApJ
561, 146 (2001)
Davidson, K.: ApJ 218, 20 (1977)
Di Matteo, T., Allen, S.W., Fabian, A.C., Wilson, A., Young, A.J.: ApJ
582, 133 (2003)
Dopita, M.A., Sutherl, R.S.: ApJS 102 16 (1996)
Elvis, M., Wilkes, B.J., McDowell, J.C., Green, R.F. et al.: ApJS 95,1
(1994)
Ferland, G.J., Baldwin, J.A., Korista, K.T., Hamann, F., Carswell, R.F.,
Phillips, M., Wilkes, B., Williams, R.E.: ApJ 461, 683 (1996)
Ferrarese, L., Pogge, R.W., Peterson, B.M., Merritt, D., Wandel, A.,
Joseph, C.L.: ApJ 555, L79 (2001)
Filippenko, A.V., Ho, L.C.: ApJ 588, L13 (2003)
Gebhardt K., et al.: ApJ 539 (2000)
Gebel, J.R., et al.: ApJ 595, 120 (2003)
Hamann, F., Ferland, G.J.: ARAA 37, 487 (1999)
Hamann, F., Barlow, T.A., Chaffee, F.C., Foltz, C.B., Weymann, R.J.:
ApJ 550, 142 (2001)
Hamann, F., Korista, K.T., Ferland, G.J., Warner, C., Baldwin, J.: ApJ
564, 592 (2002)
Hamann, F., Sabra, B.: In: Richards, G.T., Hall, P.B. (eds.), ASP Conf
Series 311, San Francisco ASP, p. 203 (2004)
Heckman, T.M.:A&A,87, 142 (1980)
Heckman, T.M.: In: L.C. Ho (ed.), Coevolution of Black Holes and
Galaxies, from the Carnegie Observatories Centennial Symposia.
Published by Cambridge University Press, p. 359 (2004)
Ho, L.C., Filippenko, A.V., Sargent, W.L.W.: ApJ 487, 568 (1997)
Ho, L.: In L.C. Ho (ed), Coevolution of Black Holes and Galaxies,
from the Carnegie Observatories Centennial Symposia. Published
by Cambridge University Press, p. 292 (2004)
Horne, K., Peterson, B.M., Collier, S.M., Netzer, H.: PASP, 116, 465
(2004)
Kaspi, S., Smith, P.S., Netzer, H., Maoz, D., Jannuzi, B.T., Giveon, U.:
ApJ 533, 631 (2000)
Kollatschny, W., Bischoff, K.:A&A,386, L19 (2002)
Kollatschny, W.:A&A,407, 461 (2003)
Koratkar, A., Blaes, O.: PASP, 111, 1 (1999)
Korista K.T., et al.: ApJS 97, 285 (1995)
Kriss, G.A., Davidsen, A.F., Blair, W.P. et al.:A&A,392, 485
(1992)
Kriss, G.A., Blustin, A., Branduardi-Raymont, G., Green, R.F.,
Hutchings, J., Kaiser, M.E.:A&A403, 473 (2003)
Laor, A., Fiore, F., Elvis, M., Wilkes, B.J., McDowell, J.C.: ApJ 477,
93 (1997)
Magorrian, J., et al.: AJ, 115, 2285 L13 (1998)
Mathews, W.G., Ferland, G.J.: ApJ 323, 456 (1987)
Maoz, D., et al.: AJ, 116, 55 (1999)
Narayan, R., Mahadevan, R., Grindlay, J.E., Popham, R.G., Gammie,
C.: ApJ 492, 554 (1998)
Netzer, H.: In: T.J.-L., Courvoisier et al. (eds), Active galactic nuclei,
Saas-Fee advanced course 20, p. 57 (1990)
O’Brien, P.T. et al.: ApJ, 509, 163O (1998)
Osmer, P.S., Smith, M.G.: ApJS, 42, 333 (1980)
Peterson, B.M., Ferrarese, L., Gilbert, K.M., Kaspi, S. et al.: ApJ 613,
682 (2004)
Peterson, B.M., Wandel, A.: ApJ 521, L95 (1999)
Quataert, E., di Matteo, T., Narayan, R., Ho, L.C.: ApJ 525, L89
(1999)
Rees M.: ARA & A 22, 471 (1984)
Scott, J.E., Kriss, G.A., Brotherton, M., Green, R.F., Hutchings, J.,
Shull, J.M., Zheng, W.: ApJ 615, 135 (2004)
Shields, G.A.: ApJ 204, 330 (1976)
Telfer, R.C., Zheng, W., Kriss, G.A., Davidsen, A.F.: ApJ 565, 773
(2002)
Urry, C.M., Padovani, P.: PASP 107, 803 (1995)
Veron, M.P., Veron, P.: Astron. Astrophys Rev. 10, 81 (2000)
Wang, T.G., Brinkmann, W., Yuan, W., Wang, J.X., Zhou, Y.Y.: ApJ
545, 77 (2000)
Wang, T.G., Zhang, X.G.: MNRAS 340, 793 (2003)
Welsh, W.F., Horne, K.: ApJ 379, 586 (1991)
Weymann, R.J., Morris, S.L., Foltz, C.B., Hewett, P.C.: ApJ 373,23
(1991)
White, R.L., Helfand, D.J., Becker, R.H., Gregg, M.D., Postman, M.,
Lauer, T.R., Oegerle, W.: AJ 126, 706 (2003)
Springer

Astrophys Space Sci (2006) 303:133–145
DOI 10.1007/s10509-006-9057-4
ORIGINAL ARTICLE
Fundamental Problems in Astrophysics
Ana I. G´omez de Castro ·Willem Wamsteker ·
Martin Barstow ·Noah Brosch ·Norbert Kappelmann ·
Wolfram Kollatschny ·Domitilla de Martino ·
Isabella Pagano ·Alain Lecavelier des ´
Etangs ·
David Ehenreich ·Dieter Reimers ·
Rosa Gonz´alez Delgado ·Francisco Najarro ·
Jeff Linsky
Received: 21 February 2006 / Accepted: 14 March 2006
C
Springer Science +Business Media B.V. 2006
Abstract Progress of modern astrophysics requires the ac-
cess to the electromagnetic spectrum in the broadest energy
range. The Ultraviolet is a fundamental energy domain since
it is one of the most powerful tool to study plasmas at tem-
peratures in the 3,000–300,000 K range as well as electronic
transitions of the most abundant molecules in the Universe.
Moreover, the UV radiation field is a powerful astrochemical
and photoionizing agent.
The objective of this review is to describe the crucial issues
that require access to the UV range. A summary has been
added to the end with a more classic view of UV needs by
astronomical object type; this approach is followed at length
in the rest of the contributions of this issue.
A. I. G. de Castro ()
Instituto de Astronom´ıa y Geodesia (CSIC-UCM), Universidad
Complutense de Madrid, Madrid, E-28040, Spain
W. Wamsteker (†)
INTA-LAEFF, Apartado 50.727, E-28080 Madrid, Spain
M. Barstow
Dept of Physics and Astronomy, University of Leicester
University Road, Leicester LE1 7RH UK
N. Brosch
The Wise Observatory, Tel Aviv University, Tel Aviv 69978, Israel
N. Kappelmann
Institut f¨ur Astronomie und Astrophysik T¨ubingen (IAAT),
Universit T¨ubingen, Germany
W. Kollatschny
Institut f¨ur Astrophysik, Universit¨at G ¨ottingen,
Friedrich-Hund-Platz 1, D-37077 G¨ottingen, Germany
D. de Martino
INAF-Osservatorio Astronomico di Capodimonte Napoli, Via
Moiariello 16, I-80131, Italy
I. Pagano
INAF-Catania Astrophysical Observatory, via Santa Sofia 78,
95125 Catania, Italy
A. L. des ´
Etangs ·D. Ehenreich
Hamburger Sternwarte, Universitt Hamburg, Gojenbergsweg 112,
D-21029 Hamburg, Germany
D. Reimers
Institut d’Astrophysique de Paris,UMR7095 CNRS, Universit´e
Pierre & Marie Curie, 98bis boulevard Arago, F-75014 Paris,
France
R. G. Delgado
Instituto de Astrof´ısica de Andaluc´ıa (CSIC), Apdo. 3004, 18080
Granada, Spain
F. Najarro
Instituto de Astrof´ısica Molecular e Infrarroja, Instituto de
Estructura de la Materia, CSIC, Serrano 121, E-28006 Madrid
J. Linsky
JILA/University of Colorado and NIST/Boulder, CO 80309-0440
USA
Keywords UV astronomy
1. Introduction
Access to the UV range is fundamental for the progress of
astrophysics since UV spectroscopy is the most powerful
tool to study plasmas at temperatures in the 3,000-300,000 K
range. Also, the electronic transitions of the most abundant
molecules in the Universe (H2, CO, OH, CS, CO+
2,C
2...) are
in this range. Moreover, the UV radiation field is a powerful
astrochemical and photoionizing agent.
The impact of UV instruments in modern astronomy can
be clearly traced through the considerable success of the
Springer

134 Astrophys Space Sci (2006) 303:133–145
International Ultraviolet Explorer (IUE) observatory and suc-
cessor instruments such as the GHRS and STIS spectrographs
on-board the Hubble Space Telescope (HST), or the FUSE
satellite operating in the far UV (90–120 nm range). Of par-
ticular importance has been access to high resolution R
40,000–100,000 spectra providing an ability to study the dy-
namics of hot plasma and separate multiple galactic, stellar or
interstellar spectral lines. Furthermore, the GALEX satellite
is providing new exciting views of UV sources. As a result,
UV facilities are in high demand; observing time on HST
remains heavily oversubscribed (a factor ∼6 in 2004), but
its UV spectroscopic capabilities were hampered by STIS
closure. Far-UV observations with FUSE also take a large
share. This success has an interesting consequence: while
astrophysicists world-wide are used to have a observatory-
like access to the space telescopes working in this range, the
BIG funding required to create/maintain large space facili-
ties is driven by key scientific projects. The objective of this
review is to describe briefly the crucial problems of modern
astrophysics that require access to the UV range. A summary
has been added to the end with a more classic view of UV
needs by astronomical object type; this approach is followed
at length in the rest of the contributions of this issue.
2. Crucial problems in modern astrophysics that
require access to the UV range
Modern astrophysics is a mature science that has evolved
from its early phase of discovery and classification to a
physics-oriented discipline focussed in finding answers to
fundamental problems ranging from cosmology to the ori-
gin and diversity of life-sustainable systems in the Universe.
This evolution is not uniform; research in fields like compact
objects or cosmology is clearly at this stage but the detection
of extrasolar planets or the identification of the sources of
γ-rays bursts are still at early stages. This diversity can be
nicely traced in several recent collections of articles devoted
to the identification of “unsolved problems in astrophysics”
or to the “fundamental problems in astrophysics” (see, for
instance, Bahcall and Ostriker, 1997). Though a much wider
science case can be drawn, we have identified three key fields
in astrophysics that cannot progress without easy and wide-
spread access to modern UV instrumentation; these are:
1. Extrasolar planetary atmospheres and astrochemistry in
the presence of strong UV radiation fields.
2. Chemical evolution of the Universe and the diffuse bary-
onic content.
3. The physics of accretion and outflow: the astronomical
engines.
This list is by no means complete, but it certainly includes
the most exciting and active problems that the majority of the
astrophysical community would like to see solved. We detail
each of these below.
2.1. Extrasolar planetary atmospheres and
astrochemistry in young planetary disks
Since the mid 1990’s, more than one hundred extrasolar
planets (hereafter called “exoplanets”) have been discovered.
Since the unexpected discovery of the first hot-Jupiter extra-
solar planet by Mayor and Queloz (1995), it is clear that
exoplanets are an extremely diverse group. With the discov-
ery of more than one hundred exoplanets, this diversity is
clearly demonstrated by their orbital properties. We have
“hot-Jupiters” with orbital periods as short as 3 days, and
several “very hot-Jupiters” with orbital periods even shorter
than 2 days but also exoplanets with periods of months to
years. Less massive exoplanets have also recently been dis-
covered (Santos et al., 2004; McArthur et al. 2004; Butler
et al., 2004, Rivera et al., 2005), and the discussions on the
true exoplanet nature show that a large variety is certainly
possible.
The same variety is also expected for the atmospheres
of these exoplanets. A quick look at the atmospheric content
and history of the solar system’s inner planets shows that with
four terrestrial planets, we find four very different possibili-
ties: Mercury has almost no atmosphere, Mars’ atmosphere
is tenuous with atmospheric pressure at ground level about
one-hundredth that of the Earth, and Venus is the extreme
opposite with more than ninety times the atmospheric pres-
sure of the Earth with the same physical size of the planet.
Note that Titan, although much smaller than the Earth, also
has an atmosphere of 1.5 Bar and is very different from other
giant planet satellites lacking atmospheres.
This diversity shows how difficult it is to predict what
should be the content of an exoplanet’s atmosphere. In the
solar system, the terrestrial atmosphere is unique with abun-
dant O2and O3produced by biological activity though traces
of O3have also been detected on the Jupiter’s satellite Europa.
Another important characteristic of the terrestrial atmosphere
is the significant amount of water. The Earth and Titan have
both much N2in their atmospheres, but Titan contains more
methane and no O2. Mars and Venus have similar atmo-
spheric composition, but they differ in total amount by a
ratio of more than 104.
Thus, there is no simple answer to the question of the ex-
pected characteristics of planets and their atmospheres. The
solar system planets provide a first hint of the expected di-
versity of the exoplanets and their atmospheres. Observations
of exoplanets and the detailed characterization of their atmo-
spheres will help us understand better the physical processes
at work in the building of a planet and its atmosphere, and in
the further evolution of such a system.
Springer

Astrophys Space Sci (2006) 303:133–145 135
It is clear that the detailed processes that created the solar
system planets are still a matter of debate and the impact of
many processes must still be clarified. In short, we do not yet
know the key physical parameters that govern the formation,
evolution and fate of a given planet and its atmosphere.
How do properties such as effective temperature, stellar
type, high-energy particle environment, and metallicity of the
central star alter the evolution of its planetary system? What
effects do a planet’s orbital parameters (orbital distance and
eccentricity) have on its size, mass and potential migration
during the formation process? Are there volatile-rich planets
like the proposed “Ocean-planets”? (Kutchner, 2003; L´eger
et al., 2004) How do interactions with other planets and plan-
etesimals in their environment influence the evolution of a
planet? This last question is undoubtedly related to the origin
of water on the Earth. Are water-rich planets in the “habitable
zone” common, rare, or exceptions?
Several processes believed to play key roles in building a
planet can now be identified. To begin with, we can look at
the best known planet, our Earth. Although still controversial,
it is generally accepted that the Earth’s original atmosphere
was accumulated simultaneously with the planet’s forma-
tion. However, the heating of the atmosphere by the young
Sun’s UV and X-ray flux, and the pressure of the strong so-
lar wind at this period, led to the hydrodynamical escape
of this primary atmosphere (as observed on HD 209458b
in Lyαλ1216 ˚
A, OIλ1305 ˚
A and CIIλ1330 ˚
A, Vidal-Madjar
et al., 2003, 2004). Tectonic activity, volcanism and planet
out-gasing then formed the secondary atmosphere in which
we now live. Late bombardment by planetesimals in the
young planetary system contributed a large fraction of the
terrestrial water but the fraction of water originating from
the Earth itself vs. the external contribution is still a matter
of debate. Finally, photosynthetic plants enriched the atmo-
sphere in O2and ozone, which are poisons to the first proto-
life and are therefore considered as atmospheric bio-markers
for advanced life forms. The observation of O2and ozone in
the atmosphere of the Earth or of any exoplanet can lead to
the conclusion that something very particular is happening
there. This something could suggest the presence of life.
In the coming decade, several ground and space-based ob-
serving programs will lead to the discovery of an extremely
large number of exoplanets, in particular, near-future space
missions including Corot, Kepler or GAIA will discover large
numbers of exoplanets transiting their parent stars. To acquire
a revealing picture of these new worlds, we need to charac-
terize the planetary atmospheres of a large sample of these
exoplanets. The observation of UV and optical absorptions
occurring when an exoplanet transits its parent star are a
very powerful diagnostic technique; in fact, the most pow-
erful technique for detecting Earth-like life-bearing planets
because of the strong absorption of stellar UV photons by
the ozone molecule in the planetary atmosphere (see G´omez
de Castro et al., in this book). We cannot predict what will
be discovered, but this will be an unprecedented opportunity
to better understand the key processes at work in the shaping
of planets and, in particular, to better understand the origin
of our own Earth.
In addition, ultraviolet radiation plays a very important
role in the evolution of the primary atmospheres of plan-
etary embryos through photoionization and photochemical
reactions (Watson et al., 1981; Lecavelier des Etangs et al.,
2004). Thus, UV spectroscopy will allow the study of the in-
teractions between the stellar UV field with the atmospheres
and, as important, with the young planetary disks. Very re-
cent chemical models are showing that the penetration of UV
photons coming from the central engine in a dusty disk could
produce an important change in the chemical composition of
the gas allowing the growth of large organic molecules. In
this context, UV photons at λ>1500 ˚
A photodissociating
organic molecules could play a key role in the chemistry of
the inner regions of the proto-planetary disk, while those pho-
todissociating H2and CO would control the chemistry of the
external layers of the disk directly exposed to the radiation
from the star. The radiation field can produce a rich pho-
tochemistry on timescales shorter than the dynamical evolu-
tion time scales, leading to the formation of large carbon-rich
molecules such as CnH2,HC
(2n+1)N, and Cn. Reactions be-
tween these species and H and H2may maintain their high
abundances in spite of the strong radiation field emerging
from the central star (see e.g., Cernicharo, 2004).
2.2. Chemical evolution of the Universe
The gas and stars are the dominant baryonic components of
the Universe which can be understood in terms of a two-
fluids system interacting through gravitation, starbirth and
death; the massive stars life cycle controls the chemical en-
richment of the Universe. Key parameters in the evolution of
this system are the relative contributions to the energy and
chemical input from the various possible sources to the gas
phase (SNe, massive star winds and radiation fields, mass
infall from the halo, galactic fountains and gas ejection in
the intergalactic medium (IGM), galactic dynamics, cosmic
rays and magnetic fields); also the roles of magnetohydro-
dynamical (MHD) turbulence and shocks in the energy cas-
cade and structure formation need to be determined. During
the last few years, a very efficient feed-back loop has been
operating between radio observations and numerical simu-
lations to study the role of MHD turbulence in the energy
cascade within the densest regions of the galactic ISM (H I
and molecular clouds). A similar feed-back loop needs to
be established with UV observations to understand the heat-
ing/cooling processes and the overall thermal and dynamical
evolution of the two-fluids system, including the formation
Springer

136 Astrophys Space Sci (2006) 303:133–145
of molecular clouds and massive stars clusters (starbursts).
This loops needs to be established at two scales:
At galactic scale where the details of the physics of the pro-
cess can be tested. The dynamical evolution of the ISM
concentrates cold matter in dense shells and filaments in
the disk, while the halo acts as a pressure-release valve
for the hot (T>105.5K) phase, thereby controlling its
volume-filling factor. Here a large-scale fountain is set
up by hot ionized gas injected from either the gas stream-
ing out of the thick disk or directly from superbubbles
inflated in the disk underneath. The gas then escapes
in a turbulent convective flow enriching the halo with
warm-hot gas. The detection of O VI, C IV and Si IV
absorption in many High Velocity Clouds (HVCs) of our
Galaxy indicates that they have hot, collisionally-ionized
envelopes (Danly et al., 1992, Tripp et al., 2003). Under-
standing the ionization of such envelopes will constrain
the properties of the Galactic corona and the Local Group
medium. UV absorption lines are also the most sensitive
probes for determining the abundances (and hence their
Galactic or extragalactic origin) of the HVCs (see e.g.,
Richter et al., 2001). Note that the most robust specie
for constraining the metallicity of HVCs is O I, since
oxygen is only slightly depleted by dust grains (Moos
et al., 2002) and the ionization potential ofOIisvery
similar to that of H I. Thus, oxygen abundances based on
the O I/H I ratio, depend only slightly on the ionization
of the gas in substantially ionized plasmas.
At low redshifts (z ∼0.1–0.2) where it is feasible to re-
solve the starbursts and thus understand the violent star
formation processes in galaxies and the variation of the
Initial Mass Function (IMF) across the Universe. Be-
cause most of the massive stars form in starburst sites,
starburst galaxies play a significant impact on the cosmic
evolution of galaxies. Starbursts are responsible for the
thermal and kinetic heating of the interstellar medium,
and they are the factory where most of the heavy ele-
ments form. These elements are dispersed throughout the
interstellar medium when massive stars explode as su-
pernovae, and they can escape from the galaxy to the in-
tergalactic medium through high velocity outflows gen-
erated by the violent star formation events occuring in
these galaxies.
UV observations are relevant not only because this range
is very sensitive to the star formation history of galaxies,
but also because it contains valuable tracers of the cold
and warm phases of the ionized interstellar medium in
starbursts that allow us to investigate the physics of the
feedback and its consequences. Thus, high-spatial reso-
lution UV spectra and imaging of nearby starbursts are
crucial to further progress in understanding the violent
star formation processes in galaxies, the interaction be-
tween the stellar clusters and the interstellar medium, and
the variation of the IMF. High-spatial resolution spectra
are also needed to isolate the light from the center to
the disk in the UV luminous galaxies found by GALEX
at z=0.1−0.3. Observations at high spectral resolu-
tion (R≥10000) are required to isolate the galactic, the
stellar and the interstellar components of several ions to
perform a quantitative characterization of the outflows.
A significant increase in spectral sensitivity (≥10–100)
with respect to HST +STIS is required to character-
ize superstellar stellar clusters of 105–106Mbeyond
Virgo, nuclear starbursts at z=0.1–0.2, and to probe the
starburst galaxy environments out to tens of kpc using
background quasars.
In addition, it is fundamental to map the distribution
and metallicity of diffuse baryonic matter and radiation in
the Universe. Independent of the different proposed mod-
els of the early Universe, the major baryonic component
at z<3 must be associated with the InterGalactic Medium
(IGM). Recent studies suggest that the Warm-Hot Intera-
galactic Medium (WHIM) at low zcontains more baryonic
mass than stars and galaxies (Richter, 2005). These observa-
tions have been done in the UV with FUSE (the OVI triplet)
and HST/STIS (broad Lyαabsorption (BLA)) and imply
cosmological mass densities of b(OVI) 0.0021h−1
70 and
b(BLA) 0.0027h−1
70 (Sembach et al., 2004; Richter et
al., 2004). These results have tremendous implications for
our understanding of the intergalactic medium and galaxy
formation.
Further out, looking into the past, the HeIIλ304 ˚
A ef-
fect provides the most sensitive tool to detect and analyzed
the properties of the intergalactic medium. From theoretical
modeling of the IGM we know that after the HI reionization,
the IGM cools by expansion, is reheated by the delayed HeII
reionization at z=3, and continues to cool with decreas-
ing redshift. Observations of the HeIIλ304 ˚
A forest over the
redshift range 2.1<z<2.9 will test this model in the most
direct way. Besides observingthe evolution of the mean IGM
temperature, the characteristic scale of the density fluctua-
tions of the IGM and its relation to the fluctuationg ionizing
radiation field at a spatial resolution of less than 1 Mpc (co-
moving distance) will be observed (see Wamsteker et al., this
book).
Spectroscopic observations of the HeIIλ304 ˚
A forest with
HST and FUSE in two bright QSOs have shown that the HeII
reionization phase of the universe ends at roughly z=2.9
(Reimers et al., 1997), i.e., we observe a transition from
optically-thick HeIIλ304 ˚
A absorption (the Gunn-Peterson
trough) to a resolved HeII 304 ˚
A forest below z=2.8. While
FUSE was able to resolve the HeII 304 ˚
A forest in only two of
the brightest high redshift QSOs in the sky (HE2347-4342,
Kriss et al., 2001; HS1700 +6416, Reimers et al., 2004), the
Springer

Astrophys Space Sci (2006) 303:133–145 137
true potential of these fundamental observations could not be
exploited due to the very low S/N of the HeII FUSE spectra.
Future observations of the HeIIλ304 ˚
A forest at high spectral
resolution and better S/N have the potential to map both the
intergalactic matter and radiation field in much detail.
Due to the possibility of observing HI and HeII simulta-
neously, the redshift range 2.1<z<2.9 is the only cosmic
epoch where the evolution of the fluctuating IGM can be
compared in detail with predictions of theoretical models of
large-scale structure formation. Knowledge of the shape of
the ionizing UV background is also necessary for the deter-
mination of heavy element abundances in more than 90% of
the baryonic component. The reason is that from the few ions
observable from the ground (CIV, SiIV, OVI,...) the state of
ionization and therefore, the element abundances cannot be
determined quantitatively. Most of the relevant lines formed
in the highly ionized component are in the intrinsic EUV at
rest wavelengths between 300 and 900A (OIII-OV, NeIII-
NeVII, SIII-SVI,....). The combination of HeIIλ304 ˚
A for-
est observations with high resolution EUV metal-line spec-
tra and optical spectra of laboratory quality from 10m-class
ground-based telescopes in a few strategic objects, such as
HS 1700 +6416 with its rich metal line spectrum (Reimers
et al., 1992), will lead to a more quantitative understanding
of the evolution of matter composition, radiation field and
structure formation in the strategic redshift regime between
2 and 3.
No further progress is feasible without high spectral
resolution/high sensitivity UV spectroscopy.
2.3. Astronomical engines
Astronomical engines (stars, black holes, etc...) can accel-
erate large masses to velocities close to the speed of light
or generate sudden ejections of mass as observed in Super-
nova explosions. They are also able to produce significantly
milder winds, as seen in the Sun, or to eject gas shells induced
by pressure pulsations in the stellar atmosphere. All of these
phenomena transform energy of various forms (gravitational,
thermal, radiative, magnetic) into mechanical energy to pro-
duce outflows in conditions very different than those tested in
Earth laboratories. Mass ejections are hot, since a fraction of
the mechanical energy involved in the acceleration heats the
gas. The ejected matter is also diffuse, since it emerges from
rarefied environments and the plasma confinement there is
weak. Thus, the study of the thermal and kinetical properties
of the ejected matter most astronomical engines need to be
studied in the UV, with the only possible exception being
very dense and slow outflows where molecules and dust can
form.
The least conventional engines are those generating highly
collimated bipolar outflows and jets. These are thought to be
driven by a combination of gravitational energy, differential
rotation and magnetic fields. They are among the most ex-
citing objects in nature; however, their underlying physics is
poorly known. This physical regime affects all of the many
scales of Astrophysics; it determines the luminosity of the
AGNs and the re-ionization of the Universe at z3. It also
determines the properties of planetary systems, which are
just angular momentum reservoirs left over when the engine
is turned off in pre-main sequence stars.
The physics of accretion-based engines, i.e., the way by
which gravitational energy is transformed into radiation and
mechanical energy (outflow) within accretion disks, is poorly
known. Recently, linear instability analyses have demon-
strated that keplerian hydrodynamical disks are stable; how-
ever, magnetohydrodynamic (MHD) disks are quite gener-
ally turbulent, and transport angular momentum outwards
quite effectively. Thus, accretion disks ought to be mag-
netized in order to be turbulent and thus be able to dump
gravitational energy into heat, as predicated in the standard
α-disk model. After the recognition of this fact, accretion
physics research is now focused on the study of the impli-
cations of magnetic fields both for the physics of the disks
and for the disk interaction with the gravitational source. To-
day, this process is identified in many astrophysical objects
spanning a range of 1010 in mass (from protostars, to white
dwarfs, neutron stars, black holes and supermassive black
holes). There are three common properties to all of these
phenomena:
1. At very high energies, there is excess energy compared
with the expected radiation from the central object and
the thin accretion disk model.
2. When jets are generated, their velocity is similar to the
keplerian velocity at the inner disk radius; e.g., ranging
from a few hundred kilometers per second in protostars
to velocities comparable to the speed of light in QSOs
and micro-QSOs.
3. Violent ejections, eruptions, and rapid flux variations are
detected. Knots are detected in the outflows, indicating
that these contain a significant non-stationary compo-
nent.
This physical behaviour applies to phenomena ranging
from the formation of the Solar System, to interacting bina-
ries, microquasars, Seyfert galaxies and quasars.
Gravity is the driving force in this process thus, the key
to understand the underlying physics lies deep inside the
gravitational potential well, in the interaction region between
the dominant source of gravity (star, white dwarf, neutron star
or black hole) and the inner disk. The radiative output from
this region is produced in the UV-range for the vast majority
of sources:
1. In AGN’s and microquasars far UV radiation (λ∼
1500 ˚
A) is produced by the accretion disk, however UV
Springer

138 Astrophys Space Sci (2006) 303:133–145
photons are energized to the X-rays range by inverse
Compton scatering with the ambient highly relativistic
electrons and the observed UV radiation is dominated
by the reprocessing of the inner UV and X-ray pho-
toionizing spectrum in the circum-nuclear matter: the gas
clouds of the Broad Lines Region (BLR). As accretion is
not stationary, the reverberation of the variations is ob-
served in the UV range providing a powerful method to
study the gas distribution around such sources allowing
the determination of the characteristic scales and masses
(e.g., Wandel, Peterson and Malkan, 1999; Kaspi et al.,
2000).
2. In accreting white dwarfs (WD), UV radiation is produced
in the atmosphere of the accretion disk (and in the WD
itself providing a useful tool to identify its characteris-
tics). The propagation of the heating fronts generated in
disk instabilities through the inner disk is tracked in the
UV providing detailed information on the inner disk struc-
ture: disks seem to be strongly depleted during quiescence.
The UV spectral energy distribution (SED) is crucial to
asses temperature profiles and extension of the disk down
to the magnetospheric radius in magnetized cataclysmic
variables (CVs). Moreover, in magnetized CVs, the ac-
cretion flow is channelled by the field to the poles where
the gravitational energy is released in a shock that heats
the flow to 106–107K. These X-ray photons are repro-
cessed into the UV in the infalling gas column; thus UV
monitorings allow tracking the shape and properties of the
funnel.
3. In T Tauri Stars (TTSs), UV radiation is produced in an
extended magnetosphere, in accretion shocks alike the ob-
served in accreting WD and in the outflow. Though many
properties of the TTSs systems are alike the observed in
WD, there are two fundamental differences: the central
object is not compact and the accretion rate is controlled
by the evolution of the accretion disk itself (instead of
mass transfer from a companion star). It also adds an im-
portant extra motivation: understanding how dynamos are
set in cool stars.
In TTSs, the magnetic interaction between the star and
the keplerian disk transforms angular momentum into
magnetic field. Differential rotation in the disk, generates
toroidal flux and the corresponding pressure push the field
lines outwards and inflate them. The dissipation of mag-
netic energy through reconnection heats up the plasma
to very high temperatures (see e.g. von Rekowski and
Branderberg, 2005) producing a magnetosphere that ex-
tends up to 4-5R∗becoming a major contributor to the UV
radiation flux. In a sense, the mediation of the magnetic
field heats up the accretion process. This also have impor-
tant implications for the radiative environment in proto-
stellar disks and young planetary systems (see G´omez de
Castro et al., 2006).
The most general physics controlling accretion-based en-
gines is non-stationary and highly non-linear, since magnetic
fields and relativity are involved. This implies an enormous
mathematical complexity that can only be addressed in two
manners: either by working with simplified models, or by de-
signing good numerical experiments (which, in turn, require
the simplified models to be properly understood). Thus, from
the physics point of view, non-relativistic objects represent
the very best laboratory to test our understanding of accre-
tion.
Key questions that remain open concern:
1. What controls the efficiency of accreting objects as
gravitational engines?, is the magnetic field needed to
guarantee that outflows are fast?, what are the relevant
timescales for mass ejection?
2. How does the accretion flow proceed from the disk to the
source of the gravitational field in the presence of moder-
ate magnetic fields?, which fraction of the gravitational
energy lost in this process is deposited on the stellar sur-
face?, which fraction is lost in amplification/dissipation
of magnetic flux?
3. Which is the role played by radiation pressure in this
whole process?
4. What role do disk instabilities play in the whole accre-
tion/outflow process?, which are the key mechanisms
driving these instabilities?
Though interacting binaries and AGNs have been studied
by the main UV missions for many years there are still many
problems to be studied because as our understanding of the
underlying physics improves, new observations are required
to test the improved theory (see Gaensicke et al., 2006; Kol-
latschny and Ting-Gui, 2006). A major breakthrough in our
understanding of these objects will come from UV spectro-
scopic observations of the pre-main sequence systems be-
cause:
TTSs represent an intermediate class of objects, where the
field plays a significant role but it is not as strong as
observed in magnetic cataclysmic binaries or in neutron
stars.Yet TTSs produce strong bipolar outflows and jets
lasting a long fraction of their pre-main sequence evolu-
tion (from some 1000 years to 107years) with velocities
comparable to the keplerian velocity at 0.01 AU (or 2.1
R). Thus, TTSs are the most efficient, accretion-based
engines, in the non-relativistic regime.
As accretion progress, the configuration of the TTSs field
evolves and the stellar dynamo sets-in. This evolution
also provides fundamental information on how the solar
dynamo was formed and evolved in the early phases.
A significant fraction of the radiation that keeps the gas ion-
ized (and the field coupled to it) is produced by magnetic
Springer

Astrophys Space Sci (2006) 303:133–145 139
reconnection associated with the performance of the en-
gine.
In addition, TTSs are unique to study the environment (ra-
diation, high energy particles, dynamical processes) in which
planetary systems, like ours, grow. Notice that recent theories
proposed that the inner, Earth-like, planets begin to build-up
some 106after the star begin to form and, at this stage, the
accretion-based engine is still operating. The radiation pro-
duced by the engine ougth to have an important effect on
the inner disk evolution and the evaporation of the primary
atmospheres of the planets-embryos.
As shown in G´omez de Castro et al. (2006), UV spec-
troscopy carried out with HST/STIS has shown that this work
is feasible from observations of the brightest TTSs. The emis-
sion from the accretion flow in CIV, SiIII, CIII has been de-
tected as the contribution of the wind to the CIII, SiIII, CII
lines. High resolution spectroscopy with an instrument 20
times more efficient than HST/STIS will allow to reach the
major factories of stars in the nearby Lupus or Taurus-Auriga
complexes. An additional advantage of this improved sensi-
tivity is that it will allow the carrying out of short-term vari-
ability studies; these are essential for studying properly the
non-stationary components. This type of study has proven to
be very valuable to distinguish the different sources of non-
stationary phenomena such as flares or shocks (G´omez de
Castro, 2002).
3. The ultraviolet Universe
In the following, a brief summary is presented on the major
issues raised by the astronomical community when asked
about whether and why access to the UV range is important
for the progress of the various research fields in astrophysics.
All these points are discussed at length in the subsequent
articles of this special volume:
3.1. The solar system
Our Solar System serves as the nearest laboratory for planet
formation and evolution and the detailed studies of its mem-
bers are applied to the understanding of other, distant plan-
etary systems. One of the basic questions in modern astro-
physics is how planets “work”, how planetary systems origi-
nated, and how life emerged on Earth. By studying the large
and the small bodies in our system, we link “local” studies to
the issue of the existence and properties of Earth-like extra-
solar planets. UV observations, along with data collected in
other spectral bands, are necessary and in some cases essen-
tial to understand the nature of our neighbours in the Solar
System.
While many objectives of solar system research can be
achieved by optical and near-IR (nIR) imaging, topics from
surface mineralogic characterization to auroral activity re-
quire the combination of information spanning a wide spec-
tral range including the UV.
Planetary studies require synoptic observations over peri-
ods of time ranging from a single revolution (hours to days)
to many years (to span at least a full solar cycle). For a given
aperture size, UV Astronomy from space can achieve much
higher spatial resolution than from the ground because of the
absence of the smearing effect of the Earth’s atmosphere and
because of the smaller diffraction limit of UV telescopes.
We identify two immediate programmatic requirements:
the establishment of a mineralogic database in the ultravio-
let for the characterization of planetary, ring, satellite, and
minor planet surfaces, and the development and deployment
of small orbital solar radiation monitors. The former would
extend the methods of characterizing surfaces of atmosphere-
less bodies by adding the UV segment and permit the study
of volatile transport on bodies with atmospheres. The latter
are needed to establish a baseline against which contempo-
raneous UV observations of Solar System objects must be
compared.
We identify two types of UV missions that would be two
stages in a single process: one requires a two-meter-class
telescope using almost off-the-shelf technology and could
be launched in the next few years. The other requires a much
larger (5–20 meter class) instrument that would provide the
logical follow-up after a decade of utilizing the smaller facil-
ity. The very large UV telescope will offer angular resolution
at par with that of the 100-m OWL telescope allowing coarse
mapping much beyond the Kuiper-Edgeworth belt.
3.2. Cool stars
Emission in the UV is an essential probe for studying impor-
tant physical processes related to the production and transport
of magnetic energy in plasmas. Our understanding of such
processes is closely related to our ability to predict the evolu-
tion of the solar magnetic activity and, therefore, to simulate
the conditions in which life has evolved on Earth and how the
solar emission of radiation will change due to the evolution
of its magnetic dynamo. Future UV missions will advance
the study of the consequences of stellar magnetic activity on
planets orbiting around them.
Cool star atmospheres represent, undoubtly, a laboratory
in which magnetic activity phenomena can be studied under a
large variety of conditions, placing strong constraints on our
knowledge of the fundamental processes involved. The con-
sequences of both large and small magnetic activity can these
be studied extensively. The UV range is unique as it permits
the study of cool star atmospheres from the chromosphere
to the corona using powerful diagnostics. Recent techniques
Springer

140 Astrophys Space Sci (2006) 303:133–145
have used the hydrogen Lyαline profile to study, for the first
time, the wind from cool stars through its interaction with
the interstellar gas (Wood, et al., ).
A 2m class UV telescope with high-resolution spec-
troscopy and monitoring capabilities would allow important
discoveries in this field. A larger aperture telescope (from
4 to 6m) would permit the study of the plasma dynam-
ics and the chromospheric – transition region structures of
fainter magnetic active stars, like brown dwarfs and stars
in clusters. This is requird to characterize the outer atmo-
spheres of parent stars of extrasolar planets that will be dis-
covered by future space missions like COROT,Kepler, and
Darwin.
3.3. Massive stars
Massive stars and their descendants are important con-
stituents of galaxies. Because of their high luminosities (up
to 106L) and their massive winds ( ˙
M=10−8to 10−4
Myr−1,v
∞=100 to 2000 km s−1)theyhaveanex-
tremely important influence on the dynamics and energet-
ics of the interstellar medium. They also enrich the inter-
stellar medium in nuclear processed material. This enrich-
ment occurs via mass loss (a massive star can lose 2/3 or
more of its mass via a stellar wind) or during the SN ex-
plosion. They directly influence star formation by disrupting
molecular clouds via SN explosions, or conversely they can
initiate star formation through massive wind-blown bubbles
and Sn shells compressing nearby molecular clouds. Massive
stars are also thought to be responsible for the reionization
of the early Universe. More recently, it has been proposed
that the most massive stars are the progenitors of gamma-ray
bursts.
The UV constitutes an optimum spectral window as the
spectra energy distributions of massive stars reach their max-
ima within this wavelength range. Appart from this efficient
coincidence established by nature, massive stars decorate the
UV spectral region with a number of key diagnostics to our
understanding of the nature of these objects and their inter-
action with the surrounding media.
High spectral resolution spectroscopy provides unique in-
formation about massive stars winds (P-Cygni profiles pro-
duced by the resonance transitions of CIV, NV, SiIV, OVI,
etc). In addition, the unsaturated line profiles from ionized
species trace very efficiently the mass-loss rate characterizing
the stellar wind. Further, when combined with ρ2sensitive
diagnostics at other wavelengths they may be used to cali-
brate the presence of inhomogeneities (“clumping”) in the
wind.
The next step is to extend this work to external galax-
ies. The optimized spatial/spectral resolution achieved at UV
wavelengths is fundamental for this purpose.
3.4. Star formation: From the ISM to planets
Planetary systems are angular momentum reservoirs gener-
ated during star formation. Solutions to three of the most im-
portant problems in contemporary astrophysics are needed to
understand the entire process of planetary system formation:
The physics of the ISM. Stars form from dense molec-
ular clouds that contain ∼30% of the total interstellar
medium (ISM) mass. The structure, properties and life-
times of molecular clouds are determined by the overall
dynamics and evolution of a very complex system – the
ISM. Understanding the physics of the ISM is of prime
importance not only for Galactic but also for extragalac-
tic and cosmological studies. Most of the ISM volume
(∼65%) is filled with diffuse gas at temperatures be-
tween 3000 K and 300,000 K, best observed in the UV,
representing about 50% of the ISM mass.
The physics of accretion and outflow. Powerful outflows are
known to regulate angular momentum transport during
star formation, the so-called accretion-outflow engine.
Elementary physical considerations show that, to be ef-
ficient, the acceleration region for the outflows must be
located close to the star (within 1 AU) where the gravita-
tional field is strong. According to recent numerical sim-
ulations, this is also the region where terrestrial planets
could form after 1 Myr. One should keep in mind that to-
day the only evidence for life in the Universe comes from
a planet located in this inner disk region (at 1 AU) from
its parent star. The temperature of the accretion-outflow
engine is between 3000 K and 107K. After 1 Myr, dur-
ing the classical T Tauri stage, extinction is small and the
engine becomes naked and can be observed at ultraviolet
wavelengths.
The physics of planet formation. Observations of volatiles
released by dust, planetesimals and comets provide an
extremely powerful tool for determining the relative
abundances of the vaporizing species and for studying
the photochemical and physical processes acting in the
inner parts of the protoplanetary disks. This region is
illuminated by the strong UV radiation field produced
by the star and the accretion-outflow engine. Absorp-
tion spectroscopy provides the most sensitive tool for
determining the properties of the circumstellar gas as
well as the characteristics of the atmospheres of the in-
ner planets transiting the stellar disk. UV radiation also
pumps the electronic transitions of the most abundant
molecules (H2, CO,...) that are observed in the UV. See,
for instance, the HST and FUSE observations of the Beta
Pictoris disk which led to conclusion that CO is produced
by an extremely large number of comets orbiting in this
young planetary system (Jolly et al., 1998; Lecavelier
des Etangs et al., 2001)
Springer

Astrophys Space Sci (2006) 303:133–145 141
A rather modest UV telescope (2-m telescope with state-
of-the-art optics, instruments and detectors) would produce
an extraordinary scientific return as outlined above. A large,
50-m, UV-optical instrument would provide an efficient
mean for measuring the abundance of ozone in the atmo-
sphere of the thousands of transiting planets expected to be
detected by the next space missions (GAIA, Corot, Kepler...).
Thus a follow-up UV mission would be optimal for identi-
fying Earth-like candidates.
3.5. Structure and evolution of white dwarfs and their
interaction with the ISM
The development of far-UV astronomy has been particularly
important for the study of hot white dwarf stars. A signifi-
cant fraction of their emergent flux appears in the far-UV and
traces of elements heavier than hydrogen or helium are, in
general, only detected in this waveband or at shorter wave-
lengths that are also only accessible from space. Therefore,
high-resolution far-UV spectroscopy has been essential for
measuring white dwarf composition, to delineate the evo-
lution of their atmospheres and to examine the relationship
between the various physical processes that determine the
appearance of these stars. In addition to highlighting photo-
spheric material, the strong blue continua of hot white dwarfs
also act as a backdrop to absorption lines from the inter-
stellar medium. Consequently, observations of white dwarfs
also provide an important probe of the interstellar space with
which they interact, their progenitors supplying material and
possibly accreting from interstellar clouds as they age. High-
resolution spectra can also provide dynamical information
on white dwarfs in binaries from which stellar masses can be
estimated. UV imaging yields complementary information
by resolving these systems, allowing direct detection of hot
white dwarfs that might otherwise be hidden in the glare of
much brighter companions at visible wavelengths.
Although white dwarfs have been studied in the far-UV
throughout the past 25 years, since the launch of IUE, only
a few tens of objects have been studied in great detail and
a much larger sample is required to gain a detailed under-
standing of the evolution of hot white dwarfs and the physical
processes that determine their appearance. Many outstand-
ing problems remain, including the origin and relationships
of the H and He-rich groups, the initial-final mass relation for
white dwarfs and their progenitors and the 3D structure of the
ISM. All white dwarfs that have ever been studied in the UV
reside within our own galaxy and must have emerged from
stellar populations with different ages and environments. To
solve the outstanding problems and make significant further
progress in the study of white dwarfs requires a substantial
enlargement of the sample, to properly examine the full range
of temperatures, gravities and possible environmental condi-
tions by probing deeper into our own galaxy and extending
studies to co-eval populations in globular clusters, the Mag-
ellanic clouds and nearby galaxies.
To achieve these goals there is a need for dramat-
ically enhanced instrument sensitivity, providing high
(R 50,000–100,000) and low resolution spectroscopy, with
diffraction limited imaging. Coupled with advances in in-
strument and detector design, a 2-m class telescope would
be able to address many of the science goals relating to ob-
servation of white dwarfs in our own galaxy, but in the time
frame beyond 2015, it is absolutely essential that a large UV
facility is constructed to reach outside the galaxy.
3.6. Interacting binaries
Interacting binaries (IBs) are among the most intriguing and
exotic stellar systems, since the stellar components interact
each other affecting their physical status and evolution. IBs
consist of a variety of stellar objects in different stages of evo-
lution and those containing accreting compact objects still
represent a major challenge to our understanding of not only
close binary star evolution but also of the chemical evolution
of the Galaxy. These end-points of binary star evolution are
showcases of wide variety of processes including mass accre-
tion and outflow, stellar wind interaction with plasma condi-
tions spanning a wide range of physical conditions including
relativistic enviroments and extreme magnetic field strenghts.
Consequently, IBs are also extremely versatile plasma physic
laboratories.
Despite their great importance for a vast range of astro-
physical questions, our understanding of close binary stars
and their evolution is still very fragmentary. The ultraviolet
is of outmost importance in the study of IBs, as a large part
of their luminosity is radiated in this wavelength range, and,
more importantly, as the UV hosts a multitude of low and high
excitation lines of a large variety of chemical species. These
transitions can be used both as probes of the plasma condi-
tions, as well as tracers of individual components within the
binary through time-resolved spectroscopy. Moreover, the
physical status of the binary components and in particular
the accreting white dwarf primaries in cataclysmic variables
(CVs), symbiotic stars, and double-degenerate binaries can
be easily isolated and studied in the UV range.
Even though substantial scientific progress has been
achieved throughout the last three decades, primarily us-
ing the International Ultraviolet Explorer (IUE), the Hub-
ble Space Telescope (HST), and the Far Ultraviolet Spectro-
scopic Explorer (FUSE), there are still many open problems.
Among them, key issues are: (i) the nature of SNIa pro-
genitors exploring both single and double-degenerate chan-
nels, (ii) the physics of accretion discs, in particular the role
of viscosity and its time-dependence, and the development
of winds, (iii) the fundamental properties of white dwarfs
in CVs, as these are strongly affected by accretion and its
Springer

142 Astrophys Space Sci (2006) 303:133–145
associated angular momentum and (iv) the nature of the
IB population in globular clusters. These can be efficiently
achieved by means of UV observations surveying much
larger samples than done so far. In particular the first three
goals require medium (R2000) to high (R20000) FUV
(possibly down to 912 ˚
A) resolution spectroscopy with high
temporal resolution capabilities (time-tag) to allow phase–
resolved studies along the binary orbit as well as with a high
duty cycle to monitor outburst evolution. The latter goal in-
stead requires a large (10 arcmin) field-of-view imager with
diffraction–limited spatial resolution using broad band FUV
and NUV filters with accurate timing capabilities.
A large collecting area is relevant to deeply investigate fast
UV variability which is an ubiquitous feature in IBs. Fast non-
periodic and quasi-periodic variations on timescales from
seconds to tens of minutes are commonly observed in CVs.
Quasi-periodic-oscillations (QPOs) of a few seconds were
discovered in the optical in the eighties and in the UV range
in the early nineties in a few bright strongly magnetized CVs.
They are believed to arise from shock oscillations though the
driving mechanism is still unclear. A proper knowledge of
their energy distribution and of the variations of amplitudes
and phases is of great potential to diagnose the magnetic field
and cooling process in the radiative shocks. Furthermore, os-
cillations during dwarf novae outbursts (DNOs) and QPOs
from a few seconds to thousands of seconds were detected
for the first time in the UV and now have been recently rec-
ognized as parallel to the high- and low-frequency QPOs
observed in X-ray binaries (Warner, 2004) with an origin
likely residing in the magnetic nature of the accreting white
dwarf. Also, flickering on timescales of minutes are believed
to be associated to fluctuations in the mass accretion.
3.7. Active Galaxies
Active Galaxies emit their maximum flux in the UV/FUV.
The overall continuum flux peaks in-between the optical and
soft X-ray spectral range. More than half of the bolometric
luminosity of an (un-obscured) AGN is emitted in this big
blue bump. Models of hot accretion disks – surrounding the
central super-massive black hole in AGN – cannot reproduce
in a simple way the observed spectral shape.
This rest frame EUV continuum of highly redshifted AGN
is important for our understanding of the evolution of the
early universe. The UV continuum of quasars ionizes the
intergalactic medium at the end of the dark ages.
Furthermore, the central continuum source in AGN ion-
izes the circumnuclear gas in the so called broad line region
(BLR) and narrow line region (NLR). The overall continuum
distribution as well as the UV spectral lines (narrow emission
lines, broad emission lines, absorption lines) are tracers of
the physical conditions of those regions where these emission
lines originate. Most important diagnostic lines for studying
the physical conditions and metallicities in the central regions
of AGN are emitted in the UV. It is possible to derive some
information for distant (z≥2) as well as luminous quasars
when the diagnostic lines are shifted into the optical range
with ground-based telescopes. But it is necessary to observe
the UV-spectra of ’nearby and present-day’ AGN for study-
ing their cosmological evolution as well as the evolution of
the universe. The UV spectra of the class of low luminous
AGN can be observed only in the local universe because of
their faintness.
The emission line region of the narrow lines is spatially
resolved in some nearby objects. They originate at distances
of pc to kpc from the central ionizing source. However, the
broad emission lines originate at distances of light days to
light months only from the central ionizing source. This BLR
is unresolved by orders of magnitudes even for the nearest
AGN. But variability studies of the ionizing continuum flux
and the emission line intensities/profiles give us information
on the structure and kinematics of the surrounding of the cen-
tral supermassive black hole in AGN as well as on their mass
itself. The monitoring of highly ionized UV lines in AGN en-
ables us to study the physics of the immediate environment
of black holes nearest to the center.
3.8. Starbursts
Starbursts are systems with very high star formation rates
per unit area. They are the preferred places where massive
stars form, the main source of thermal and mechanical heat-
ing in the interstellar medium, and the factory where the
heavy elements form. Thus, starbursts play an important role
in the origin and evolution of galaxies. The similarities be-
tween the physical properties of local starbursts and high-z
star-forming galaxies highlight the cosmological relevance
of starbursts. On the other hand, nearby starbursts are lab-
oratories for studying violent star formation processes and
their interaction with the interstellar and intergalactic media,
in detail and deeply. Starbursts are bright at ultraviolet (UV)
wavelengths, as they are in the far-infrared, due to the ’picket-
fence’ interstellar dust distribution. After the pioneering IUE
program, high spatial and spectral resolution UV observa-
tions of local starburst galaxies, mainly taken with HST and
FUSE, have made relevant contributions to the following is-
sues:
–The determination of the initial mass function (IMF) in
violent star forming systems in both, low and high metal-
licity environments, and in dense (e.g. in stellar clusters)
and diffuse environments: A Salpeter IMF with high-
mass stars constrains well the UV properties.
–The modes of star formation: Starburst clusters are an
important mode of star formation. Super-stellar clusters
have properties similar to globular clusters.
Springer

Astrophys Space Sci (2006) 303:133–145 143
–The role of starbursts in AGN: Nuclear starbursts can
dominate the UV light in Seyfert 2 galaxies, having bolo-
metric luminosities similar to the estimated bolometric
luminosities of the obscured AGN.
–The interaction between massive stars and the interstel-
lar and intergalactic media: Outflows in cold, warm and
coronal phases leave their imprints on the UV interstellar
lines. Outflows of a few hundred km s−1are ubiquitous
phenomena in starbursts. These metal-rich outflows and
the ionizing radiation can travel to the halo of galaxies
and reach the intergalactic medium.
–The contribution of starbursts to the reionization of the
universe: In the local universe, the fraction of ionizing
photons that escape from galaxies and reach the inter-
galactic medium is of a few percent. However, in high-z
star-forming galaxies, the results are more controversial.
Despite the very significant progress over the past two
decades in our understanding of the starburst phenomenon
through the study of the physical processes revealed at satel-
lite UV wavelengths, there are important problems that still
need to be solved. High-spatial resolution UV observations
of nearby starbursts are crucial to further progress in under-
standing the violent star formation processes in galaxies, the
interaction between the stellar clusters and the interstellar
medium, and the variation of the IMF. High-spatial reso-
lution spectra are also needed to isolate the light from the
center to the disk in UV luminous galaxies at z=0.1–0.3
found by GALEX. Thus, a new UV mission containing an
intermediate spectral resolution long-slit spectrograph with
high spatial resolution and high UV sensitivity is required to
further progress in the study of starburst galaxies and their
impact on the interstellar and intergalactic media.
3.9. Supernovae (SNe)
UV observations of SNe are required not only for understand-
ing of the SN phenomenon itself, such as the kinematics and
the metallicity of the ejecta, but also for providing exciting
new findings in Cosmology, such as the tantalizing evidence
for “dark energy” that seems to pervade the Universe and
to dominate its energetics. SNe are bright events that can be
detected and studied even at very large distances. Ultraviolet
spectroscopy is crucial in order to:
Study the metallicity of individual SNe
Study the metallicity of the intervening ISM/IGM
Study the kinematics of the fast moving (i.e. the outermost
layers) of the ejecta through the analysis of strong UV
lines with P-Cygni profiles.
Study the overall energetics of SNe explosion at early
phases (from shock breakout to optical maximum for
types of SNe, but most importantly for all Type II SNe)
Study of the strong emission lines produced in the inter-
action of the ejecta with pre-SN circumstellar material,
e.g. NV 1240 ˚
A and collisionally excited CIV 1550 ˚
A,
NIV] 1470 ˚
A, OIII] 1665 ˚
A, NIII] 1750 ˚
A, CIII] 1908 ˚
A.
SNIa are very good standard candles (e.g. Macchetto and
Panagia, 1999) to measure distances to distant galaxies, cur-
rently up to redshift z1 and, considerably more in the
foreseeable future. This is a challenging proposition, both
for technical reasons (observations in the near IR of increas-
ingly faint objects) and for more subtle reasons, i.e. one must
verify that the discovered SNe are indeed SNIa and that these
SNa share the same properties of their local Universe rela-
tives. One can only discern Type I from Type II SNe on the
basis of the overall properties of their UV spectral energy
distributions (Panagia, 2003, 2005).
4. Summary
This review outlines the scientific reasons behind the need
for an ultraviolet observatory. Most of science described here
could be carried out with two basic instruments:
1. A high-resolution (50,000–100,000) spectrograph cov-
ering the whole 90–320 nm spectral range.
2. A low-spectral resolution (1000–5000) high-sensitivity
spectrograph allowing integral field spectroscopy (long-
slit in its simplest version) with spatial resolution (50
mas) and wavelength coverage from 110–450nm.
These instruments should provide an improvement by a fac-
tor of ∼20 in effective area over the HST/STIS capabilities.
This improvement is rather conservative from the technolog-
ical point-of-view since it could be achieved by improving
optical designs and coatings and make use of MCP detec-
tors with enhanced sensitivity, bigger size and improved dy-
namic beeing related to new fast read-out electronics. It is
amazing the large progress that could be achieved with a
relatively modest investment; a good example of this is the
proposedWorld Space Observatory project.
Looking into the far future, it is clear that the frontier is
building larger facilities that increase the effective collecting
surface by, at least, 2 orders of magnitude. A properly instru-
mented 4–6 m telescope in space would be very useful for
future UV observations.
A larger, 20 m size, telescope in space represents a huge
technological defy. Coordinated constellations of 1-m size
telescopes seem to be the most realistic manner to get large
collecting surfaces. In turn, this would allow carrying out
UV interferometry. The potential of high spatial resolution
(milliarcsecond scale) instruments is enormous. One promis-
ing concept to get micro-arcsecond resolution imaging is
the Stellar Imager mission under study at NASA GSFC, a
Springer

144 Astrophys Space Sci (2006) 303:133–145
kilometer scale interferometer composed of around 30 small
telescopes formation, flying in space (Carpenter et al., 2004).
Another possibility would be building large ground-based
telescopes in ozone depleted areas. The discussion of the
’ozone hole’ due to human activity on the one hand, and
the realization that photon absorption by ozone in the UV is
one of the important sources of opacity in the atmosphere,
would argue that the location of a ground-based telescope
underne ath an ozone hole may extend the spectral range
accessible from the ground into the UV. Assuming this effect
to be present and significant, the best place for such a ground-
based UV telescope would be in the Antarctic, possibly at the
Concordia station at Dome C. This location shows the best
seeing for any Earth based observatory, but as far as we are
aware no long-term study of the atmospheric transparency
at short wavelengths has been conducted there. The study
should also consider the atmospheric emissions at a location
relatively close to the South Magnetic Pole, and the influence
of sunlight scattered into the telescope when the Sun is below
the horizon.
A detailed accounting on the scientific requirements to UV
observatories can be found in Kappelmann et al. contribution
to these proceedings.
Acknowledgements During the editing of the present volume, our dear
collegues Willem Wamsteker, Michael Festou and Marcello Rodon´o
suddenly passed away. We dedicate this work to them with the hope that
their worldwide contribution to UV science will be kept recognized in
the future of UV astronomy.
References
Avillez, M., Breitschwerdt, D.: Volume filling factors of the ISM phases
in star forming Galaxies. I. The role of the disk-halo interaction.
Astronomy and Astrophysics, in press (2005)
Avillez, M., Breitschwerdt, D.: Global dynamical evolution of the ISM
in star forming Galaxies – I. high resolution 3D HD and MHD sim-
ulations: effect of the magnetic field. Astronomy and Astrophysics
425, 899 (2004)
Bahcall, J.N., Ostriker, J.P.: Unsolved problems in astrophysics prince-
ton university press (1997)
Carpenter, K.G., Schrijver, C.J., Allen, R.J., et al.: SI- The stellar imager
SPIE astronomical telescopes and instrumentation. SPIE paper
# 5491, V28
Cernicharo, J.: The polymerization of acetylene, hydrogen cyanide, and
carbon chains in the neutral layers of carbon-rich proto-planetary
nebulae. Astrophysical Journal 608, L41 (2004)
Ehenreich D., Tinetti G., Lecavelier des Etangs A., Vidal-Madjar A.,
Selsis F.: The transmission spectrum of earth-size transiting plan-
ets. A& A, in press (2005)
G´omez de Castro, A.I.: On the source of the flaring activity in AB
Doradus: the UV spectral signatures. Monthly Notices of the R.A.S
332, 409 (2002)
G´omez de Castro, A.I.: Magnetic activity and the interaction between
the stellar magnetosphere and the accretion disk. Astrophysics and
Space Science 292, 561 (2004)
Jolly, A., McPhate, J.B., Lecavelier, A., Lagrange, A.M., Lemaire,
J.L., Feldman, P.D., Vidal Madjar, A., Ferlet, R., Malmasson, D.,
Rostas, F.: HST – GHRS observations of CO and CI fill in the
beta – Pictoris circumstellar disk. Astronomy and Astrophysics
329, 1028–1034 (1998)
Kriss, G.A., Shull, J.M., Oegerle, W., Zheng, W., Davidsen, A.F.,
Songaila, A., Tumlinson, J., Cowie, L.L., Deharveng, J.-M.,
Friedman, S.D., et al.: Resolving the structure of ionized helium
in the intergalactic medium with the far ultraviolet spectroscopic
explorer. Science 293, 1112 (2001)
Kaspi, S., Smith, P.S., Netzer, H.M.D., Jannuzi, B. T. Giveon, U.:
Reverberation measurements for 17 quasars and the size-mass-
luminosity relations in active galactic nuclei. Astrophysical Jour-
nal 533, 631 (2000)
Kutchner, M.J.: Volatile-rich earth-mass planets in the habitable zone.
Astrophysical Journal 596, L105–L108 (2003)
L´eger, A., et al.: A new family of planets? “Ocean-Planets”. Icarus 169,
499–504 (2004)
Lebouteiller, V., Kunth, D., Lequeux, J., Lecavelier des Etangs, A.,
D´esert, J.-M., H´ebrard, G., Vidal-Madjar, A.: Abundance differ-
ences between the neutral and the ionized gas of the dwarf galaxy
IZw 36. Astronomy and Astrophysics 415, 55–61 (2004)
Lecavelier des Etangs, A., et al.: Deficiency of molecular hydrogen in
the disk of βpictoris. Nature 412, 706–708 (2001)
Lecavelier des Etangs, A., Vidal-Madjar, A., McConnell, J.C., H´ebrard,
G.: Atmospheric escape from hot Jupiters. Astronomy and Astro-
physics 418, L1 (2004a)
Lecavelier des Etangs, A., D´esert, J.-M., Kunth, D., Vidal-Madjar, A.,
Callejo, G., Ferlet, R., H´ebrard, G., Lebouteiller, V: FUSE ob-
servations of the H I interstellar gas of I Zw 18. Astronomy and
Astrophysics 413, 131–137 (2004b)
Mayor, M., Queloz, D.: A Jupiter-mass companion to a solar-type star.
Nature 378, 6555–355 (1995)
Moos, H.W., Sembach, K.R., Vidal-Madjar, A., York, D.G., Friedman,
S. D. et al.: Abundances of deuterium, nitrogen, and oxygen in the
local interstellar medium: overview of first results from the FUSE
mission. Astrophysical Journal, Suppl. Series 140, 3 (2002)
Reimers, D., Vogel, S., Hagen, H.-J., Engels, D., Groote, D., Wamsteker,
W., Clavel, J., Rosa, M.R.: The O/C abundance ratio in absorbing
gas clouds at high redshift. Nature 350, 561 (1992)
Reimers, D., Koehler, S., Wisotzki, L., Groote, D., Rodriguez-Pascual,
P., Wamsteker, W.: Patchy intergalactic He II absorption in HE
2347–4342. II. The possible discovery of the epoch of he-
reionization. Astronomy and Astrophysics 327, 890 (1997)
Richter, P., Savage, B.D., Wakker, B.P., Sembach, K.R., Kalberla,
P.M.W.: The FUSE spectrum of PG 0804 +761: a study of atomic
and molecular gas in the lower galactic halo and beyond. Astro-
physical Journal 549, 281 (2001)
Richter, P., Savage, B.D., Tripp, T.M., Sembach, K.R.: FUSE and STIS
observations of the wram-hot intergalactic medium toward PG
1259 +593. Astrophysical Journal Suppl 153, 165 (2004)
Richter, P.: Perspectives for hunting the missing baryons in the local
universe. Proc. 39th ESLAB Symp, Noorwijk 19–21 (2005), F.
Favata, J. Sanz-Forcada, A.Gim´enez (eds.), p. 157, (2005)
Rivera, E.J., Lissauer,J.J., Butler, R.P.,Marcy, G.W., Vogt, S.S., Fischer,
D.A., Brown, T.M., Laughlin, G., Henry,G.W.: A 7.5 Moplus; Planet
Orbiting the Nearby Star, GJ 876. Astrophysical Journal 634, 625–
640 (2005)
Sembach, K.R., Tripp, T.M., Savage, B.D., Richter, P.: Physical prop-
erties and baryonic content of low-redshift intergalactic Lyαand
OVI absorption line systems: the PG 1116 +215 sight line. The
Astrophysical Journal Suppl 155, 351 (2004)
Tripp, T.M., Wakker, B.P., Jenkins, E.B. Bowers, C.W., et al.: Complex
C: A low-metallicity, high-velocity cloud plunging into the milky
way. Astronomical Journal 125, 3122 (2003)
Vidal-Madjar, A., Lecavelier des Etangs, A., D´esert, J.-M., Ballester,
G.E. et al.: An extended upper atmosphere around the extrasolar
planet HD209458b. Nature 422, 143 (2003)
Springer

Astrophys Space Sci (2006) 303:133–145 145
Vidal-Madjar, A., D´esert, J.-M.: Lecavelier des Etangs, A.,
H´ebrard, G. et al.: Detection of oxygen and carbon in
the hydrodynamically escaping atmosphere of the extraso-
lar planet HD 209458b. The Astrophysical Journal 604, L69
(2004)
Wandel, A., Peterson, B.M., Malkan, M.A.: Central masses and broad-
line region sizes of active galactic nuclei. I. comparing the
photoionization and reverberation techniques. The Astrophysical
Journal 526, 579 (1999)
Warner, B.: Rapid oscillations in cataclysmic variables. PASP 116, 115
(2004)
Watson,A.J., Donahue, T.M., Walker, J.C.G.: The dynamics of a rapidly
escaping atmosphere – applications to the evolution of earth and
Venus. Icarus 48, 150 (1981)
Springer

Astrophys Space Sci (2006) 303:147–151
DOI 10.1007/s10509-006-9058-3
ORIGINAL ARTICLE
Guidelines for Future UV Observatories
Norbert Kappelmann ·J¨urgen Barnstedt
Received: 13 March 2006 / Accepted: 14 March 2006
C
Springer Science +Business Media B.V. 2006
Abstract Ultra-violet image sensors and UV optics have
been developed for a variety of space borne UV astronomy
missions. Technology progress has to be made to improve
the performance of future UV space missions. Throughput is
the most important technology driver for the future. Required
developments for different UV detector types – detectors are
one of the most problematic and critical parts of a space born
mission – and for optical components of the instruments are
given in these guidelines. For near future missions we need
high throughput optics with UV sensors of large formats,
which show simultaneously high quantum efficiency and low
noise performance.
Keywords UV astronomy
1. Introduction
Within the last years a lot of workshops have taken place in
which the need for new instruments for future space-borne
UV astronomy missions was discussed (e.g. “Report of
UV/O Working Group to NASA”, 1999, “Hubble’s Sci-
ence Legacy: Future Optical-Ultraviolet Astronomy from
Space”, 2002,“Innovative Design for the next large aper-
ture optical/UV Telescope”, STSI, 2003). In a large number
of publications the characteristic parameters for new devel-
opments and associated technologies for future UV instru-
mentations are described ( eg. Proc. SPIE,Volume 2999,1997
“Photodetectors: Materials and Devices II”). Therefore this
paper is not a comprehensive review of future UV technology
developments, but focuses on the requirements for new UV
N. Kappelmann ()·J. Barnstedt
Institut f¨ur Astronomie und Astrophysik, Abteilung Astronomie
(IAAT), Universit¨at T¨ubingen, T¨ubingen, D-72076, Germany
instruments which will give the astronomical UV commu-
nity the possibility to meet the science goals, described in
the papers of this book.
2. General remarks
An improvement on current UV capabilities of the HST in
the order of magnitudes could be made by high quantum
efficiency UV detectors, high efficiency optics and the use
of 3–4 m telescopes. The technology needed for advanced
UV-astronomy can benefit from advances being made in
other wavelength regimes, in particular from developments
in the visible, infrared and x-ray wavelength bands. In gen-
eral throughput is the most important technology driver for
future UV space astronomy missions, especially for UV
spectroscopy. UV technology development has to be made in
the areas of detectors, optical components and their coatings
and large light weighted mirrors.
2.1. UV sensors
Many astronomical objects produce orders of magnitudes
more photon fluxes at optical wavelengths than they do
in the vacuum UV. In order to eliminate this huge back-
ground contribution and substantial source of noise solar-
blind detector and imaging systems are required. A com-
prehensive overview about UV imaging detectors is given
by Joseph (2000). Furthermore the reader is referred to
Welsh and Kaplan (1992), Ulmer (2002) and references
therein.
Several specific detector types can be used for vacuum UV
astronomy in the future, multidimensional detectors, semi-
conductive array (e.g. CCDs) and microchannel plate de-
tectors (MCP). The further development of technologies for
Springer

148 Astrophys Space Sci (2006) 303:147–151
Fig. 1 Solar-blind detective
quantum efficiencies obtained
by various UV detectors
adopted from Joseph (2000)
these detectors is the basis to enhance the performance for
UV applications. The current technology can be classified in
two categories: Solid-state devices based on silicon or wide
bandgap semiconductors and photoemissive devices, cou-
pled with a gain component and an electron detector. The
detective quantum efficiencies (DQE) for various UV detec-
tors as MCPs, CCDs, Electron-Bombarded CCDs and for
the expected DQEs for future AlGaN solid-state sensors are
shown in Fig. 1.
“3D” energy-resolving detectors such as photon count-
ing superconducting tunnel junction (STJ) or transition-edge
detectors (TES) have the potential to replace the detectors
which are now in use for UV mission in the far future. The
development of these energy resolving detectors will improve
efficiency and reduce the number of optical elements.
MCPs have a good potential to stay for the next years
the default detectors in the UV regime especially below 200
nm, due to the fact that they have a very good flight his-
tory, they are solar blind, have a very low readout-noise and
are radiation hard. A review about the actual image detec-
tor technology, including different readout anode types, like
image readout anodes, delay line image readout anode, in-
tensified CCD/CIDs (CCD devices couples to an MCP im-
age intensifier), cross strip anodes and pixel array anodes, is
given by Siegmund (2000). A review about the performance
of lager format CCDs and future directions of the develop-
ments of CCDs suitable for wide field UV imaging is given by
Clampin (2000). The primary technical problems of CCDs
are the high QE in the visible, low QE in the UV and the
radiation tolerance of these imaging sensors. Improvement
of solarblind UV imaging CCDs with excellent radiation tol-
erances have been made by the production of CCDs made
out of SiC (Sheppard, 1996) and GaN.
Dynamic range and linearity technology investments
should be made to ensure that potential flight detectors have
large dynamic range. It is very important that detectors are
stable and provide a linear response to the signals. It is essen-
tial to reduce substantially the background rates: the detector
background noise should not become the limiting factor and
should not determine the sensitivity of the measurements.
Fig. 2 GaN photocathodes compared with other photocathodes
adopted from Ulmer (2002)
2.2. Photocathodes
The development of new photocathodes is the approach to
lead to devices for UV space mission for the next decade.
Significant impact on future capabilities can be made by ad-
vances in new materials for the photocathodes as GaN, di-
amond or GaAs (e.g. Ulmer, 2003). For example diamond
photocathodes show efficiencies of 50% in the 200–1200 ˚
A
band. For MCP detectors based on silicon substrates diamond
can be used as a direct opaque cathode (e.g Beetz, 2000). A
comparison of GaN photocathodes with other photocathodes
is shown in Fig. 2.
Walker (2000) has reported the fabrication and charac-
terization of solar-blind AlxGa1−xN photodiodes (x∼0.70)
grown on sapphire with an internal quantum efficiency which
is greater than 90%. The device response drops four orders
of magnitude at 275 nm and remains at low response for the
entire near-ultraviolet and visible spectrum. Promising future
technology is based on the production of high quality films
made of wide bandgap semiconductors. For UV detectors it
is absolutely necessary to suppress the sensitivity in the red,
and wide-band gap semi-conductors fulfill this requirement
very well. Coupling the resulting photocathode to a device
Springer

Astrophys Space Sci (2006) 303:147–151 149
such as a micro-channel plate (MCP) is necessary to produce
imaging (Ulmer, 2002). Because UV coatings are extremely
sensitive to highly absorbing molecular contaminations care-
ful attention must be paid to possible contamination sources
and investments have to made in the long term stability of
the coatings.
2.3. Optical components: Gratings
Throughput can be maximized with intelligent optical sys-
tem designs and by judicious use of optical materials and
coatings. Effective coatings for the optical elements are
needed to maximize throughput: high reflectivity particu-
lar below 200 nm is required. One part of the optical ele-
ments are gratings and therefore it is essential to develop
gratings which have high efficiency, efficient grove shapes
and produce very low scatter. Holographic gratings are to-
day’s standard gratings for most UV spectroscopic instru-
ments. This is largely due to their scattering (typically around
10−5˚
A−1) and to their large sizes. Interesting developments
are made by direct-writing technologies. Today they are still
in an early stage of development, but this technology is able
to produce gratings with very low scatter, efficient groove
shapes, and excellent aberration correction. Due to low rul-
ing densities direct-write technologies cannot be used in the
current stage of development in high-resolution UV spec-
troscopy but within the next years large, corrective gratings
with high groove densities will be available (e.g. Wilkinson,
1999).
2.4. Optical components: Filters
Several designs of filters for use in vacuum UV imaging sys-
tems are discussed. These designs incorporate all reflective
optics, and are characterized by comparatively high in-band
throughout and very low out-of-band transmission. Filters
which can be tuned over ranges useful for vacuum UV astro-
nomical observations will be a very good tool for imaging.
Adjustable broad band and small band imaging will give
access to a bunch of diagnostic lines over the whole UV
wavelength regime.
Acousto-optical-tuneable-filters (AOTFs) have the poten-
tial of providing a bandwidth selectable across 1 octave. A
tunable RF signal is applied across a birefringent crystal, re-
sulting in a selectable output wavelength. AOTFs have been
built and are in use at IR wavelengths on ground-based in-
struments. To use these filters at vacuum UV wavelengths in
the future, considerable work must be done to characterize
birefringent UV-transmitting crystals, followed by extensive
prototype development (e.g. Voloshinov, 2004).
Preliminary research by Jelinski (2000) indicates that al-
kali halides, as CaF2, BaF2, LiF, MGF2, may be excel-
lent candidates for FUV tunable transmission crystal filters.
The transmission curves of these halides shift to longer
wavelengths as the temperature is increased and to shorter
wavelengths if the halides are cooled down (Davis, 1966).
This temperature dependent optical behavior can be utilized
in the design of tunable filters in two different ways, with
tunable bandwidth or tunable center wavelength.
2.5. Optical Components: Micro-mirrors, fibres
Very efficient spectroscopic observation can be done by ob-
serving multiple objects or positions simultaneously in the
focal plane. In ground based instruments more than 1000
objects can be observed. Micro-mirror arrays can be used
as programmable masks. In this arrangement, micro-mirrors
are reflecting/blocking unwanted light and open a path for
the desired rays. This concept shows great promise for UV
applications although it is in an early phase of development.
In another arrangement the incoming light is partly reflected
into a spectrograph (from the desired objects) and the other
part is reflected into a light baffle. To use micro-mirror ar-
rays in the UV wavelength regime, UV coatings should be
applicable to these arrays but it needs to be verified that
the mirrors can be manufactured with the required surface
quality.
In ground-based multiple objects instruments light-
transmitting fibres feed light from selected targets into a spec-
trograph. However, development of new fibres with excellent
transmission throughout the UV would greatly simplify this
concept since the entire focal plane would be accessible
to a single spectrograph with fibre reformatting. Unfortu-
nately optical fibres are opaque in the UV. Progress is made
for example with silica-core fibres (e.g. Wang, 2005) and
one type of fused silica fibres will transmit light to wave-
lengths as short as 180 nm. Developments have to be made
to create transmissive fibres for use at shorter wavelengths
down to 91 nm. Considerable development work also needs
to be done to couple these fibers into a powered optical
system, to bundle them, and to understand their flexibil-
ity properties and their behavior under extreme thermal
conditions.
3. Requirements for UV mission in the near future
The detailed needs for future UV space-born instruments
and the justification for the required parameters are given
for every science case described in the papers of this book.
High throughput is a request from all science cases (e.g S/N
=10 within 10 minutes for a target with a flux of 10−16
ergs cm−2sec−1) The specific requirements to the wave-
length range and the spectral resolution for spectroscopic
(R=λ/δλ) and imaging observations are summarized as
follows:
Springer

150 Astrophys Space Sci (2006) 303:147–151
1. Medium spectral resolution:
R≤5000 with a simultaneous wavelength coverage
from 91 nm to 450 nm.
Additionally a spatial resolution in the order of ∼0.01
arcsec is required.
2. High spectral resolution:
R≥30.000 with a simultaneous wavelength coverage
from 91 nm to 350 nm.
High spectral resolution of R∼50.000 or even R≥
100.000 is required for example for a galactic white
dwarf spectroscopic survey (Barstow and Werner, this
book), for resolving plasma velocities in a number of
environments such as cometary comae (Brosh et al. this
book), for measurements of thermal broadening of ab-
sorption lines in exospheres (Gomez et al., this book),
for studies of cool winds and astrospheres (Pagano et
al, this book) and for detailed ISM studies (Wamsteker
et al., this book).
3. Efficient spectroscopic observation should be performed
by integral-field spectroscopy, for example multiple ob-
ject spectroscopy of white dwarfs in the 90–130 nm
range with R ∼1.000 and a limiting flux of around
10−20 ergs cm−2sec−1(Barstow and Werner, this
book).
4. For imaging at least a spatial resolution of 0.01 arcsec
and field of views with up to 2 arcmin is required
(Brosh et al., this book). For a white dwarf survey of the
LMC/SMC and of globular clusters a field of view of
10 arcmin is required (Barstow and Werner, this book).
Additionally tunable filters will be a very good tool to
select spectral lines of interest.
5. Furthermore some science cases need time-tagged ob-
servations with a time accuracy down to fractions of a
second (e.g. G¨ansicke et al., this book).
Similar science questions given in this book can be
addressed with a planned 2 m – class UV mission, the
WSO/UV project. The spectroscopic capabilities of this mis-
sion are a spectral resolution of R50.000 and a cov-
erage of the whole wavelength regime(102–310 nm) with
two observations(102–178 nm and 275–310 nm). A de-
tailed overview of this project is given by Barstow et al.
(2003).
4. Summary
In summary it is shown, that with modest developments the
scientific objectives outlined in this book can be achieved
with 3–4 m class telescopes. Most of the described science
in this book can be achieved basically by two spectrographs,
which should have a simultaneous wavelength coverage from
90 nm to 450 nm and a high throughput.
1. A high spectral resolution spectrograh with R∼
50.000–100.000
2. A medium spectral resolution spectrograh with R∼
1.000–5.000 allowing integral field spectroscopy with
a spatial resolution of ∼0.01 arcsec
To fulfill these requirements for near future UV missions es-
pecially developments of microchannel-plate detectors and
semiconductor arrays with high quantum efficiency, large dy-
namic range, low background noise and with large formats
are necessary. This should be accompanied by improving the
technology of optical components to provide high through-
put and low scatter and by the development of large, preci-
sion, lightweight mirror surfaces with good micro-roughness
properties.
References
Hubble’s Science Legacy: Future Optical/Ultraviolet Astronomy from
Space ASP Conference Proceedings 291, (2002)
Barstow, M., et al.: Proc. SPIE 4854 364–374 (2003)
Beetz, C.P., Boerstler, R., Steinbeck, J., Lemieux, B., Winn, D.R.:
Nucl. Instr. and Meth. in Phys. Res., A442, Issue 1–3, 443–451
(2000)
Clampin, M.: UV-Optical CCDs, proceedings of the space astrophysics
detectors and detector technologies conference held at the STScI.
Baltimore, June 26–29 (2000)
Davies, R.J.: J. Opt. Soc. Am. 56, 837 (1966)
Jelinski. P., Siegmund, O.H.W., Welsh, B.: Narrow-band tunable fil-
ters for use in the FUV region. Proceedings of the Space Astro-
physics Detectors and Detector Technologies Conference held at
the STScI. Baltimore, June 26–29
Joseph, C.: UV Technology Overview, in proceedings of the space as-
trophysics detectors and detector technologies conference held at
the STScI. Baltimore, June 26–29 (2000)
Joseph, C.: Proceedings of the space astrophysics detectors and detector
technologies. ASP Conf. Ser. 164, 420 (1999)
Ulmer, M.P., et al.: Proc SPIE 4650 (2002)
Ulmer, M.P., Wessels, B.W., Han, B., Gregie, J., Tremsin, A., Sieg-
mund, O.H.W.: Advances in wide-bandgap semiconductor base
photocathode devices for low light level applications Proc. SPIE
5164, 144–154 (2003)
Ulmer, M.P.: Requirements and design considerations of UV and x-
ray detectors for astronomical purposes, Proc. SPIE 2999,p.259
(1997)
Ulmer, M.P., Razeghi, M., Wessels, B.W.: Development of GaN-based
films for use in UV sensitive but visible blind detectors in Pro-
ceedings of the Space Astrophysics Detectors and Detector Tech-
nologies Conference held at the STScI, Baltimore, June 26–29
(2000)
Siegmund, O.H.W.: MCP imaging detector technologies for UV in-
struments, in proceedings of the space astrophysics detectors and
detector technologies conference held at the STScI. Baltimore,
June 26–29 (2000)
Sheppard, S.T., Melloch, M.R, Cooper, J.A.: IEEE Electron Device
Letters 17, 4 (1996)
Voloshinov, V., Gupta, N.: Applied Optics IP, 43(19), 3901–3909 (2004)
Walker, D., Kumar, V., Mi, K., Sandvik, P., Kung, P., Zhang, X.H.,
Razeghi, M.: Solar-blind AlGaN photodiodes with very low cutoff
wavelength. Appl. Phys. Lett. 76 403 (2000)
Springer

Astrophys Space Sci (2006) 303:147–151 151
Wang, T., Guo, X., Chen, Z.: Proc. SPIE 5623, 145–150 (2005)
Welsh, B.Y., Kaplan, M.: NASA’s ultraviolet Astrophysics Branch: the
next decade, in EUV, X-Ray and Gamma-Ray intrumentation for
astronomy III 452–463 (1992)
Wilkinson, E.: Maturing and developing technologies for the next gen-
eration of UV gratings ultraviolet-optical space astronomy beyond
HST. Conference held at the STScI, ASP Conference Series 164,
420 (1999)
Springer

Astrophys Space Sci (2006) 303:153–170
DOI 10.1007/s10509-006-9047-6
ORIGINAL ARTICLE
Massive stars in the UV
F. Najarro ·A. Herrero ·E. Verdugo
Received: 14 February 2006 / Accepted: 27 February 2006
C
Springer Science +Business Media B.V. 2006
Abstract We emphasize in this paper the importance of the
UV range for our knowledge of massive stars and the funda-
mental role played by past and present space-based UV ca-
pabilities (IUE, HST, FUSE and others). Based on a review
of the work developed in the last years and the state of the art
situation for quantitative spectroscopy of massive stars, we
present crucial advances which could be addressed by hy-
pothetical future space-based UV missions. Advantages and
unique data that these missions could provide are explained
in the context of our present knowledge and theories on mas-
sive stars in the Milky Way and nearby galaxies. It is argued
that these studies are our key to a correct interpretation of
observations of more distant objects.
Keywords UV astronomy .Massive stars .Winds .
Abundances
1. Introduction
Massive stars and their descendants are important con-
stituents of galaxies. Because of their high luminosities (up
to 106L) and their massive winds (M=10−8to 10−4
F. Najarro ()
Departamento de Astrof´ısica Molecular e Infrarroja, Instituto de
Estructura de la Materia, CSIC, Serrano 121, 28006, Madrid,
Spain
e-mail: najarro@damir.iem.csic.es
A. Herrero
Instituto de Astrof´ısica de Canarias, C/ V´ıa L ´actea s/n,
E-38200 La Laguna, Tenerife, Spain ; Departamento de
Astrof´ısica, Avda, Astrof´ısico Francisco S´anchez 2, E-38271 La
Laguna
E. Verdugo
ESAC-ESA, P.O. Box 50727, 28080 Madrid, Spain
Myr−1,v
∞=100 to 2000 km s−1) they have an extremely
important influence on the dynamics and energetics of the in-
terstellar medium. They also enrich the interstellar medium
in nuclear processed material. This enrichment occurs via
mass loss (a massive star can lose 2/3 or more of its mass
via a stellar wind) or during a SN explosion. They directly
influence star formation by disrupting molecular clouds via
SN explosions, or conversely they can initiate star forma-
tion through massive wind-blown bubbles and SN shells
compressing nearby molecular clouds. Massive stars are
also thought to be responsible for the reionization of the
early Universe. More recently it has been proposed that
the most massive stars are the progenitors of gamma-ray
bursts.
Being crucial in many relevant aspects of astrophysics,
detailed knowledge of massive stars have been hampered
by the presence of strong stellar winds which dominate the
resemblance of the atmospheres, the yields of ionizing ra-
diation (Gabler et al., 1989; Najarro et al., 1996) and the
evolution of these objects, leading to significant modifica-
tions in their observable spectra. Thus, we find stellar wind
signatures at all wavelength ranges, from the UV ionized res-
onance transitions through the optical (Hα) to the Infrared and
Radio.
Nevertheless, the correct interpretation of their wind
lines in terms of radiation driven wind theory has pro-
vided the spectroscopist with a wonderfull tool to investi-
gate the physics of galaxies. As they can be observed in
medium resolution spectra as individual objects in galax-
ies out of the Local Group or as integrated spectra of
starburst regions in galaxies with significant redshifts (see
Fig. 1), the correct knowledge of the physics of mas-
sive stars will yield information about the energy budget
and chemical composition of galaxies along the cosmos
history.
Springer

154 Astrophys Space Sci (2006) 303:153–170
Fig. 1 UV astronomy as key to
understand high redshift
galaxies. Optical spectrum of a
galaxy at very high redshift
(z=2.732) compared to a local
starburst (adapted from
Kudritzki, 1998)
1.1. Importance of UV observations of massive stars
The UV constitutes an optimum spectral window as the spec-
tral energy distributions of massive stars reach their maxima
within this wavelength range. Appart from this efficient co-
incidence established by nature, massive stars decorate the
UV spectral region with a number of key diagnostics to our
understanding of the nature of these objects and their inter-
action with the surrounding media. The relevance of such
diagnostics in the UV can be easily recognized since:
– The presence of P-Cygni profiles produced by the reso-
nance transitions of C IV, N V, Si IV, O VI, etc, provides
unique information about the stellar winds associated with
massive stars.
– The bulk of blocking lines over several ionization stages
from elements from the iron group enable to estimate
metallicity and other stellar properties such as effective
temperature.
– The unsaturated line profiles from ionized species trace
very efficiently the mass-loss rate characterizing the stellar
wind. Further, when combined with ρ2sensitive diagnos-
tics at other wavelengths they may be used to calibrate the
presence of inhomogeneities (“clumping”) in the wind.
– They encompass enough number of X-Ray sensitive diag-
nostic lines such as O VI, C IV and N V.
– The optimized spatial resolution achieveval at UV wave-
lengths allows to resolve individual stars in starforming
regions in external galaxies.
In this paper we review recent progress within the field
of massive stars and discuss open problems which could be
addressed by future UV missions.
2. Temperature scale for massive OB stars
A calibration of the effective temperature of O supergiants
is a key issue for the correct description of the radia-
tion hardness in the EUV and UV spectral ranges and
the corresponding ionizing photons released by the stars.
This can have enormous consequences on the energy bud-
get of their surroundings, as well as of starburst regions
and galaxies whose UV spectra are characterized by these
stars.
Since NLTE models are available, the temperature scale
of massive OB stars has been determined using model atmo-
spheres to fit a given ionization equilibrium. Helium, being
the atom used to spectroscopically classify the earliest stars,
has become the traditional temperature indicator for these
objects, although other ionization balances are formally pos-
sible. Of course, the resulting temperature scale is model
dependent.
Very recently a number of calculations from different au-
thors have strongly changed the temperature scale of massive
OB stars. These calculations have been based on new families
of model atmospheres that include sphericity, mass-loss and
line -blanketing, in addition to NLTE. The combination of
all these factors results in temperatures that are much cooler
than those hitherto assumed.
Vacca, Garmany and Shull (1996) presented a compilation
of the spectroscopic determinations of effective temperatures
of massive OB stars. They gave preference to the most recent
calculations, that at that time were mostly based on plane
parallel, hydrostatic, unblanketed model atmospheres. Their
temperature scale for dwarfs and supergiants can be seen in
Fig. 2.
Springer

Astrophys Space Sci (2006) 303:153–170 155
Fig. 2 The temperature scale
for Galactic O dwarfs (left) and
supergiants (right). The solid
lines are for the Vacca et al.
(1996) scale, the dashed line for
the scale defined by Martins et
al. (2005) (the one the authors
define as the theoretical scale),
filled symbols are data from
Repolust et al. (2004) and open
symbols are data from Herrero
et al. (2002)
The first calculations pointing to a cooler temperature
scale were those from Martins et al. (2002), who used CM-
FGEN (Hillier and Miller, 1998), a code with all the im-
provements indicated above. These authors limited their
calculations to OB dwarfs, so that the influence of mass-
loss effects were negligible. Therefore, the main differences
with Vacca, Garmany and Shull were clearly due to line-
blanketing. These differences could reach up to 4000 K for
early types, and decreased towards O9 and B0 types, as can
be appreciated in Fig. 2.
The work from Martins et al. was followed by a series of
papers with similar results. Herrero, Puls and Najarro (2002)
gave a temperature scale for supergiants in Cyg OB2 us-
ing FASTWIND (Santolaya et al., 1997; Puls et al., 2005),
another code with NLTE, sphericity, mass-loss and (in this
case approximated) line-blanketing. They found differences
up to 8000 K. In this case both mass-loss and line-blanketing,
played a role. These authors also showed that two stars with
the same spectral type and luminosity class may have differ-
ent effective temperatures if their wind densities are different.
While the results from Herrero, Puls and Najarro were based
on analyses of only seven Cyg OB2 supergiants, Repolust,
Puls and Herrero (2004) presented an analysis of 24 stars
(17 giants and supergiants and 7 dwarfs), based on a slightly
improved version of FASTWIND that confirmed the same
trends. These same trends have also been confirmed by Mar-
tins, Schaerer and Hillier (2005) who have calculated models
with CMFGEN and have given a new temperature scale for
massive OB stars of different luminosity classes. These tem-
perature scales agree quite well with those from Repolust,
Puls and Herrero, and confirm that new models result in ef-
fective temperatures that are several thousands Kelvin cooler
for early and intermediate spectral types, decreasing towards
late spectral types. The different temperature scales can be
seen in Fig. 2
OB stars in the Magellanic Clouds have been analyzed
by Massey et al. (2004, 2005) using FASTWIND. Their re-
sults confirm that supergiants are 3000–4000 K cooler than
dwarfs of the same spectral type at any metallicity. How-
ever, a clear trend of temperature with metallicity at a given
spectral type and luminosity class is not seen: authors obtain
that SMC dwarfs and supergiants are hotter than Milky Way
counterparts, but while LMC dwarfs seem to extend the SMC
scale, LMC supergiants seem to extend the Milky Way scale.
In addition, SMC dwarfs of types O3–O4 do not fit into the re-
lation indicated by the later O types in that galaxy, but have
a temperature closer to their Milky Way analogues. When
comparing to other authors we find further inconsistencies:
O4O5 SMC dwarfs analyzed by Massey et al. (2004, 2005)
are hotter than Milky Way ones, but O4–O5 SMC dwarfs
analyzed by Bouret et al. (2003) are cooler than similar ob-
jects in our Galaxy. While the analyses by Massey et al.,
2004, 2005 constitute a major step forward to understand the
metallicity dependence of the temperature scale of massive
OB stars, the low number of objects and the large scatter in
the relations indicate the necessity of further analyses, par-
ticularly because the scatter could be related to differences
in wind density at a given spectral type.
We should finally mention the temperature scale ob-
tained by Bianchi and Garc´ıa (2002) and Garc´ıa and Bianchi
(2004). These authors analyse stars by means of UV spectra
(FUSE and IUE) using WM-basic (Pauldrach, Hoffmann and
Lennon, 2001). The temperatures they find are much cooler
than those from other authors. Wind clumping (see Section 5)
affecting their results is a plausible explanation still requiring
investigation.
For types O9–B0 there is no clear difference between tem-
perature scales from different authors or metallicities. It is
then not strange that a comparison of temperatures scales
for B supergiants results in no apparent difference between
Galactic (from McErlean, Lennon and Dufton (1999)), LMC
(Evans et al., 2004) and SMC analyses (Trundle et al., 2004;
Trundle and Lennon, 2005) (see Fig. 3; the figure includes a
few giants, but this does not change the conclusions that fol-
low). We can see in Fig. 3 that the scatter at a given spectral
type is very large. At spectral type B1, for example, we find a
difference of 3500 K between the hottest and coolest B1 su-
pergiants (both from the SMC), and the difference increases
to 4500 K if we include the hottest object, a B1 III star. This
large difference can be attributed to the different wind densi-
ties: the coolest B1 supergiant (AV78) has a mass-loss rate of
2.29 ±0.34 ×10−6Myr−1, while the hottest one (AV242)
has 0.84 ±0.13 ×10−6Myr−1.The larger radius of the
first star (79.0 versus 36.6 R) also contributes to a cooler
Springer

156 Astrophys Space Sci (2006) 303:153–170
Fig. 3 The temperature scale for early B supergiants in the Milky Way
and the Magellanic Clouds. Spectral types are coded so that 9 means
O9, 10 means B0 and so on. Solid circles are data by McErlean et al.
(1999) for the Galaxy; triangles are data from Trundle et al. (2004) and
Trundle et al. (2005) for the SMC; and squares are data from Evans
et al. (2004) for the LMC
temperature. For details of the analysis, see Trundle et al.
(2004).
2.1. Potential of UV temperature diagnostics
We have seen above the impact of the new generation of at-
mospheric models on the temperature scale of massive OB
stars. One of the immediate advantages of blanketed models
is the simultaneous use of different ionization equilibria to
determine the effective temperature of the star. Thus, appart
from the “traditional” He I/He II ionization balance, we may
utilize unsaturated C II–IV,NII–V,OII–V,SiIII–IV, etc diag-
nostic lines from which we, ideally, should obtain consistent
values with the helium ionization equilibrium. Besides, the
UV range encompasses a whole forest of iron group lines
and enables to use the strength ratios of combined features
of different ionization stages of the same ion to constrain
the effective temperature of the star. Recently, several stud-
ies have been performed identifying key diagnostic lines to
trace effective temperature of OB stars in the UV (Heap et al.,
2004, Bouret et al., 2005 and references therein). Fig. 4-left
shows the sensitivity of the so called WFE54 index to tem-
perature and gravity (Heap et al., 2004), where WFE54 rep-
resents a weighted ratio of the equivalent widths of the the Fe
Vand Fe IV around λ1370 and 1620˚
A respectively. We note
that, although this index displays a relatively high sensitivity
to stellar temperature, it also depends strongly on gravity.
Therefore, if gravity is determined from other diagnostics
(optical spectrum), this index constitutes an excellent tem-
perature indicator for Teff ≥35,000K. On the other hand,
the ratio of the carbon C III1175-IV1169 line strengths (see
Fig. 4-right) is essentially only dependent on the effective
temperature (Heap et al., 2004). The reader is referred to
Heap et al. (2004) for a detailed identification of temperature
UV diagnostics in O dwarfs.
For supergiants, the situation is very similar. However, the
stellar wind may play a fundamental role. Thus, the sensitiv-
ity of some strategic UV lines to effective temperature will
be shifted in the parameter domain compared to the O dwarfs
case. Fig. 5 shows how the C III1175 and Si IV1400 lines react
considerably to changes of barely 1000 K in a mid type O
supergiant.
We may then conclude that the new generation of blan-
keted models for massive stars provide powerfull diagnostics
to obtain reliable effective temperatures for massive OB stars
by means of ionization equilibria of metals other than H and
He, allowing this way direct estimates of temperature from
UV observations. This opens an important window to study
massive stars in star forming regions in external galaxies us-
ing the unique spatial and spectral capabilities of future UV
missions.
3. Abundances
With the advent of new blanketing codes, quantitative spec-
troscopic studies of the UV forest of metal lines has become
reality. These codes allow to fit not only the outshining UV
saturated and unsaturated profiles from metal ions but also
the underlying blanket of lines from iron group elements.
Therefore, the new generation of models provides, for the
first time, direct estimates for abundances of elements such
as C, N, O, Si or S by fitting individual unsaturated lines
while robust estimates of abundances of iron group elements
may be obtained fitting integrated features in strategical UV
wavelength regions. Of course, the analysis has to provide as
well stellar parameters such as Teff, stellar mass, M,L∗and
clumping.
Figure 6 displays the potential of new blanketed models
to perform quantitative analyses of UV spectra for massive
stars. It displays the excellent agreement for the Luminous
Blue Variable (LBV) P Cygni between the model and the
observed IUE SWP region (Najarro, 2001). Interestingly, we
see from Fig. 6 that the main contributor to blanketing in
this zone is Fe III. However, we can see as well that there
are some regions (e.g. 1420–1500 ˚
A, 1570–1730 ˚
A) where the
model clearly underestimates the blanketing. Before tenta-
tively blaming it on the behavior of the extinction law, it is
necessary to investigate the effects of other iron-group ele-
ments (Ni, Co, Cr) which are normally not included in many
new models from computational saving reasons. Recomput-
ing the same model but adding Ni and Co Najarro (2001)
found that, indeed, Ni III provides most of the missing blan-
keting as shown in Fig. 6. Further, Co III also contributes sig-
nificantly in the 1750–1800 ˚
A region. When these two species
are included, the agreement of the model with the observa-
tions is excellent throughout the whole IUE SWP range.
Springer

Astrophys Space Sci (2006) 303:153–170 157
Fig. 4 Temperature diagnostics
in the UV for O dwarfs. left Fe
V/Fe IV index as function of
temperature and gravity. Right C
III/C VI line sensitivity as a
function of effective temperature
(adapted from Heap et al., 2004)
Fig. 5 Temperature sensitivity of C III and Si IV in mid-type supergiants.
HST-spectra of CygOB2#11 and model fits with T=1000K
In O stars, the need of UV spectra to determine metal
abundances becomes a must as the optical spectra no longer
show the strong metal features present y B and A stars and
we run out direct diagnostics to constrain the abundance of
iron group elements. Fig. 7 displays the potential of the UV
to determine metal abundances in O type stars. We see from
Fig. 7-top the strong dependence with metallicity and spectral
type of the UV C IV line, while Fig. 7-bottom demonstrates
the possibility to estimate the iron abundance in O stars.
We also see that as metallicity decreases (Fig. 7-bottom)
the strengths of metal features in the UV become weaker,
and enhanced S/N is required to perform reliable metal
abundance determinations. Future missions with enhanced
UV sensitivity will play a key role in our ability to derive
accurate abundances in low-metallicity environments
4. X-rays
UV spectra of massive stars also trace the presence of X-
rays in their stellar winds. X-ray emission in the wind alters
significantly the wind ionization structure, enhancing the
populations of the so called “superions” such as N Vand
OVI. In fact, X-rays were utilized more than ten years ago to
explain the observed strength of O VIλ1036 in galactic and
LMC O supergiants Pauldrach et al. (1994). For O and B
dwarfs (low density winds) Macfarlane et al. (1994) showed
that the effects of X-rays were enhanced when compared to
supergiants (high density) as for lower wind densities we get
significantly less recombinations to compensate the Auger
ionization. They also found that these effects increased to-
wards later spectral types. This effect is driven by the drop
of the photospheric to X-ray flux ratio as the lower effec-
tive temperature for a later spectral type reduces drastically
the bolometric stellar luminosity. Recent studies, confirming
these findings have been carried out for a extended sample
of O stars by Bouret et al. (2003, 2005) and Martins et al.
(2005). Fig. 8 displays the crucial effects of including X-rays
on key diagnostic features such as N Vλ1240 and several Fe
IV lines around λ1640 for αCam, a late O supergiant (Na-
jarro et al., in prep). Given the distance limited sample of
massive stars from which current X-ray missions may pro-
vide direct measurements of X-rays, it is evident that future
UV missions with enhanced sensitivity will constitute our
ideal tools to probe the presence of X-rays in the winds of
massive stars in external galaxies.
5. Clumping
Recent evidence indicates that currently accepted mass-loss
rates may need to be revised downwards by as much as a
factor of ten or more, because the most commonly used mass-
loss diagnostics are severely affected by small-scale density
inhomogeneities (“clumps”) in the wind, redistributing the
matter into regions of enhanced and depleted, almost void
density.The amount of clumping is quantified by the so-called
clumping factor, fcl. Diagnostics sensitive to the square of the
density, ρ2, will tend to overestimate the mass-loss rate of a
clumped wind by a factor √fcl. Considering that numerous
Springer

158 Astrophys Space Sci (2006) 303:153–170
Fig. 6 Comparison of the
blanketed model (dashed) and
the averaged observed (solid)
UV IUE-SWP spectrum of P
Cygni. A model including
Nickel and Cobalt
(dashed-dotted) is also displayed
in the IUE-SWP region (adapted
from Najarro, 2001)
stellar evolution calculations have found that changing the
mass-loss rates of massive stars by even a factor of two has a
dramatic effect on their evolution (Meynet et al., 1994), it is
clear that such revisions would have enormous implications.
Indeed, strong clumping has been claimed to be present
in stellar winds of early type stars (e.g., Eversberg et al.,
1998; L´epine et al., 1999; Crowther et al., 2002; Hillier et al.,
2003; Bouret et al., 2003). Herrero, Puls and Najarro (2002),
Repolust, Puls and Herrero (2004) and Markova et al. (2004)
report that clumping may cause mass-loss rates for O-stars
with Hαin emission to be overestimated by factors of 2.5.
Detailed modelling of the UV and optical spectra of selected
O stars by Bouret et al. (2005) indicates that not only are the
winds strongly clumped, but that the clumping seems to begin
very close to the wind base so that all mass-loss rates may
need to be revised downwards, by factors of 7. Prinja et al.
(2005) analyzed unsaturated wind lines in lower luminosity
B supergiants and showed that their mass-loss rates may be
factors of 10 or more less than theoretical expectations.
A compelling, independent indication for clumping has
come from analyses of the Far-UV wind lines due to the P
vλλ1118,1128 doublet (Crowther et al., 2002; Hillier et al.,
2003; Bouret et al., 2005; Massa et al., 2003; Fullerton et al.,
2006), which has only become widely accessible since the
launch of FUSE. Because Phosphorus has a low cosmic
abundance, this doublet never saturates, providing useful
Springer

Astrophys Space Sci (2006) 303:153–170 159
Fig. 7 (Top) Metallicity
dependence of metal ions UV
lines. Compilation of UV C iv
stellar wind profiles of O dwarfs
and giants in the Galaxy, LMC
and SMC (adapted from
Kudritzki, 1998). (Bottom)
Model fits with different
metallicities of the observed UV
Fe iv spectrum of an O star in
the SMC (adapted from Haser
et al., 1998)
Fig. 8 Effects of X-Rays in
α-Cam O9If. The strong
changes induced by including
X-rays (dashed-dotted line) in
the ionization structure of N and
Fe in the wind are reflected in
the observed N Vλ1240 and Fe
IV lines around λ1640
estimates of Mfor those cases where the ionization frac-
tion is computed consistently (Crowther et al., 2002; Hillier
et al., 2003; Bouret et al., 2005) or to Mq, where qis the ion-
ization fraction of Pv. At least for mid-O stars Massa et al.
(2003) and Fullerton et al. (2006)) have shown that P4+is a
dominant ion. Therefore, for these stars the P vλλ1118,1128
doublet constitutes a usefull constrain to determine Mitself.
This is demonstrated in Fig. 9 which displays model fits from
Crowther et al. (2002) to a SMC O supergiant for different
combinations of clumping, mass-loss rate and phosphorus
abundance. Interestingly, the mass-loss rates derived by all
these authors turned out to lie considerably below those in-
ferred from other diagnostics.
The most reasonable way to resolve this discordance, un-
less nature does not like phosphorus to participate in massive
stars, is to invoke extreme clumping in the wind. In a clumped
wind, the continuum and Hαmeasurements (both sensitive
to ρ2) overestimate the actual M, whereas the line strengths
of PV, being ∝ρ, are not directly affected by clumping. If
the winds of OB stars are indeed substantially clumped, then
Springer

160 Astrophys Space Sci (2006) 303:153–170
Fig. 9 Clumping diagnostics in O stars. The P vλλ1118, 1128 as tracer
of low M (adapted from Crowther et al., 2002)
the actual mass-loss rates are much lower than previously
thought, by a factor of 10 or even more. Further, in cases of
strong clumping the ionization structure of relevant species
may be altered as well affecting key diagnostic lines at all
wavelength ranges.
The downward mass-loss rate revisions suggested above
would have dramatic consequences for the evolution of and
feed-back from massive stars. Therefore, it is important to
have robust and precise determinations of the mass-loss rates.
All clumping diagnostics are subject to some degree of uncer-
tainty which can be reduced by combining suitable diagnos-
tics, scanning different portions of the wind, from close to the
base (Hα) and near IR lines (Najarro et al., 2004) over inter-
mediate regions (Brα, mid-/far-IR continua) to the outermost
(radio) region. A consistent analysis will severely constrain
the radial stratification of the clumping, expressed in terms
of the clumping-factor, fcl(r). If used in combination with
other diagnostics from (F)UV wind lines and state-of-the-
art model atmospheres allowing for a decent description of
the required ionization balance, the “true” mass-loss rates
can be uniquely derived. Fig. 10 shows that, appart from
thePvλλ1118,1128 doublet there are other UV lines from
different metals which can trace very efficiently the presence
of clumping, especially for those regions of the parameter
domain where the lines are unsaturated. Therefore, future
UV missions with enhanced sensitivity will provide enough
number of observations of OB stars over different spectral
types and allow this way, to constrain the presence of clump-
ing in these objects.
6. The wind momentum luminosity relation (WLR)
The winds of massive OB stars are driven by the absorption
of photons from the radiation field by thousands of spectral
lines from many atomic elements. These photons accelerate
the atoms towards the empty surrounding space in such a
way that the atom velocity increases with the distance to the
stellar surface, until the material no longer absorbs them.
Since the winds of hot stars are driven by radiation, we
expect a tight relationship between the mechanical momen-
tum of the stellar wind and the photon momentum. Actually,
the theory of radiative driven winds (Castor et al., 1975;
Kudritzki et al., 1989) predicts that the “modified stellar
wind momentum” depends directly on luminosity through
the Wind Luminosity relation (WLR)
logDmom =logD0+xlog(L/L)Dmom
=˙
Mv∞(R∗/R)0.5(1)
where the coefficients D0and xare a function of spectral type
and luminosity class. Further, the coefficient controlling the
dependence on luminosity, x, is determined by the statistics
of the thousands of metal lines driving the wind, so a differ-
ent WLR has to be established for each metallicity environ-
ment. The weak dependence on stellar radius arises from the
competition of the accelerated stellar wind against the grav-
itational potential. We immediately see that, once calibrated
on stars with known distance, the WLR may constitute a
powerfull indicator of extragalactic distances (see Kudritzki
and Puls, 2000, for a thoroughfull discussion). During the
last decade an enormous effort has been made to calibrate
the WLR as function both of spectral type and metallicity
and compare the results with the predictions by theory (Puls
et al., 2000; Vink et al., 2000, 2001).
Fig. 11 shows the different relationships obtained by
Kudritzki et al. (1999) for different spectral types in the
galaxy. Of concern is the shift by more than one decade in
modified momentum between the relation of early and mid
B-supergiants. Though not totally, blanketing may provide
the key to partially reduce the discrepancy between the ob-
served WLR for O and early B supergiants and that obtained
for mid Bs. In fact, Urbaneja (2004) found a significant cor-
rection on the modified momenta of these stars with respect
Springer

Astrophys Space Sci (2006) 303:153–170 161
Fig. 10 Unsaturated UV lines
as tracers of clumping in O stars
(adapted from Bouret et al.,
2005)
to the values derived by Kudritzki et al. (1999) as shown in
Fig.12-right.
Results on the metallicity dependence of the WLR for O
stars and B supergiants and their comparison with theoretical
predictions are displayed in Fig. 12. From Fig. 12-right we
see how the modified momenta of galactic B-supergiants ob-
tained with blanketed models (Urbaneja, 2004) agree much
better with theoretical predictions than those obtained with
unblanketed models (Kudritzki et al., 1999), although a sig-
nificantly different slope is obtained. The whole worsens in
the case of the SMC, where the WLR obtained from blan-
keted models (Trundle and Lennon, 2005) differs severely
from theoretical predictions (Vink et al., 2001). In the case of
O stars, the situation is significantly improved. From Fig. 12-
left (see Massey et al., 2005, and references therein) we
see how the theoretically predicted WLRs for 0 stars in the
Galaxy, LMC and SMC are reasonably well reproduced by
spectroscopic studies.
The WLR provides a strong test for the theory of radia-
tively driven winds, and the UV provides the best region
to test the WLR, as it allows derivations of the mass loss
rate, the terminal velocity and the metallicity. Stellar param-
eters can also be derived from the UV (Bianchi and Garcia,
2002; Garcia and Bianchi, 2004), although further work is
needed for some of them (mainly gravity and radius). What
we need here are observations of high resolution and high
S/N at different metallicities. Inspection of figures 11 and 12
tell us that the number of observed stars is too low to pro-
vide a really constraining test of the theory. The Magellanic
Clouds, together with the Milky Way, provide a wide range
in metallicity, and other galaxies in the Local Group offer ad-
ditional opportunities. Thus, M31 may extend the observed
metallicities towards supersolar values, Sex A or Leo A may
extend them below SMC values and M33 offers a system
Fig. 11 The WLR as a function of spectral type. Galactic WLR for
O, B and A supergiants (Kudritzki et al., 1999). Note the considerable
offset between the WLR for early and mid B supergiants
with a strong metallicity gradient seen nearly face-on. The
ability to get UV spectra for stars in all Local Group galax-
ies is the key to our correct understanding of the spectral
type and metallicity dependence of the WLR. To that end,
an UV telescope with more collecting power than HST is
required.
6.1. The thin winds problem
Puls et al. (1996) showed that the observed WLR for O-
type dwarfs exhibited a severe curvature toward very low
wind momenta for luminosities lower than logL/L=5.3.
Kudritki and Puls (2000) reviewed three effects present in
the physics of thin winds in dwarfs which may introduce
deviations from the standard WLR. The first one is re-
lated to the decoupling of metal ions with the rest of the
Springer

162 Astrophys Space Sci (2006) 303:153–170
Fig. 12 Metallicity scaling of the WLR for O and B supergiants.
(left) Z scaling for O supergiants (adapted from Massey05). (Right)
Z scaling for B supergiants (Kudritzki et al., 1999; Urbaneja, 2004;
Trundle and Lennon, 2005). (dash-dot): linear regression to galactic
stars with blanketed models (Urbaneja, 2004); (dash-dot): unblanketed
galactic (Kudritzki et al., 1999); (solid): blanketed SMC (Trundle and
Lennon, 2005); (upper dotted) theoretical galactic predictions (Vink
et al., 2000); (lower dotted) theoretical SMC predictions (Vink et al.,
2001). Note the considerable correction to the modified wind momenta
for B-supergiants due to line blanketing
plasma. This will happen if the density falls bellow the
range for which coulomb-collisions are able to redistribute
the photon momentum absorbed by the metal ions to the
bulk of wind plasma (Springmann and Pauldrach, 1992;
Babel, 1995). The second one is caused by the shadow-
ing of photospheric lines which considerably lower the line
force, resulting on a net reduction of the mass-loss rate, espe-
cially in the case of B-dwarfs (Babel, 1996). Finally, for low
mass-loss rates where the continuum is thin throughout the
transonic region, curvature terms of the velocity field may
lead to line-accelerations much smaller than in the standard
computations, resulting again on reduced mass-loss rates
(Puls et al., 1998; Owocki and Puls, 1999))
Recent spectroscopic analysis using unified models seem
to confirm the presence of this turnover. Herrero, Puls and
Najarro (2002) obtained a very low value for the mass-
loss rate of the O9.5V star lOLac. Their result, displayed
in Fig. 13 showed that in order to match the observed UV
spectra of the star, the mass-loss rate hat to be reduced more
than one order of magnitude bellow the theoretical predic-
tion. It must be stressed that, although the presence of X-
rays could introduce a significant uncertainty in the absolute
M value, the upper limit is still well bellow the predictions
from radiatively wind theory (Vink et al., 2000). Recently
Bouret et al. (2003) and Martins, Schaerer and Hillier 2004,
2005 have obtained similar results on analysis of O dwarfs
in the SMC and the Galaxy (see Fig. 14). One important
conclusion from the above studies is that the discrepancy
in the mass loss rates obtained seems not to be related to
Fig. 13 The weak wind problem in O dwarfs. Model fits with different
mass-loss rates of the UV C IVλ1550 for the O9V star 10 Lac. ˙
Mvalues
are 1,2 and 8 ×10−10 Myr−1
metallicity, as is present on both the Galactic and SMC
stars.
Figure 14 clearly shows a breakdown of the WLR for low
luminosity O dwarfs. Although present results based on UV
studies may suffer from effects such as X-rays, advection or
adiabatic cooling, results from other indicators like Hαare
not appropriate because of insensitivity to very low mass-loss
rates. Moreover, other indicators in the IR and radio may be
affected by clumping. Detecting this clumping in thin winds
may be extremely difficult, and therefore high resolution,
high S/N in the UV becomes the best (if not the only) way
to derive the correct mass loss rate.
Springer

Astrophys Space Sci (2006) 303:153–170 163
Fig. 14 The problem of thin winds. (Left) Mass loss rates derived for O
dwarfs in NGC 346 (Bouret et al., 2003, LMC). (Right) Modified wind
momenta as a function of stellar luminosity for O stars (from Martins
et al., 2005). Filled (open) symbols are Galactic (LMC, SMC) stars.
Triangles and stars (squares, circles) correspond to luminosity class V
(III,I). Note the low momenta of the SMC objects and the galactic star
lOLac (large triangle)
7. Wind terminal velocity
In the previous section we have seen that winds in massive
OB stars are driven by photon absorption by numerous metal
spectral lines. Atoms are accelerated until they cannot absorb
photons anymore. From that moment on, atoms move freely
into space asymptotically approaching a maximum velocity,
which is formally reached at infinity. This is the so-called
wind terminal velocity,v∞. While the structure of the velocity
field resulting from this process is very difficult to determine,
the terminal velocity can be derived in a comparatively easy
way.
Determination of the terminal velocity is one of the key
aspects in studying the stellar wind. As explained in the in-
troduction to this chapter it enters the expression for the mod-
ified Wind momentum-Luminosity Relationship (WLR), the
product of the wind momentum times the square root of the
radius, which is proportional to a power of the stellar lumi-
nosity. Therefore, reproducing the correct terminal velocity
with models is crucial to obtain the correct mass-loss rate for
a given luminosity, and to use the modified WLR as distance
indicator. Moreover, one of the strong predictions of the the-
ory of radiatively driven winds is the metallicity dependence
of the terminal velocity. Therefore, a determination of v∞in
regions of different metallicity constitutes a strong test for
the theory.
Some of the spectral lines in the winds of OB stars are so
optically thick, that they absorb photons even at large dis-
tances, when the terminal velocity has been reached and the
wind density has fallen by orders of magnitude. We should
note that these absorbed photons do not accelerate the wind
significantly further, as their number and integrated momen-
tum is very low compared to that of the whole wind. But
leave their signature in the stellar spectrum, an thus we can
derive the terminal velocity.
Atoms moving at a given velocity in the wind will absorb
photons shifted to the blue with respect to their laboratory
wavelength (at which they have been emitted at the stellar
surface). Atoms moving at the largest velocity reached by the
wind will absorb photons with the largest blueshift. There-
fore, determining the maximum blueshift in an optically thick
line we will obtain the wind terminal velocity. Of course, the
number of absorbing atoms has to be large enough to absorb
photons at the low densities present in the layers moving
at the terminal velocity. Blue saturated P-Cygni profiles are
therefore ideal to determine the wind terminal velocity as
they fulfill this requirement. In addition, a number of effects
that we describe below have to be taken into account.
While we could directly try to derive the terminal velocity
from this maximum blueshift, the fitting of the whole P-Cygni
profile formed in the wind gives us much more information.
Fortunately, there is a simple method that allows us to calcu-
late this profile in an approximated way, but precise enough
to derive a number of interesting physical magnitudes, like
the terminal velocity or the number of absorbing atoms. It
is the Sobolev plus Exact Integration (SEI) method. We use
here the formulation as described by (Herrero et al., 2001)
(see also Haser, 1995, Lamers et al. (1999)). The velocity
stratification is usually parameterized with a β-law,
w(x)=(1 −b/x)β(2)
with
w(x)=v(x)
v∞
,x=r/R∗,b=1−w1/β
min 
Springer

164 Astrophys Space Sci (2006) 303:153–170
Fig. 15 Effects of the global
parameters on the SEI synthetic
profiles of the strongest UV
wind lines in OB stars. X-axis
units expressed as velocities, in
units of the wind terminal
velocity, referred to the rest
wavelength of the blue
component of each doublet. In
each plot, three synthetic
profiles are shown, one that
defines the central value (solid
line) and two others with a
variation of ±10% (dashed
lines) of the central value. From
Urbaneja (2004)
R∗being the stellar photospheric radius and wmin the ratio of
the velocity at x=1 to the terminal velocity, fixed at a value
of 0.01.
This is consistent with more exact hydrodynamical calcu-
lations. However, the indetermination in the exponent of the
velocity field, β, produces an additional uncertainty in the
terminal velocities.
When deriving wind terminal velocities, it is important to
account for the velocity dispersion vturb (usually termed as
“turbulent velocity”) present in those winds, to correctly re-
produce the position of the emission peak, the blue through
and the slope of the blue absorption, as originally proposed
by Hamann (1981). Following (Haser, 1995) we adopt a pa-
rameterization of the form
vturb =atv(r)+bt,(3)
i.e., the turbulent velocity is assumed to be (roughly) propor-
tional to the local wind-speed v(r), and the coefficients are
defined by
at=vta −vti
1−vmin
;bt=vta −at,
where vti =vturb(v=vmin ) is the minimum turbulent veloc-
ity (chosen to be of order sound-speed) and vta =vturb(v=
v∞) the maximum one. These parameters affect the overall
appearance of all synthetic lines. Fig. 15 shows their effects
on the synthetic profiles for a saturated case.
We also have to correct for the underlying photospheric
components, which we do in an approximate way, by using
IUE spectra of hot stars with weak winds and low projected
rotational velocities as templates. Selection of photospheric
template and continuum rectification has a big impact when
fitting emission peaks, but has little effect in the determined
terminal velocities. For Galactic stars a sample of Milky Way
dwarfs with appropriate spectral types is usually selected
(i.e., from the INES database). Of course, when working with
Springer

Astrophys Space Sci (2006) 303:153–170 165
Fig. 16 Fit to the C IV profile in Cyg OB2#10. The terminal velocity
v∞is 1650 km s−1and the turbulent velocity has been set to 300 km
s−1. The exponent βin the velocity law is 0.8 (from Herrero et al., 2001)
extragalactic stars we have to use photospheric templates
with the correct metallicity. Differences in metallicity may
be a small source of error.
Because terminal velocities are derived from broad ab-
sorption features we can use modest resolution to obtain
them. This allows us to look at relatively faint stars. These
may be stars in other galaxies or Galactic stars that are opti-
cally bright but UV faint because of extinction in the Galactic
plane. This extinction limits seriously our ability to observe
massive stars in very young Galactic clusters.
Figures. 16 and 17 give an example of profile fitting for
Cyg OB2#10, an 09.5 I Galactic star, and M33-0900, a M33
B0.5-B1 Ia star. Both spectra have been observed with HST-
STIS at an approximate resolution of 1200 during one and
half hours, reaching the same S/N ratio (25). In both cases
the wind terminal velocity was derived with an accuracy of
5%, although Cyg OB2#10 is a relatively bright star with
V= 9.88 and M33-0900 has V= 17.3. This accuracy is of the
order of that given by the radius term in the WLR expression
(the total error is actually dominated by the mass-loss rate
uncertainty).
These numbers therefore may be taken as a lower limit
in terms of v∞observations, and indicate that most of the
early OB stars in young Galactic star-forming (or recently
star-formed) regions arc hidden to us by dust obscuration in
the UV. In Cyg OB2, for example, one of the richest Galactic
OB associations, only six O stars have been observed in the
UV with HST, and only another 3 might be observed under
the same conditions.
The number of stars for which v∞has been determined
is therefore small. For example, Howarth and Prinja (1989)
list 203 Galactic O stars that were observed with IUE and
for which they determined the wind terminal velocity. This
work still constitutes the main source of data for Galactic
v∞, but only a few stars belong to a given cluster. More
recently, Prinja and Crowther (1998) list terminal velocity
determinations for 31 stars between O3 and B1 in the LMC
(including six O3 stars in R136) and 9 stars in the SMC.
More scarce are data for M31 and M33 stars. Bresolin et al.
Fig. 17 Fit to the C IV profile in M33-0900. The terminal velocity v∞
is 950 km s−1and the turbulent velocity has been set to 170 km s−1.
The exponent βin the velocity law is 1.0 (from Urbaneja et al., 2002)
(2002) and Urbaneja et al. (2002) list six early B supergiants
each, in M31 and M33 respectively.
While this allows us to study the behavior of the terminal
velocity within our Galaxy, it is barely enough to study its
behavior as a function of the spectral type and luminosity
class for a given metallicity. or to study it within a single
young cluster.
With a sensitivity 20 times larger, as expected for new UV
missions we could see a qualitative change in the present
situation. In our Galaxy we could gain a few magnitudes,
depending on extinction conditions, reaching deeper in the
dusty young clusters. However the real gain would be in the
extragalactic stars. Even at the distance of M31 and M33 we
could observe the O dwarfs and determine their terminal wind
velocities. With modern ground-based optical telescopes we
can secure their optical spectra and derive their stellar param-
eters. This would allow us to dissentangle the v∞– spectral
type – metallicity relation, providing us with a very tight test
for the theory of radiatively driven winds.
8. A-type supergiants in the ultraviolet
A-type Supergiants are evolved massive objects (∼9−
25M) located in a region of the H-R diagram where evo-
lution is very rapid. Therefore, they are few in number:
only about one hundred galactic stars are classified as such.
Among these supergiants there is a clear gap between spec-
tral types A5 and F0, which could be related to the evolution
of these objects. The evolutionary stage of A-supergiants is
still unclear. Although the most extensive work on their abun-
dances suggests that these stars have evolved directly from
the main sequence (Venn, 1995a), the uncertainties in these
studies are very high, due to the well recognized difficulty
of modeling their atmospheres (Venn 1995b; Verdugo et al.
1999a). A recent work by Przybilla et al. (2005) analyses the
various effects involved in the modelling of A-supergiants
atmospheres (line blanketing, non-LTE, helium abundances,
spherical extension, velocity fields, variability). In this work,
Springer

166 Astrophys Space Sci (2006) 303:153–170
accurate stellar parameters are determined from a hybrid non-
LTE spectrum synthesis technique for four BA supergiants.
From the abundances analysis these authors found that three
of the stars studied appear to have evolved directly from the
main sequence but for the AIb star, ηLeo, a blue-loop sce-
nario is derived.
A-type Supergiants are intrinsically the brightest stars
at visual wavelengths, and therefore the best potential ex-
tragalactic distance indicators using the wind momentum-
luminosity relation (Kudritzki et al., 1999). This relation is
derived from the radiation-driven wind theory and mainly
based on Balmer line fits.
Radiation pressure is adopted as the dominant driving
mechanism for the mass loss of A-supergiants. Unified wind
models, which include a solution of the spherical trans-
fer equation in the comoving frame and a non-LTE treat-
ment of hydrogen and helium, were developed by Santolaya
et al. (1997). These models succeed in fitting a num-
ber of profiles of the Balmer series for the brightest A-
supergiants (Kudritzki et al., 1999). but cannot reproduce
neither all the Hαprofiles observed, nor their variability.
An even more sophisticated model developed by Aufden-
berg (2000), which includes non-LTE line blanketing for
several metallic lines, failed to fit a typical HαP-Cygni pro-
file for the brightest A-supergiant Deneb (Aufdenberg et al.,
2002).
Stellar winds in A-type supergiants can be studied us-
ing the optical spectrum or/and the ultraviolet (UV) spec-
tral range. In the optical all lines seem to be photospheric
except Hαwhich shows a variety of very different profiles:
symmetric absorption, P-Cygni, double-peaked or pure emis-
sion profiles (Verdugo et al., 1999). It is in the UV range
and particularly in the ultraviolet resonance lines where the
presence of a stellar wind is cleanly traced. However, com-
pared to the amount of work devoted to OB stars, the UV
spectra of A-supergiants have scarcely been examined (e.g.
Lamers et al., 1995, 1978; Praderie et al., 1980; Underhill
and Doazan 1982; Hensberge et al., 1982). The most com-
prehensive studies of A supergiants in the ultraviolet range
were performed by Talavera and G´omez de Castro (1987) and
Verdugo, Talavera and G´omez de Castro (2006, 2003, 1997)
from the observations taken by the International Ultraviolet
Explorer (IUE) satellite.
The UV spectrum of A-supergiants is characterized by the
presence of variable discrete absorption components (DACs;
see some examples in Verdugo et al., 1999b) associated with
the resonance lines of different ions, mainly Mg II,AlII,
Si II,CII and Fe II. The appearance of these DACs is also
related to the luminosity of the star. In Fig. 18 we show three
typical observed Mg II[uv1] profiles: symmetric absorption
profiles in Ib stars, profiles formed by several components,
and a classical radiative-wind profile (without emission) in
the Ia and Iab stars. The same behavior is observed in the
other lines cited above as is also shown in Fig. 18 for the Fe
II[uv1] lines.
It may therefore seem that the less luminous A-supergiants
do not show any perceptible trace of mass motion in their
spectrum, but a variability analysis showed the presence of
DACs in the ultraviolet Mg II[uv1] lines of two Ib stars, which
indicates that mass outflow exists. DACs in the UV spectrum
of A-supergiants were initially found only in the brightest
A-supergiants. The time scales of variability of these com-
ponents are of the order of several months. However, a mon-
itoring programme performed with the IUE satellite in two
Ib A-supergiants revealed the appearance and evolution of a
single blueshifted component in a much shorter time scale
(∼1 month; see Fig. 19).
The DACs are stronger and more steady in luminous A-
supergiants, whereas the Ib stars exhibit these features in a
smoother but more variable way.
These two groups are also found from the analysis of the
optical spectrum of A-supergiants (mainly from the Hαpro-
file; Verdugo et al., 1999a, 2003). The existence of these two
groups must lie in a different extension, density and proper-
ties, in general, between the envelope of the Ala/lab super-
giants and the one of the AIb supergiants. These differences
suggest a different evolution history of these stars. In fact,
Przybilla et al. (2005) found from the abundance analysis of
four A-type supergiants (one Ib star and three Ia/lab stars), a
blue-loop scenario for the only Ib star studied because of a
first dredge up abundance ratios, while the other three objects
appear to have evolved directly from the main sequence.
Another very interesting finding from the UV spectrum
of A-supergiants which can also be linked with the evolution
history of these stars is that there are some luminous stars
which present a shortward shifted component at high velocity
(∼−150 km s−1) but there is not a component at zero velocity
(except the interstellar components of the resonance lines)
or this component is less intense than the high velocity one.
Such behavior has been detected in the Fe II lines of some
bright A-supergiants (see Fig. 8). In principle, the absence
of a component at 0 km s−1could be due to the fact that the
lines of Fe II are only formed in the wind, which would have
a lower degree of ionization than the photosphere. However,
this phenomenon is only detected in a few stars of our sample.
Most of the A-supergiants show a zero velocity component
for the lines of this ion. It is possible that this component
at0kms
−1is formed also in the wind. In a low density
envelope where shocks could be occurring, the spectra would
present a pre-shock component (high velocity) and a post-
shock component (low velocity). Therefore, the presence or
not of such zero velocity component would be related to the
density of the wind.
One of the main predictions of the radiatively driven wind
theory is that the terminal velocity of the wind should in-
crease with the escape velocity of the star. However, as shown
Springer

Astrophys Space Sci (2006) 303:153–170 167
Fig. 18 Mg II[uv1] (top) and Fe
II[uv1] (bottom) in
A-supergiants.
in Fig. 21, the opposite behavior is found in A-supergiants
(Talavera and G´omez de Castro, 1987; Verdugo et al., 1997,
2003).
The terminal velocity, v∞, is the mean velocity reached
by wind material in regions far away from the star, where
acceleration has effectively ceased but interaction with the
interstellar medium has not yet become important. The ter-
minal velocities of the winds of A-type supergiants can be
measured directly from the UV P Cygni profiles.
Traditionally, the terminal velocity of a stellar wind has
been observationally defined as the modulus of the largest
negative velocity seen in absorption in the P Cygni profiles of
UV resonance lines. For a P Cygni profile with a deep absorp-
tion (saturated profiles) through and a nearly vertical violet
edge, the measured edge velocity (vedge) was considered the
terminal velocity of the wind (Abbot 1978). However chaotic
motions in the winds may extend and soften the vertical vi-
olet edge resulting in an overestimation of v∞. The differ-
ence vedge −v∞arising from a local velocity field, which has
been parameterized as “microturbulence” by Hamann (1980,
1981) and Groenewegen et al. (1989). However, Howarth and
Prinja (1989) demonstrate that for saturated ultraviolet line
profiles the maximum velocity at which zero residual in-
tensity is recorded (vblack) provides an accurate measure of
the wind terminal velocity. For stars without saturated pro-
files, but with identifiable discrete absorption components,
the final central velocity reached by these components, vDAC(t
→∞), also provides a good indicator of v∞(e.g. Howarth
and Prinja, 1989). However, estimating this quantity observa-
tionally requires frequent UV spectra taken over a sufficiently
extensive period, and such a data are only available for a very
few stars. Therefore from a single UV spectrum DACs only
provide a lower limit to v∞.
In order to determine terminal velocities of A-type super-
giants is required to analyze several UV spectra taken over a
large period. The wavelength or velocity, vedge, where the vi-
olet edge of the Mg II profiles reach the continuum provides
an upper limit for the terminal velocities while in most cases
the vDAC is a lower limit.
Radiative winds are known to be unstable against small
perturbations of the radiative force. However, such small per-
turbations in the wind cannot account for the aforementioned
spectral features. The existence of magnetic fields is a more
viable option and has been suggested by many different au-
thors to explain the observations: (1) co-rotating interaction
regions have been suggested to explain the presence of the
UV DACs (Mullan 1984), (2) Co-rotating weak magnetic
surface structures could explain the observed Hαvariability
Springer

168 Astrophys Space Sci (2006) 303:153–170
Fig. 19 Cross-correlation function for the Mg II resonance lines of
HD46300 (top) and HD87737 (bottom). It is clear The appearance and
evolution of a blueshifted component (∼−200 km s−1) which migrates
through the symmetric profile of the line.
Fig. 20 Fe II [uv62],[uv63] lines in two A-supergiants showing no
components at rest (dashed vertical line). The position of a high velocity
component (∼−150 km s−1) is marked with a solid vertical line
(Kaufer et al., 1996), and (3) The existence of extended cool
loops could account for another phenomenon observed in BA
supergiants: High Velocity Absorptions (HVA; see Israelian
et al., 1997). All these facts have motivated us to under-
take a search for magnetic fields in the atmospheres of A-
supergiants (Verdugo et al., 2005, 2003). Spectropolarimetric
techniques has been drastically improved in the last few years
allowing to detect weak magnetic fields (of a few hundred
gauss) in massive stars. Magnetic fields have been discovered
in 5 OB stars (see Henrichs et al., 2005 for a recent review).
Specific behavior of variable stellar wind lines belongs to the
well-known indirect indicators of a magnetic field in early-
type stars. Typical cyclical variability of the DACs associated
Fig. 21 Terminal velocity of the wind vs. escape velocity for A-type
Supergiants. Empty circles correspond to stars for which the measured
terminal velocity is uncertain due to the particular shape of the Mg
II[uv1] lines
to UV lines are thought to be caused by magnetic fields at the
base of the flow. Henceforth, analysis of the UV spectra vari-
ability is crucial to identify potential magnetic massive stars.
UV high resolution spectroscopy is therefore instrumental
to make progress in the different open questions addressed
above:
Is the radiation driven wind theory fully applicable to A-
type supergiants? In addition, it seems to exist two differ-
ent groups of A-type supergiants based on luminosity class
but only a few UV spectra of luminosity class Ib stars are
available. Therefore, high resolution spectra are needed to
measure reliable wind parameters and to confirm the possi-
ble existence of two different types of A-supergiants. High
resolution spectra are also needed to confirm the lack of UV
Fe II resonance lines at rest in some of the brightest stars.
Moreover, studies of UV variability are decisive to analyse
the stellar winds properties, as well as the relevance of surface
magnetic fields.
Acknowledgements
We would like to thank M. Urbaneja for providing some of
the figures. F.N. acknowledges PNAYA2003-02785-E grant
and the Ramon y Cajal program. This work has been partly
supported by the Spanish MEC through PNAYA projects
AYA2004-08271-C02-01 and 02.
References
Aufdenberg, J.P.: Ph.D. Thesis, Arizona State Univ. (2000)
Aufdenberg, J.P., Hauschildt, P.H., Baron, E., et al.: ApJ 570, 344 (2002)
Springer

Astrophys Space Sci (2006) 303:153–170 169
Babel, J.: Astronomy and Astrophysics 301, 823 (1995)
Babel, J.: Astronomy and Astrophysics 309, 867 (1996)
Bianchi, L., Garc´ıa, M.: The effective temperatures of Mid-O stars.
Astrophysical Journal 581, 610 (2002)
Bouret, J.C., Lanz, T., Hillier, D.J. et al.: Quantitative Spectroscopy of
O Stars at Low Metallicity: O Dwarfs in NGC 346. Astrophysical
Journal 595, 1182 (2003)
Bouret, J.-C., Lanz, T., Hillier, D.J.: Lower mass loss rates in O-
type stars: Spectral signatures of dense clumps in the wind of
two Galactic O4 stars. Astronomy and Astrophysics 438, 301
(2005)
Bresolin, F., Kudritzki, R.P., Lennon, D.J. et al.: Space Telescope Imag-
ing Spectrograph Ultraviolet Spectroscopy of Early B Supergiants
in M31. Astrophysical Journal 580, 213 (2002)
Castor, JI., Abbott, D.C., Klein, R.I.: Radiation-driven winds in of stars.
Astrophysical Journal 195, 157 (1975)
Crowther, P.A., Hillier, D.J., Evans, C.J., et al.: Astrophysical Journal
579, 774 (2002)
Evans, C.J., Crowther, P.A., Fullerton, A.W. Hillier, D.J.: Astrophysical
Journal 610, 1021 (2004)
Eversberg, T., L´epine, S., Moffat, A.F.J.: Astrophysical Journal 494,
799 (1998)
Fullerton, A.W., Massa, D.L., Prinja, R.K.: Astrophysical Journal 637,
1025 (2006)
Gabler, R., Gabler, A., Kudritzki, R.P., Puls, J., Pauldrach, A.: Unified
NLTE model atmospheres including spherical extension and stellar
winds – Method and first results. Astronomy and Astrophysics 226,
162 (1989)
Garc´ıa, M. Bianchi, L.: The effective temperatures of hot stars. II. The
Early-O Types Astrophysical Journal 606, 497 (2004)
Groenewegen, M.A.T., Lamers, H.J.G.L.M., Pauldrach, A.W.A.:
Astronomy and Astrophysics 221, 78 (1989)
Hamann, W.-R.: Astronomy and Astrophysics 84, 342 (1980)
Hamann, W.R.: Line formation in expanding atmospheres – on the va-
lidity of the Sobolev approximation. Astronomy Astrophysics 93,
353 (1981)
Haser, S.M., Pauldrach,A.W.A., Lennon, D.J., Kudritzki, R.-P., Lennon,
M., Puls, J., Voels, S.A.: Quantitative UV spectroscopy of
early O stars in the Magellanic Clouds. The determination of
the stellar metallicities. Astronomy and Astrophysics 330, 285
(1098)
Haser, S.: Ph.D. Thesis, Munich (1995)
Heap, R.S., Lanz, T., Hubeny, I.: Fundamental properties of O Stars
astro-ph0412345 (2004)
Henrichs, H.F., Schnerr, R.S., Ten Kulve, E.: ASPC 337, 114 (2005)
Hensberge, H., Lamers, H.J.G.L., de Loore, C., Bruhweiler, F.C.: As-
tronomy Astrophysics 106, 137 (1982)
Herrero, A., Puls, J., Corral, L.J., Kudritzki, R.P., Villamariz, M.R.: An
Analysis of HST UV Spectra of Cyg OB2 stars. Astronomy and
Astrophysics 366, 623 (2001)
Herrero, A., Puls, J. Najarro, F.: Fundamental parameters of Galac-
tic luminous OB stars VI. Temperatures, masses and WLR
of Cyg OB2 supergiants Astronomy and Astrophysics 396,
949 (2002)
Hillier, D.J., Lanz, T., Heap, S., et al.: Astrophysical Journal 588, 1039
(2003)
Hillier, D.J. Miller, D.L.: Astrophysical Journal 496, 407 (1998)
Howarth, I.D. Prinja, R.K.: The stellar winds of 203 Galactic O stars - A
quantitative ultraviolet survey, Astrophysical Journal Supplement
69, 527 (1989)
Israelian, G., Chentsov, E., Musaev, F.: MNRAS 290, 521 (1997)
Kaufer, A., Stahl, O., Wolf, B., et al.: Astronomy and Astrophysics 305,
887 (1996)
Kudritzki, R.P., Pauldrach, A., Puls, J., Abbott, D.C.: Radiation-driven
winds of hot stars. VI – Analytical solutions for wind models
including the finite cone angle effect. Astronomy Astrophysics
219, 205 (1989)
Kudritzki, R.-P.: Quantitative Spectroscopy of the Brightest Blue
Supergiant Stars in Galaxies Stellar astrophysics for the local
group: VIII Canary Islands Winter School of Astrophysics 149
(1998)
Kudritzki, R.-P., Puls, J.: Winds from hot stars Annual Review of As-
tronomy and Astrophysics 38, 613 (2000)
Kudritzki, R.P., Puls, J., Lennon, D.J., Venn, K.A., Reetz. J., Najarro,
F., McCarthy, J.K., Herrero, A.: The wind momentum-luminosity
relationship of galactic A- and B-supergiants. Astronomy and As-
trophysics 350, 970 (1999)
Lamers, H.J.G.L.M., Haser, S., de Koter, A., Leitherer, C.: The Ion-
ization in the Winds of O Stars and the Determination of Mass-
Loss Rates from Ultraviolet Lines. Astrophysical Journal 516, 872
(1999)
Lamers, H.J.G.L.M., Snow, T.P., Lindholm, D.M.: ApJ 455, 269 (1995)
Lamers, H.J.G.L.M., Stalio, R., Kondo, Y.: ApJ 223, 207 (1978)
Lamers, H.J.G.L.M., Snow, T.P.: ApJ 219, 504 (1978)
L´epine, S., Moffat, A.F.J. Astrophysical Journal 514, 909 (1999)
Macfarlane, J.J., Cohen, D.H., Wang, P.: X-ray induced ionization in the
winds of near-main-sequence O and B stars Astrophysical Journal
437, 351 (1994)
Massa, D., Fullerton, A.W., Sonneborn, G., et al.: Astronomy and As-
trophysics 586, 996 (2003)
Markova, N., Puls, J., Repolust, T., et al.: Astronomy and Astrophysics
413, 693 (2004)
Martins, F., Schaerer, D., Hillier, D.J.: Astronomy and Astrophysics
382, 999 (2002)
Martins, F., Schaerer, D., Hillier, D.J., Heydari-Malayeri, M.: Puzzling
wind properties of young massive stars in SMC-N81 Astronomy
and Astrophysics 420, 1087 (2004)
Martins, F., Schaerer, D., Hillier, D.J.: Astronomy and Astrophysics
436, 1049 (2005)
Martins, F., Schaerer, D., Hillier, J., Meynadier, F., Heydari-Malayeri,
M., Walborn, N.O.: stars with weak winds: the Galactic case As-
tronomy and Astrophysics 441, 735 (2005)
Massey, P., Bresolin, F., Kudritzki, R.P., Puls, J. Pauldrach, A.W.A.:
Astrophysical Journal 608, 1001 (2004)
Massey, P., Puls, J., Pauldrach, A.W.A. et al.: The Physical Properties
and Effective Temperature Scale of O-Type Stars as a Function of
Metallicity. II. Analysis of 20 More Magellanic Cloud Stars and
Results from the Complete Sample Astrophysical Journal 627, 477
(2005)
McErlean, N.D., Lennon, D.J. Dufton, P.L.: Astronomy and Astro-
physics 349, 553 (1999)
Meynet, G., Maeder, A., Schaller, G., et al.: Astronomy and Astro-
physics 103, 97 (1994)
Najarro, F., Kudritzki, R.P., Cassinelli, J.P., Stahl, O., Hillier, D.J.: Stel-
lar winds and the EUV continuum excess of early B-giants. As-
tronomy and Astrophysics 306, 892 (1996)
Najarro, F.: Spectroscopy of P Cygni ASP Conf. Ser. 233: P Cygni
2000: 400 Years of Progress p 133 (2001)
Najarro, F., Figer, D.F., Hillier, D.J., Kudritzki, R.P.: Metallicity in the
Galactic Center: The Arches Cluster Astrophysical Journal Letters,
611, L105 (2004)
Owocki, S.P., Puls, J.: Astrophysical Journal 510, 355 (1999)
Pauldrach, A.W.A., Kudritzki, R.P., Puls, J., Butler, K., Hunsinger, J.:
Radiation-driven winds of hot luminous stars. 12: A first step
towards detailed UV-line diagnostics of O-stars. Astronomy and
Astrophysics 283, 525 (1994)
Pauldrach, A.W.A., Hoffmann, T.L. Lennon, M.: Astronomy and
Astrophysics 375, 161 (2001)
Praderie, F., Talavera, A., Lamers, H.J.G.L.M.: Astronomy and Astro-
physics 86, 271 (1980)
Springer

170 Astrophys Space Sci (2006) 303:153–170
Prinja, R.K., Massa, D., Searle, S.C.: Astronomy and Astrophysics 430,
L41 (2005)
Prinja, R.K. Crowther, P.A.: HSTUV measurements of wind structure
and velocities in Local Group OB stars, Monthly Notices of the
Royal Astronomical Society 300, 828 (1998)
Przybilla, N., Butler, K., Becker. S.R., Kudritzki. R.P.: Astronomy and
Astrophysics, astro-ph/05099669 (2005)
Puls, J., Kudritzki. R.P., Herrero, A. et al.: O-star mass-loss and wind
momentum rates in the Galaxy and the Magellanic Clouds Obser-
vations and theoretical predictions Astronomy and Astrophysics
305, 171 (1996)
Puls, J., Kudritzki, R.P., Santolaya-Rey, A.E., Herrero, A., Owocki,
S.P.: Proc. 2nd Boulder-Munich Workshop, PASP 131, 245
(1998)
Puls, J., Springmann, U., Lennon, M.: Radiation driven winds of
hot luminous stars. XIV. Line statistics and radiative driv-
ing Astronomy and Astrophysics Supplement Series 141,23
(2000)
Puls, J., Urbaneja, M.A., Venero, R., et al.: Astronomy and Astrophysics
435, 669 (2005)
Repolust, T., Puls, J., Herrero, A.: Astronomy and Astrophysics 415,
349 (2004)
Santolaya-Rey,A.E., Puls, J., Herrero, A.: Astronomy and Astrophysics
323, 488 (1997)
Springmann, U.W.E., Pauldrach, A.W.A.: Astronomy and Astrophysics
262, 515 (1992)
Talavera, A., G´omez de Castro, A.I.: Astronomy and Astrophysics 181,
300 (1987)
Trundle, C., Lennon, D.J., Puls, J. Dufton, P.L.: Astronomy and Astro-
physics 417, 217 (2004)
Trundle, C., Lennon, D.J.: Understanding B-type supergiants in the
low metallicity environment of the SMC II Astronomy and Astro-
physics 434, 677 (2005)
Underhill, A.B., Doazan, V.: B stars with and without emission lines,
NASA SP-456 (1982)
Urbaneja, M.A.: Ph.D. Thesis, La Laguna (2004)
Urbaneja, M.A., Herrero, A., Kudritzki, R.P. et al.: An analysis of STIS
HST UV spectra of M 33 early B supergiants. Astronomy and
Astrophysics 386, 1019 (2002)
Vacca, W.D., Garmany, C.D., Shull, J.M.: ApJ 460, 914 (1996)
Venn, K.A.: ApJ 449, 839 (1995a)
Venn, K.A.: ApJSS 99, 659 (1995b)
Verdugo, E., Talavera, A., G´omez de Castro, A.I.: ESA SP-413 293
(1997)
Verdugo, E., Talavera, A., G´omez de Castro, A.I.: Astronomy and As-
trophysics 346, 819 (1999a)
Verdugo, E., Talavera, A., G´omez de Castro, A.I.: Astronomy and As-
trophysics SS 137, 351 (1999b)
Verdugo, E., Talavera, A., G´omez de Castro, A. I., Henrichs, H. F.,
Geers, V. C., Wiersema, K.: ASPC 305, 364 (2003)
Verdugo, E., Henrichs, H.F., Talavera, A., G´omez de Castro, A.I., Schn-
err, R.S., Geers, V.C.: ASPC 337, 324 (2005)
Verdugo, E., Talavera, A., G´omez de Castro, A.I.: Astronomy and As-
trophysics, in preparation (2006)
Vink, J.S., de Koter, A., Lamers, H.J.G.L.M.: New theoretical mass-
loss rates of O and B stars. Astronomy and Astrophysics 362, 295
(2000)
Vink, J.S., de Koter, A., Lamers, H.J.G.L.M.: Mass-loss predictions
for O and B stars as a function of metallicity. Astronomy and
Astrophysics 369, 574 (2001)
Springer