Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://hea-www.harvard.edu/QEDT/Papers/bjw_tenerife.ps Äàòà èçìåíåíèÿ: Thu Feb 27 21:47:43 1997 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 00:58:05 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: m 5

ISO Observations of Quasars and Quasar Hosts
Belinda J. Wilkes
Smithsonian Astrophysical Observatory, 60 Garden St., Cambridge, MA 02138, USA
Abstract. The Infrared Space Observatory (ISO), launched in November 1995, allows
us to measure the far­infrared (far­IR) emission of quasars in greater detail and over
a wider energy range than previously possible. In this paper, preliminary results in
a study of the 5--200 ¯m continuum of quasars and active galaxies are presented.
Comparison of the spectral energy distributions show that, if the far­IR emission from
quasars is thermal emission from galaxian dust, the host galaxies of quasars must
contain dust in quantities comparable to IR luminous galaxies rather than normal spiral
galaxies. In the near­IR, the ISO data confirm an excess due to a warm `AGN­related'
dust component, possibly from the putative molecular torus. We report detection of the
high­redshift quasar, 1202­0727, in the near­IR indicating that it is unusually IR­bright
compared with low­redshift quasars.
1 Introduction
Quasars are multi­wavelength emitters, emitting roughly equal amounts of ra­
diation throughout the whole electromagnetic spectrum from far­IR through to
fl­ray energies. 10% are also strong radio emitters. To understand the energy
generation mechanisms at work, it is first essential to obtain multi­– data cover­
ing the full spectral range of the emission. We now have a good understanding
of the spectral energy distributions (SEDs) of low­redshift quasars and active
galaxies. However, in the far­IR this has been limited by the short lifetime and
wide beam of the ground­breaking IRAS satellite. Now, more than 10 years later,
ISO is providing us with the chance for a second, more detailed look at the far­
IR sky, allowing us to extend our knowledge to IR­fainter and higher redshift
sources and to longer and shorter wavelengths (5--200¯m).
To this end we are observing a sample of quasars and active galaxies with
the photometer on ISO (ISOPHOT). The sample was originally designed to in­
clude ¸ 130 quasars and active galaxies covering the full range of redshift and of
known SED properties. With the reduced in­flight sensitivities of ISOPHOT, our
program has been reduced significantly and will likely include ¸ 50 objects, not
all with full wavelength coverage. The sample will include full wavelength cover­
age for a well­defined subset of optically­selected, PG quasars (Laor et al. 1996),
along with a few high­redshift quasars, X­ray selected Seyfert 1 galaxies, and
red quasars.
One question that the ISO data will address is particularly relevant to this
conference, namely the contribution of the host galaxy in the far­IR. Figure 1
shows SEDs of spiral and elliptical galaxies superposed on the SED of a median
low­redshift quasar (from Elvis et al. 1994). The plot clearly shows the near­IR

2 Belinda J. Wilkes
(¸ 1 \Gamma 2¯m ``window'' on the host galaxy which has been used to great advantage
(McLeod & Rieke 1994, Dunlop et al. 1993). Although the strength of the far­IR
peak, due to cool dust, is as yet unknown, Figure 1 demonstrates that this is the
most likely wavelength range for a second ``window'' on the host galaxy.
Fig. 1. The Median SED of a low­redshift quasar superposed on the SED of a spiral
galaxy showing the well­explored, near­IR ``window'' (¸ 1¯m) and the potential far­IR
``window'' on a quasar's host galaxy (courtesy Kim McLeod).
2 ISO Observing Program
ISOPHOT observations are being made in eight broad bands covering the full
energy range of the instrument, 5--200 ¯m. The detector/filter combinations
are: P1:5,7,12 ¯m; P2: 25 ¯m; C100: 60,100 ¯m and C200: 135,200 ¯m. Until
October 1996, AOTs P03 and P22 were used in rectangular chopping mode
with a chopper throw of 180'' in all cases. The apertures, chosen to match the
instrument field of view or the point spread function as applicable, are 52'' for
– Ÿ 12¯m, 120'' for – = 25¯m and the array size for the long wavelength points.
For the largest and/or brightest sources, we use staring mode with a separate sky
observation. Following the recommendations of the ISOPHOT team, we have re­
specified our remaining time to observe a smaller subset of objects, re­observing
where necessary, using small rasters whose dimensions depend on the detector in
use. These observations are scheduled to begin in early 1997 and should provide
the reliable long­wavelength data which is currently lacking.

ISO Observations of Quasars and Quasar Hosts 3
3 Analysis of ISOPHOT data.
ISOPHOT suffers from several well­known problems which complicate the data
analysis and limit (currently) the accuracy with which fluxes can be determined
(for details see ISOPHOT Observers Manual and associated updates):
-- The responsivity of all detectors drifts significantly following a change in
the incident signal, for example when pointing to a new source or changing
filters. This drift is difficult to calibrate and the analysis software does not
yet support fitting for chopped observations such as ours. We have thus
concentrated our efforts on analysing objects with observations sufficiently
long that the detector reaches a stable portion of the drift curve, generally
?
¸ 128 secs total time.
-- The internal (FCS) calibrators needed to be re­calibrated in­orbit following a
change of state of FCS1 in February 1996. The combination of the shortness
of the FCS observations (16/32 secs) taken during an individual observation
and the lack of a definitive re­calibration led us to use the default responses
for all detectors in this analysis. There is a drift ¸ 30% in the detector
responsivity as a function of time during an orbit so our flux normalisation
errors are expected to be of this order.
-- At long wavelengths (C200 detector in particular) the two adjacent beams are
large enough that part of the detector lies within a part of the telescope beam
which is significantly vignetted. This leads to an asymmetry in the derived
fluxes across the detector and a ¸ 20% flux correction. The correction for
this affect has not yet been released and has not been applied to our data.
Our analysis was performed using the PHOT Interactive Analysis Package
(PIA), an IDL­based system provided by the PHOT team. We carried out the
following steps: non­linearity correction, read­out de­glitching, 1st order fit to
ramps, dark current subtraction, de­glitching to delete highly discrepant points,
deletion of data during strong detector drifts and of remaining highly­discrepant
points, background subtraction, and calibration using the default responsivities.
Background subtraction is done using the average of the background in the
chopper plateaux before and after each source plateau. On a non­linear drifting
curve, this is not accurate and adds noise to the signal which could be reduced by
fitting the background and source curves separately and then subtracting them.
Currently points on the drift curve are deleted, reducing the potential S/N of
the observations once more sophisticated analysis is carried out.
From the subset of our sources with relatively long exposure times, we chose
four objects covering a range of luminosity and redshift: PG1244+026, a radio­
quiet Seyfert 1 galaxy; PG1543+489, a radio­quiet quasar; 3C249.1 (PG1100+772),
a radio­loud quasar; and 1202­0727, a high­redshift, radio­quiet quasar. The ob­
servational details are provided in Table 1. For these sources we found reliable
detections in most/all the short wavelength bands (5­25 ¯m). In the long wave­
length bands, however, only two of the sources are detected, the other two have
negative ``detections'' in all four long wavelength bands.

4 Belinda J. Wilkes
Subsequent analysis of additional sources not reported here has shown a large
number of negative signals at long wavelengths, particularly 135,200 ¯m. We
are currently investigating the cause and, since the negative signals are often at
levels similar to the positive ones, are treating all our long wavelength data with
scepticism. The most likely cause is cirrus confusion and would imply significant
structure on the scale of ¸ 3 0 (our chop distance). However further investigation
is required to confirm this.
Table 1. Details of the ISOPHOT observations.
Name z ISO date AOT 1 Filter Time Filter Time Filter Time Filter Time
¯m sec 2 ¯m sec 2 ¯m sec 2 ¯m sec 2
PG1244+026 0.048 14/07/96 PHT03 4.85 16 7.3 256 12 256 25 256
PHT22 60 64 100 64 135 64 200 64
PG1543+099 0.400 30/05/96 PHT03 4.85 256 7.3 128 12 128 25 256
PHT22 60 64 100 64 135 64 200 128
3C249.1 0.313 17/06/96 PHT03 4.85 256 7.3 256 12 256 25 256
(PG1100+772) PHT22 60 128 100 128 135 256 200 256
1202\Gamma0727 4.69 19/07/96 PHT03 4.85 512 7.3 512 12 512 25 512
PHT33 60 512 100 128 135 512 200 512
1: AOT: Astronomical Observation Templates
2: on­source time
4 Spectral Energy Distributions
We have combined our ISO results with data from other wavelengths collected by
ourselves and from the literature to generate SEDs of the four objects (Figure 2).
To investigate possible contributions from the quasar host galaxy, particularly in
the far­IR, SEDs for several kinds of galaxies were generated using data from the
literature (McLeod, private communication). The IRG and ULIRG templates
correspond roughly to L IR ¸ 10 10\Gamma11 L J and 10 11:5 L J respectively. Super­
posed on each quasar SED, we have shown various galaxy SEDs as labelled and
a median SED for low­redshift quasars (Elvis et al. 1994) for direct comparison
(Figure 2).
PG1244+026 is a low luminosity active galaxy (L¸ 10 44 erg s \Gamma1 ), officially
classified as a Seyfert 1 galaxy and with low redshift (z=0.048). Figure 2 shows
an L \Lambda spiral galaxy which is consistent with the AGN SED ¸ 1¯m and with
the cool dust contribution in the far­IR. Between 5 and 100 ¯m the quasar
SED is dominated by an ``AGN'' component believed to originate in warm dust
within the putative molecular torus. This component peaks ¸ 25 \Gamma 60¯m in
PG1244+026.
PG1543+489, also an optically selected PG quasar, has a luminosity¸ 10 45:5
erg s \Gamma1 (a bona fide quasar) and a redshift of 0.400. In this case an L \Lambda galaxy

ISO Observations of Quasars and Quasar Hosts 5
Fig. 2. The far­IR -- ultra­violet SEDs of a: PG1244+026, b: PG1543+489, c: 3C249.1,
d: 1202­0727, with various galaxy and quasar SEDs (as labelled) superposed. Each
different dataset uses a different symbol, the ISO points are always indicated by open
triangles.
would make no significant constribution in the near­IR. A galaxy with the max­
imum host galaxy luminosity seen to date (McLeod & Rieke 1994) is a factor
¸ 4 too low at 1¯m. An amount of dust comparable to an IR­bright galaxy is
necessary to explain the far­IR emission (L IR ¸ 10 10\Gamma11 L J ). Once again a
mid­IR bump due to warm dust is apparent, peaking ¸ 100¯m.
3C249.1 is a lobe­dominated, radio­loud quasar at a reshift 0.389. The ISO
short wavelength detections once again show a typical mid­IR bump with a peak

6 Belinda J. Wilkes
Fig. 3. The radio--far­IR SED of 3C249.1 showing the current limits on the slope of
the far­IR turnover.
! 100¯m. The IRAS upper limits (Elvis et al. 1994) are very low (¸ 20 mJy at
100 ¯m) and inconsistent with the ISO detections. Re­analysis of the IRAS
survey data using the current software (XSCANPI) yields upper limits which
are largely consistent, as shown in Figure 2c. Unfortunately we currently have
only weak upper limits on the far­IR emission of this source from ISO indicating
that the data could be consistent with an L \Lambda , IR­bright galaxy but not one
with maximal luminosity in the near­IR. However since we have no estimates
of the host galaxy from the near­IR, this provides a very weak constraint on
the amount of dust. Figure 3 shows the far­IR SED with a marginal detection
at 1mm (Antonucci et al. 1990). Assuming that the 1mm flux represents the
long wavelength tail of the far­IR emission, our current data give an upper
limit to the slope of the far­IR turnover of ff ! 2:2 (f š / š ff ). The discontinuity
between the far­IR and radio emission in the SEDs of lobe­dominated, radio­loud
quasars is generally interpreted as evidence for differing emission mechanisms
and thus thermal IR emission (Antonucci et al. 1990). However, the flat lower
limit to the far­IR turnover in this source prevents us from ruling out non­thermal
synchrotron emission.
1202­0727 is one of the highest redshift quasars known (z=4.690). It is an
extremely interesting source, mentioned a number of times in these proceedings
(Barvainis, Yamada). It is a double source with 4'' separation at mmwavelengths,
has strong CO emission (Ohta et al. 1996, Omont et al. 1996), a sub­mm spec­
trum which suggests emission from 50--100 K dust (Isaak et al. 1994) and a
Lyff emission companion 2'' away with the same redshift (Hu et al. 1996). The
source has a very high luminosity, ¸ 10 47 erg s \Gamma1 , such that the contribution

ISO Observations of Quasars and Quasar Hosts 7
from its host galaxy in the rest­frame optical and near­IR is several orders of
magnitude below that seen by ISO (12, 25 ¯m observed frame, Figure 2d). The
mid­IR bump, with a broad peak from 4--80 ¯m in the rest frame, is two orders
of magnitude stronger than that of the low­redshift median (Fig. 2d). The rest­
frame far­IR emission determined by the sub­mm observed frame data (Isaak
et al. 1994) would require a host galaxy similar to an ultra­luminous IR galaxy
(ULIRG) in a pure dust scenario. The crude far­IR upper limits from ISO suggest
that our planned re­observation could strongly constrain the mid­IR SED.
5 Conclusions
Although the current status of the ISO far­IR data limits the usefulness of ISO to
study the host galaxy dust contribution, we can already demonstrate that, if the
far­IR emission of bona fide quasars (L? 10 44 ergs \Gamma1 ) is from the host galaxy,
these galaxies are unusually far­IR bright, comparable to IR­bright galaxies or
ULIRGs (L IR ¸ 10 10\Gamma11:5 L J ). The ISO data also allow us to investigate the
mid­IR ``AGN'' bump in quasars covering a range of redshift and luminosity. We
plan to use these data to test and constrain current models of emission from a
molecular torus (Pier & Krolik 1992, Efstathiou & Rowan­Robinson 1995).
We have detected the z=4.69 quasar, 1202­0727, in the rest­frame near­IR at
a level far above that seen in typical low­redshift quasars. Observations of more
high­redshift quasars are necessary to determine whether the near­IR emission
is unusual, as are many other aspects of this source. For pure host galaxy far­IR
emission, the host must be comparable to an ULIRG to explain the mm data
(Isaak et al. 1994).
Acknowledgements
I would like to thank all my collaborators on this project, in particular Drs. Kim
McLeod, Jonathan McDowell and Martin Elvis at CfA and our other ISO co­Is.
Thanks are also due to the ISOPHOT team in Heidelberg and the ISO centers
at VILSPA and IPAC for their prompt and invaluable help in response to my
frequent email messages. The financial support of NASA grant NAGW­3134 is
gratefully acknowledged.
References
Antonucci, R., Barvainis, R. & Alloin, D. 1990, ApJ, 353, 416
Dunlop, J. S., Taylor, G. L., Hughes, D. H. & Robson, E. I. 1993, MNRAS 264, 455
Elvis, M., Wilkes, B. J., McDowell, J. C., Green, R. F., Bechtold, J., Willner, S. P.,
Cutri, R., Oey, M, S., and Polomski, E.
Hu, E. M., McMahon, R. G. & Egami, E. 1996, ApJ, 459, L53
Isaak, K. G., McMahon, R. G., Hills, R. E. & Withington, S. 1994 MNRAS, 269, L28
``ISOPHOT Observers Manual'' and associated updates found on WWW:
http://isowww.estec.esa.nl:80/ISO/iso manuals.html

8 Belinda J. Wilkes
Laor, A., Fiore, F., Elvis, M., Wilkes, B.J., & McDowell, J.C. (1996) ApJ. in press
McLeod, K.K. & Rieke, G. 1995, ApJ, 441, 96
McLeod, K.K. private communcation
Ohta, K, Yamada, T., Nakanishi, K., Kohna, K., Akiyama, M. & Kawabe, R, 1996,
Nature, 382, 4260
Omont, A., Petitjean, P., Guilloteau, S., McMahon, R. G., Solomon, P. M. & Pecon­
tal, E. 1996, Nature, 382, 4280
Pier, E. A. & Krolik, J. H. 1992, ApJ 401, 99
Efstathiou, A. & Rowan­Robinson, M. & 1995, MNRAS 273, 649
XSCANPI, Interactive software for the analysis of IRAS data provided by IPAC via
``telnet xscanpi.ipac.caltech.edu''