Download exploring the universe with the low frequency array a scientific case

EXPLORING THE UNIVERSE WITH
THE LOW FREQUENCY ARRAY
A SCIENTIFIC CASE
Version 1.00 – 25-September-2002
This Science Case was compiled by the astronomical community in the
Netherlands, and is based on contributions to the March 2001 International
LOFAR Science Case. In places it refers to the 2002 baseline design. In
particular the sensitivities and resolutions are no longer appropriate for the
(first phase) design for which funding has was secured in 2003/4. The plan is
still to realise the full design when sufficient funds become available.
For an accurate description of the current LOFAR Plan visit: www.lofar.org
This Science Case was compiled by:
A.G. de Bruyn (ASTRON), R.P. Fender (Amsterdam), J.M.E. Kuijpers (Nijmegen), G.K. Miley
(Leiden), R. Ramachandran (Amsterdam/ASTRON), H.J.A. Röttgering (Leiden), B.W.
Stappers (Amsterdam/ASTRON), M.A.M. van de Weygaert (Groningen), M.P. van Haarlem
(ASTRON)
Contributors to the Original International Science Case include:
J.B.G.M. Bloemen (SRON Utrecht), K.M. Blundell (Oxford), F.H. Briggs (Groningen), R.
Braun (ASTRON), A.G. de Bruyn (ASTRON), G.S. Bust (ARL – UTexas), J.R. Dickel
(Illinois), S. Doeleman (MIT/Haystack), T.A. Ensslin (MPI Garching), R.P. Fender
(Amsterdam), J.C. Foster (MIT/Haystack), B.M. Gaensler (MIT), M.A. Garrett (JIVE), T.L.
Gaussiran II (ARL – UTexas), J.N. Hewitt (MIT), B. Isham (EISCAT, Tromsø, Norway), F.P.
Israel (Leiden), N.E. Kassim (NRL), T.J.W. Lazio (NRL), F.D. Lind (MIT/Haystack), G.K.
Miley (Leiden), R. Ramachandran (Amsterdam/ASTRON), P. Rodriguez (NRL), H.J.A.
Röttgering (Leiden), J.E. Salah (MIT/Haystack), B.W. Stappers (Amsterdam/ASTRON), B.
Thidé (Swedish Institute of Space Physics, Uppsala), M.P. van Haarlem (ASTRON), J.M.
van der Hulst (Groningen), K.W. Weiler (NRL)
Page 1 of 59
1
2
3
4
5
6
Introduction.............................................................................................................................................. 3
1.1
Background - Exploring the Low-Frequency Universe ................................................................. 3
1.2
Scientific Objectives of LOFAR...................................................................................................... 5
1.3
Low-Frequency Radio Radiation: Unique Physical Diagnostics .................................................. 5
1.4
A Census of the Low-Frequency Sky ............................................................................................ 7
1.5
Seeking Variable Sources.............................................................................................................. 9
REIONIZATION OF THE UNIVERSE.................................................................................................. 10
2.1
Introduction and key questions .................................................................................................... 10
2.2
Constraints on the Epoch of Reionization ................................................................................... 11
2.3
Constraints on the Sources of Reionization. ............................................................................... 12
2.4
Optimizing LOFAR for Studying Reionization ............................................................................. 13
2.4.1
Intensity................................................................................................................................. 13
2.4.2
Expected Angular Scales..................................................................................................... 13
2.4.3
Implications for LOFAR ........................................................................................................ 14
2.5
Answers from LOFAR to Key Questions ..................................................................................... 15
FORMATION AND EVOLUTION OF GALAXIES, CLUSTERS AND ACTIVE NUCLEI ................... 17
3.1
Introduction and Key Questions................................................................................................... 17
3.2
Distant Radio Galaxies: Probes of Massive Galaxy and Cluster Formation ............................. 17
3.3
Distant starburst galaxies: Probes of galaxy evolution............................................................... 20
3.4
Diffuse Radio Sources: Probes of Intergalactic Gas Evolution .................................................. 23
3.4.1
Cluster Radio Halos: Tracing Over-dense Regions. .......................................................... 24
3.4.2
Giant radio sources: Tracing under-dense regions. ........................................................... 25
3.5
Gamma Ray Bursters – Prompt Emission and Afterglows......................................................... 26
3.6
Answers from LOFAR to key questions ...................................................................................... 26
THE HIGHEST ENERGY PHENOMENA ............................................................................................ 29
4.1
LOFAR as a Cosmic Ray Detector.............................................................................................. 29
4.1.1
Background Information: Cosmic Rays............................................................................... 29
4.2
Gravitational Waves - LIGO events ............................................................................................. 33
4.2.1
An Unbiased Survey for Transient Astronomical Radio Sources (STARE) ...................... 33
4.2.2
The Physics of Collapse and Explosion.............................................................................. 33
4.3
LOFAR as an All-Sky Monitor ...................................................................................................... 35
4.3.1
Particle acceleration in cosmic explosions ......................................................................... 36
4.3.2
New emission regimes at LOFAR frequencies................................................................... 39
THE MILKY WAY AND NEIGHBOURING GALAXIES ....................................................................... 42
5.1
Testing the Standard Shock Acceleration Model........................................................................ 42
5.2
Thermal and Nonthermal Emission in Nearby Galaxies............................................................. 43
5.3
Supernova Remnants and the Distribution of Ionized Gas in the ISM ...................................... 44
5.4
Discovering New SNRs ................................................................................................................ 44
5.4.1
Spectral Studies ................................................................................................................... 44
5.4.2
Absorption & ISM Studies .................................................................................................... 44
5.4.3
Extragalactic SNR Studies................................................................................................... 45
5.5
H II Regions .................................................................................................................................. 46
5.6
Interstellar Propagation Effects.................................................................................................... 46
5.7
Polarimetry .................................................................................................................................... 48
5.8
Recombination Lines .................................................................................................................... 49
5.9
Neutron Stars and Pulsars ........................................................................................................... 49
5.9.1
Searching for new pulsars ................................................................................................... 49
5.9.2
Emission physics.................................................................................................................. 50
5.10
Jupiter and Extrasolar Jupiters ................................................................................................ 52
5.10.1
Jupiter Magnetosphere .................................................................................................... 52
5.10.2
Decameter Bursts ............................................................................................................ 53
5.10.3
Extrasolar Planets ............................................................................................................ 53
THE SUN AND SOLAR-TERRESTIAL RELATIONSHIPS................................................................. 55
6.1
Introduction ................................................................................................................................... 55
6.2
Space Weather ............................................................................................................................. 56
6.2.1
LOFAR and Coronal Mass Ejections .................................................................................. 56
6.2.2
The Solar Magnetic Cycle and the Terrestrial Climate ...................................................... 56
6.2.3
LOFAR and CME warning ................................................................................................... 57
6.2.4
LOFAR, Active Radar and Space Weather ........................................................................ 58
Page 2 of 59
1 Introduction
1.1 Background - Exploring the Low-Frequency Universe
During the last half century our knowledge of the Universe has been revolutionized by the
opening of observable windows outside the narrow visible region of the spectrum. Radio
waves, infrared and ultraviolet radiation and X- and gamma rays have provided new and
completely unexpected information about the nature and history of the Universe and have
resulted in the discovery of a cosmic zoo of strange and exotic objects. One of the few
spectral windows that still remain to be explored is at the low radio frequencies, the lowest
energy extreme of the accessible spectrum. LOFAR, the Low Frequency Radio Array, is a
large radio telescope that will open this virgin territory to a broad range of astrophysical
studies.
The mission of LOFAR is to survey the Universe at frequencies of from ~ 10 – 240 MHz
(corresponding to wavelengths of 1.5 – 30 m). Radio astronomy was born at these
wavelengths in 1931, when Karl Jansky investigated the background noise that was plaguing
transatlantic short-wave communications. Since then, low frequency radio astronomy has
been neglected because of the poor resolving power of the available facilities and the
disturbing effects of the ionosphere on observations.
Because the spatial resolution of a telescope is proportional to its operating frequency, radio
telescopes such as the Westerbork Radio Synthesis Telescope have extremely poor spatial
resolution when operated at low frequencies. The radio images obtained by low frequency
radio facilities (resolutions of arcminutes) are blurred by a factor of several thousand
compared with optical pictures of the sky. The blurred images result in a so-called "confusion"
1
effects that have limited the sensitivity at low radio frequencies to the level of ~1 Jansky and
the number of objects that can be studied to the brightest few hundred.
The resolution of a radio telescope can be improved by enlarging the aperture, or in the case
of a Westerbork-type array of antennas, by increasing the maximum distance between the
elements of the array - i.e. the “baseline”. At low frequencies, to achieve useful resolutions
comparable with visible images of the sky, maximum baselines of several hundred kilometres
are needed.
Until now, one of the most important limitations in achieving such long baselines at low radio
frequencies has been the complicated structure of the ionosphere and its variation over time.
Just as the atmosphere causes stars to twinkle, the irregularities in the ionosphere produce
jittering in the radio images.
There have been several recent technological developments that make it possible to
compensate for the jittering effects of the ionosphere. First, the large increase of computer
power makes it possible to create and process images of very wide fields on short enough
timescales to monitor and correct for the ionospheric jitter. Second, progress in antenna
design have made it possible to construct low-frequency antennas with several simultaneous
beams that can be pointed to and used to monitor different regions of the sky simultaneously.
Combining these developments with interferometer calibration algorithms that have been
developed during the past 3 decades in the Netherlands and the USA will allow the
construction of a radio array with baselines extending to a few hundred kilometres, operating
at frequencies down to ~10 MHz. Such an array will for the first time probe the distant
Universe at the low-energy extremity of the electromagnetic spectrum.
LOFAR will consist of an array of antennas spread over a ~ 400 km region, that will provide
sufficient resolution to allow radio sources to be identified with visible objects, even at low
frequencies. Occupying its central 2 km will be a more densely filled "virtual" core (VC), that
will allow more effective calibration of the instrument and optimize its sensitivity for a special
experiment to study the reionization phase of the Universe. The design will provide fast (~1
ms) frequency selection and pointing, capable of rapidly imaging radio sources across the sky
1
-26
1 Jansky (Jy) = 10
-2
W m Hz
-1
Page 3 of 59
and spectrum. Multiple independent beams will herald a new technological approach to
observing, yielding unprecedented flexibility when compared with higher cost, higher
frequency ground or space-based systems.
The power of LOFAR is specified in Table 1.1 and illustrated in Figure 1.1, where sensitivity
and spatial resolution are plotted against frequency. The limits for previous low-frequency
facilities are compared with that of LOFAR.
Figure 1.1 Comparison of LOFAR capabilities with those of previous facilities as a function of frequency.
The sensitivity after one hour of integration (left) and angular resolution (right) of LOFAR are shown in
red together with those of previous low frequency radio survey facilities.
Frequency Range
Effective Collecting Area
Configuration
Low
High
Low (10-90 MHz)
High (110-200 MHz)
High (200-240 MHz)
Virtual Core (VC) (25%)
Extended Array
Sensitivity
10 MHz
30 MHz
75 MHz
120 MHz
200 MHz
Point Source
VC
Full Array
12 mJy
3.0 mJy
6.5 mJy
1.6 mJy
4.0 mJy
1.0 mJy
0.5 mJy
0.13 mJy
0.14 mJy 0.03 mJy
Greatest Instantaneous
Sky Coverage
10-90 MHz
110-240 MHz
6.1
)-2
2
10 (freq./15 MHz m
5.1
2
10 m
5.1
)-2
2
10 (freq./200 MHz m
< ~ 2 km
400 km
Brightness Temperature
VC
Full Array
5.3
20 K
10 K
5.1
12 K
10 K
4.6
7.3 K
10 K
4.0
0.9 K
10 K
3.4
0.25 K
10 K
0.1 sterad/beam at 10 MHz
4 sterad (virtual core)
0.64 arcsec at 240 MHz
Highest Angular
Resolution (FWHM)
Number of Spectral
Channels
Time Resolution
Polarization
4096
1 ms
full Stokes
# Sensitivity calculations are for a single polarization, 1 hour integration, 4 MHz bandwidth
Table 1.1 Specifications for LOFAR
Page 4 of 59
1.2 Scientific Objectives of LOFAR
The sensitivities and spatial resolutions attainable with LOFAR will make possible several
fundamental new studies of the Universe as well as facilitating unique practical investigations
of the environment of the earth.
•
•
In the very distant Universe (7 10), LOFAR can search for the signature
produced by the reionization of neutral hydrogen. This crucial phase change is
predicted to occur at the epoch the formation of the first stars and galaxies, marking
the end of the so-called “dark ages”. The redshift at which reionization is believed to
occur will shift the 1420 MHz line of neutral hydrogen into the LOFAR observing
window.
In the distant “formative” Universe (1.5 < z < 7), LOFAR will detect the most
distant massive galaxies and will study the processes by which the earliest structures
in the Universe (galaxies, clusters and active nuclei) form and probe the intergalactic
gas.
•
In the nearby Universe, LOFAR will map the 3-dimensional distribution of cosmic
rays and global magnetic field in our own and nearby galaxies.
•
The High Energy Universe, LOFAR will detect the ultra high energy cosmic rays as
they pierce the Earth’s atmosphere.
•
Within our own galaxy, LOFAR will detect flashes of low-frequency radiation from
pulsars and short-lived transient events produced by stellar merging and interactions
and will search for Jupiter-like extra-solar planets.
•
Within our solar system, LOFAR will detect coronal mass ejections from the Sun
and provide continuous large-scale maps of the solar wind. This crucial information
about solar weather and its effect on the Earth will facilitate predictions of costly and
damaging geomagnetic storms.
•
Within the Earth’s immediate environment, LOFAR will map irregularities in the
ionosphere continuously, detect the ionizing effects of distant gamma-ray bursts and
the flashes predicted to arise from the highest energy cosmic rays, whose origin of is
unclear.
•
By exploring a new spectral window LOFAR is likely to make unexpected
"serendipitous" discoveries. Detection of new classes of objects and/or new
astrophysical phenomena have resulted from almost all previous facilities that open
new regions of the spectrum, or pushed instrumental parameters, such as sensitivity
by more than an order of magnitude (e.g. Harwit, M. Cosmic Discovery. Harvester
Press, 1981)
The key scientific drivers will be described in detail below. Additional LOFAR applications will
include the study of supernova remnants and their interaction with the interstellar medium,
interstellar propagation effects, interstellar radio recombination lines. Much LOFAR science
builds on fundamental areas of research that have been pursued intensively or pioneered
within the Netherlands during the last half century.
1.3 Low-Frequency Radio Radiation: Unique Physical
Diagnostics
Before discussing the scientific drivers of LOFAR in detail, we draw attention to the special
aspects of the radiation at the extremely low frequencies at which LOFAR will observe. Such
processes will provide diagnostics that LOFAR will exploit in pursuing its scientific goals.
Astronomy in the visible is dominated by thermal radiation emitted by hot gas in stars and
galaxies. However, radio astronomy operates in a regime in which the dominant radiation is
due to other mechanisms. Several unique astrophysical diagnostics can be derived from
Page 5 of 59
radiation processes that occur at frequencies accessible to LOFAR. The following are
amongst the relevant radiation mechanisms.
Synchrotron Emission. Synchrotron radiation is produced by electrons moving close to the
speed of light in a magnetic field. This is the dominant radiation mechanism encountered in
classical radio astronomy. Because the objects that emit luminous radio radiation (e.g. distant
active galaxies, e.g. Fig. 2) are different from those than luminous thermal radiation (e.g.
bright stars and nebulae), the radio Universe appears very different from the visible Universe.
During the last half century opening the radio Universe to astronomy therefore resulted in
many unexpected discoveries. Synchrotron sources that will be observed by LOFAR include
lobes and jets emitted by the nuclei of most distant galaxies, and the cosmic rays and
supernova remnants produced by stars in normal galaxies.
Figure 1.2 The radio galaxy Cygnus A, one of the brightest objects in the radio sky appears radically
different when observed at radio or optical wavelengths. Cygnus A is located at a distance of 600
million light years. The lower frame shows a picture of the radio emission. On the upper frame this radio
image (red) is superimposed on a (blue) optical picture of the same part of the sky. Most of the optical
objects are foreground stars. The elliptical galaxy responsible for producing the two enormous radio
lobes is the fuzzy object close to the centre of the picture. The radio source is due to relativistic plasma
produced by a rotating massive black hole deep at the centre of the galaxy. The radio size of ~140 kpc
(500,000 light-years) is an order of magnitude larger than that of the central galaxy. LOFAR will make
observations of the sky at wavelengths that are longer by a factor of ~100 than those at which this radio
picture was taken.
Page 6 of 59
Relatively little is known about the spectra of most synchrotron sources at frequencies lower
than ~30 MHz. Because synchrotron-emitting electrons lose energy as they grow older,
8
radiation at the lowest frequencies is produced by the oldest relativistic electrons (age ~10
y). LOFAR will therefore be able to study the death pangs of "fossil" radio sources and
provide new constraints on radio source evolution.
Coherent Plasma Emission Although the proposed new facility will observe large numbers
of synchrotron sources, one of the most exciting aspects of LOFAR is that it will operate in a
low-energy region of the spectrum that is sensitive to other little-studied radiation processes,
such as coherent plasma emission and cyclotron emission. Coherent plasma emission is
known to be important in the Sun and Jupiter, the two brightest objects in the low-frequency
radio sky. Astronomical surveys made in regions unexplored spectral regions dominated by
unstudied emission mechanisms have always revealed unexpected populations of objects.
This is also likely to be the case with LOFAR.
Absorption Processes Many synchrotron sources have spectra that decline sharply at low
radio frequencies. This decline at low frequencies is usually attributed to absorption of the
synchrotron radiation either within the emitting object itself or in the path between the emitter
and the earth. A study of this absorption can provide diagnostics about the densities and
geometry of gas and plasma inside the radio sources, the surrounding environment and the
path between the emitting source and the earth.
There are three main relevant processes.
• Synchrotron self absorption, an internal effect whose cut-off frequency constrains the
sizes, geometries and magnetic field strengths of compact extragalactic and galactic
radio sources.
• Thermal absorption, whose cut-off frequency constrains the density, temperature and
geometry of the interstellar medium in galaxies.
• The Razin-Tsytovich Effect, that operates when gas densities are sufficiently high and
frequencies are sufficiently low that the index of refraction is less than unity. The
relevant cut-off frequency constrains the density, magnetic field strength and
geometry of the interstellar medium.
Extending observations of such phenomena to the lowest frequencies will improve the
sensitivity of such diagnostic studies by up to an order of magnitude.
1.4 A Census of the Low-Frequency Sky
One of the most important techniques that will be used to pursue LOFAR's scientific
objectives will be to conduct unbiased surveys of large regions of the accessible sky at
several frequencies. Large-sky surveys are the bedrock on which much of modern
astrophysical knowledge has been built and have traditionally been major drivers of Dutch
astronomy. For example, surveys play an important strategic role by providing catalogues of
unique objects that act as leverage in obtaining time at highly oversubscribed ground-based
and space facilities. Surveys such as the optical Palomar Observatory Sky Survey (POSS)
and the radio 3C, Parkes and Westerbork surveys have provided material for at least 2 or 3
decades of high-impact astrophysical discoveries.
In the early eighties the Netherlands played a major role in producing the highly successful
IRAS all-sky infrared survey that opened up a new spectral window, just as LOFAR will do.
The IRAS surveys have had a fundamental impact over a broad range of galactic (e.g. Fig. 3)
and extragalactic astrophysics and the early success of IRAS stimulated ESA to fund the
follow-up ISO mission. Early access to the IRAS survey played a major role in shaping Dutch
astronomy during the subsequent 2 decades. LOFAR will be the IRAS of low-frequency radio
astronomy.
Page 7 of 59
Figure 1.3 Infrared picture of the Milky Way as seen by the highly successful Dutch-US-UK survey
satellite IRAS that opened up the far IR region of the spectrum to a broad range of astrophysical studies
(Beichman et al 1988). LOFAR will be a similar survey facility for exploration of the lowest energy
accessible region of the electromagnetic spectrum. Both IRAS and LOFAR use the extreme demands of
fundamental research to provide test-beds for frontier technological developments.
Although they did not open new spectral windows, recent productive radio surveys include the
Westerbork surveys at 325 MHz (WENSS and WISH) and the NRAO VLA Sky Surveys at
1400 MHz (NVSS and FIRST). Between them, these three surveys have resulted in the
detection of millions of radio sources. These surveys so far contributed to seven Dutch Ph.D.
theses on topics such as studies of the large-scale structure of the Universe, the nature of
extra-galactic radio sources and their relation to galaxy formation and the use of galactic
foreground polarization as a probe of the interstellar medium in our own galaxy.
LOFAR will exploit the experience gained in conducting such surveys to open up the low
frequency radio window. By realizing a combined jump in sensitivity and resolution (see Fig.
1), LOFAR will take radio astronomy into hitherto unexplored regions of parameter space.
Just as IRAS sources were the basis of many successful follow-up proposals on large optical
telescope, we expect that the properties of LOFAR sources will be studied by the highly
oversubscribed present and planned optical and infrared facilities, such as the VLT, Keck,
HST, ALMA, SIRTF and the NGST.
LOFAR will survey and monitor the galactic and extragalactic sky at several frequencies. With
the unique capability to observe with a number of synthesized beams simultaneously, it will
be possible to observe large areas of sky to unprecedented depths. Since the location of
parameter space that LOFAR will observe is unknown territory, an estimate of the number of
sources that LOFAR will observe is necessarily based on surveys of small regions of sky that
have been carried out at higher frequencies. Using such radio surveys as carried out with the
Westerbork radio telescope at 1.4 GHz (Katgert et at 1988 and Garrett et al 2000), the
number of sources per unit surface area on the sky can be calculated (see Table 2). If the
sky is observed at 5 different illustrative LOFAR frequencies (10, 30, 75, 120 and 200 MHz)
using 5 different beams, then after one year of continuous observing, significantly more than
one million radio sources will be detected at each frequency. We note that LOFAR surveys
will have a different character at low and high frequencies. At low frequencies, large parts of
the sky can be surveyed, yielding important samples of very distant radio galaxies and cluster
halos. Repeated observations of these areas will allow for a systematic search for important
variable objects including Jupiters orbiting nearby stars and afterglows of the mysterious
gamma-ray bursts. At high frequencies, the depth will be such that a large fraction of the radio
sources will identified with distant starbursting galaxies. This will yield well defined samples
for studying the evolution and spatial distribution of this important class of galaxy.
Page 8 of 59
One of the unique features of LOFAR will be its ability to survey the sky simultaneously with
several thousand different frequency channels and a spectral resolution of < 1 KHz. This will
allow the spectral signal of the reionization of the Universe to be studied and facilitate
serendipitous detection of spectral lines in the LOFAR band. The LOFAR large-sky surveys
and dedicated observations directed at special regions of the sky will form the basis for
studies in several key areas of science that will be described below in this document.
Observing
Frequency
(MHz)
Angular
Resolution
(arcsec)
Survey
1
limit
(10 )
(mJy)
Surface density
of sources
(No. per arcmin)
Area covered
after 1 year
(sq. deg)
Total number
of sources
after 1 year
10
30
75
120
200
15
5.2
2.1
1.3
0.8
30
2
0.3
0.1
10
0.5
4
25
66
125
3000
3000
60
62
7.5
5.4 million
4.3 million
5.3 million
15 million
3.3 million
Table 1.2 Surveying the Sky with LOFAR
1
The survey limit is determined either by confusion or the feasible integration time. It is
assumed that 5 independent LOFAR beams will be available.
1.5 Seeking Variable Sources
Many celestial objects are variable. Besides charting the static low-frequency sky, one of the
main goals of LOFAR will be to search for and study variable objects. LOFAR variable
sources are laboratories for studying some of the highest energy phenomena in the Universe.
The ability to make large-sky surveys with LOFAR is therefore an important driver for the
LOFAR design.
Page 9 of 59
2 REIONIZATION OF THE UNIVERSE
2.1 Introduction and key questions
According to present views of the early Universe (e.g see Figure 2.1), hydrogen recombined
into its neutral state about half a million years after the Big Bang, when the primordial matter
had cooled to a temperature of about 3000 K. The Universe then entered a period of
“darkness”, when its temperature became steadily lower due to the universal expansion.
These “Dark Ages” (Rees 1996) ended many hundreds of million years later, when the first
stars were formed and assembled into proto-galaxies. Ionizing radiation from these stars and
embyo galaxies then began to warm the Universe and produce observable radiation. After a
sufficient number of ionizing sources had formed, the temperature and the ionized fraction of
the Universe increased rapidly and most of the neutral hydrogen eventually disappeared. This
period, in which the Universe went from a phase in which almost all the hydrogen was neutral
to a state in which it was almost completely ionized, is referred to as the “reionization” of the
Universe. The moment that marked the time when 50% of the hydrogen in the Universe was
ionized is termed the “Epoch of Reionization" (E-o-R) (e.g. see Gnedin, 2001). The seeds of
several properties of our present Universe, such as the observed web-like distribution of
galaxies (Figure 2.2) are likely to have been embedded during this crucial period in history.
Figure 2.1 Cartoon of the likely development of the early Universe. About 500,000 years after the Big
Bang (z ~ 1000) hydrogen recombined and remained neutral for a few hundred million years (the “dark
ages”). At a redshift, z ~ 10, the first stars galaxies and quasars began to form, heating and reionizing
the hydrogen gas. LOFAR will be able to observe the global spectral signature of neutral hydrogen close
to this epoch of reionization (E-o-R).
Page 10 of 59
Figure 2.2 The cosmic web. Matter in the local Universe (hot and warm baryons) is known to lies in a
large-scale filamentary structuresand this cosmic web is believed to extend to the highest redshifts. This
picture was taken from Cen & Ostriker 1999 and illustrates the results of numerical N-body
hydrodynamical models. The seeds of the cosmic web were implanted during the dark ages, just prior
to reionization. This primeval era will be accessible to LOFAR.
Three of the most important questions need to be addressed about this fundamental change
in the state of the early Universe are:
1. When did the reionization of the Universe occur?
2. How rapidly did the reionization of the Universe progress?
3. What were the dominant sources of ionization and how did they affect the global
progression of the reionization?
Neutral hydrogen comprises ~ 5% - 10% of the 'critical' density of the Universe (reference ?
check) and is potentially one of the most important diagnostics of the final stages of the dark
ages of and the beginning stages of reionization (e.g. Rees, 1999).
2.2 Constraints on the Epoch of Reionization
Computer simulations imply that the process of reionization was rapid and took place at some
epoch between redshifts z ~ 15 and z ~ 6 (Gnedin 2000, Tozzi et al. 2000). There are
several observational constraints are consistent with these predictions and allow the E-o-R to
be further refined.
The spectra of the most distant quasars allow an upper limit to be placed on the epoch at
which reionization occurred. It has long been known that the reionization of the Universe must
have occurred at epochs corresponding to a redshift of z > 5, because of the absence of
absorption blueward of Ly in the spectra of quasars out to that redshift. Such a cut-off in the
continuum spectra would be produced by neutral hydrogen in the Universe prior to
reionization (Gunn – Peterson effect – reference). During the last two years several quasars
and a few galaxies having 6.5 > z > 5 have been discovered with the Sloan Digital Sky
Survey (Fan et al, 2001, Becker et al, 2001 Pentericci et al, Hu et al.2002). The spectra of
these objects showed the long predicted “Gunn-Peterson” absorption. The implication of this
exciting result is that in the period corresponding to 6.5 > z > 5, the fraction of neutral
Page 11 of 59
-4
-5
hydrogen in the Universe decreased dramatically from > 10 to ~10 . The epoch of
reionization is likely to have occurred in the period just before these quasars were observed.
The earliest time that reionization can have occurred can also be constrained observationally.
First, the epoch of reionization must have taken place at redshifts, z < 30. Otherwise there
would be a suppression of the first Doppler peak in the angular fluctuation spectrum of the
Cosmic Microwave Background (Tegmark & Zaldarriaga 2000; De Bernardis et al. 2000). This
is not observed. Secondly, measurements of the temperature of the intergalactic medium
(Lyman forest clouds) at redshift z ~ 3, imply that the reionization must have occurred after
z ~ 10. Because of the (adiabatic) expansion of the Universe, an earlier epoch of reionization
would have resulted in a temperature of the z ~ 3 intergalactic medium that would be cooler
than observed (Theuns et al, 2002; Hui et al, 2001).
In summary, the observational constraints and numerical simulations all imply that most of the
neutral hydrogen in the Universe must have become ionized during the epoch corresponding
to a redshift range of between z ~ 10 and z ~ 6.
If the 21 cm line emitted by neutral hydrogen (van der Hulst 1948??) just before reionization
will be redshifted by the expansion of the Universe from its rest frequency of 1420 MHz to
within the frequency range between 100 and 200 MHz (corresponding to redshifts of 6-13).
Observations at LOFAR radio frequencies can therefore offer a direct view of the primeval
hydrogen gas, which eventually formed into 'visible' stars, galaxies and clusters. LOFAR will
be the first telescope to operate in the frequency range of 100-200 MHz that has enough
sensitivity and dynamic range to detect and study this neutral hydrogen signal. The ability to
study the signal of reionization has therefore played an important role in driving the array
configuration design of LOFAR as well as the frequency range to be covered by its receivers.
2.3 Constraints on the Sources of Reionization.
At least three different types of sources have been proposed as contributing to the radiation
that reionized the Universe: (i) emission from the first generation of stars, (ii) radiation
released in the collapse of the first gas clouds that formed sub-galactic fragments, and (iii)
emission from an early generation of active galactic nuclei and quasars.
The powerful ionizing ultra-violet and X-ray continua of quasars were long thought to play the
dominant role in the ionization of the Universe, because they appear to be responsible for the
bulk of the ionizing background emission at low redshifts (z < 4). However, the space density
of bright quasars at high redshifts is now known to be insufficient by a large factor to make a
significant contribution to the global reionizing flux (e.g. Madau, 2000). Nevertheless, they
may have played a role in “preheating” the intergalactic medium.
Present evidence favours emission from stars as the dominant factor in reionization. In recent
years systematic deep multicolour surveys for have uncovered a population of objects at
redshifts of up to ~ 6.5 (e.g. Steidel et al 200X, Malhotra and Rhoads, 2001). These objects
are termed “Lyman-break” galaxies, because of the characteristic cutoff in their ultraviolet
spectra shortward of Ly-, due to absorption by intrinsic neutral hydrogen and are known to
be undergoing star formation.
The first stars were probably more massive (> 100 Msun, e.g. Abel et al. 2002) and hotter than
the bulk of the stars forming in present-day galaxies. These young stars would produce
copious amounts of ultraviolet photons, that would ionize the surrounding hydrogen, in socalled “Stromgren spheres”. When the Stromgren spheres from collections of young stars in
the first mini-galaxies start to overlap, a dense patchwork of ionized regions, separated by
neutral gas, will develop (Gnedin and Ostriker, 1997; Gnedin, 2000). This ionized mosaic will
gradually spread until the whole Universe becomes ionized.
A complication in this scenario for reionization is the possible role played by gas clouds
around the first galaxies (“mini-halos”) in shielding the Universe from exposure to the first
generation of ionizing stars (Barkana and Loeb, 2000; Shapiro et al, 2002). Such shielding
may well have delayed the global reionization process.
Page 12 of 59
2.4 Optimizing LOFAR for Studying Reionization
LOFAR will detect the spectral signature emitted by neutral hydrogen in the period before full
re-ionization. To optimize the design of LOFAR for the study of this neutral hydrogen, it is
important to discuss the properties of the expected signal.
2.4.1
Intensity
The signal is expected to be similar in all directions, i.e. it is a global signal. The cosmic
baryon density and the scale factor of the Universe are now sufficiently well established that
the maximum strength of the HI signal can be predicted to within about 50%. If all neutral
hydrogen at high redshifts was distributed uniformly and if the epoch of reionization
corresponded to z ~ 8.5, the global spectral emission signal at ~ 150 MHz should have a
brightness temperature of ~10-20 mK. This assumes the gas emits the 21cm line, that
requires the spin temperature of the gas to be above that of the CMB, or about 2.73 x (1+z) ~
25 K. This signal can easily be detected by LOFAR after an integration of a few hundred
hours (e.g. see Table 1.1).
Besides a global spectral feature due to emission, in some regions a radio absorption signal
might also be detected immediately prior to full reionization. When the first ionizing sources
appeared, before (e.g. stars, quasars or collapsed mini-halos), the hydrogen spin temperature
decoupled from the emission temperature of the cosmic microwave background (CMB)
(Madau et al, 1997; Meiksin, 1999) and may have approached the kinetic temperature of the
gas. If such a source was sufficient to heat, but not ionize the gas, the intergalactic medium
would acquire a spin temperature below that of the CMB temperature. The neutral gas might
then be seen in absorption against the CMB. The expected signal may well exceed the
intensity of the emission signal (e.g. Tozzi et al, 2000),
Figure 2.3 Cartoon of the expected brightness temperature of the cosmic background in the vicinity of
the hydrogen reionization edge as a function of observing frequency. Three cases are shown for the HI
step, corresponding to different models of how the process took place. All three produce an effect that is
capable of being detected by LOFAR (from Shaver et al.1999).
2.4.2
Expected Angular Scales
Prior to full re-ionization the intergalactic medium was a mixture of neutral, partially ionized,
and fully ionized structures. A patchwork of neutral hydrogen emission, and possibly
absorption against the cosmic background radiation will result in fluctuating sky brightness
levels with structures up to a degree in size. Simulations suggest that the brightness
temperature will decrease when the angular scale increases (cf Tozzi et al, 2000).
The low-surface brightness HI signal itself can only be detected by smoothing the data to a
resolution of at least a few arcmin. This means that only the inner part of the LOFAR array,
out to baselines of a few km, will be effective in mapping the HI signal.
Page 13 of 59
2.4.3
Implications for LOFAR
LOFAR will be equipped with dipoles optimized for the middle of the 110-220 MHz band. At ~
150 MHz LOFAR would be sensitive to HI at z ~ 8.5 which is about in the middle of the
favoured redshift range for reionization.
The size of the virtual core is optimized for detecting a signal from structures of 10’ to 15’ at ~
150 MHz, i.e. the expected global signal from the first stages of reionization.
Because the epoch of reionization, and therefore the frequency of the spectral signal, is
uncertain a very wide backend, at least 32 MHz and preferably 64 MHz, will be required. The
adopted correlator design should cover such a wide band and allow a finely spaced grid of
frequencies to execute the search.
The spectral resolution and instantaneous frequency range is determined by the number of
available frequency channels. Taking the strawman design of 4096 frequency channels over
a frequency range of ??, an instantaneous frequency resolution of ?? will be attained for
each of the ?? beams. This is well matched to the expected line profile expected from the
global signal.
Although the collecting area and sensitivity of LOFAR is more than sufficient to detect the
expected reionization signal, need to detect and the reionization experiment imposes
stringent requirements on dynamic range and calibration. An expected global reionization
signal of ~15 mK must be detected within a total noise temperature for the virtual core due to
the 150MHz sky and the system of 0.5K (see Table 1.1), i.e. a dynamic range of 50.
Calibration of the faint expected signal of reionization dictates a telescope with a substantial
collecting area (cf. Shaver et al., 1999) and having long baselines (hundreds of kilometers).
Collecting area and resolution are needed to deal with a crucial aspect of this experiment,
namely to identify, model, and remove the foreground emission that will contaminate the
reionization signal. One obvious contaminant is the population of discrete radio sources. At
the flux density levels that LOFAR can reach within acceptable integration times (to noise
levels well below 10 µJy at 150 MHz) starburst galaxies will comprise the bulk of the radio
source population. The long baselines of LOFAR (> 100 km) are crucial for isolating and
spectrally characterizing these discrete sources, so that they can be removed from the
visibility data on the shorter baselines.
A second even more important contaminant is the diffuse nonthermal galactic foreground
emission which is responsible for the bulk of radio noise from the sky at frequencies at which
LOFAR will operate. Fortunately, this diffuse emission has very little structure on the 5'-60'
angular scales at which the signals will be sought. Faint galactic foreground fluctuations that
do exist should have no spectral features, allowing them to be disentangled from the
reionization signal.
Page 14 of 59
Figure 2.4 Comparison of the predicted brightness fluctuations as a function of angular scale with the
sensitivity attainable after integrations of 100 and 1000 hrs with a central aperture of LOFAR. The
simulated fluctuation spectra show the amplitude of the 3 sigma peaks in sky brightness as computed by
Tozzi et al (1999), for two different cosmological models. The sensitivity lines represent a confidence
level of 5 times the noise level attained after the indicated integration time. The flat portion of the
sensitivity curves at the right side of the plot indicate the response of a fully filled aperture of diameter
equal 300m to brightness fluctuation on angular scales greater than 22 arcminutes. Dispersing the same
number of elements over a wider area to obtain better angular resolution causes the sensitivity to
surface brightess to decrease, as denoted by the
diagonal line, so that once the aperture is diluted over a 1.5 km diameter area, the instrument should
still detect the background peaks at 5 times the noise level on angular scales of 6 arcminutes after 1000
hours of integration.
Any area in the galactic halo, where the foreground noise is minimal, can be selected to
search for the E-o-R signal. Initially several suitable areas, optimized for ionospheric stability
(long night-time integrations) and minimum dynamic range requirements (far from strong
confusing radio sources).
Long integration times, approaching weeks or more, will be required and devoted to this
important experiment. LOFAR's multi-beaming capability will enable the simultaneous
imaging of large areas of sky, increasing the effective integration times. The output of E-o-R
mapping experiments will be a large set of narrow-band images over a wide area of the sky
each encompassing a few hundred square degree of solid angle, over a wide frequency
range. These 'image cubes' will be analyzed at various spectral and angular resolutions via
standard power spectrum techniques, similar to those used in experiments that studied
fluctuations in the cosmic microwave background (reference). The results will be compared
with theoretical simulations for a range of cosmological models and sources of re-ionization.
2.5 Answers from LOFAR to Key Questions
LOFAR will provide a unique tool for investigating the dark ages of the Universe and the
epoch of reionization. It will provide answers for the fundamental questions posed in Section
2.1.
By detecting and measuring the frequency of the global spectral emission signal from neutral
hydrogen just prior to reionization, LOFAR will confirm that reionization occurred in the
presently favored redshift range (10 > z > 7) and accurately determine the epoch of
reionization (Question 1).
Page 15 of 59
By measuring the width of the global spectral emission feature and the amplitude of spatial
fluctuations in the signal, LOFAR will determine how gradually the reionization of the Universe
occurred (Question 2).
By measuring the amplitude and scale of the emission fluctuations and by searching for local
absorption features, LOFAR will constrain the sources responsible for the reionization and the
scale of structures that formed the first galaxies (Question 3).
If no signal from the reionization is detected by LOFAR, this in itself will be a fundamental
result. It will push the epoch at which reionization occurred to well beyond a redshift of z ~ 10
and mean that present views about the early evolution of the Universe will need to be revised
radically.
References
Abel et al, Science (2002)
Barkana and Loeb, (2000)
Barkema and Loeb 2001, Physics Reports 349, 125
Loeb and Barkema, 2001, Ann Rev. Astr. Astrophys., 39, 19.
Barkana and Loeb 2002
Briggs, de Bruyn, Vermeulen, A&A (2001)
De Bernardis et al., Nature 404, 995 (2000)
Fan et al. Ap.J. (astro-ph 0111184)
Gnedin & Ostriker Ap. J. 486, 581 (1997)
Gnedin, N. Ap.J., 542, 535-541 (2000)
Gnedin, 2001
Hui L. et al
Kamionkowski & Liddle Phys.Rev.Lett. 84, 4525-4528 (2000)
Madau, P., Meiksin, A. & Rees, M.J.: Ap.J. 475, 429-444 (1997)
Madau, P. 2000
Meiksin, A.: In: "Perspectives on radio astronomy: Science with Large Antenna Arrays",
published by ASTRON, Ed. M. van Haarlem (1999)
Rees, M.J.: In "Perspectives ...", (199)
Shaver et al. A&A 345, 380 (1999)
Shaver, P. and de Bruyn, A.G.: In "Perspectives ...." (1999)
Spergel & Steinhardt: Phys.Rev.Lett. 84, 3760-3763 (2000)
Tegmark & Zaldarriaga astro-ph/0004393
Tozzi,P., Madau,P., Meiksin,A. & Rees, M.J. Ap J. 528, 597 (2000)
Zheng et al. astro-ph/0005247
Page 16 of 59
3 FORMATION AND EVOLUTION OF GALAXIES,
CLUSTERS AND ACTIVE NUCLEI
3.1 Introduction and Key Questions
One of the most intriguing problems in modern astrophysics concerns the processes by which
galaxies and groups of galaxies formed and emerged, from the extremely smooth Universe
indicated by the microwave remnant of the original Big Bang (e.g. Lahav et al. 2000 and
references therein). A related problem concerns the formation of the massive black holes,
believed to be located at the centres of most galaxies and to power the enormously energetic
explosive events that are observed as quasars and radio galaxies.
Although there are various scenarios for the development of structure in the Universe, the
epoch and mechanism of the formation of galaxies, quasars and clusters are still open. Some
important questions are:
1. When and how did the first galaxies form?
2. When and how did the first active galactic nuclei (massive black holes) form?
3. When did the first large-scale structure form and how did the galaxies and intergalactic
gas evolve into rich clusters that we observe in the present Universe?
4. What is the relative time sequence for the emergence of galaxies, massive black holes
and proto-clusters? In particular, what is the star formation history of the earliest known
galaxies and how is this related to the birth of massive nuclear black holes and the
formation of galaxy clusters?
The metamorphosis that resulted in the changes mentioned above occurred when the
Universe was between 5 and 30% of its present age, corresponding to the redshift range 7 >
z > 1.5. Because the space density of luminous quasars and radio galaxies was several
orders of magnitude larger than now, this period of the Universe is often termed the “quasar
era”. During the quasar era the global star formation rate was a factor of 4 - 10 greater than at
present, massive elliptical galaxies formed and clustering of galaxies began to be formed out
of the primeval gas (e.g. Kauffmann et al 1996).
There are three classes of targets in this quasar era that will be observed by LOFAR with the
goal of investigating some of the above questions. These key types of objects are (i) luminous
radio sources, produced by black holes in the nuclei of massive forming galaxies, (ii)
"starburst" galaxies, i.e. infant galaxies observed to be undergoing a vigorous episode of star
formation and (iii) diffuse radio emission that can be used to probe intergalactic gas.
3.2 Distant Radio Galaxies: Probes of Massive Galaxy and
Cluster Formation
Half a century ago, the discovery of galaxies with radio luminosities of up to 5 orders of
magnitude greater than our own galaxy inaugurated a new era in observational cosmology.
Distant radio galaxies are still unique cosmic probes (e.g. Miley 1999, 2000). Because of their
huge optical and infrared luminosities and their lumpy appearances they are believed to be
massive galaxies undergoing formation as demonstrated in Fig. 3.1 (e.g. Röttgering and Miley
1996, Pentericci 1999, Rengelink 1999, Röttgering et al 1999).
Page 17 of 59
Figure 3.1 Distant radio galaxies are massive galaxies undergoing formation. This picture taken with the
Hubble Space Telescope shows that the distant radio galaxy 1138–262 (z=2.2) is extremely clumpy.
The structure bears a remarkable resemblance to the predicted images from computer models depicting
the birth of massive galaxies in clusters, through the merging together of large number of small galaxies
(from Pentericci et al 1999). LOFAR will locate and study the most distant forming massive galaxies.
Recent Dutch work has shown that luminous distant radio galaxies tend to be located in
regions of significant galaxy overdensity, that probably mark the early stages of rich clusters
of galaxies (Kurk et al 2000, Pentericci et al 2000). Observations of the environment of one
such radio galaxy has recently revealed the most distant grouping of galaxies known, a
presumed proto-cluster of >20 galaxies at a redshift of z ~ 4.1, corresponding to a distance of
~13 billion light years, or ~10% of the present age of the Universe (Venemans et al 2002).
Figure 3.2 The most distant known group of galaxies discovered with the VLT in Chile (from Venemans
et al. 2002). LOFAR will pinpoint the most distant forming proto-clusters. (Left) Spatial distribution of 20
Ly emitting galaxies located in a forming cluster associated with the powerful radio galaxy 1338-193 at
z ~ 4.1. (Right) Spectra of the 20 Ly emitting galaxies in the proto-cluster.
Luminous radio galaxies are therefore excellent laboratories for investigating the birth of the
first massive galaxies and the formation of the first massive black holes. In addition,
they signpost the earliest stages in the formation of rich galaxy clusters.
Page 18 of 59
LOFAR will detect luminous radio galaxies out to unprecedented distances. The most
efficient method for finding distant radio galaxies was developed in the Netherlands and uses
an empirical correlation between radio spectral steepness and distance (Blumenthal and
Miley 1979,Tielens et al 1979, Röttgering 1993, van Ojik 1995, de Breuck 1999). Because the
synchrotron-emitting electrons age and lose energy, the nonthermal radio spectrum steepens
at frequencies greater than ~ 1 GHz. For distant objects this "break frequency" is redshifted to
longer wavelengths in the observer's frame. Measurement of spectra at the lowest
frequencies (<100 MHz) is needed to filter out radio sources with steep spectra at large
distances.
Figure 3.3 Pinpointing the most distant active galaxies using their steep radio spectra.
(Left) Radio spectral index plotted against redshift (distance) for various samples of radio sources
illustrating that the most distant radio sources tend to have extremely steep nonthermal spectra.
Concentrating on objects with the steepest radio spectra is therefore a highly effective technique for
finding the most distant radio galaxies (Blumenthal & Miley 1979).
(Right) An optical spectrum of TN J0924-2201, taken with the Keck telescope in Hawaii showing the
Lyman spectral line of hydrogen (van Breugel et al. 1999). This object has a redshift of z ~ 5.1 and is
the most distant radio galaxy presently known. It is located at a distance of ~ 13.5 billion light years,
corresponding to an epoch when the Universe less than 10% of its present age. By measuring such
spectra at lower frequencies, LOFAR will extend such studies to larger distances.
Large-sky LOFAR surveys at 2 frequencies (e.g. 3000 sq. degrees at 30 MHz and 120 MHz)
will easily detect and determine radio spectra for more than 10 million radio sources with flux
densities larger than 1 mJy at 30 MHz (see Table 1.2). Simple extrapolation of spectra and
redshift statistics from existing radio surveys (e.g. Jarvis et al. 2001) shows that the LOFAR
surveys will contain more than 1000 radio galaxies with extremely steep radio spectra,
located at distances larger than the most distant radio galaxy presently known (Fig. 3.4).
Page 19 of 59
Figure 3.4 Redshift distribution expected for sources with a spectral index steeper than
= 1.3 detected in a 3000 sq. degree LOFAR survey at 30 MHz with a flux density limit of 1 mJy. The
distribution was calculated by extrapolation from data on samples of very bright radio sources compiled
by Jarvis et al (2001), under the assumption of a constant comoving source density.
The objects with extreme spectra from the LOFAR survey will be unique targets for following
up with large optical and infrared telescopes. Once their distance (redshift) has been
measured, optical and infrared diagnostics will be used to compare with models and
investigate the processes by which massive galaxies and clusters form in the early Universe.
Besides being able to pinpoint the most distant forming massive galaxies, LOFAR will be an
important tool for studying them. Absorption of continuum emission by neutral hydrogen at
1420 MHz is a well-known technique for probing gas at high redshifts. By using the ultrasteep spectrum technique to pinpoint suitable continuum "background" targets and by
providing a sensitive (multi-beaming) spectrograph at frequencies < 240 MHz, LOFAR will
extend the HI absorption technique to the redshifts > 5. In this way LOFAR can study
neutral hydrogen in the first clumpy proto-galaxies, predicted by hierarchical models of
galaxy formation.
3.3 Distant Starburst Galaxies: Probes of Galaxy Evolution.
A crucial stage in the emergence of galaxies is the epoch at which the first stars formed.
Galaxies where large numbers of stars are being produced are important laboratories for
studying galaxy formation. The IRAS satellite advanced the knowledge of such objects
substantially. One of the most important results of IRAS was the discovery of a class of
galaxies that are exceedingly bright in the infrared (Soifer et al. 1984). The excess infrared
emission is known to be due to radiation by dust produced during enormous bursts of star
formation.
Page 20 of 59
Figure 3.5 Radio emission from M82, the best studied "local" starburst galaxy (from Muxlow et al
1995). This galaxy is forming new stars at a rate of about 10 solar masses per year. The many clumps
are due to the remnants of exploding stars ("supernovae”). LOFAR will locate and study an
unprecedented number of distant starburst galaxies.
One of the most remarkable properties of galaxies is the existence of a very sharp relation
between the infrared emission (thermal) and the radio emission (synchrotron) emission.
Although van der Kruit (1971) had pointed out the existence of such a relation for nearby
galaxies, a decade later data from the IRAS satellite showed that this correlation extends over
several orders of magnitude of luminosity, from nearby spiral galaxies to luminous IRAS
starburst galaxies (See Figure 3.6 and e.g. de Jong et al. 1984; Sanders & Mirabel, 1985).
The infrared-radio relation is explained as a direct consequence of star formation. The
infrared emission is produced by warm dust, heated by energetic photons produced by young
stars, while the radio emission is due to synchrotron emission produced by the interaction of
supernovae with the magnetic field of the galaxy.
Figure 3.6 The tight correlation between the 1.4 GHz radio luminosity and the 60 micron IRAS infrared
luminosity as observed for a sample of nearby galaxies (from Condon 1991). The infrared-radio relation
is explained as a direct consequence of star formation. The infrared emission comes from warm dust,
heated by energetic photons produced by young stars, while the radio emission is due to synchrotron
emission produced by the interaction of supernovae with the magnetic field of the galaxy. LOFAR will
extend such studies to the era of the Universe when galaxy formation was rampant.
Page 21 of 59
Figure 3.7 The spectral energy distribution of the well-known nearby starburst galaxy M82. The spectral
shape is characteristic for all starburst galaxies. The spectrum is a composite one that arises from two
main components: synchrotron emission from supernova remnants (solid line) and thermal emission
from dust heated by young stars (dotted line). The dot-dashed line is the predicted spectral component
due to free-free emission.
The infrared-radio relation is the tightest known relationship involving the global properties of
galaxies. It contains important information about the distances of star-forming galaxies and
about the processes responsible for producing the stars.
For high redshift objects, the spectrum (see Fig. 3.7) will be shifted redwards and observed at
longer wavelengths. Because of the characteristic and constant rest-frame spectrum, the
radio/sub-mm continuum flux density ratio provides a sensitive redshift indicator for starbust
galaxies (cf. Carilli and Yun, 1999). Using a combination of the radio fluxes with data from
far-IR and millimeter surveys, LOFAR will measure the distances of an unprecedented
numbers of starburst galaxies.
The radio flux densities measured by LOFAR will provide an estimate for the star
formation rate of starburst galaxies independent of those derived from thermal radiation
from stars, at optical, infrared and sub-millimetre wavelengths. These (thermal)
measurements of the star formation rate show that star formation increases dramatically with
increasing distance. Because radio measurements are insensitive to dust obscuration, they
are needed in order that a complete census of the cosmic star-formation history can be made.
LOFAR will also search for evolution in the infrared - radio relation. Although changes in
the balance of thermal to nonthermal radiation as a function of redshift would complicate the
use of this relationship as a distance indicator, it would also imply that processes and/or
timescales of star formation in galaxies were different in the early Universe than from now.
Unraveling such effects would be of fundamental importance in understanding galaxy
evolution.
The depth of the LOFAR surveys, the excellent angular resolution and above all the large
simultaneous sky coverage will make LOFAR a powerful tool for studying large numbers of
starburst galaxies out to the redshifts at which the bulk of galaxy formation is believed to
occur. The "Rosetta stones" of nearby star-forming galaxies are M82 (see Fig. 3.5 and 3.7)
which has a star forming rate of 10 solar masses per year and the "ultra-luminous infrared
galaxy" Arp 220 (z=0.018) which forms stars at a rate of 250 solar masses per year. LOFAR
operating at ~200 MHz would be able to detect M 82 out to a redshift of 1.1 and Arp 220 out
Page 22 of 59
to z=3.3 (e.g. Garrett 2002). The vast majority of radio sources in LOFAR surveys near 200
MHz will be distant starburst galaxies. The number of starburst galaxies detected in a 3-year
survey with LOFAR should be about 10 million.
To carry out such studies, LOFAR surveys at ~200 MHz will be complemented by a new
generation of optical and infrared surveys that are already being conducted and will be further
developed during the next decade, such as Sloan, SIRTF, OMEGACAM on the VST and
VISTA. Additional data from radio surveys at high frequencies will provide information about
the radio morphologies and spectra, allowing the (more compact) starburst galaxies to be preselected and providing lists of targets whose redshifts will be measured using multi-object
spectroscopy. Such multi-object spectrographs will be capable of taking spectra of many
hundreds of galaxies simultaneously. LOFAR will thereby produce large numbers of starburst
galaxies that will be important objects for further study by the Atacama Large Millimeter Array
(ALMA) and the Next Generation Space Telescope (NGST).
3.4 Diffuse Radio Sources: Probes of Intergalactic Gas
Evolution
Extragalactic radio sources have long been used as probes of the surrounding intergalactic
gas (e.g. Miley 1980). The radio sources are shaped by pressure from the gas in which they
are embedded (see Fig 3.8). In addition, shocks in this gas (e.g. at the boundaries of galaxies
or clusters) can reaccelerate the older electrons and modifying the radio spectra. In clusters,
the motion of the galaxies through the gas during the lifetime of the sources can radically
distort the shapes of the radio sources, resulting in long tails of trails of radio emission (Miley
et al. 1972). These radio sources provide a "fossil" record. Under various assumptions, the
history of the nuclear activity and properties of the surrounding gas can be disentangled (e.g.
Ensslin et al. 1998, 2000).
Figure 3.8 Examples of cluster radio sources showing interaction with the intra-cluster gas.
Small panels - left. Radio images of 4 diffuse radio sources in nearby clusters showing the filamentary
structure of the synchrotron emitting plasma shaped by shocks in the cluster gas (Slee et al 2001).
Larger panel - right. Simulation of interaction between a cluster radio source and the intracluster
medium. This simulation takes account of the appropriate gas dynamics and magnetic field configuration
during the evolution of shocks in the cluster gas (Ensslin and Brüggen 2002).
The large numbers of cluster radio halos that will be observable with LOFAR will be used to probe
evolution in the intra-cluster gas due to e.g. merger-induced shocks out to a redshift of z ~1.
The low frequencies at which LOFAR will operate are optimized for detecting the oldest radio
sources with steep spectra and thereby for studying such fossils. The ages of the
synchrotron-radiating electrons are at most 100 million years, i.e. small compared with the
age of the Universe (~1.5 billion years). There is increasing evidence that radio galaxies go
through recurrent episodes of nuclear activity. A study of the statistics of the oldest (steepest
spectra) radio sources as a function of redshift will provide new information about the dutycycle of nuclear activity.
Page 23 of 59
The excellent sensitivity for steep radio spectra and diffuse features uniquely equips LOFAR
for studying such radio fossils and using them to probe intergalactic gas out to high
redshifts. The two most important classes of targets for such studies, cluster radio halos and
giant radio sources, probe gas in what are likely to be the densest and sparsest regions of the
Universe.
3.4.1
Cluster Radio Halos: Tracing Over-dense Regions.
An important but relatively neglected building block of galaxy clusters are the extended
diffuse halos of radio emission, sometimes found to inhabit the centres of rich clusters. These
halos have typical sizes of ~1 Mpc and steep radio spectra (< -1), well suited to
observations with LOFAR. The halo synchrotron emission provides diagnostics for studying
the magnetic field and plasma distribution within clusters, important inputs to models of
cluster evolution.
Prime examples of known cluster halos are the Coma cluster (e.g. Wilson, 1970; Jaffe, 1977),
Abell 2256 (e.g. Bridle et al., 1979; Röttgering et al 1994) and A3667 (Röttgering et al 1997;
see Fig 3.9). Clusters found to contain radio halos appear to have large X-ray luminosities
and galaxy velocity dispersions (e.g. Hanish 1982), both characteristics of dynamic activity
that would be expected if the clusters are composed of merging sub-clusters as predicted by
some cluster evolution models (Röttgering et al. 1994 and references therein).
Figure 3.9 Cluster radio sources as tracers of turbulence in intracluster gas.
Radio and X-ray observations of the rich galaxy cluster Abell 3667 (Röttgering et al 1997, Johnsten et al
in prep.). The X-ray emission (data from the XMM satellite shown with false colours) is a tracer of the
hot cluster gas. The radio emission (contours from the MOST radio telescope) traces shocks and
magnetic field structure in the cluster gas. The radio arcs observed in the outer region of the cluster are
presumed due to turbulence and shocks caused by the merging of 2 sub-clusters to form Abell 3667.
The shocks reaccelerate remnant relativistic plasma previously ejected by cluster galaxies to produce
the arc-like radio structures. LOFAR will extend such studies to high redshifts.
Page 24 of 59
Several models have sought to explain the origin of the relativistic electrons emitting the halo
radio radiation. These include (i) superposition of large numbers of relic radio sources (Harris
and Miley 1978), (ii) galactic wakes (Roland 1981), (iii) intra-cluster turbulence and shocks
induced by small clusters falling into the large potential well of the main cluster (Miniati et al
2001 and references therein) and (iv) charged pion decay after hadronic interactions with the
thermal hot cluster gas (Dolag and Ensslin 2000, and others). Since these models predict
different radio morphologies, magnetic field configurations, and spatial variations in radio
spectra, LOFAR observations complemented by X-ray images of the hot gas (e.g. XMM and
Chandra) can help distinguish between these models.
Because of their diffuseness, low luminosities, and steep-spectra, radio halo sources are
difficult to detect with conventional facilities, such as Westerbork (Kempner and Sarazin
2001). However, their properties are well matched to LOFAR. The number of cluster halos
that can be detected with LOFAR can be estimated using a simple model proposed by
Ensslin and Röttgering (2002). This model takes into account the locally observed fraction of
cluster radio halos, the observed relation between radio and X-ray luminosity, and a PressSchecter description of the merging rate of massive clusters as a function of redshift. A
LOFAR survey at ~120MHz, covering half the sky to a 5-sigma flux limit of 0.1 mJy (1 hour
per pointing) is feasible on the time scale of a year and would detect ~1000 halos at the 10
sigma level, of which 25 % are expected to be at redshifts larger than z ~ 0.3.
A more direct strategy to find cluster radio halos would be to use the large future cluster
catalogues as a target list for deep integrations with LOFAR at frequencies ~100 - 200 MHz.
New cluster surveys are being made with the XMM X-ray telescope, the Planck satellite and
the Sloan Digital Sky Survey should catalogue as many as 500,000 new clusters (e.g. Pierre
et al 2001, Barthelmann and White 2002). LOFAR has sufficient sensitivity to detect existing
radio halos in all these clusters.
Observations of the morphologies of large numbers of such distant radio halos will study the
effects of cluster formation and evolution on the cluster gas, e.g. due to shock waves
produced by the cluster mergers. If halo sources are indeed associated with the aggregation
of sub-clusters, the frequency of their occurrence might well be much higher in more distant
proto-clusters, where merging is predicted to be rampant. By measuring the statistics of
cluster halos and their properties as a function of redshift, LOFAR will therefore constrain
models of cluster formation and evolution.
A fascinating effect related to the diffuse cluster radio sources that may well also be
observable with LOFAR is emission from the large Thompson scattering halos that are
expected to occur around some clusters. Radiation from a bright radio source near the cluster
centre should undergo Thomson scattering by the intra-cluster medium that will result in a
scattered halo on a much larger scale. The scattering process leads to a significant and
characteristic polarization pattern allowing it, in principle, to be separated from any intrinsic
halo emission (Sunyaev, 1982; Wyse and Sarazin, 1990). The shape and spectrum of the
Thompson halo carries important information on the time-evolution of the central radio source,
possible beaming of the central source and the properties of the intra-cluster magnetic field.
For a 10 Jy source, (not uncommon at 100 MHz for a central cluster source), a scattered halo
with flux density of 10 - 100 mJy is expected. This should easily be detectable by LOFAR.
Larger scattering halo are expected to be produced around older sources. The Thompson
7
halo around a 10 year-old source will extend out to a projected radius of 3 Mpc and the
8
scattering halo around a 10 year-old source will be 30 Mpc in extent. Such scales will be
readily mapped in the LOFAR surveys.
3.4.2
Giant radio sources: Tracing under-dense regions.
A small fraction of extragalactic radio sources have sizes that exceed 1 Mpc. These giant
radio galaxies (e.g. Willis, Wilson and Strom 1974, Saripalli et al 1986, Subrahmanyan et al
1996, Schoenmakers 1999) are amongst the largest known structures in the Universe. To
8
become so large, the giant radio sources must be relatively old (> ~10 y) and be imbedded in
a very tenuous intergalactic medium. It is notoriously difficult to obtain good and reliable
samples of such giants since their surface brightness at high radio frequencies is so very low.
Page 25 of 59
Because of its sensitivity at low frequencies, LOFAR can easily distinguish the very steep
spectrum extended emission of giant radio galaxies from confusing background and
foreground sources that have flatter spectra. Studies of nearby and distant radio galaxies with
LOFAR will pinpoint special regions in the Universe where the intergalactic medium is sparse,
i.e. a contrasting extreme from rich clusters. Comparison of the environment of giant radio
sources with that of in rich clusters can provide unique information about the effect of largescale structure on galaxy formation and evolution.
Because of their extreme sizes giant radio sources are also an important class of objects for
testing unification scenarios in which the observed properties of different types of active
galaxies and quasars are determined by the orientation of the radio jets with respect to the
observer (e.g. Barthel 1989).
3.5 Gamma Ray Bursters – Prompt Emission and Afterglows
Many of the currently viable models of gamma-ray bursts (GRBs) postulate NS-NS or NS-BH
coalescense as the energy source for the burst. In these models, the possibilities for radio
burst generation by neutron star motion through a magnetic field are similar to those
described for the LIGO events above. Another possible source of radio bursts is in the
magnetized wind that may flow from the binary system (Usov and Katz 2000). The very large
magnetic field associated with these winds is supported by a surface current at the boundary
of the wind and the ambient medium, and variations in this current may drive coherent
emission of low frequency electromagnetic waves. This radiation peaks at frequencies
fundamentally set by the proton gyrofrequency, and for typical burst parameters the peak
occurs around 1 MHz. The high-frequency tail may be detectable by LOFAR; Usov and Katz
estimate that flux densities as high as 100 Jy at 30 MHz may be produced.
Radio emission following GRB's has been detected in about 40% of the objects for which
observations have been possible. This radio emission is usually in the form of “afterglows”,
which last for a number of days after the GRB. Shorter lasting (one or two day) “flares” have
also been seen in a few objects. Both radio phenomena show strong absorption at low
frequencies which indicate that only the very brightest radio events might be detectable with
LOFAR. These detections would test flare and afterglow models involving forward and
reverse shocks in a frequency range where the models have not yet had to confront data.
The data gathered to date suggest that the gamma rays are beamed and the radio afterglows
are not. Constraints on the size of the opening angle and the observed rate at which GRB's
show radio afterglows (40%) indicate that for every gamma-ray burst beamed in our direction
there should be 40 radio afterglows. Since, these afterglows last for days, the transient survey
area covered during their time “on” should be at least half the sky, meaning about 20 radio
afterglows per day will fall at some time in the LOFAR field of view. The difficulty will be in
detecting the afterglows. Absorption greatly reduces the flux at 150 MHz, from about 100
microJy for the typical afterglow to about 2 microJy. The reduced flux means that only about
0.3% of the volume out to a redshift of 1 (the typical redshift of a GRB) will be surveyed by
LOFAR. Assuming a uniform and isotropic distribution of GRB sources, this brings the
number of detectable afterglows to about 22 per year. These LOFAR-detected afterglows
would provide a valuable sample for GRB afterglow studies.
3.6 Answers from LOFAR to key questions
The above discussion shows that LOFAR will provide a wealth of information relevant to the
fundamental cosmological questions posed in Section 2.1.1. To summarise the most relevant
LOFAR observations:
1. LOFAR will detect radio galaxies at larger distances than hitherto possible. This will
constrain the epoch of formation of the first massive black hole formation (Question 2)
Study of these LOFAR distant radio galaxies at other wavelengths will provide information
about the formation of massive galaxies and links between nuclear activity and star formation
in galaxies at the earliest epochs (Questions 1, 4).
Since distant radio galaxies pinpoint proto-clusters, study of the environment of LOFAR
distant galaxies will constrain the formation of clusters at the earliest epochs (Question 3).
Page 26 of 59
2. With its unprecedented sensitivity to nonthermal radio emission from star formation,
LOFAR will detect large numbers of star-forming galaxies in the early Universe. Studies of
these objects will provide new information on the evolution of star formation at the earliest
epochs of the Universe. Comparison of the LOFAR properties of distant field galaxies with
galaxies in groups (e.g. close to distant radio galaxies) will provide unique information about
galaxy and cluster evolution (Questions 1, 2, 3, 4).
3. The statistics and properties of the diffuse steep-spectra radio sources in clusters and
around giant radio sources will provide information about the importance of merging galaxy
sub-clusters as a function of cosmic epoch and constrain hierarchical models for the
formation and evolution of clusters (Question 4).
References:
Bartelmann M., White S. D. M., 2002, A&A, 388, 732
Barthel P. D., 1989, ApJ, 336, 606
Beichman C. A., Neugebauer G., Habing H. J., Clegg P. E., Chester T. J., eds 1988, Infrared
astronomical satellite (IRAS) catalogs and atlases. Volume 1: Explanatory supplement, Vol. 1
Blumenthal G., Miley G., 1979, A&A, 80, 13
Bridle A., Fomalont E., Miley G., Valentijn E., 1979, A&A, 80, 201
Carilli C. L., Yun M. S., 1999, ApJ, 513, L13
Condon J. J., 1992, ARA&A, 30, 575
De Breuck C., 1999, Ph.D. thesis, Univ. Leiden, (1999)
de Jong T., Clegg P. E., Rowan-Robinson M., Soifer B. T., Habing H. J., Houck J. R., Aumann H. H.,
Raimond E., 1984, ApJ, 278, L67
Dolag K., Enßlin T. A., 2000, A&A, 362, 151
Ensslin T. A., Biermann P. L., Klein U., Kohle S., 1998, A&A, 332, 395
Enßlin T. A., Br¨uggen M., 2002, MNRAS, 331, 1011
Ensslin T. A., Röttgering H. J. A., 2002, A&A, in press
Garrett M. A., 2002, in Proceedings of the 6th European VLBI Network Symposium, Ros, E., Porcas,
R.W., Lobanov, A.P., & Zensus, J.A. (eds.), MPIfR, Bonn, Germany, astro-ph/0206270
Garrett M. A., de Bruyn A. G., Giroletti M., Baan W. A., Schilizzi R. T., 2000, A&A, 361, L41
Hanisch R. J., 1982, A&A, 116, 137
Harris D., Kapahi V., Ekers R., 1980, A&AS, 39, 215
Harris D., Miley G., 1978, A&AS, 34, 117
Jaffe W., 1977, ApJ, 212, 1
Jarvis M. J., Rawlings S., Willott C. J., Blundell K. M., Eales S., Lacy M., 2001, MNRAS, 327, 907
Katgert P., Oort M. J. A., Windhorst R. A., 1988, A&A, 195, 21
Kauffmann G., 1996, MNRAS, 281, 487
Kempner J. C., Sarazin C. L., 2001, ApJ, 548, 639
Kurk J. D., Röttgering H. J. A., Pentericci L., Miley G. K., van Breugel W., Carilli C. L., Ford H., Heckman
T., McCarthy P., Moorwood A., 2000, A&A, 358, L1
Lahav O., Bridle S. L., Percival W. J., Peacock J. A., Efstathiou G., Baugh C. M., Bland-Hawthorn J.,
Bridges T., Cannon R., Cole S.,
Colless M., Collins C., Couch W., Dalton G., De Propris R., Driver S. P., Ellis R. S., Frenk C. S.,
Glazebrook K., Jackson C., Lewis I.,
Lumsden S., Maddox S., Madgwick D. S., Moody S., Norberg P., Peterson B. A., Sutherland W., Taylor
K., 2002, MNRAS, 333, 961
Miley G., 1999, Ap&SS, 269, 219
Miley G., 2000, in From Extrasolar Planets to Cosmology: The VLT Opening Symposium, Proceedings
of the ESO Symposium held at Antofagasta, Chile, 1-4 March 1999. Edited by Jacqueline
Bergeron and Alvio Renzini. Berlin: Springer-Verlag, 2000. p. 32., pp 32
Miley G. K., 1980, ARA&A, 18, 165
Miley G. K., Perola G. C., van der Kruit P. C., van der Laan H., 1972, Nature, 237, 269+
Miniati F., Jones T. W., Kang H., Ryu D., 2001, ApJ, 562, 233
Muxlow T. W. B., Pedlar A., Sanders E. M., 1995, p. 74
Pentericci L., 1999, Ph.D. thesis, , Univ. Leiden
Pentericci L., Kurk J. D., Röttgering H. J. A., Miley G. K., van Breugel W., Carilli C. L., Ford H., Heckman
T., McCarthy P., Moorwood A., 2000, A&A, 361, L25
Pierre M., Alloin D., Altieri B., Birkinshaw M., Bremer M., Böhringer H., Hjorth J., Jones L., Le F`evre O.,
Maccagni D., McBreen B., Mellier Y., Molinari E., Quintana H., Röttgering H., Surdej J., Vigroux
L., White S., Lonsdale C., 2001, The Messenger, 105, 32 Rengelink R. B., 1999, Ph.D. thesis,
Leiden University
Roland J., 1981, A&A, 93, 407
Röttgering H. J. A., 1993, Ph.D. thesis, University of Leiden
Page 27 of 59
Röttgering H. J. A., Best P. N., Lehnert M. D., eds 1999, The Most Distant Radio Galaxies, Proceedings
of the colloquium, Amsterdam, 15-17 October 1997, Royal Netherlands Academy of Arts and
Sciences
Röttgering H. J. A., Miley G. K., 1996, in Bergeron J., ed., The Early Universe with the VLT, pp 285–299,
Springer-Verlag
Röttgering H. J. A., Snellen I., Miley G., de Jong J., Hanisch B., Perley R., 1994, ApJ, 436, 654
Röttgering H. J. A., Wieringa M. H., Hunstead R., Ekers R. D., 1997, MNRAS, 290, 577
Sanders D. B., Mirabel I. F., 1985, ApJ, 298, L31
Saripalli L., Gopal-Krishna Reich W., Kuhr H., 1986, A&A, 170, 20
Schoenmakers A. P., 1999, Ph.D. thesis, Utrecht University
Slee O. B., Roy A. L., Murgia M., Andernach H., Ehle M., 2001, AJ, 122, 1172
Soifer B. T., Neugebauer G., Helou G., Lonsdale C. J., Hacking P., Rice W., Houck J. R., Low F. J.,
Rowan-Robinson M., 1984, ApJ, 283, L1
Subrahmanyan R., Saripalli L., Hunstead R. W., 1996, MNRAS, 279, 257
Sunyaev R. A., 1982, Soviet Astronomy Letters, 8, 175+
Tielens A., Miley G., Willis A., 1979, A&AS, 35, 153
van Breugel W., de Breuck C., Stanford S. A., Stern D., Röttgering H. J. A., Miley G. K., 1999, ApJL,
518, 61
van der Kruit P. C., 1971, A&A, 13, 405+
van Ojik R., 1995, Ph.D. thesis, University of Leiden
Venemans B. P., Kurk J. D., Miley G. K., Röttgering H. J. A., van Breugel W., Carilli C. L., De Breuck C.,
Ford H., Heckman T.,
McCarthy P., Pentericci L., 2002, ApJ, 569, L11
Willis A., Strom R. G., Wilson A. S., 1974, Nature, 250, 625
Willson M. A. G., 1970, MNRAS, 151, 1
Wise M. W., Sarazin C. L., 1990, ApJ, 363, 344
Page 28 of 59
4 THE HIGHEST ENERGY PHENOMENA
4.1 LOFAR as a Cosmic Ray Detector
20.5
The acceleration of particles as observed in Cosmic Rays up to a particle energy of 10
eV
is one of the outstanding challenges in High Energy Astrophysics. In particular, in the range of
15
20.5
energies between 10 - 10
eV, both the acceleration site and the physical process that
causes particles to gain such extreme energies are unknown. For comparison, the energy in a
20.5
10
eV particle is comparable to the kinetic energy in a tennis ball launched by a world
champion player. The energy density in CR nuclei is an order of magnitude larger than in CR
electrons in the Galaxy. The CR nuclei are responsible for most of the diffuse gamma-ray
background whereas the CR electrons mainly emit synchrotron radiation at radio
15
wavelengths. At energies below 10 eV, it is now generally agreed on that the acceleration
process is a universal first order Fermi-type acceleration, so-called diffusive shock
acceleration, in strong shocks that occur in galactic Supernova Remnants (SNRs), and
possibly superbubbles and pulsars as well. At these "medium" energies models exist that can
reproduce the observed "universal" particle energy spectrum. The theoretical results are
consistent with observations of a few SNRs both at radio and X-ray wavelengths although a
detailed agreement between theory and observation is still lacking. It should be noted that
there may exist an injection problem for electrons which require preheating and are therefore
less easy to accelerate in the standard shock model than atomic nuclei.
20.5
At 10
eV, however, the particle cyclotron radius in the galactic magnetic field is much
larger than the dimensions of the Galaxy, and simply no container inside the Galaxy is known
to exist which can host and accelerate particles up to such extreme energies. Up to now, only
20
a few tens of particles have been observed at energies above 10 eV and only a few events
19
per square km per century occur above 4.10 eV. Possible source candidates of these
UHECR (ultra high energy cosmic rays) are the shocks in radio lobes of powerful radio
galaxies, intergalactic shocks (IGSs) created during the epoch of galaxy formation, the
sources of Gamma Ray Bursts (GRBs), or, alternatively, the particles are decay products of
supermassive particles such as topological defects from phase transitions in the early
universe.
It is clear then that the study of the highest-energy Cosmic Rays is one of the most
interesting problems in High Energy Astrophysics as it relates to such fundamental problems
as the nature of the intergalactic medium, the epoch of structure formation in the universe, the
so-called Hypernovae and merging binary neutron stars in GRBs, and/or fundamental high
energy physics in the early universe.
LOFAR will contribute directly to the study of CRs both at medium and at the highest
energies:
1. by the efficient detection of coherent radio emission from UHECRs, and the accurate
determination of the directions of their sources which can then be identified. The coherent
radio emission in the form of "antenna radiation" and emitted by the thin (less than the
radiated wavelength at the low observing frequencies of LOFAR) slab of secondary
particles in the electromagnetic cascade that is generated when the primary CR travels
through the Earth's atmosphere;
2. by active radar echo detection of UHECR-induced extensive air showers in the VHF
frequency range (30 - 100 MHz). This method of detection is based on scattering of a
short radar pulse off the column of ionized air which is produced in an air shower from a
primary CR as it travels through the earth's atmosphere.
4.1.1
Background Information: Cosmic Rays
The term Cosmic Rays (CRs) stands for energetic particles, both atomic nuclei and electrons
20.5
with energies up to 10
eV. They have been detected directly in our Galaxy, and indirectly in
our Galaxy as well as other galaxies from their electromagnetic signature in gamma rays
down to the radio domain, e.g. in Supernova Remnants (SNRs), radio pulsars, solar flares
and flare stars, protostars, planetary magnetospheres, X-ray binaries, jets in radio galaxies
and quasars, Active Galactic Nuclei (AGN) and Gamma Ray Bursts (GRBs). Their differential
energy spectrum is observed to follow a (combination of) power-law(s) over a wide range in
Page 29 of 59
8
20.5
energies between 10 eV and 10
eV with a power-law index between -2 - -3.8 and, for a
large part, with a typical index -2.7 - -2.8, eg. at 20 - 200 GeV an index of -2.71 ± 0.04 for
protons, and of -2.79 ± 0.08 for helium nuclei (Menn et al. 2000). Small changes in slope
15.5
17.8
19
appear at "the knee" around 10
eV, the second "knee" at 10
eV, and the ankle at 10
eV. CRs with energies above the ankle ("the foot") are called UHECR, ultrahigh energy
cosmic rays (Sakaki et al. 2001; see Nagano & Watson 2000 for a review). Because of their
universal occurrence and comparable properties, much effort has gone into identifying a
universal acceleration process. Nowadays, it is believed that diffusive shock acceleration
(Axford et al. 1977; Krimsky 1977; Bell 1978), a first-order Fermi-type acceleration process, is
this universal mechanism (for a recent review see Achterberg 2001). It operates in strong
collisonless shocks such as occur in a multitude of explosive objects in the universe - SNRs,
superbubbles, radio pulsars, accreting compact objects and black holes in AGN, jets in X-ray
binaries and radio galaxies - and produces a differential power law spectrum in energy with
power law index -2, close to and somewhat flatter than is observed, for any shock as long as
the shock is both strong and non-relativistic,.
4.1.1.1 Composition
12
As for the composition of CRs, it should be noted that electrons are observed up to 10 eV
3
-3
only, and that in our Galaxy the energy density in electrons is only 6 . 10 eV m as compared
5
-3
to 5 . 10 eV m for atomic nuclei. The total energy density in CRs near the Sun is about 1
-3
-3
-3
MeV m (number density 10 m ) comparable to that in thermal energy of the Galactic gas,
to the energy density of the Galactic magnetic field, and to that in the Cosmic Microwave
15
Background (CMB). Up to the "knee" in the spectrum at 10 eV the composition is
dominated by protons but at higher energies the composition is essentially unknown. As the
atomic number of a primary CR increases, the muon multiplicity in the atmospheric cascade
created by the primary particle increases. A recent attempt to determine the composition of
primary particles around the knee in the L3+C(osmics) experiment at CERN through a
combination of the muon detector - the most accurate detector of particle momenta in its time
- and an Extensive Air Shower Array on top of the roof of the L3 main hall has demonstrated
that either the primary particles all have a high atomic number or that the simulation codes for
the cascade production are not correct at these energies (Wilkens 2002). The latter possibility
is a realistic one as the cross-sections for forward particle production which go into the codes
are not accessible to measurements in the lab.
4.1.1.2 TeV Gamma Rays
TeV gamma-rays have been detected in a small number of cases. There is a strong interest
in detecting these and other kinds of high-energy particles such as high-energy neutrinos.
These could either be produced bottom-up by conventional processes or top-down as decay
products from "exotic" and "relic" supermassive particles, a collective term used for
theoretically proposed particles in extensions of the Standard Model (e.g., "WIMPS" for
24
weakly interacting massive particles, axions, neutralinos) or topological defects (10 eV)
associated with the boundaries of fields which have undergone a phase transition and
symmetry breaking. High-energy primary gamma rays can be efficiently detected from their
secondary muons which are produced when the gamma-ray interacts in Earth's atmosphere
(Alvarez-Muñiz & Halzen 1999). Detection of high-energy neutrinos is important as a
diagnostic of the acceleration process, e.g. in AGN radio lobes. Detection of exotic particles is
of direct relevance to determine the nature of the dark matter component in the universe.
A first attempt to detect TeV gamma rays from 8 GRBs detected simultaneously with
BATSE on board CGRO has been performed with the muon-detector at L3+C at CERN (van
den Akker et al. 2002). Only an upper limit was obtained, and there is a clear need for more
sensitive detectors of such high-energy particles because of their fundamental significance.
4.1.1.3 Neutrino Oscillations
A handful of TeV gamma ray sources have, at present, been detected beyond doubt.
Essentially, two mechanisms are known to produce such gamma-rays: inverse Compton
scattering of high-energy electrons off their own synchrotron radiation, or alternatively, decay
of pions produced by energetic protons colliding with thermal gas in the source. Note that,
according to standard shock acceleration theory, such highly energetic protons are expected
to be present at the same location as the electrons. To discriminate between both possibilities
Page 30 of 59
detection of high-energy neutrinos is important (Waxman & Bahcall 2000) as for example will
be possible with AMANDA. The efficiency of neutrino detection in its turn depends on the
neutrino rest masses; if non-zero, oscillations between different kinds of neutrinos occur. One
input parameter to the solution of the neutrino oscillation problem is the absolute flux of
secondary muons at sea level. This has been one of the motivations for the application of the
L3 detector at CERN for CR experiments in L3+C as described by van Mil (2001). The
absolute flux of muons at sea level is expected to be determined by the L3+C measurements
with an accuracy of 4 % at an energy of 100 GeV (Petersen 2002).
4.1.1.4 Energy Budget in Radio galaxies
There are a number of untested predictions of diffusive shock wave acceleration theory, and
a number of remaining problems. According to theory electrons are more difficult to
accelerate than ions. This is known as the injection problem. The observed underabundance
of electrons in the galaxy with respect to nuclei is in qualitative agreement with theory, but it
should be noted that if a similar situation applies to AGN and jets of radio galaxies the
minimum energy present in accelerated particles in these objects is enhanced dramatiaclly,
and can pose a severe problem to the energy budget.
4.1.1.5 LOFAR and Origin of the Highest Energy CRs
Existing model computations of shock wave acceleration agree that CRs can be accelerated
in SNRs up to the knee, either in individual remnants or in a concerted way. However, beyond
the knee, the cyclotron radius of a particle becomes too large for trapping in a SNR, and
acceleration of particles at these energies must be of extragalactic origin (Norman et al. 1995;
Gallant & Achterberg 1999; Bahcall & Waxman 2000) . Possible sites include the shocks at
the termination of jets in radio lobes of galaxies, intragalactic shocks from the epoch of galaxy
formation and the environs of GRBs. Alternatively, they could be decay products of
supermassive particles (Bhattacharjee & Sigl 2000). As the cyclotron radius of a particle of
20
energy 10 eV becomes comparable to the thickness of the Galaxy accurate determination of
the direction of motion of the particle would allow discrimation between its possible sources.
So far, the celestial distribution of observed UHECR is cosnistent with an isotropic distribution
(Dova 2001; Takeda et al. 2001). It is here that LOFAR is expected to play a decisive role
based on the coherent radio pulses that such energetic particles emit as has been proposed
first by Jelley et al. (1965; see also Smith et al. 1965) and recently by Falcke & Gorham
(2001). A primary CR induces a particle cascade in the atmosphere that is aligned along the
direction of motion of the primary particle. A substantial part of the cascade is leptonic and
gives rise to various kinds of radiation, such as Cherenkov emission, transition radiation and
also synchrotron emission in the terrestrial magnetosphere. At the high Lorentz factors
considered here the cascade is in fact confined to a slab of meters thickness. As a result, at
long wavelengths the particles act coherently in an antenna-like fashion and the emission is
enhanced over the incoherent level by a factor N, N being the number of electrons in a
bunch. It should be noted that it is the net number of particles - the net difference between
number of electrons and positrons - which counts (Alvarez-Muñiz et al. 2000; Gorham et al.
2000; Saltzberg et al. 2001). From the arrival times and intensities of the radio pulse at
various antennas of LOPES (and at a later stage LOFAR) the direction of the primary particle
can then be determined with an accuracy of 1 - 2 degrees. This mechanism forms the core of
the proposal LOPES for a LOFAR Prototype Station by Falcke & Gorham (2001).
A related problem is the cut-off energy of the CR spectrum. One would expect that
19
photo-pion production on the CMB would lead to a cut-off at energies of 5.10 eV, the socalled GZK (Greisen-Zatsepin-Kuzmin) cut-off. The observed particles at higher energies
imply that the source regions are at a distance less than 100 Mpc. Clearly, it will be important
to actually detect a cut-off, and LOFAR is expected to contribute significantly.
4.1.1.6 LOFAR, active radar and UHECR
An UHECR can be detected also by interrogating the ionized air column produced by the
UHECR with radar pulses (Gorham 2001). The UHECR creates an air shower as the primary
particle passes through Earth's atmosphere which is leptonic for a large part, and ionizes the
air. The resulting ionized trail can be detected by scattering of an impinging radar pulse. This
Page 31 of 59
radar technique is expected to be competitive to fluorescence detection at near-UV
18
wavelengths for primary energies above 10 eV. Meteors of about 1 µg have a comparable
kinetic energy and also generate ionization trails which are detected by radar. The main
difference between the two ionization trails is that a meteor travels at a speed of 30 km/s (at
least below 70 km/s) while the trail from an UHECR appears practically instantaneously. The
19
technique is well suited to detect neutrinos above 10 eV for which air showers develop
much deeper in the atmosphere. A unique signature of their presence would be the detection
of highly-inclined, near-horizontal showers.
References
Achterberg, A.: 2001 IAU Symposium 195, 291
Aharonian, F.A., Atoyan, A.M.: 2000 A&A 362, 937
Alvarez-Muñiz, J. & Halzen, F.: 1999 ApJ 521, 928
Alvarez-Muñiz, J, Vázquez, R.A., Zas, E.: 2000 Phys. Rev. D 62, 619
th
Axford, W.I., Leer, E., Skadron, G.: 1977 EOS 57, 780 (Proc. 15 Int. Cosmic Ray Conf.,
Plovdiv)
Bahcall, J.N. & Waxman, E.: 2000 ApJ 542, 542
Bell, A.R.: 1978 MNRAS 182, 147
Bhattacharjee, P. & Sigl, G.: 2000 Phys. Rep. 327,
Biermann, P.L. et al.: 2001 A cosmic ray ground array combined with a radio array in the
Netherlands and
Germany
Boggs, S.E., Lin, R.P., Slassi-Sennon, S., Coburn, W., Pelling, R.M.: 2000 ApJ 544, 320
Breitschwerdt, D., Dogiel, V.A., Völk, H.J., Kassim, N.E., Weiler, K.W.: 2002 A&A 385, 216
Dova, M.T.: 2001 astro-ph/0101379
Duric, N.: 2000, in Radio Astronomy at Long Wavelengths, eds. R.G. Stone, K.W. Weiler,
M.L. Goldstein,
J.-L. Bougeret, Geophysical Monograph 119, American Geophysical Union,
Washington, DC, p.277
Gallant, Y.A. & Achterberg, A.: 1999 MNRAS 305, L6
Gorham, P.W., Saltzberg, D.P., Schoessow, P. et al.: 2000 Phys. Rev. E 62, 8590
th
Jelley, J.V.: 1965, in Proc. of 9 ICRC vol I, 698
Kassim, N.E. & Yusef-Zadeh, F.: 2000, in Radio Astronomy at Long Wavelengths, eds. R.G.
Stone, K.W.
Weiler, M.L. Goldstein, J.-L. Bougeret, Geophysical Monograph 119, American
Geophysical Union, Washington, DC, p.277
Krimsky, G.E.: 1977, Doklady Akad. Nauk. SSR 242, 1306
Longair, M.: 1990, in Low Frequency Astrophysics from Space, eds. N.E. Kassim & K.W.
Weiler (Berlin:
Springer-Verlag) p. 227
Menn, W. et al.: 2000 ApJ 523, 281
Nagano, M., Watson, A.A.: 2000 Rev. Mod. Phys. 72, 689
Norman, C.A., Melrose, D.B., Achterberg, A.: 1995 ApJ 454, 60
Petersen, B.G.: 2002, The Cosmic Ray Induced Muon Spectrum Measured with the L3
Detector, PhD
Thesis, University of Nijmegen, in preparation
Sakaki, N. et al.: 2001 Proc. of ICRC, 333
Saltzberg, D., Gorham, P., Walz, D. et al.: Phys. Rev Letters 86, 2802
th
Smith, F.G., Porter, N.A., Jelley, J.V.: 1965 in Proc. of 9 ICRC, vol I, 701
Takeda, M. et al.: 2001 Proc. of ICRC, 341
van Mil, A.: 2001, Cosmic-Ray Muons in the L3 Detector, PhDThesis, University of Nijmegen
Waxman, E. & Bahcall, J.N.: 2000 ApJ 541, 707
Webber, W.R.: 1990, in Low Frequency Astrophysics from Space, eds. N.E. Kassim & K.W.
Weiler (Berlin:
Springer-Verlag) p. 217
Webber, W.R.: 1997 Space Sci. Rev. 81, 107
Wilkens, H.G.S.: 2002, Multimuon Events at L3+Cosmics, in Proc. of the XVIII ECRS,
Moscow
Page 32 of 59
4.2 Gravitational Waves - LIGO events
Mainly for practical reasons, most existing radio sky surveys have been made at relatively low
frequencies. From these surveys have come both serendipitous discoveries (for example,
pulsars) and the catalogues of sources used for high-resolution mapping at high frequencies.
Serendipitous discoveries often provide the greatest opportunities for advances in our
understanding of the physics of astronomical objects. LOFAR would be an excellent
instrument for the study of bursting and transient radio sources. The large instantaneous field
of view means that LOFAR can effectively monitor a large fraction of the sky at all times,
averaging on many different time scales. This makes possible for the first time a sensitive
unbiased survey for astronomical transients. The factor of 100 increase in collecting area over
previously operated instruments and the array's capability for data buffering and rapid
electronic pointing, assisted at the lower frequencies by plasma delays, will also for the first
time make possible prompt follow-up of transients detected with LOFAR and other
instruments.
4.2.1
An Unbiased Survey for Transient Astronomical Radio Sources (STARE)
The inner compact configuration of LOFAR will consist of 25% of the array's collecting area
arranged in a circle of 2 km diameter. All signals from the individual receptors will be brought
to the central processor, allowing for synthesis mapping of a substantial fraction of the entire
field of view of the receptors (the receptor field of view is an area of about 100 degrees
diameter for the low frequencies, and about 25 degrees diameter for the high
frequencies).These images will be examined in real time for transient events, and averaging
will provide information on a range of time scales. This activity will represent the first unbiased
survey for bursting and transient sources carried out with a radio telescope with a large
collecting area and with sophisticated interference mitigation capabilities. Large-area monitors
for transient sources at other wavelengths in astronomy have been very productive. For
example, such observations led to the discovery of gamma-ray bursts and of samples of
supernovae useful for cosmological studies. A low-frequency radio survey is likely to reveal
processes giving rise to coherent emission, and has great discovery potential.
4.2.2
The Physics of Collapse and Explosion
4.2.2.1 Radio Supernovae
Radio studies of SNe have shown that multi-frequency monitoring of the rapid radio turn-on is
critical to determining the early phases of the explosion, the physical mechanisms at work
(e.g., SSA vs. f-f absorption) and the structure and density of the final, pre-supernova stellar
mass-loss stages. For very rapidly evolving SNe, only at low frequencies does one have
sufficient warning to obtain high quality turn-on information, since a transition seen at 100
MHz takes ~15 times as long from explosion to peak flux density as it does at 5 GHz. For
example, SN 1987A reached 5 GHz peak already at age ~1 day and 840 MHz peak at age
~3 days, which led to very poor radio turn-on sampling of this once in 400 year event.
Follow-up, long-term, frequent (at least initially) monitoring at multiple frequencies is then
needed to establish the properties of the last periods of stellar mass-loss before explosion. At
cm wavelengths the presupernova mass-loss evolution for periods of ~20,000–30,000 years
before the explosion can be measured. However, only at long wavelengths does the radio
emission remain sufficiently bright that one can study the mass-loss history from even earlier
epochs.
Unfortunately, due to the competition between rapidly declining optical depth, slowly
decreasing flux density with time, and increasing flux density at lower frequencies due to the
nonthermal spectrum, there is a rough equality of peak flux density at all frequencies. Thus,
27
the peak spectral luminosity even at LOFAR frequencies is unlikely to exceed L R ~ 10
-1
-1
erg s Hz (flux densities of a few mJy for objects at or closer than the Virgo Cluster) except
for unusual cases. Also, because of rapid evolution in the early phases, sensitivity of <1 mJy
(5 sigma) in ~30 minutes is desirable with resolution of 1 arcsec to reduce confusion with
background emission from the parent galaxy.
Page 33 of 59
Radio supernovae are also candidates for targeted searches with LOFAR since they are
known to occur in dusty starburst galaxies (exemplified by the enigmatic objects in M82) with
no optical counterpart. To find and study such objects one must carry out continuous, or at
least frequent, monitoring of the radio structure of many nearby galaxies to look for the
appearance of new radio point sources. Such observations would improve the rather poorly
established supernova rates in different galaxy types.
4.2.2.2 Giant Pulses
The strong magnetic fields associated with neutron stars in pulsars are believed to be
responsible for the nonthermal processes that produce the observed radio pulses. In two
cases, these pulses are exceedingly strong, and appear as “giant pulses”. Giant pulses have
been detected from the Crab pulsar and from the millisecond pulsar PSR B1937+21, both in
our Galaxy. In the case of the Crab pulsar, a giant pulse occurs on average once every ten
seconds. The flux density distribution of the giant pulses has a power-law distribution above
200 Jy. Although the lower flux density cutoff has been determined to be about 50 Jy, no
upper flux density cutoff has been seen. Individual giant pulses with flux densities exceeding
1000 Jy have been detected, and some giant pulses can outshine the entire Crab Nebula.
Pulsars emitting giant pulses with these properties could be detected at least to the galaxies
in the Local Group and potentially even further. Indeed, giant pulse-emitting pulsars offer the
best hope for detecting pulsars beyond the Magellanic Clouds.
Detecting more giant pulse-emitting pulsars would have a number of scientific applications.
First, detecting a new Galactic pulsar(s) with giant pulses may make the study of such pulses
easier. The strong intensity of the Crab Nebula makes study of the individual “normal” pulses,
and some giant pulses, difficult. Second, the two known giant pulse-emitting pulsars are quite
dissimilar: The Crab pulsar is a young, strong-field pulsar while PSR B1937+21 is a
(presumably) old, recycled, millisecond pulsar. It is not clear to what extent conclusions
determined for one can be applied to the other. Third, existing observations suggest that the
giant pulses from the Crab pulsar occur in the same region as the “normal” pulses, but result
from a short modulation of the pulse emission process. Additional pulsars would both test this
model as well as provide additional constraints on the nature of this modulation mechanism.
Finally, the dispersion and rotation measures for pulsars in other galaxies would allow us to
probe their interstellar media and potentially place constraints on the intergalactic medium in
the Local Group.
4.2.2.3 LIGO Events
The detection of gravitational waves will represent one of the most important measurements
in the history of physics. Evidence for the reality of such signals, and of their astrophysical
origin, may be provided by other instruments operating in coincidence. Evidence for
coincidence is, of course, not necessary to prove the existence of gravitational waves, but
has the potential to strengthen greatly the case for astrophysical gravitational wave sources,
and to provide complementary information on the sources.
The most promising sources for the gravitational wave detectors under construction are the
coalescence of two compact objects in a binary system, including those composed of two
neutron stars (NS-NS binaries), a neutron star and a black hole (NS-BH binaries), and two
black holes (BH-BH binaries). NS-NS binaries have been most extensively studied
theoretically. Predicted event rates are highly uncertain, but it has been estimated that
3
coalescence will occur at a rate of 3 per year (distance / 200 Mpc) for the initial LIGO
system. Predictions for LIGO II are even more uncertain since at the LIGO II levels of
sensitivity NS-BH and BH-BH events are expected to dominate. Current estimates give
LIGO II rates of once per week for NS-BH events and once per hour for BH-BH events. For a
typical coalescing system, the time during which gravitational radiation is emitted within the
detector bandwidth is approximately 15 minutes, during which time there are approximately
16,000 cycles. Hansen and Lyutikov (2000) have examined the possibility of radio emission
from NS-NS coalescence. They model the system as a conducting sphere (one neutron star)
moving through an external magnetic field (due to the other neutron star). They compute
induced currents and the acceleration of charged particles drawn off the neutron star, and by
assuming the energy of the charged particles is converted to radio waves with the same
Page 34 of 59
-2
2/3
efficiency as in pulsars, they predict a flux density of F 2.1 mJy (/0.1) ( D/100 Mpc) B15
-5/2
a7 where is the efficiency of energy conversion, D is the distance, B15 is the magnetic field
15
7
in units of 10 G, and a7 is the semi-major axis in units of 10 cm. Again, we emphasize this
prediction is highly uncertain. However, it does give a plausible mechanism for the generation
of radio waves in NS-NS coalescence.
One would expect any radio emission to be modulated at the orbital period of the NS-NS
binary. This is roughly given by the dynamical time scale of about a millisecond, though the
signal will be chirped as the binary merges. In analogy with pulsars, one would expect the
radio emission to be polarized, and the time dependence of the polarization would provide
important information on the geometry of the system.
4.3 LOFAR as an All-Sky Monitor
LOFAR will detect low-frequency radio emission associated with a very broad spectrum of
objects whose output in this band is variable and transient.
There are two key aspects to LOFAR studies of the variable sky:
•
•
Large field of view -- LOFAR as an all-sky monitor – the low frequency, central
concentation of dipoles (`virtual core') and multiple beams provide LOFAR will the
potential to monitor two thirds of the sky daily
Low-frequency emission -- in addition to, and coupled with, the potential of LOFAR
as an all-sky monitor, is the possibility to detect for the first time variable emission at
very low frequencies, probing the low-energy tail of particle acceleration and/or
previously unobserved coherent radiation processes which may be the signatures of
exotic phenomena}
In the following we will discuss these two complementary aspects in turn.
The combination of a central concentration of about one quarter of the LOFAR dipoles within
a region ~10 km across, combined with the low observing frequency and potential for multiple
beams allows LOFAR to be used as an all-sky monitor (Fig. 4.1).
Figure 4.1 One concept of LOFAR as an all-sky monitor. Four `virtual core' beams observe the zenith
as the sky tracks past, monitoring up to two-thirds of the sky daily. A newly-identified event can then be
localised with arcsecond accuracy using the full array.
Page 35 of 59
4.3.1
Particle acceleration in cosmic explosions
All of the most energetic phenomena known in our Universe, namely Gamma Ray Bursts
(GRBs), Supernovae (SNe), accretion onto supermassive black holes in Active Galactic
Nuclei (AGN) and accretion onto stellar-mass black holes and neutron stars in X-ray Binary
systems (XRBs) result in radio emission. This is a direct consequence of the energy input into
the surrounding medium resulting in a shock which accelerates particles to high enough
energies (Lorentz factors may exceed 1000) that they emit synchrotron radiation as they
spiral around the magnetic field lines (which are also probably enhanced in the shock region).
In many cases (AGN, XRBs, probably also GRBs) this emitting region is associated with a
collimated outflow, or `jet'. By the very nature of these explosive events, this energy input and
consequential particle acceleration are variable phenomena -- recurrent in some cases -- and
as a result the radio emission from these sources is also variable.
In more detail, all of the aformentioned phenomena produce a phase of particle acceleration
followed by energy losses which may be energy dependent (`radiative' synchrotron or
inverse-Compton losses dominate) or energy-independent (adiabatic expansion). These
processes are convolved with the evolving optical depth to synchrotron self-absorption within
the source region itself, which is generally high initially at the low frequencies observable with
LOFAR. The general pattern of behaviour is remarkably similar for the different kinds of
sources described above, due to the same underlying physics -- see Fig 2. The violent,
relativistic phenomena provide are the lighthouses of the Universe and inject a huge amount
of energy into the interstellar, and intergalactic medium. For the first time LOFAR as an allsky monitor will allow us to detect and monitor these events wherever they are in the sky,
providing us with an unparalled study of their properties, distributions etc.
Furthermore, LOFAR will be key to providing alerts of these energetic phenomena to other
facilities. These violent phenomena currently occupy a large fraction of the world's high
energy astrophysics community, utilising e.g. Chandra, XMM-Newton, HST orbiting
observatories and a vast array of ground-based telescopes, to record data while these events
are occuring. The goals of such observations are to probe the laws of physics under extremes
of gravity, pressure and density in ways which can never be achieved in a terrestrial
laboratory. However, these observatories, with their smalls fields of view, require triggering by
wider-field detectors. Currently the high-energy astrophysics community is supplied in this
respect only by the soft X-ray all-sky monitor (ASM) onboard RXTE, and the (less successful)
instruments onboard HETE. Neither mission is expected to survive throughout the operational
lifetimes of the new class of orbiting observatories, removing from them a key aspect of their
operation. LOFAR will fill this void, providing notice of activity of all these classes of events
through detection of their radio synchrotron emission, and as such will fulfill a key role for the
high-energy astrophysics community.
Page 36 of 59
Figure 4.2 (a): Simulation of a single synchrotron `bubble' event -- a generic representation of the
processes underlying the radio emission from GRBs, SNE, AGN and XRBs. A shock produces a phase
of particle acceleration of finite, typically short, duration. Subsequently the relativistic gas expands until it
becomes optically thin at progressively lower frequencies. In this simulation the timescales are typical
for GRBs or XRBs -- at GHz frequencies the emission peaks within a few days and begins to decay. At
LOFAR frequencies the source evolution is slower, due to the greater internal optical depth, and the
source peaks about one month after particle injection, but from this point onwards is stronger than at
higher frequencies and may remain visible for a year or more subsequently. This model is based on van
der Laan (1966).
Figure 4.2 (b, left): Radio observations of the X-ray transient source CI Cam. The time and frequency
evolution of the event is very similar to that predicted from the simple model presented in Fig 4.2(a).
Furthermore, note that at the lowest frequencies (in this case 300 MHz) the rising phase of the event
was detectable within days of the high frequency emission (which was already peaking). An expanding
emitting region (right) was found to be associated with this transient event, which was accompanied by a
bright X-ray flare and attracted world-wide attention.
How many of these different types of explosive events we expect to detect can be
approximated based on our knowledge of the coupling between higher-energy emission and
radio emission, coupled with the statistics of the properties of these sources at e.g. X-rays. In
table 1 we present estimates for the different types of transient object.
Page 37 of 59
Predicted detection numbers -- transient sources
Class of object
GRB afterglows (extragalactic)
Supernovae (mostly extragalactic)
Persistent Galactic X-ray binaries
Transient Galactic X-ray binaries
Transient X-ray binaries in M31
Expected rate per year
~20
~1
~5
~10
~1/3
In Figure 4.3 we illustrate the observed radio fluxes of persistent and transient neutron-star
and black hole XRBs, clearly demonstrating that most will be readily detectable with LOFAR.
In fact, we will probe the luminosity function of such sources deeper than has been achieved
in X-rays.
Figure 4.3 Radio vs. X-ray emission for persistent and transient neutron star and black hole X-ray
binaries, scaled to a distance of 1 kpc and with X-rays corrected for absorption (Gallo & Fender, 2002).
All the sources in this figure would be detected with LOFAR in all-sky-monitor mode (admittedly the
weaker sources would take up to a week), whereas the current X-ray all sky monitor onboard RXTE has
a sensitivity of only tens of milliCrab - ie. LOFAR will detect more of these sources than has been
possible in X-rays.}
Detection of radio emission from sources in M31 and the rest of the local group would be a
notable `first'. Furthermore, popular models for the `Ultraluminous X-ray sources' in many
external galaxies sugegst that they should be significantly beamed towards us (relativistic
aberration, see e.g. Koerding, Falcke \& Markoff 2002) in which case we should also detect
several of these systems, providing important support for this hypothesis.
It should be noted that the variability associated with the stellar-mass events (GRBs, SNe,
XRBs) occurs on considerably shorter timescales than that associated with AGN. Thus, while
AGN may prove to constitute the largest class of significantly variable sources observed with
LOFAR, they will be easily distinguishable from the other classes of event.
In addition to the transient, outbursting, sources, quasi-persistent emission associated with jet
production in X-ray binaries will also be monitored. Currently six systems (Cyg X-3, SS 433,
GRS 1915+105, LSI +61 303, Cyg X-1 and Sco X-1) are known which are bright enough that
LOFAR will provide {\em daily} radio light curves. A further, fainter, population of possibly {\em
Page 38 of 59
fifty} black hole and neutron star XRBs, accreting at lower rates, should be detected within
one year.
In both transient and persistent studies we will probe the jet formation characteristics of XRBs
far deeper than ever before. Furthermore, scaling the radio and X-ray fluxes from Fender \&
Kuuklers (2001), LOFAR will actually probe the population of X-ray transients considerably
fainter than has been possible with X-ray surveys. For the first time we will be discovering
more of these phenomena at radio wavelengths than in any other band.
LOFAR as an all-sky monitor, in opening up a new window on the universe, will also provide
us with the opportunity to observe coherent low-frequency radio emission from exotic sources
for the first time. In all likelihood by monitoring the whole sky at radio wavelengths for the first
time we will discover not only predicted sources of low-frequency radio emission (next
section), but also previously unsuspected objects and phenomena -- the prospects are truly
exciting.
4.3.2
New emission regimes at LOFAR frequencies
In addition to its capabilities as an all sky monitor, observing at low frequencies opens up the
possibility of exploring the low-energy distribution of particles in the emitting plasma, with
important consequences for our understanding of mass-flow in these objects. As well as the
synchrotron emission associated with shock-wave phenomena and consequential particle
acceleration, there are also mechanisms suggested for the production of transient lowfrequency radio emission via coherent radiation processes associated with exotic events. We
consider these possibilities further below:
4.3.2.1 Low-frequency synchrotron emission
The synchrotron emission frequency of a relativistic electron in a magnetic field scales as
2
, where is the Lorentz factor of the electron. In observing synchrotron emission below
100 MHz, we are probing considerably further towards the low-energy tail of the electron
distribution than has been previously possible. This is important for the total mass in the
emitting region, as the number of electrons of a given Lorentz factor resulting from shock
-p
accleration can be considered as a power-law, e.g. $N()d , where p~2. As an example,
the number (=mass) of emitting particles responsible for an optically thin synchrotron
spectrum between 1-15 GHz (a typical broad observing range) is four times less than that
required to explain the same spectrum observed to 100 MHz, and ten times less than that
required to explain the spectrum observed to 20 MHz. Quantitatively constraining the mass
outflow and comparing with the mass inflow rate is a key question for models of the accretion-outflow coupling in XRBs and AGN.
Furthermore, low-frequency imaging of radio-jet XRBs may be expected, in analogy to AGN,
to reveal low-surface brightness radio nebulae powered by the jet, providing an independent
measure of the energy budget of the outflow and its interaction with the ISM (see Fig 4).
Page 39 of 59
Figure 4.4 A large-scale (> 5 arcmin) radio nebula powered by an inner relativistic jet from the neutron
star XRB Cir X-1, observed at 1 GHz with the Australia Telescope Compact Array (from Fender 2002).
These observations are at 1 GHz; already at 2.3 GHz the nebula is resolved out - at LOFAR frequencies
many more of these structures should be detected, revealing the interaction of the jet with the ISM, and
providing an independent measure of the mass and energy output from the jet.
4.3.2.2 Coherent processes
Beyond synchrotron emission, coherent radiation processes may be detectable at LOFAR
frequencies. Predictions include LIGO Events and prompt emission from GRBs:
The detection of gravitational waves will represent one of the most important measurements
in the history of physics. Evidence for the reality of such signals, and of their astrophysical
origin, may be provided by other instruments operating in coincidence. Evidence for
coincidence is, of course, not necessary to prove the existence of gravitational waves, but
has the potential to strengthen greatly the case for astrophysical gravitational wave sources,
and to provide complementary information on the sources.
The most promising sources for the gravitational wave detectors under construction are the
coalescence of two compact objects in a binary system, including those composed of two
neutron stars (NS-NS binaries), a neutron star and a black hole (NS-BH binaries), and two
black holes (BH-BH binaries). NS-NS binaries have been most extensively studied
theoretically. Predicted event rates are highly uncertain, but it has been estimated that
3
coalescence will occur at a rate of 3 per year (distance / 200 Mpc) for the initial LIGO
system. Predictions for LIGO II are even more uncertain since at the LIGO II levels of
sensitivity NS-BH and BH-BH events are expected to dominate. Current estimates give LIGO
II rates of once per week for NS-BH events and once per hour for BH-BH events. For a typical
Page 40 of 59
coalescing system, the time during which gravitational radiation is emitted within the detector
bandwidth is approximately 15 minutes, during which time there are approximately 16,000
cycles. Hansen and Lyutikov (2000) have examined the possibility of radio emission from NSNS coalescence. They model the system as a conducting sphere (one neutron star) moving
through an external magnetic field (due to the other neutron star). They compute induced
currents and the acceleration of charged particles drawn off the neutron star, and by
assuming the energy of the charged particles is converted to radio waves with the same
efficiency as in pulsars, they predict detectable radio emission at LOFAR frequencies.
One would expect any radio emission to be modulated at the orbital period of the NS-NS
binary. This is roughly given by the dynamical time scale of about a millisecond, though the
signal will be chirped as the binary merges. In analogy with pulsars, one would expect the
radio emission to be polarized, and the time dependence of the polarization would provide
important information on the geometry of the system.
Many of the currently viable models of GRBs postulate NS-NS or NS-BH coalescense as the
energy source for the burst. In these models, the possibilities for radio burst generation by
neutron star motion through a magnetic field are similar to those described for the LIGO
events above. Another possible source of radio bursts is in the magnetized wind that may flow
from the binary system (Usov \& Katz 2000). The very large magnetic field associated with
these winds is supported by a surface current at the boundary of the wind and the ambient
medium, and variations in this current may drive coherent emission of low frequency
electromagnetic waves. This radiation peaks at frequencies fundamentally set by the proton
gyrofrequency, and for typical burst parameters the peak occurs around 1 MHz. The highfrequency tail may be detectable by LOFAR; Usov & Katz estimate that flux densities as high
as 100 Jy at 30 MHz may be produced.
References
van der Laan H., 1966, Nature, 211, 1131
Fender R., 2002, in prep
Fender R., Kuulkers E., 2001, 324, 923
Gallo E., Fender R., 2002, in prep
Hansen B.M.S., Lyutikov M., 2001, MNRAS, 322, 695
Koerding E., Falcke H., Markoff S., 2002, A\&A, 382, L13
Usov V.V., Katz J.I., 2000, A\&A, 364, 655
Page 41 of 59
5 THE MILKY WAY AND NEIGHBOURING GALAXIES
LOFAR will contribute directly to the study of CRs both at medium and at the highest energies
by detailed imaging of the galactic distribution at low frequencies where the emission is
dominated by sunchrotron radiation from CR electrons. These observations will allow a
detailed confrontation with theory and permit the ultimate test of the standard model of
diffusive shock acceleration. Determination of the CR electron distribution will be possible by
detailed investigation of absorption "holes" in the synchrotron radio emission from the Galaxy
at low frequencies. Such absorption is imprinted by intervening ionized thermal gas around
young hot stars in the Galaxy (HII regions) with known distances on the radio emission on its
way to Earth. From the amount of absorption the synchrotron emissivities can be derived, and
from it, given the magnetic field strength, the spatial distribution of the electron component in
CRs. These can then be compared directly to the electron distribution obtained from the soft
(< 1 MeV) Galactic gamma-ray background which are caused by bremsstrahlung and inverse
compton scattering of CR electrons (< 1 GeV). Comparison with the nuclear component in
CRs as obtained from the diffuse galactic gamma-ray background above 100 MeV provides
an important test of the diffusive shock acceleration mechanism. This technique requires
observations at low frequencies at high spatial resolution (of 1') and a high sensitivity (down
2
to 0.1 mJy/arcsecond );
5.1 Testing the Standard Shock Acceleration Model
SNRs, both in our Galaxy and in neighbouring galaxies appear to be the sources of most of
the accelerated electrons observed in their radio synchrotron emission. This is corroborated
by gamma-ray observations which reveal several strong sources associated with SNRs in
our Galaxy and which are caused by CR nuclei (Webber 1997). X-ray spectra of SNRs which
still have a pulsar inside ("plerions") are compatible with Fermi acceleration at relativistic
shocks (Gallant 2002; van der Swaluw 2002). The fact that acceleration models of SNRs can
15
produce the required number of particles up to an energy of 10 eV is not a proof that this is
how the Galaxy accelerates particles. A definite test will be to determine the actual spatial
distribution of CR electrons and CR nuclei. In principle this can be done by combining radio
observations from the Galaxy with the galactic diffuse gamma-ray emissions as follows
(Longair 1990; Webber 1990; Duric 2000). At the low frequencies observable with LOFAR the
galactic synchrotron emission dominates. Further, HII regions become optically thick below 30
MHz and absorption caused by the ionized gas in galactic HII regions with known distances
shows up in high resolution observations as holes ("swiss cheese" appearance). A detailed
spatial distribution of the synchrotron emissivity as a function of frequency (and polarization)
can then be determined. Together with magnetic field determinations, eg. from radio pulsar
measurements, the spatial distribution of energetic electrons between 200 - 300 MeV can
then be determined, and compared with electron distributions independently obtained from
the galactic soft (30 - 800 keV) gamma-ray background which is a combination of relativistic
bremsstrahlung and Inverse Compton scattering of CR electrons below 300 MeV (which
cannot be measured near Earth due to solar modulation) off the interstellar radiation field
(Boggs et al. 2000). Further, this CR electron distribution can be compared to the spatial
distribution of energetic protons inferred from the diffuse hard gamma-ray (100 MeV - 100
0
GeV) galactic background which comes from secondary mesons which are produced when
CR protons collide with the ambient interstellar gas (Aharonian & Atoyan 2000). These data
then constitute a strong test on the acceleration models. A similar procedure can be applied
to nearby galaxies. Conversely, the data give a handle on the existence of a galactic wind.
For instance, the discrepancy between the radial dependence of Galactic CR nuclei as
inferred from EGRET diffuse gamma-rays above 100 MeV and the most likely CR distribution
from SNRs, superbubbles and pulsars, can be explained by propagation effects in a galactic
wind where CR transfer by diffusion is taken over by advection in the wind at a certain height
above the disk (Breitschwerdt et al. 2002).
Finally, reliable determinations of spatial variations in the spectral index of the radio
continuum emission in a number of selected well-resolved SNRs at low frequencies such as
will be possible by LOFAR allow testing of the diffusive shock acceleration hypothesis and
study of the problem of the injection energies (Kassim & Yusef-Zadeh 2000).
Page 42 of 59
5.2 Thermal and Nonthermal Emission in Nearby Galaxies
Radio continuum emission from galaxies provides significant information on the properties of
the interstellar medium. At low frequencies radio emission will be dominated by the
synchrotron mechanism. At frequencies between 20 and 300 MHz one predominantly
observes the old population of relativistic electrons present in the disks and haloes (or thick
disks) of galaxies, i.e. the aged version of the relativistic electrons which were supposedly
generated by supernovae and associated events (such as the supernova shocks ploughing
through the interstellar medium).
Imaging the radio emission of galaxy disks at low frequencies will enable us to separate better
the thermal and non-thermal emission. This is done best by comparing many images
spanning a large range in frequency rather than images at only a few frequencies as has
been done until now. Escpecially high resolution imaging below 300 MHz will provide a much
more robust determination of the non-thermal component, and by implication the thermal
component in higher-frequency images. This is important, because it provides us with one of
the least biased views of star formation in galaxies. Low-frequency imaging at the same time
provides detailed information on spectral index variations across galaxies, which in turn
function as important clues to location-dependent aspects of relativistic electron production
and evolution.
Only a few low-frequency surveys of galaxies presently exist at 80 and 160 MHz (Slee
1972a,b; 1977) and at 57.5 MHz (Israel and Mahoney 1990), all with limited resolutions too
poor to resolve any spatial structure. However, they do imply the frequent occurrence of
significant spectral flattening also at the low-frequency end, with spectral turnovers in the 100
- 1000 MHz range. This flattening may be explained by relativistic electron energy losses or
free-free absorption of synchrotron emission by the ionized interstellar medium. Israel &
Mahoney (1990) have shown that in the latter case, galaxies must have a pervasive presence
of well-mixed clumpy thermal/nonthermal gas, with an ionized component at temperatures
3
well below Te = 10 K. This component might correspond to the ionized fraction of the cold
neutral interstellar medium. Alternatively, the observed spectral turnovers could be caused by
relativistic electron energy losses in large-scale galactic winds (Hummel 1991; Pohl et al
1991). At least in the case of the Local Group galaxy M33, Israel et al. (1992) found that freefree absorption is the most probable cause of the observed low-frequency spectral flattening.
Either explanation is extremely interesting. If free-free absorption is the dominant process,
further low-frequency observations provide one of the few ways, if not the only one, of
studying a major but hitherto virtually inaccessible component of the interstellar medium. If
galactic winds apply, low-frequency observations again provide one of the few ways of
studying the large-scale energetics and disk-halo interactions of spiral galaxies. Further
progress in this field requires accurate determination of integrated galaxy spectra in the
frequency range 20 – 1000 MHz, with good spectral sampling. LOFAR would play the
essential role at the lower end of this range. In addition, low-frequency imaging of e.g. edgeon galaxies is also powerful discriminating tool as galactic wind models predict a spectra
steepening away from the galactic plane, whereas the free-free absorption model predicts
little or no spectral change perpendicular to the plane. In the plane itself there will be strong
absorption by ionised gas in HII regions.
References
Hummel K., A&A, 251, 442, 1991
Israel F.P., Mahoney M.J., 1990 ApJ, 352, 30
Israel F.P. et al. 1992, A&A, 261, 47
Pohl et al., 1991 A&A, 250, 302
Slee O.B., 1972a ApL, 12, 75
Slee O.B., 1972b Proc Astr Soc Austral, 2, 159
Slee O.B., 1977 Austral. J. Phys. Suppl., 43, 1
Page 43 of 59
5.3 Supernova Remnants and the Distribution of Ionized Gas
in the ISM
The nonthermal spectra and extended morphologies of Galactic SNRs ideally match LOFAR’s
unprecedented, low frequency surface brightness sensitivity. The naturally large fields
encompass the largest, individual Galactic SNRs and efficiently survey whole populations in
nearby galaxies. Continuum spectrum studies are more accurate because of wide frequency
coverage, and thermal absorption and scattering effects can only be measured at LOFAR
frequencies.
5.4 Discovering New SNRs
Catalogs are severely selection effect limited and incomplete at surface brightness levels
-20
-2
-1
~10 W m Hz sr (Green 1991), which LOFAR will exceed by two orders of magnitude.
Catalogs under-sample both older SNRs and bright, compact, young SNRs, required for
understanding SN/SNR birthrates, statistics, ISM energy input, and for comparison with
progenitor populations. Sensitive (< 1 mJy/beam) LOFAR Galactic plane surveys will easily
reveal many (>100) new SNRs and PSR/SNR associations, critical to studies of the radio
lifetime of SNRs and PSR birth kinematics (Frail et al. 1994). LOFAR will also reveal unique
sources such as the SS433 jet system within W50 and as seen in the Galactic center. Recent
330 MHz VLA images have shown an unprecedented combination of resolution and surface
brightness sensitivity and has detected many new exotic sources (Kassim & Frail 1996, Lang
et al. 1999, LaRosa et al. 2000). However it falls short of detecting emission on scales large
enough to connect with structure seen on single dish maps (Sofue 2000) and linked to
Galactic center black hole activity. LOFAR’s dense uv coverage will achieve this and provide
equally spectacular inner Galaxy maps in many more directions.
5.4.1
Spectral Studies
SNR radio continuum spectra trace the energy spectra of shock accelerated relativistic
electrons and test predictions of shock acceleration theory, probing the complex modulation
of blast-wave physics by the pre-existing environment into which remnants evolve. Theory is
presently very poorly constrained, since models predict varieties of subtle variations which
current observations are too inaccurate to test (Jones et al. 1998). Increasingly believable
evidence of spectral variations is emerging, but the poor accuracy of the primitive lowest
frequency data remains the weakest link and guarantees that LOFAR will have a powerful
impact.
LOFAR will revolutionize the accuracy of SNR spectra where Fermi theory predicts curvature
and strongly constrains magnetic fields. Present spectra are inaccurate at low frequencies,
and believable curvature is suggested towards only a few sources (Reynolds & Ellison 1982).
Integrated spectra also test Van der Laan theory where controversial bends are linked to
cosmic ray gas compression (Green 1990). Spatial spectral variations test theory in greater
detail, e.g. steep spectrum breakout regions interpreted as weaker shocks (Pineault et al
1998). But curved electron spectra, magnetic field gradients, particle acceleration/diffusion
and other effects weave a complex relationship which present maps are too crude to
constrain. This is despite the maturity of spectral analysis (Anderson & Rudnick 1993, Zhang
et al 1997) utilizing new techniques (e.g. spectral tomography, Katz-Stone et al. 2000) robust
against calibration errors. But they cannot overcome the poor resolution and sensitivity of the
lowest frequency images. LOFAR spectra will distinguish remnant morphologies (e.g.
composites and barrels, Dwarakanath 1991, Gaensler, 1998) and reveal the influence of
pulsar winds (Frail et al 1994). A preeminent mystery is the missing Crab shell, demanding
explanation in context of blast wave physics/ISM properties. LOFAR will improve the best limit
(Frail et al 1995) by two orders of magnitude.
5.4.2
Absorption & ISM Studies
The surprising discovery of SNR internal absorption at 74 MHz implies LOFAR will detect
thermal material inside numerous remnants. The Cas-A absorption (Error! Reference
source not found.) is only the second detection of unshocked ejecta inside a young SNR
Page 44 of 59
(Kassim et al 1995). It is expected from theory and provides a crucial probe of reverse shock
physics in young SNRs. The Crab also shows thermal absorption from its foreground thermal
filaments, constraining their location relative to the pulsar powered nonthermal emission
(Bietenholz et al 1997). Extending SNR thermal absorption measurements beyond these
pathologically bright sources requires LOFAR.
SNR integrated spectra reveal low frequency turnovers (Kassim 1989) due to extrinsic
thermal absorption. This offers a powerful constraint on the distribution of ionized ISM gas,
-3
including hard upper limits on the Warm Ionized Medium density ( 0.26 cm ) which can be
greatly improved by lower frequency ( 50 MHz) LOFAR measurements. The patchy
absorption is consistent with Extended HII Region Envelope absorbers suggested by
stimulated low frequency recombination lines (Anantharamaiah 1986) which LOFAR will also
be uniquely sensitive too. However existing spot frequency data are at primitive resolution
(>15’), with the 74 MHz VLA image in Error! Reference source not found. being the first
spatially resolved detection of ISM absorption against a Galactic SNR (Lacey 2001). A >100
times more sensitivity, frequency versatile LOFAR will redefine this field of study. For inner
Galaxy complexes the contrast between absorption and emission will disentangle the relative
superposition of HII regions and SNRs as in W30 (Kassim & Weiler 1990) and the Galactic
center (Pedlar et al 1989), but in most cases is impossible without LOFAR.
Scattering measurements test blast wave physics which predicts turbulence near SNR
shocks. These are accessible to LOFAR through measurements of background sources, with
broadband coverage distinguishing between competing absorption effects. VLBI has explored
for such effects, e.g. the heavily scattered line of sight towards 1849+005 passing
coincidentally ~10’ near SNR G33.6+0.1 (Spangler et al 1986), but connections remain
2
unproven. The myriad, sub-mJy background sources and the dependence of scattering will
allow many more measurements by LOFAR.
5.4.3
Extragalactic SNR Studies
Nearby galaxies provide powerful statistical samples of co-distant SNRs (Jones et al 1998).
LOFAR will find new SNRs and anchor spectral and morphological identifications made at
higher frequencies. Many SNRs show extrinsic turnovers at higher frequencies ( 100 MHz)
than towards Galactic SNRs. At 408 MHz in M82 this implies ionized gas with emission
6
-6
measures ~10 cm pc, comparable to giant HII regions (Wills et al. 1997). This places the
discrete sources relative to the host galaxy ISM, and measures particle acceleration
efficiency, relates cosmic ray origin to specific SNe types, and measures SN rates, star
formation, and ISM properties (Duric et al 1995, Jones et al 1998). Lower frequency
observations are crucially required but without LOFAR have insufficient resolution.
In summary, LOFAR will offer unique insights into SNRs and their interaction with the ISM,
with its surface brightness sensitivity and field of view well matched to SNR spectra and
morphology. Intrinsic absorption seen by the 74 MHz VLA imply LOFAR will trace thermal
material within many more SNRs, and it will complete SNR catalogs by uncovering many
(>100) new remnants and SNR-PSR associations. Its broad-band response and high
resolution will empower shock-acceleration studies where sensitive, spatially resolved
spectral index maps trace synchrotron emitting relativistic electron energy spectra.
Comparisons with higher frequency images will measure spectral variations at unmatched
accuracy to test theoretical predictions, and will re-define integrated spectra. LOFAR will be a
powerful diagnostic of ISM gas, with SNRs and background sources providing a grid against
which to map the ISM through absorption and scattering processes. LOFAR SNR applications
extend to nearby galaxies, mapping the ionized ISM with respect to the discrete sources.
Powerful, co-distant sampled LOFAR SNR images and spectra will provide powerful
constraints on the star formation history, ISM energetics, and cosmic ray gas properties of
nearby galaxies.
References
Anantharamaiah, K.R., 1986, JA&A., 7, 131.
Anderson, M.C. & Rudnick, L. 1993, ApJ, 408, 514.
Page 45 of 59
Bietenholz, M.F., Kassim, N.E., Frail, D.A., Perley, R.A., Erickson, W.C., & Hajian, A.R., 1997,
ApJ, 490, 291.
Duric, N., Gordon, S.M., Goss, W.M., Viallefond, F., & Lacey, C., 1995, ApJ, 445, 173.
Dwarakanath, K.S., 1991, JA&A, 12, 199.
Frail, D.A., Kassim, N.E., Cornwell, T.J., & Goss, W.M., 1995, ApJ, 454, L129.
Frail, D.A., Kassim, N.E., & Weiler, K.W., 1994, ApJ, 107, 1120.
Frail, D.A., Goss, W.M., & Whiteoak, J.B.Z., ApJ, 437, 781.
Gaensler, B.M., 1998, ApJ, 493, 781.
Green, D.A., 1990, AJ, 100, 1928.
Green, D.A., 1991, PASP, 103, 209.
Jones, T.W., Rudnick, L., et al. 1998, PASP, 110, 125.
Kassim, N.E., 1989, ApJ, 347, 915.
Kassim, N.E., Perley, R.A., Dwarakanath, K.S., & Erickson, W.C., 1995, ApJ, 455, L59.
Kassim, N.E., & Frail, D.A., 1996, MNRAS, 283, L51.
Kassim, N.E. & Weiler, K.W., 1990, ApJ, 360, 184.
Katz-Stone, D.M., Kassim, N.E., Lazio, T.J.W., & O'Donnell, R., 2000, ApJ, 529, 453.
Lacey, C.K., Lazio, T.J.W., Kassim, N.E., & Dyer, K., 2001, ApJ, (in press).
Lang, C., Anantharamaiah, K.R., Kassim, N.E., & Lazio, T.J.W., 1999, ApJ, 521, L41.
Larosa, T.,N., Kassim, N.E., Lazio, T.,J.,W., & Hyman, S.D., AJ, 119, 207.
Pedlar, A., Anantharamaiah, et al. 1989, ApJ, 342, 769.
Pineault, S., Landecker, T.L., Swedlyk, C., & Reich, W., 1998, JRASC, 92, 31.
Reynolds, S.P. & Ellison, D.C., 1992, 399, L75.
Sofue, Y. 2000, ApJ, 540, 224.
Spangler, S.R., Mutel, R.L., Benson, J.M., & Cordes, J.M., 1986, ApJ, 301, 312.
Wills, K.A., Pedlar, A. Muxlow, T.W.B., & Wilkinson, P.N., MNRAS, 291, 517.
Zhang, X., Zheng, Y., Landecker, T. & Higgs, L.A., 1997, A&A, 324, 641.
5.5 H II Regions
Observations at 330 MHz have demonstrated that measurements of optically thick H II
regions can constrain source electron temperatures, emission measures, and filling factors. At
lower frequencies these regions appear as cooler regions against a much hotter Galactic
background, allowing kinematic distance ambiguities to be resolved and the superposition of
thermal and nonthermal sources to be separated along complex lines of sight through the
Galaxy. A classic example is the 330 MHz VLA observation that revealed the thermal Galactic
center source Sgr A West in absorption against the nonthermal Sgr A East supernova
remnant, constraining the superposition of these sources along the most confused Galactic
line of sight. Along other lines of sight, kinematic-distance ambiguities resulting from radiorecombination-line measurements can be resolved using the detection, or non-detection, of H
II regions in absorption below 100 MHz, because foreground (“near”) H II regions would be
much more prominent absorption features on low-frequency LOFAR images than distant
(“far”) ones.
5.6 Interstellar Propagation Effects
All Galactic and extragalactic radio sources are observed after their radiation has propagated
through the Galactic plasma (Rickett 1990). Variations in the plasma density produce
-2
refractive index fluctuations, scaling as , which in turn scatter the radiation. The magnitude
of radio-wave scattering from the interstellar plasma is strongly direction dependent, but the
effects can remain significant at frequencies as high as 10 GHz or higher. The density
(refractive index) microstructure responsible for interstellar scattering occurs on scales of
order 1 AU. The density fluctuations, in turn, are thought to arise from velocity and/or
magnetic field fluctuations. In addition to their corrupting effects, interstellar propagation
effects are a powerful sub-parsec probe of the interstellar plasma, can provide a tracer of
energy input into the ISM, can serve as a filter to find extremely compact sources, and may
be linked to cosmic ray propagation. The strong frequency dependence of interstellar
scattering means that studies of it are optimized with high-resolution, low-frequency
observations.
LOFAR would prove useful for interstellar scattering studies in a number of respects. The
most recent compilation of scattering observations contained 223 sources, a number that will
Page 46 of 59
be greatly increased by LOFAR. Not only will LOFAR increase the number of known pulsars
by a large amount, but it will also enable scattering observations in less intense scattering
regions. The vast majority of scattering studies have been conducted using compact
extragalactic sources viewed through regions of intense scattering (e.g., Cygnus), where the
scattering effects can be detected at frequencies near 1 GHz on baselines of length 50–
5000 km, typically with VLBI. The use of VLBI has also restricted most scattering studies to
2
relatively bright sources. Because of the dependence of scattering and LOFAR's sensitivity
a much larger volume of the Galaxy will be opened for scattering studies.
Of particular interest is the distribution of the scattering material near the Sun and its spatial
spectrum. There are conflicting claims for the detection of a signature from the boundary of
the Local Bubble. The interior of the Local Bubble is a hot, nearly fully ionized gas whereas
the gas outside is cooler and perhaps only partially ionized. At the interface, one might expect
some amount of turbulence and an increased level of scattering. However, the generally
weak level of scattering near the Sun means that a signature of this interface is difficult to
detect at frequencies near 1 GHz. The magnitude of scattering observables, e.g., angular
broadening, depends not only upon the total rms electron density fluctuations toward a
source, but also upon their distribution along the line of sight. Thus, multi-frequency scattering
observations of nearby pulsars may resolve the local structure of the interstellar medium.
Current shock-acceleration theories, relevant to the origin of cosmic rays, also suggest that
upstream of a SNR should be an ideal site for the generation of the density fluctuations
responsible for interstellar scattering. High-frequency searches for the signatures of such
2
upstream turbulence have a mixed record. The dependence of interstellar scattering would
allow much more stringent tests to be applied.
Compact incoherent extragalactic sources should be limited by the inverse Compton
12
catastrophe to brightness temperatures of no more than about 10 K. Nonetheless, various
sources have been identified that have inferred brightness temperatures well in excess of the
2
3
Compton limit (by factors of 10 –10 ). The primary means for identifying these sources has
been via their variability: Low frequency variables (LFVs) are sources that vary on time scales
of order 1 year at frequencies below 1 GHz while intraday variables (IDVs) are sources that
vary on time scales of order hours at frequencies near 5 GHz. Various lines of evidence,
including a dependence on Galactic latitude, suggest that the variability is extrinsic in origin
and results, either partially or entirely, from refractive interstellar scintillation (RISS). The
variations are caused by AU-scale density fluctuations drifting past the line of sight and
focusing and defocusing the background source. The effective velocity of the density
fluctuations is determined by a combination of the Earth's velocity and motions within the
interstellar medium. Just as in the case of optical scintillation in the Earth's atmosphere (“stars
twinkle, planets don't”), in order to display RISS a source must be less than a characteristic
size. The implied diameters of LFV and IDV sources are of order 10 µas to 10 mas, well
below can be probed even with current space-based VLBI at the relevant frequencies. Thus
RISS, via LFV and IDV, can serve as a probe of the interstellar medium on AU-scales and a
filter on source diameter selecting milliarcsecond- and sub-milliarcsecond-diameter sources.
Monitoring programs typically indicate that only a small fraction, ~1%, of sources are compact
enough to display LFV. IDV is a more recently recognized phenomenon, but initial monitoring
programs suggest that the proportion of IDV sources might be much higher. Although no
detailed comparison of IDV and LFV has been done, other than the possible proportion of
variable sources, sources displaying either of the two phenomena have a number of
similarities. Most notable is that flat-spectrum sources are most likely to display either IDV or
LFV. With notable exceptions, LFV monitoring programs have consisted of measuring the flux
density of sources at only a few epochs. Sampling a large number of sources relatively
frequently might reveal that a much larger fraction of sources really are LFVs. Such a
monitoring program would also elucidate motions of the Earth relative to the local interstellar
medium potentially as well as motions within the interstellar medium. An exemplar of this kind
of monitoring program is that of Bondi et al. (1994) who were able to determine the Earth's
motion relative to the local interstellar medium by finding annual variations in the level of
variability. Their analysis relied on only 43 sources; with a sufficiently large number of sources
one could consider differences in the level of annual variations in different directions, which
would presumably indicate motions within the medium.
Page 47 of 59
With its relatively large beams LOFAR would be a powerful monitoring instrument. It is not
implausible that any given region of the sky (and the sources within it) would be observed on
average once a week. Over the lifetime of the instrument, 5–15 year light curves with nearly
weekly sampling could be produced on a considerable number of sources. Such light curves
would be powerful diagnostics of RISS-induced (and intrinsic) variability. The results of a
monitoring program could also be used to define a homogeneous sample from which to probe
other aspects of LFV. For example, is there a redshift dependence on the likelihood for a
source to display LFV? The presence (or absence) of such a dependence would probe the
evolution of radio sources. The long time scales of LFV, as opposed to those of IDV, make its
identification easier in large samples of sources. In an individual source, IDV can be identified
relatively quickly, but with current single-beam, high-frequency instruments monitoring a large
number of sources is demanding on telescope resources. Conversely multi-beam, highfrequency instruments, such as the proposed Square Kilometer Array, are scheduled to
become operational around 2010, after LOFAR is projected to have been operational for
approximately 5 years. Light curves from a combined SKA and LOFAR monitoring program
would be complimentary means of exploring and contrasting IDV and LFV.
LOFAR may also probe novel regimes of scattering. Typical analyses of scattering assume
that the scattering medium extends an infinite distance transverse to the line of sight. At low
frequencies, as the scattering diameter increases, the scattering diameter may begin to
approach the size of the scattering region. If so, the assumption of an infinite scattering region
is no longer appropriate. Such spatially-limited scattering, if detected, would result in the
apparent structure of the source being determined by the scattering region. To date,
scattering observations have focused on the Galactic plasma. At low frequencies,
intergalactic scattering may become detectable. The dominant source of intergalactic
scattering toward distant sources is probably the interstellar media of intervening galaxies,
however, at high redshifts or low frequencies effects from dense Ly clouds may also become
detectable.
References
Bondi, M., Padrielli, L., Gregorini, L., et al. 1994, A&A, 287, 390
Rickett, B. J. 1990, ARAA, 28, 561
5.7 Polarimetry
The magnetic field is a major source of pressure in the ISM and controls the flow of charged
particles in and out of the Galactic disk. Radio continuum polarization data carry important,
and often unique, information about the strength and topology of the large scale Galactic
magnetic field, plasma turbulence and the ionized components through the measurement of
Faraday rotation. As such, these observations complement the line-of-sight measurements of
interstellar scattering. Moreover, additional observational constraints would help motivate
further MHD modelling, particularly now that computational power is approaching what is
needed to do the complex simulations. Polarization studies could provide insight in the
following areas: (1) Disk/halo emission (Parker instabilities, bubbles); (2) Cosmic ray origin
and propagation; (3) Particle acceleration via reconnection processes; (4) Origin of the
Galactic magnetic field (wound up by differential rotation from a primordial seed field?); (5)
The existence of large-scale currents (e.g., those thought to be required by the Galactic
3
7
center filaments); and (6) Distribution and temperature of ionized gas (10 –10 K).
A recent additional motivation for the study of the diffuse polarized emission of disk galaxies
is the recognition of its importance for high-redshift galaxies. The microJansky population of
radio sources (at a few GHz) appears to be dominated by starforming galaxies at high
redshift. The radio luminosity of these galaxies is generated probably by synchrotron emission
from their disks. The sources of particles responsible for this emission seem clear (massive
star formation, OB stars, SNe and pulsars). However, it is less clear how the magnetic field
was formed and how strong it is. Understanding the field in our Galaxy (i.e., a prototypical
z = 0 galaxy) is essential to infer correctly the properties of high-z, young disk galaxies.
Page 48 of 59
5.8 Recombination Lines
In the 20-100 MHz range both hydrogen and carbon recombination lines occur from high
quantum number transitions (in contrast to the n < 200 transitions observed at centimeter
wavelengths). At such high quantum numbers, atoms have radii of order 0.1 mm and are
extremely sensitive to their surroundings, making them excellent probes of the ambient
physical conditions. Furthermore, at these low frequencies, most of the observed
recombination lines are due probably to carbon rather than hydrogen. Sufficiently broad
backends will allow simultaneous observations of various line sequences. At the lowest
frequencies, the level populations are determined by collisions with free electrons rather than
by radiative transitions; the lines are thermalized and can be seen in absorption against
background sources. The quantum number at which the lines change from being in emission
to absorption depends on the balance of collisional excitation (temperature and density) and
radiative excitation (local radiation field).
A number of Galactic regions that produce these lines have been found, including a large
region that stretches 40° along the Galactic plane in the inner Galaxy. Recombination line
observations in the direction of individual sources, notably Cas A, have yielded important
3
information on the condition of the diffuse interstellar medium at low densities, ne < 1 cm ,
and temperatures, Te < 100 K. However, all observations of these lines have been made with
filled apertures with extremely low angular resolutions. The complexity of the analysis
requires resolutions better than 1” for further progress. This is needed in particular to
discriminate between two sets of models, one corresponding to a cold ISM phase associated
with molecular hydrogen, and one corresponding to a slightly warmer ISM phase associated
with neutral atomic hydrogen. The very central portion of LOFAR will have sufficient surface
brightness sensitivity for detection of these lines and would provide much needed angular
resolution to map their distribution. The sensitivity of the LOFAR will also be sufficient to
extend such studies to nearby galaxies.
5.9 Neutron Stars and Pulsars
LOFAR provides an excellent opportunity to re-open the study of pulsars at low frequencies
with greatly enhanced sensitivity and modern analysis techniques. Despite the fact that
pulsars were originally discovered at low frequencies (Hewish et al 1968) the vast majority of
subsequent studies have been carried out at higher frequencies and presently there are few
observatories which are able to make low frequency observations of pulsars. The move away
from low frequencies has been due mainly to the problem of dispersion of the emission from a
pulsar during its traverse through the interstellar medium. As this effect, which causes pulses
to be smeared out, scales with frequency to the power three, low-frequency observations are
particularly affected. However recent improvements in storage capacity and computational
power mean that techniques which completely compensate for this dispersion can now be
implemented (Hankins 1974). While pulse broadening due to interstellar scattering will still be
an issue at the lowest frequencies, a large fraction of pulsars will be able to be studied with
unprecendeted time resolution at low frequencies.
The flux density of pulsars typically increases steeply as we go to lower frequencies before
they turn over somewhere in the range 100 to 250 MHz (e.g. Malofeev et al. 1994) while there
are even some pulsars for which no such break in the spectrum has yet been seen down to
frequencies as lows 50 MHz. Thus all these sources are either brightest in the LOFAR band
or below. Combined with the excellent sensitivity of LOFAR this means that LOFAR will be
able to make some of the most sensitive measurements ever of pulsars. It also indicates that
LOFAR will make an excellent instrument for searching for low-luminosity or extremely steep
spectrum sources.
5.9.1
Searching for new pulsars
LOFAR's sensitivity and unique frequency range combined with its sky coverage make it an
ideal instrument for undertaking pulsar surveys. The only caveat being the limitations imposed
by scattering in the interstellar medium restricting our detection capabilities for the more
distant sources.
Page 49 of 59
The apparent detection of the Geminga pulsar only at the specific frequency of around 100
MHz (Malofeev & Malov 1997; Kuzmin & Losovskii 1997), and the existence of pulsars like
B0943+10, which have a flux-density spectrum with a spectral index steeper than -4.0
(Deshpande & Radhakrishnan 1994), suggests that there may be quite a number of pulsars
which are presently only detectable at lower frequencies. This strange behaviour could be
either intrinsic to the emission mechanism of pulsars, or due to other geometrical effects. The
most likely reason may be the latter, where the pulsar emission cone width increases as we
go to lower frequencies as seems to be indicated by the the radius-to-frequency mapping
seen at higher frequencies and the dipole geometry of the pulsars magnetic field. This effect
becomes very significant at low frequencies where the ``beaming fraction'' of pulsars
increases considerably. Thus, pulsars which may not be beamed towards us at higher
frequencies may be detectable only at the lower frequencies. Discovering these pulsars will
not only be vital for understanding the geometry of pulsar emission regions but will improve
our knowledge on the total number of radio pulsars in the Galaxy.
Even with our present understanding of luminosity model of pulsars we expect that LOFAR
will be able to detect a large number of pulsars. With the projected sensitivity of (~55 K/Jy at
150 MHz) Monte Carlo simulations of pulsars ``detectable'' by LOFAR indicate that $\sim$
1500 could be discovered (above the declination of -25 degrees). The sensitivity of LOFAR
will thus enable us to greatly extend our knowledge of the pulsar luminosity function to lower
luminosities. Improved understanding of the luminosity function is vital for determining how
many pulsars there are in the Galaxy, this has important implications for our understanding of
both the formation mechanism of pulsars but also for other experiments like gravitational
wave detectors.
LOFAR will also open up the possibility to serach for low-frequency emission from more
15
exotic neutron stars. A sub-class of neutron stars with magnetic field strengths ~10 G which
are currently only known to emit at X-ray and -ray wavelengths has recently been
discovered, these are the so-called magnetars (e.g. Kouveliotou et al. 1998) and anomalous
X-ray puslars (e.g. Israel et al. 1999). A recent claim of a detection of one of these sources as
a radio pulsar at 100 MHz (Shitov 1999) challenges our ideas about whether the radio
emission mechanism of pulsars works at such high magnetic field stengths (e.g. Baring &
Harding 2001). Furthermore there are a large number of supernova remnants (SNRs) which
have no compact counterparts, as we believe that there is a strong link between the formation
of neutron stars and supernova remnants, the nature of the sources produced by these
explosions is thus important. For example do they all become pulsars or is it simply that the
beam does not cross our line of sight. These remnants could potentially contain lowluminosity or steep-spectrum radio pulsars and recent very long integrations have revealed
two low-luminosity pulsars in SNRs (Camilo et al. 2002a,b). With a sensitive instrument such
as a LOFAR many more such sources may be discovered thereby improving our
understanding between the relationship of the supernova remnants and the compact objects
they form, and the percentage of so formed neutron stars that become pulsars. A similar
issue is the nature of the unidentified EGRET gamma-ray sources many of which are
presumed to be neutron stars (e.g. Halpern et al. 2002). A survey of SNRs as well as
unidentified EGRET sources and the vast number of -ray sources expected to be discovered
with GLAST (e.g. Baring & Harding 2000) with LOFAR would certainly provide an excellent
window on the pulsar population.
It may also be possible with LOFAR to detect pulsar's in M31. Assuming a pulsar luminosity
function similar to that of our own Galaxy (Lyne et al. 1998) and a typical spectral index of -2
we find that the brightest pulsars will have a flux ~0.15 mJy at 120 MHz and should be
detectable by a LOFAR with sensitivity as given above, in an observation of 4 hours.
5.9.2
Emission physics
The instantaneous sensitivity of LOFAR makes it a very attractive instrument for studying
individual pulses from pulsars. Studies of single pulses on timescales ranging from a fraction
of a pulse period (i.e. microseconds) to many pulse periods (up to hours) are vital to our
understanding of the pulsar emission mechanism.
Page 50 of 59
Pulsars show a number of variations in their pulse to pulse characteristics which are believed
to be associated with the emission mechanism: nulling is seen in some pulsars and
corresponds to periods when the pulsar stops emitting completely; mode changing occurs in
some pulsars and corresponds to a completely distinct emission mode leading to a different
average pulse profile from that normally observed; drifiting subpulses, the average pulse
profiles of pulsars are often made up of many subpulses, in a small number of pulsars these
pulses exhibit a regular drift pattern through the average pulse window (Deshpande & Rankin
1999). In general LOFAR will enable us to expand the number of pulsars for which we can
study individual pulses by an order of magnitude. As well as greatly increasing the number of
pulsars and thereby increasing out statistical sample, the low-freqeuncy nature of the
observations are important also.
The most popular model of these subpulses is that they lie on a ring centered on the
magnetic axis of the pulsar and rotate around like a “carousel” (Ruderman & Sutherland
1974). Deshpande & Rankin (1999), in observations made at at 400 MHz, show that using
extremely high signal-to-noise observations they were able, for the first time, to make a map
of the pulsar emission region or “carousel” of PSR B0943+10. Follow-up work has been
carried out by Asgekar & Deshpande (2001) at the very low-frequency of 35 MHz. At 35 MHz
our sightline probes a different sight line and thus provides significant new information about
the emission region. Future studies of this phenomenon in this pulsar and others with LOFAR
would not only have greater sensitivity but could also be made with simultaneous frequencies,
giving us information on the emission map over a larger radius. Present studies have shown
that the drift rate is less and that the drifting behaviour is less stable at low frequencies.
However because long observations are required to study these phenomenon and the
majority of low-frequency arrays presently working are meridian transit instruments, LOFAR's
ability to track pulsars will allow it to make a significant contribution to clarify these issues.
Another area where the raw sensitivity of LOFAR will provide a leap forward is in the study of
millisecond pulsars. Millisecond pulsars are old neutron stars which have been spun up to
rapid rotation rates (< 25 ms) by accretion from a binary companion star. This accretion
process has reduced their magnetic field strengths by 4 orders of magnitude below those of
the younger pulsars we have been discussing previously. The more rapid rotation rates and
the in general lower luminosity of millisecond pulsars means that virtually no studies of single
pulses has been made of this evolutionarly distinct pulsar popultaion. LOFAR will enable us to
study single pulses from approximately 20 millisecond pulsars. These observations will be
made at the higher LOFAR frequencies to eliminate the problems of scattering. This will
represent the first chance to determine if millisecond pulsars exhibit the mode changing,
drifting subpulse and nulling characteristics of their slower spinning brethren.
The study integrated pulse profiles, an average of many individual pulses, will also be an
important program for LOFAR. Comparison of the arrival times of multiple low-frequency
pulse profiles from a particular pulsar will allow us to determine the dispersion measure, a
measure of the integrated number of electrons along the line of sight, to 1 part in ten
thousand. This will allow us to not only study the dispersive effects of the interstellar medium
but also any affects due to propagation through the pulsar's magnetic field itself.
The integrated pulse profiles also exhibit a phenomenon called radius-to-frequency mapping
that pulses broaden as we go to lower and lower frequencies. This is believed to be an
indication that the pulses are being emitted further out in the pulsars magnetosphere where
the dipole field lines are most divergent. In some pulsars this phenomenon is only observed
when we get to the lowest frequencies and thus observations with LOFAR, especially at
frequencies below 100 MHz, can provide excellent evidence of the phenomenon. A
comparison with this broadening effect and the frequency where the pulsar's spectrum turns
over will also be a constraint on the emission process.
Page 51 of 59
Figure 5.1 Average polar cap map of emission of PSR B0943+10 using a 125 pulse sequence and the
cartographic transformation procedure described in Deshpande & Rankin 2001. These observations
were made using the Gauribidanur Telescope at 35 MHz. Twenty sub-beams can be discerned in this
map which highlights the power of low-frequency observations in allowing us to greatly improve our
understanding of pulsar emission physics and the geometry of pulsar magnetospheres.
5.10 Jupiter and Extrasolar Jupiters
5.10.1 Jupiter Magnetosphere
Jupiter has powerful radiation belts which produce bright synchrotron radiation. Although
measurements at a variety of radio frequencies have been used to determine the energy
spectrum of the radiating electrons, a number of questions about their distribution and source
remain uncertain. The emission has been extensively mapped at cm wavelengths and
reconstructed in 3 dimensions using tomographic techniques (de Pater & Sault 1998; Sault et
al. 1997). This synchrotron emission was also recently detected at 74 MHz using the VLA
(dePater 1999). Above about 1 GHz, the spectrum shows a standard power law with a
spectral index of ~ -0.5 but it flattens and then turns over at lower frequencies. This spectrum
has been empirically fit with a four parameter model (van Allen 1976) but refinement of the
fitting parameters and their relation to physical conditions remain to be explored. The most
generally accepted causes of the decrease in low energy electrons are pitch angle scattering,
radial diffusion processes, and accleration at instabilities in the field and plasma. These
processes can all vary with position in the magnetosphere and the presence of satellites.
It is thus very important to image Jupiter's magnetosphere at several frequencies below 300
MHz, where the spectrum is changing, at the same time and with the same spatial frequency
coverage. This is the only way that the different effects can be separated and their
importance evaluated. The requirement for similar timing is necessary because, as well as
possible long term variations with solar distance, Jovian seasons (small but present), and
solar stimulation (Klein et al. 1998), the inclined and offset magnetic dipole significantly
changes the observed emission with Jupiter's 10-hour rotation period. In addition to helping
understand the energetics of Jupiter's magnetosphere, these data may have inplications for
Page 52 of 59
establishing how the effects of space weather may get transferred through the earth's
magnetosphere.
5.10.2 Decameter Bursts
The source of bursts of radio emission from Jupiter at decameter wavelengths still remains a
mystery more than 40 years after their discovery. Many observational parameters have been
determined but we still don't know the actual location of the source nor the detailed emission
mechanism. Until we have the high angular resolution afforded by LOFAR we will not be able
to determine the key positional information.
These bursts of decametric radio emission occur from frequencies below the ionospheric
cutoff up to 40 MHz. Their repetition is tied to the spin rate of the planet with modulation
caused by the viewing aspect from earth plus the location of the satellite Io, and to a lesser
extent some of the other inner satellites, with respect to the line of sight to earth. Many of
these data indicate that the radiation may be highly beamed. Any possible solar dependence
is unconfirmed. Early VLBI experiments showed that the source size is small, probably less
than about 500 km, but they had no phase reference and so the source locations remain
unknown (Dulk 1970; Lynch et al. 1972).
Jovian storms have typical bandwidths of a few MHz, drift either up or down in frequency with
rates of about 1 MHz/minute, and can last for tens of minutes. Individual pulses, however, can
have bandwidths of only tens of kHz and last for fractions of a second. The brightness peaks
6
at a fequency of about 10 MHz; typical flux densities at 20 MHz would be about 10 Jy (see
e.g. the review by Carr & Desch 1976).
The detailed emission mechanism is not known. The radiation is highly circularly polarized
which may suggest cyclotron radiation from energetic electrons orbiting Jupiter's magnetic
field lines in the ionospheric regions of the planet. They appear to be dumped there through
perturbations by the satellites passing through Jupiter's extensive magnetosphere. The
coupling between Io's orbital motion and Jupiter's ionosphere, the instability mechanism
which can amplify the electromagnetic waves, and the propagation of the radiation through
the plasmasphere are still problems to be resolved.
Clearly, a knowledge of specifically where the source lies plus its variation with rotation and
Io's location will greatly aid efforts to characterize the interrelated mechanisms resoponsible
th
for these decameter bursts from Jupiter. The good resolution of LOFAR – about 1/10 of
Jupiter's disk for a 400 km baseline at 30 MHz - would be very valuable for monitoring
changes in the source position with time. The emission is strong so that individual bursts can
be characterized as well as whole storms. For example, does the position shift as the
frequency drifts, how far are the feet of the active flux tubes from Io's position, etc.?
References
Carr, T. & Desch, M. 1976, in Jupiter, ed. by T. Gehrels, 693
Dulk, G. 1970, ApJ, 159, 671
Klein, M., Bolton, S., Gulkis, S., & Levin, S. 1998, BAAS, 30, 1079
Lynch, M., Carr, T., May, J., Block, W., Robinson, V., & Six, N. 1972, ApLett, 10, 153
de Pater, I. 1999, in Perspectives on Radio Astronomy: Science with Large Antenna Arrays,
ed. by M. P. van Haarlem, 327
dePater, I. & Sault, R. 1998, JGR Planets, 103, No. E9, 19, 973
Sault, R., Oosterloo, T, Dulk, G., & Leblanc, Y. 1997, A & A, 324, 1190
Van Allen, J. 1976, in Jupiter, ed. by T. Gehrels, 928
5.10.3 Extrasolar Planets
The recent detections of extrasolar planets have stimulated the imagination of many
astronomers and members of the public as well. The existence of places possibly similar to
the earth brings up the question of our uniqueness and all the implications thereof. The
results to date, however, are surprising in that all the detected planets appear to be many
Jovian masses and to lie well within 1 a.u. of their parent stars. It is difficult to understand how
they could have formed in such tight orbits without tidal disruptions or have been perturbed
Page 53 of 59
into stable orbits that close to a more massive object. It may be that such planets are rare and
have been detected by observational selection but maybe the solar system is the unusual
case. The method of detecting cyclic variations in radial velocity of a star to indicate orbital
motion, requires a large mass for the planet and a small orbital radius in order to produce a
detectable shift in the star's motion. Detection of decameter bursts similar to those from
Jupiter might offer a way of detecting lower mass planets at any distance from their stars.
Five solar planets have magnetospheres although only Jupiter produces strong decameter
bursts. The intensity of these bursts should depend on the magnetic moment of the planet
which is likely proportional to the mass and also the rotation rate. Injection from the stellar
wind is at least partly responsible for the magnetospheric particle population; this, in turn, is
proportional to the inverse square of the distance from the parent star and the stellar activity.
Finally, the passage of satellites through the magnetosphere affects the dumping of the
particles into the ionosphere. It is clear that we need to study Jupiter thoroughly with LOFAR
to find the location of its burst radiation which will help us to more thoroughly assess the
emission processes and better choose other candidates for detection of similar planetary
systems.
As a quick assessment of the relative possibilities for detection of bursts from extrasolar
planets, we note that Jupiter has the strongest magnetic field but the other planets also have
different geometrical effect: the earth is closer to the sun but does not have any satellites
within its magnetosphere; Saturn does have weak hectometric bursts detected by Voyager
but its magnetic axis is aligned with its rotational axis rather than tilted and it is further from
the sun; Uranus and Neptune have highly tilted magnetic as well as rotational axes so both
their stimulating and viewing geometries may be unfavorable.
One good planet in the solar system with decameter bursts readily detectable by LOFAR is
very promising, however, particularly if we wish to begin the search for extrasolar planets like
those detected to date by the radial velocity techniques. They have larger masses than
Jupiter (and thus likely larger magnetic fields) and are closer to their parent stars (with thus a
greater cross section for the stellar wind). A reasonable detection rate is certainly possible.
It should be noted that the sun produces noise storms of comparable brightness in the same
frequency range but they are much more sporadic and do not have the precise repetition with
rotation rate. Thus, it will be necessary to monitor each candidate for a few months to search
for periodicities appropriate for a planetary rotation, say a few hours to several days.
6
With typical bursts of about 10 Jy at 20 MHz for a planet at Jupiter's distance, (they can be a
factor of 10 higher) we can expect 0.4 mJy at 1 pc or 4 microJy at 10 pc. This may make
detections marginal as we have only about 5 MHz bandwidth and a few minutes per burst.
Page 54 of 59
6 THE SUN AND SOLAR-TERRESTIAL
RELATIONSHIPS
6.1 Introduction
The magnetic activity of the Sun is responsible for the production of Coronal Mass Ejections
(CMEs), magnetized clouds of dramatic energies hurled into space, which when hitting the
Earth's magnetosphere have severe impacts on both space-borne and ground-based
technological systems. The coupling between disturbed solar wind and terrestrial
magnetosphere which causes these magnetic storms on the Earth and which are known
collectively as "space weather" is electrodynamic in origin. Accurate prediction of the storms
is of large economic importance. A direct way to image CMEs and to determine whether they
are moving on a collision path towards Earth is by detecting the solar light scattered from the
density enhancements in CMEs, and this principle is worked out to great sophistication in
orbiting and planned satellites. So far, it has appeared difficult to predict whether or not a
particular CME is going to hit the Earth. The relatively large number of false alerts makes the
present space weather forecast of little value. For example, it is difficult to decide from
satellite data alone whether a "halo" CME is moving toward the Earth or, alternatively, away
from both Sun and Earth on the other side of the Sun.
Solar wind, terrestrial magnetosphere and ionosphere are elements of one coupled
system which causes space weather. Not only is a better understanding of this coupling
required for an accurate forecast and the mitigation of economic damage the magnetic cycle
of the Sun probably also has a direct but ill-understood effect on the terrestrial climate itself.
This fundamental problem is receiving increasing attention recently, and presents one of the
most urgent challenges in the fundamental study of solar-terrestrial relations. There is now
increasing evidence for a correlation between variations in the global mean temperature of
the Earth, solar irradiance in the UV, smoothed sunspot numbers, the open magnetic flux in
the solar wind, the fast solar wind, and auroral activity over a considerable period. Also there
is an anticorrelation with modulations in CR flux caused by the solar magnetic field. Which of
these factors is instrumental in causing the climate changes is as yet unknown. One of the
links is the ionizing potential of the CRs as they are modulated by the magnetic field in the
heliosphere and their effectiveness in creating aerosols onto which water droplets form.
Another possible link are the changes in electric current system brought about by interaction
of CMEs that have a southward directed component of the interplanetary magnetic field which
leads to magnetic reconnection between CME and terrestrial magnetosphere on the day-side
and the onset of so-called magnetic substorms. How and if these changing currents influence
the climate, eg. by setting up convection winds, remains unknown at present.
LOFAR will contribute in these areas in a number of ways:
1. by accurate predictions of Earthward-bound CMEs and estimating their arrival times from
direct imaging of the free-free and synchrotron emission in the CMEs emitted at radio
wavelengths directly. It should be noted that even at a low observing frequency of 10
MHz the corresponding plasma frequency in the solar wind is only at a distance of a few
solar radii from the Sun, a mere 5% of the distance from Sun to Earth, and therefore,
accurate determination of the future path of the CME still leaves at least a day for
measures to be taken. Also, radio detection of a CME implies that the disturbance is on
our side of the Sun and does not suffer from the ambiguity of satellite detection;
2. by detailed tomography of CMEs and related disturbances using imaging and scattering
techniques which brings within reach realistic MagnetoHydroDynamic modelling of the
disturbance with the terrestrial magnetosphere. Such fundamental research is necessary
to understand the evolution of geomagnetic (sub)storms, and more general, if and how
the solar wind magnetic field influences the terrestrial climate, either directly through the
electric circuit, or indirectly via CR modulation;
3. by active radar detection of CMEs - as distinguished from "passive" detection of a CME
by its own emission - as proposed in the active radar experiment LOIS (2002). The
detections of temporal and frequency characteristics of radar echos from the solar wind at
low frequencies (in view of the small densities) are expected to provide a sensitive
diagnostic of the solar wind and CMEs similar to the current use of radar for meteorites
and asteroids which allows their detection and accurate determination of their orbits and
Page 55 of 59
shapes. Also, active radar will permit unravelling of the plasma turbulence in the solar
wind by a variety of non-linear wave coupling processes, and reveal the microscopic
nature of effective resistivity and viscosity.
6.2 Space Weather
Space weather is the name given to the collective of magnetic storms and disturbances at the
Earth. It is governed by the strongly varying solar output that hits the Earth and its environs.
Actually, the magnetized solar wind, the terrestrial magnetosphere and the ionosphere are
elements of one electric circuit, and variations in one element is of direct consequence on the
current in the circuit. As a result global magnetic disturbances reach the Earth's surface,
electric currents are induced in conductors on the ground, and energetic particles are
produced (Daglis et al. 2001; Daglis 2001). Space weather hazards include malfunction and
damage of telecommunication including long-distance telephone cables, navigation,
surveillance satellites (spacecraft charging, single-event upsets), high-voltage power grids
and danger to astronauts and passengers of airplanes at high latitude and altitude. They also
can affect over-the-horizon radars, HF, VHF, and UHF communications, the use of GPS
satellites, increased satellite drag by atmospheric heating and corresponding inaccuracies in
tracking orbiting objects, and deterioration of magnetic-torque attitude control systems of
satellites. The main causes of space weather are strong variations in UV radiation that
change the altitude of the ionosphere, and disturbances in the solar wind such as Coronal
Mass Ejections (CMEs). These solar products are largely related to the magnetic activity of
the Sun which has a cycle of 2*11 years. Another fundamental aspect of the solar-terrestrial
relationship is a possible connection with the Earth's weather and tropospheric climate. Not
only does the magnetic activity on the Sun have a strong impact on terrestrial conditions, also
it presents the nearest laboratory for magnetic explosions (reconnection, particle acceleration,
Joule heating and electric motors) similar to processes in magnetospheres of such distant
objects as compact stars and accretion disks in X-ray binaries and AGN (Kuijpers; Pavlidou et
al. 2001; Kuijpers 2001).
6.2.1
LOFAR and Coronal Mass Ejections
Coronal Mass Ejections (CMEs) are the most energetic explosions on the Sun. On a time
6
scale of minutes a huge magnetized plasmoid of dimensions of 10 km and total mass of
12
order a few times 10 kg is propelled outwards at a speed of 500 km/s or more, representing
24
a total energy of 10 J. The nature of the explosion is believed to be MagnetoHydroDynamic
in origin, and is the result either of an MHD instability or a loss of equilibrium which arises
when the free energy in the CME becomes too large with respect to the line-tying effects of
the magnetic field at the phostospheric boundary. Apparently, magnetic reconnection causes
the magnetic structure to relax to a lower energy state thereby converting part of the free
magnetic energy into motion of the plasmoid, and another part into particle acceleration which
can be observed as gyrosynchrotron emission from the moving CME.
As the CME is a strong density enhancement in the first place it distinguishes itself by
an excess of thermal bremsstrahlung modified by gyroresonance effects, which under typical
conditions is optically thin above an observing frequency of 50 MHz. As a result, simultaneous
observations at various frequencies (imaging spectroscopy) allows tomographic imaging of its
spatial and temporal structure (Bastian & Gary 1997). This would complement the images
from the proposed Frequency-Agile Solar Radiotelescope (FASR) which will operate between
0.3 - 30 GHz (Hurford et al. 1999) and would allow a detailed solution of the actual formation
process in terms of magnetic reconnection and its consequences. Also, such observations
are important for realistic 3D MHD modelling of the interaction of a CME and the Earth's
magnetosphere (Keppens & Goedbloed 2000; Poedts 2001), and its effect on the
magnetospheric ring current and the development of magnetic storms. To sketch the
technical difficulty of a realistic MHD model: the conductivity of some of the elements can
exhibit five orders of magnitude variations. Another important diagnostic of CMEs with LOFAR
will be scattering measurements against cosmic background sources - angular broadening
and scintillation measurements - to determine its internal turbulence.
6.2.2
The Solar Magnetic Cycle and the Terrestrial Climate
The magnetic activity of the Sun as measured by sunspot numbers over the period 1880 1993 correlates well with global mean temperature variations at the Earth (Soon et al. 1996).
Page 56 of 59
The magnetic activity also correlates well with solar irradiance variations in the EUV as
measured from 1978 onwards (Solanki & Fligge 2000) which in their turn correlate well with
the amount of open magnetic flux and anticorrelates with the CR modulation at the Earth by
the solar wind magnetic field (Lockwood 2002). The flux of CRs correlates with the global
average of low cloud cover over cycle 1983 - 1994 (Marsh &Svensmark 2000). The open
magnetic flux comes mainly from coronal holes which are the sources of the fast solar wind
(Wilhelm et al. 2000). Finally, the smoothed sunspot numbers correlate well with visual
observations of auroral activity over the past 1000 years (Pulkkinen et al. 2001; Pulkkinen
2001). At present, it is not clear which of these quantities is causing the climate variations: the
UV irradiation directly (Soon et al. 1996), the time variation of the ionising potential of CRs
which modulate the number of cloud condensation nuclei and aerosols (Marsh & Svensmark
2000; see also the "CLOUD" proposal by Fastrup et al. 2000), or the magnetic activity via
disturbances such as CMEs in the solar wind. Because of its superior potential to image the
physical consditions in CMEs (see previous section) LOFAR will be in an excellent position to
contribute to the investigation of the possible relation between solar wind disturbances and
climate.
6.2.3
LOFAR and CME warning
When a fast CME happens to impact on the Earth's magnetosphere and when its prevailing
magnetic field is directed southward, i.e. opposite to the terrestrial field in the magnetosphere,
over a prolonged period of time a geomagnetic storm is created whereby accelerated
particles are dumped into the magnetosphere and large inductive voltages set up across the
field lines. The physical process underlying this development is magnetic reconnection of flux
tubes at the day-side. These effects can be hazardous to humans and satellites in space as
well as to power lines on the ground. Naturally, strong pressure has built up in the past few
years to establish a timely warning system. So far, CMEs have been detected most easily
with coronographs aboard satellites such as SOHO as their density enhancements scatter
solar light in all directions. To determine whether or not a particular CME, usually of the halo
type, was going to hit the Earth has proved not to be a simple matter. One problem has been
that it is difficult to find out if the CME is moving toward the Earth, or rather moving away both
from Sun and Earth. Finding out the "geo-effectiveness" of a CME is one of the major issues
in space weather.
At the end of 2002, NASA will launch SMEI (Webb et al. 2002) to accurately predict
the arrival of CMEs at Earth. This satellite does not use a coronograph but images scattered
sunlight and uses a tomographic "time-dependent" analysis to detect ionized clouds and their
11
motion towards Earth down to 10 kg (Jackson & Hick 2000). The cameras have a huge
15
dynamic range to reduce stray-light to less than one part in 10 that can distinguish a CME or
solar wind perturbation at large distances from the Sun. In the near future STEREO will be
launched which consists of two satellites and will produce stereoscopic images of the solar
wind, and of CMEs as they propagate. Undoubtedly, the predictive power of satellites for
CMEs hitting the Earth will improve considerably by these missions. Still, a satellite warning
system will remain a costly affair, and it is here that LOFAR is expected to contribute
significantly. Note that much of the radio emission from disturbances in the solar corona come
from near the plasma frequency. Therefore, an observing frequency as low as 10 MHz
10
-3
corresponds to an electron density of 10 m or a distance from the Sun of only a few solar
radii, still far away from Earth and allowing a warning time of at least a day. At radio
wavelengths, it will be immediately obvious if a CME is between Earth and Sun rather than
behind it. Secondly, the most massive and energetic CMEs are associated with so-called
Type II radio outbursts from the shock wave in front of the CME (Cane et al. 1987). These
Type II bursts are a specific form of coherent emission at the fundamental and second
harmonic of the plasmafrequency and can be easily recognized on a dynamic radiospectrum.
The behaviour of the images of the Type II radio emission as a function of time and frequency
enable to determ the path of the CME accurately. Further, tomographic techniques will permit
the modelling of the temporal and spatial evolution of a CME in detail from simultaneous
multi-frequency LOFAR measurements of the radio emission, generated inside the CME.
Sometimes synchrotron emission - so-called Type IV emission - from accelerated trapped
particles is observed (Bastian et al. 2001) which can be distinguished from the internal
thermal bremsstrahlung.
Page 57 of 59
6.2.4
LOFAR, Active Radar and Space Weather
Radio sounding has now become a sophisticated technique to determine electron density
structures and their dynamics, primarily in the ionosphere (Hunsucker 1991; Kohl et al. 1996)
and recently also in the magnetosphere (Green et al. 2000). The use of active radar at
frequencies above 10 MHz for a timely detection of CMEs directed towards Earth several
days in advance of possible geomagnetic storms has been advocated by Rodriguez (2000)
and a detailed proposal can be found in the proposal for LOIS (LOFAR Outrigger in
Scandinavia LOIS, Thidé 2002). The corresponding plasmafrequencies still correspond to
distances below a few solar radii away from the Sun. The radar technique allows for an early
detection and is a potentially sophisticated diagnostic of the spatio-temporal structure of a
CME or any other moving density disturbance in the solar wind. The idea is to radiate a pulse
which is narrow both in time and frequency extent and to unravel the complex dynamic radio
spectrum that is returned by the propagating disturbance. The return pulse from the Earthward propagating structure is expected to be boosted both in ampltude and in inverse
duration as compared to the echo from the surrounding corona. Recent experiments have as
yet not been conclusive. In principle, the technique is a powerful one, and sophisticated ray
tracing computations have now been started for the solar wind to demonstrate its feasibility,
using the Swedish Institute of Space physics (IRF) ray tracing system "RaTS" (Västberg
1997) .
References
Bastian, T.S. & Gary, D.E.: 1997 JGR 102, 14031
Bastian, T.S., Pick, M., Kerdraon, A., Maia, D., Vourlidas, A.: 2001 ApJ 558, L65
Cane, H.V., Sheeley, N.R., Howard, R.A.: 1987 J. Geophys. Res. 92, 9869
Daglis, I.A. (ed.): 2001 Space Storms and Space Weather Hazards, Kluwer Acad. Publ.,
Dordrecht,
Netherlands
Daglis, I.A., Baker, D.N., Galperin, Y., Kappenman, J.G., Lanzerotti, L.J.: 2001 EOS 82, 585
Fastrup, B. et al.: 2000 CLOUD Proposal, CERN/SPSC 2000-021; SPSC/P317
Green, J.L., Benson, R.F., Fung, S.F., Taylor, W.W.L., Boardsen, S.A., Reinsich, B.W.: 2000,
in Radio
Astronomy at Long Wavelengths, eds. R.G. Stone, K.W. Weiler, M.L. Goldstein, J.-L.
Bougeret, Geophysical Monograph 119, American Geophysical Union, Washington,
DC, p. 359
Hurford, G.J., Gary, D.E., Bastian, T.S., White, S.M.: 1999 BAAS 31, 956
Jackson, B.V. & Hick, P.: 2000 Adv. in Space Res. 25, 1875
Keppens, R. & Goedbloed, J.P.: 2000 ApJ 530, 1036
Kuijpers, J.: 1995 Flares in Accretion Disks, in Lect. Notes in Physics 444, 135, SpringerVerlag, Berlin
Kuijpers, J.: 2001 PASA 18, 407
Lockwood, M.: 2002 A&A 382, 678
Pavlidou, V., Kuijpers, J., Vlahos, L., Isliker, H.: 2001 A&A 372, 326
Poedts, S.: 2001, in SOLSPA 2001 Euroconference: Solar Cycle and Space Weather, ESA
SP-477, 77
Pulkkinen, T.I.: 2001 AGU #SH 52A-07
Pulkkinen, T.L., Nevanlinna, H., Pulkkinen, P.J., Lockwood, M.: 2001 Space Sci. Rev. 95, 625
Rodriguez, P.: 2000, in Radio Astronomy at Long Wavelengths, eds. R.G. Stone, K.W.
Weiler, M.L.
Goldstein, J.-L. Bougeret, Geophysical Monograph 119, American Geophysical
Union, Washington, DC, p. 155
Solanki, S.K., Fligge, M.: 2000 Space Sci. Rev. 94, 127
Soon, W.H., Posmentier, E.S., Baliunas, S.L.: 1996 ApJ 472, 891
Thidé, B.: 2002 Radio Studies of Solar-Terrestrial relationships, LOIS Science Team
Västberg, A.: 1997 Investigatioons of the Iospheric HF Channel by Ray Tracing, PhD Thesis,
Uppsala
University, Uppsala, Sweden, Apr. 1997, IRF Sci. Rep. 241
Webb, D.F., Johnston, J.C., Radick, R.R. et al.: 2002 EOS 83, 33
Wilhelm, K., Dammasch, L.E., Marsch, E., Hassler, D.M.: 2000 A&A 353, 749
Page 58 of 59
Page 59 of 59