The Early Universe - Deep Surveys
High-mass, evolved stars have their black-body peak around 2 microns, thus this range is a tracer of high-mass stars both in the Milky Way and in external galaxies. Observing in the mid-infrared the interstellar extinction is greatly reduced and such evolved stars may be detected, even in highly obscured regions. This allows massive star formation to be traced conveniently. Similarly, at high redshifts, important visible lines are translated into the near-infrared and near-infrared lines into the mid-infrared. At z=2.4, the H-alpha line is shifted to 2.2 microns, in the middle of the K window; Paschen-alpha shifts to 4.5 microns and Brackett-alpha to 9.7 microns. By z= 4.3 H-alpha has shifted as far as 3.5 microns - well into the L window - while Paschen-alpha and Brackett-alpha reach 8.0 and 17.4 microns respectively. To obtain line luminosities at a wide range of redshift, for example, to study the evolution of the Baldwin effect to high redshift, it is necessary to observe as far out into the infrared as is possible.
Recent deep surveys conducted
by ESA's Infrared Space Observatory suggest that the number of galaxies
in the distant Universe exceeds model estimates, derived from optical observations,
by a factor of 10. These surveys suggest a strong evolution in galaxy formation,
such that more - and perhaps more luminous - galaxies were formed at earlier
epochs. Still, these new data account for only a tenth of the unresolved
infrared background measured by COBE/DIRBE. At mid-infrared wavelengths,
CanariCam could explore the early Universe by imaging galaxies at redshifts
z>2 (in an analogous manner to the HDF), uncovering perhaps several thousand
galaxies per square degree, despite its limited field. From these data,
luminosity functions (LFs) can be derived. Measuring the LFs at different
redshifts will permit the evolution of the luminosity and/or density of
galaxies with cosmic time to be determined. Galaxies undergoing massive
star formation in dust-embedded environments will reveal themselves at
far-infrared wavelengths, since the dust serves to re-radiate the ultraviolet
(UV) emission from the new stars as infrared. These "starburst" galaxies,
initially revealed by IRAS, are found to be numerous in the Local Universe.
CanariCam will be able to investigate the counterparts in the distant Universe
at a range of wavelengths in the mid-infrared that can be combined with
EMIR data for the near-infrared. We can thus characterise the dependence
of star formation on redshift, and to relate the findings to the evolution
of other parameters of the galaxy population.
Ultraluminous Galaxies and Active Galactic Nuclei
The nature of IR-luminous galaxies is still open to some debate. Two models have been invoked: heating of dust by super-massive black holes in the galactic nuclei; or by large populations of massive stars. Due to the very high extinction, visible spectroscopy is an unreliable guide to the conditions in the galaxies. Spectroscopy with CanariCam will allow line ratios, particularly of the 12.8 micron [NeII], 9.0 micron [ArIII], and 10.5 micron [S IV] lines to be calculated. The ionisation potentials of Ne0 (22 eV), Ar+ (28 eV) and S++ (35eV) are ideally suited to the calculation of the hardness of the radiation field, an important diagnostic of the exciting mechanism. In Seyfert galaxies, ionisation models (e.g. Spinoglio & Malkan, 1992) predict a ratio of 1:3.7:1.3, whilst a ratio of 1:0.037:20 is predicted for Starburst galaxies. These ratios imply that, at one extreme, all three lines should be detected at roughly equal strengths but, at the other, only the 10.5 micron [S IV] line would be detected. Of course, in many cases it is likely that both radiation fields ae present, in which case the observations will serve to establish their distribution over the galaxy and relative importance. In these studies the diffraction-limited capability of CanariCam will be crucial.
The line widths are predicted to be in the range from 100-1000 km/s, well-matched to the CanariCam resolution of 210 km/s with the R~1400 first-order grating.
By observing IR-luminous
galaxies at a wide range of redshifts to distances corresponding to >90%
of the age of the Universe (z > 5), CanariCam will have an excellent capability
of studying the properties and evolution of ULGs, and the relationship
between these galaxies and AGN discovered by other techniques. The presence
of such systems would imply that the epoch of galaxy formation occurred
very early in the history of the Universe. With CanariCam we will be able
to characterise the evolution of ULGs in space-time, and test the hypotheses
linking IR galaxies and quasars. CanariCam will also have the capability
of probing the heavily extinguished nuclei of the brightest and closest
of these objects with spectroscopy to establish the excitation, ionising
source, and abundance of heavy elements in different redshift regimes.
These spectroscopic observations will help determine the physical conditions
in the optically-obscured interiors of these galaxies, thereby distinguishing
between stellar and non-thermal processes as the ultimate source of their
There are an enormous
range of exciting and fundamental projects which can be carried out in
this field in the thermal infrared in a large telescope, particularly as
observations are, at present, limited mainly to space-bound observatories
with relatively limited sensitivity and highly limited integration times.
Studies of galaxies by ISO and from the KAO have shown the presence of
cool and warm dust is very common, not just in normal galaxies, but also
in Starburst galaxies (like M82) and in quasars (such as 3C48). AGNs often
show the presence of dust with a maximum emission at around 20 microns,
whilst "Starburst" galaxies show an almost asymptotic increase in flux
beyond 10 microns due to the presence of warm dust.
A further field previously alluded to, is the study of the mid-infrared spectrum of normal and active galaxies. Previous studies have revealed the widespread presence of a feature at around 11 microns, which is presumed to be due to PAHs. This feature can, however, only be studied with good s/n and low spectral resolution in a few relatively bright objects. Part of the problem with the interpretation of the observations is the poor signal to noise ratio that can be obtained and the relatively low resolution with which these features have to be studied.
A good understanding of
the 11-micron features requires a considerable increase in instrumental
sensitivity, as can be expected from the new generation of 10-m telescopes
and state of the art instrumentation. Such an increase in sensitivity will
enable better spectra to be obtained, helping to decide the question of
whether or not the observed features are genuinely due to PAHs and, if
so, which. A second and possibly even more important point is the possibility
of greatly increasing the sample of objects in which this feature can be
detected. At present no statistical study is possible, due to the small
sample size; this situation will rapidly change with the prevision of mid-infrared
spectroscopy of good sensitivity on a 10-m telescope.
A topic of great interest
to Spanish researchers is the study of gas in extragalactic star-forming
regions. Some regions, such as High Metallicity Giant Extragalactic H II
Regions and the circumstellar regions of nearby AGNs are of very high metallicity.
These regions are poor in visible lines and are predominantly cooled through
mid-IR line emission. It is of considerable interest to make kinematic
and compositional studies of these regions, to derive abundances of metals,
to characterise the kinematic behaviour and to derive the physical properties
of the ionising regions. Although our maximum proposed resolution is rather
lower than might be desirable for detailed kinematic studies, but with
the large aperture of the telescope and the high sensitivity of CanariCam,
allied to the long-slit capability, a wide range of objects could be observed
with good signal-to-noise.
Star Formation and Feeding of AGN
The Unified Model for AGNs postulates that the energy source is the accretion of material onto a single massive object - Black Hole. Whilst the presence of a Black Hole seems to be essential in order to explain fast X-ray variability and collimated radio jets, many other properties of AGNs can be explained in terms energetic star-formation. During the 1970s Seyfert galaxies were divided into two different classes, Seyfert 1 and Seyfert 2, based on the presence or absence of broad Balmer hydrogen emission-line components. Since then, a variety of efforts have attempted to reunify these galaxies into a single underlying type, distinguished by viewing angle towards the nucleus and/or degree of obscuration of the active nucleus.
An unsolved puzzle in theories of active galactic nuclei is the connection between the accretion and energy generation in the central parsec and the processes on larger scales in the host galaxy. Over the past few years, based on mid-infrared studies of Seyfert galaxies, it has been found that the host galaxies of Seyfert 2 nuclei have substantially elevated rates of star formation compared to the hosts of Seyfert 1 nuclei and field galaxies. This behaviour is in violation of the simplest form of unified theory for Seyfert nuclei, which would require all isotropic characteristics of type 1 and type 2 to be the same. The result is based on a number of comparisons including large-beam IRAS (InfraRed Astronomical Satellite) 12 micron and 5 ground-based 10.6 micron measurements, which allow the separation of the low surface brightness extended mid infrared flux from the high surface brightness contribution centred on the AGN. This study demonstrated the difference between Seyfert 1 and 2 hosts at a high level of statistical significance, and in the direction of having greatly increased star formation in the galaxy disk. The effect has presumably not been seen in optical studies because of extinction.
It is as yet unclear how to interpret this result. Does it indicate that type 1 and type 2 Seyferts represent different evolutionary stages, the latter corresponding to a phase of an increased amount of interstellar material that both tends to increase obscuration of the nucleus and to foster increased star formation? Does enhanced star formation also enhance the transport of material into the AGN increasing its luminosity? Or are we seeing the result of some global event - perhaps a galaxy merger - that simultaneously tends to increase the nuclear accretion and triggers the star formation episode? These questions and others can be probed with high-resolution mid-infrared imaging.
The host galaxies of low luminosity quasars appear to be very similar to Seyfert hosts, at least in the visible and the near infrared. How do the far infrared host galaxy properties change with the increased nuclear luminosities of the quasars? The diffraction limit at 10-microns will give similar (or even better) resolution on quasars out to z = 0.3 as does the 5-arcsec beam used on the Seyfert galaxies. Therefore the approach used with the Seyferts can be applied directly to nearby quasars to determine if there is a similar trend in quasars.
The spectroscopic capabilities
of the instrument can be used to observe the fine structure lines of NeII
(12.8 microns), ArIII (9.0 microns), S III (18.7 microns) and S IV (10.5
microns) in order to determine excitation conditions in obscured regions
around the AGN, helping constrain the relative roles of the AGN and star
formation. In favourable cases, the line strengths can be used to constrain
the relative roles of starburst and AGN excitation in these galaxies to
complement the approach using infrared imaging.
The presence of a huge maximum of emission (typically peaking at 60 micron) is a good tracer of starburst activity. With IRAS and ISO their poor spatial resolution has only allowed one to say that this emission-component is present, without knowing where within the galaxy it is emitted. With CanariCam it would be possible to resolve this emission within the galaxy and use it as a tracer of the history of starburst activity. We should remember that the resolution of CanariCam will be at least a factor of 2 better at 8 micron than the 0.4 arcsecond seeing expected with the telescope in the visible and, at 20 microns, the resolution will be comparable with the best visible seeing. Even below 8 microns we can expect to surpass the best spatial resolution possible in the visible without the use of adaptive optics. Mid-infrared observations can tell us much about extragalactic star formation in galaxies of all kinds, particularly when one takes into account the fact the spatial resolution of the 10-m telescope will be around two orders of magnitude better than ISO's.
in various of the available mid-infrared lines, such as [NeII], [ArIII],
and [S IV] allows the hardness of the radiation field to be determined
from line ratios. These lines are difficult to detect with existing telescopes
and instruments but, with the enormous gain resulting from the new generation
of 8-10m instruments, the feasibility of detecting all three lines simultaneously
and calculating reliable line ratios out to a significant red shift will
become a realisable possibility.
Mid-infrared polarimetry of AGNs
Given that synchrotron radiation is an important component of the emission of AGNs, polarimetric observations in the thermal infrared of quasars and AGNs are of particular interest. This field has been almost impossible to exploit for ground-based telescopes (because of lack of signal) and, due to the particular calibration difficulties which ISO has presented, has been regrettably little-exploited even with ISO. Future satellites, such as SIRTF and FIRST will not have a polarimetric capability at all.
It is already known that
the polarisation of some objects, such as blazars, suffers a radical change
between the near infrared and the radio regimes. Typically the radio polarisation
angle is constant (lined-up with the relativistic jet), whilst the visible
and near infrared show large and rapid changes both in the degree and the
angle of polarisation. What is completely unknown is the behaviour of the
polarisation in the intermediate region. It is possible that, with the
option, the mid and far-infrared could show the presence of strong Faraday
rotation which, if quantified, would allow the fundamental physical parameters
of the active nucleus, such as the magnetic field and electron density
to be accurately determined.