Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.arcetri.astro.it/science/SNe/valencia.ps Дата изменения: Fri Nov 21 12:17:10 2003 Дата индексирования: Tue Oct 2 08:33:25 2012 Кодировка: Поисковые слова: active galaxy

The Infrared Supernova Rate
F. Mannucci 1 G. Cresci 2 R. Maiolino 3 and M. Della Valle 4
1 IRA-CNR, Largo E. Fermi 5, 50125 Firenze, Italy filippo@arcetri.astro.it
2 Dip. di Astronomia, Universita di Firenze, Largo E. Fermi 5, 50125 Firenze Italy
gcresci@arcetri.astro.it
3 INAF-Osservatorio Astrosico di Arcetri, Largo E. Fermi 5, 50125 Firenze Italy
maiolino@arcetri.astro.it
4 INAF-Osservatorio Astrosico di Arcetri, Largo E. Fermi 5, 50125 Firenze Italy
massimo@arcetri.astro.it
Supernovae (SNe) exploding inside dusty regions could dominate, even by a
large amount, the number of core-collapse events in the universe, as most of
the star-forming activity is hidden by dust. Nevertheless, centuries of optical
searches have discovered only very few SNe in dusty regions and no very
obscured event.
This is clearly a selection eect, and infrared or radio observations are
needed to reveal highly obscured SNe. We have therefore started an infrared
search for extincted SNe using several telescopes. The aim of this search is to
obtain a signicant sample of SNe in dusty galaxies to study their properties
and to obtain a complete estimate of the total SN rate in the local universe,
important to calibrate the SN rate at high redshift now under study. In princi-
ple the number of events could also be used to derive information on the main
energy source (starburst vs. AGN) of the galaxies when they are dominated
by a hidden central source, as for the Luminous Infrared Galaxies (LIRGS).
1 The problem
The observed rates of the core-collapse SNe, when derived from optical ob-
servations and normalized to the B luminosity of the galaxy, don't show any
signicant dependence on the galaxy type. Normal galaxies between Sa and
Sm [1], starburst galaxies [2], galaxies with an active nuclei [1], and interacting
galaxies [3], all show the same SN rate of about 1 in SN units (SNu).
This is a puzzling result, as when a new episode of star formation starts in
an old galaxy, both the B luminosity and the SN rate increase (if the obscu-
ration by the dust is neglected) but the SN rate expressed in SNu (number
of SNe per century per 10 10 solar luminosities in the B band) is not expected
to remain constant. The SN rate, barely contaminated by the underlying old

2 F. Mannucci G. Cresci R. Maiolino and M. Della Valle
population, is expected to show a sharper increase and evolve on dierent
time scales. As a result the constancy of the SN rate cannot be explained by
dust-free models of galaxy evolution.
Large amount of dust are always present in starburst galaxies, eecting
both the B luminosity and the SN rate. Extinctions of A V  10 are often
found, preventing the detection of SNe by optical observations. It is therefore
crucial to use radio or infrared observation to derived a more complete view
of the SN events. Even in the near-infrared, at 2 m of wavelength, dust
extinction is much reduced, being about 1/10 of that in V.
When dealing with dusty active galaxies, the normalization based on the
B luminosity has no clear meaning as this band is produced by both the old
and new populations and is absorbed by the dust. In this case we prefer to use
the \far infrared SN unit" SNuIR, dene as the number of SNe per century
per 10 10 solar luminosities in the Far InfraRed (FIR). This normalization is
more meaningful as the FIR luminosity is proportional to the current Star-
Formation Rate (SFR). It is actually possible to predict the number of expect
SN from the FIR luminosity [4, 5]. This prediction depends on several factors,
as the radio properties of the SN (used to estimate the intrinsic SN rate in
nearby galaxies), the relation between SFR and FIR luminosity, the Initial
Mass function (IMF), the presence of an AGN. The number of detected SN
can also be used to constrain these parameters.
2 The ground-based observations
Several groups have completed or started near-IR SN searches [6, 7, 8, 5],
but these works produced only two detections and no spectroscopic follow-up.
The reason of these negative results are probably due to a combination of low
spatial resolution, limited eld-of-view, low sensitivity and small number of
expected events.
Our campaign started in 1999. Observations up to 2001 are described in [4],
while in this contribution we present the updated results up to summer 2003.
The galaxies were selected to have large FIR luminosities, between 210 11 L
and 2  10 12
L , corresponding to about 0.3 3 expected SNe per year per
galaxy. Such high expected rates were chosen to assure signicant statistical
results even in a short period of time. The distances are below 200 Mpc,
assuring enough sensitivity and resolution to detect point sources over the
bright galaxy background. We monitored 47 starburst galaxies in the K band
(2.2m) mainly by using 4m class telescope in sites of good seeing, the TNG in
La Palma and the NTT at La Silla. Some observations were also obtained by
the UoA 61" telescope. In 2002 and 2003 we obtained 50 new images, mainly
with the NTT. The total number of observations is now 304, with an average
number of 6.5 observations per galaxy. A sample of less distant, less luminous
galaxies were also monitored with the TIRGO 1.5m telescope. The results will
be discussed in a dierent paper (Cresci et al., in preparation).

The Infrared Supernova Rate 3
The various images of the same galaxy were carefully aligned, scaled to
the same ux, reduced to the same PSF and subtracted. Typical limiting
magnitude were K17 on the nucleus and K19 at distances larger that
about 1 arcsec.
These observations produced the detection of 4 events, the rst signicant
sample of events detected in the near-infrared. For one event, SN2001db [9],
we also obtain a spectroscopic follow-up: this event is a type II SN discovered
after maximum light. The extinction, measured by the H/H and Br/H
line ratios [10] is A V  5:6. As expect, this one was the SN with the highest
extinction know at that time.
Obtaining an infrared SN rate from the data is not straightforward and
is subject to large uncertainties: this is due to the small number of detected
events, to the variability of the properties of core-collapse SNe in the near-IR
and to the dependence of the detection limit on the distance from the galaxy
nucleus. Using the same hypothesis of [4] we derive an expected number of
55 SNe if they are all out of the nucleus and 16 if they are in the central
arcsec. Reducing these numbers to the observed 4 events imply extinctions of
A V = 33 and A V = 10, respectively. The measured SN rate SN NIR
r , assuming
that 80% of the SNe explode in the nucleus as in [4], is 0.40 SNuIR.
These limits are already quite high and are based on 4 years of (sparse)
observations. To do better it is necessary to use instruments with higher sen-
sitivity to point sources, as the HST.
3 Archive HST observations
The NICMOS camera on the HST is an ideal instrument to look for SNe in
the near-infrared. Its resolution is a few times higher than from the ground
under average seeing, and the PSF is much more stable allowing for a better
subtraction even near the bright galactic nucleus.
Most nearby starburst galaxies were already observed with this camera.
Unfortunately, the HST target selection policy does not usually allow for
duplicate observations of the same object with the same instrument setting.
As a consequence, only very few archive data can be used to look for variability.
In a pilot study, we have searched the NICMOS archive for repeated ob-
servations of starburst galaxies with a long time span. We found 4 objects:
NGC34, NGC5256, Arp200 and NGC6240. For NGC34 and Arp220 a narrow-
band lter image was acquired a few months after the corresponding broad-
band one. In this case each pixel of the broad-band image can be "scaled"
to the other bandpass by interpolation over the observed broad-band colors.
NGC5356 was observed twice with the same F160W lter but with a dier-
ent camera. NGC6240 was observed twice in the same broad-band lter and
camera, probably because of problems with the PSF of the rst observation.
All the images have short exposure times, up to 4 minutes: nevertheless the
resulting limit magnitudes are between H=18.8 and 21.0. The nuclear region

4 F. Mannucci G. Cresci R. Maiolino and M. Della Valle
where the residuals of the subtraction are large is conned in the central 0.3
arcsec, but we expect to reduce it when the same instrument setting is used.
Despite the small sample of only 4 objects, the small number of observa-
tions and the far than ideal instrument settings, the galaxies should produce
about 5 observable SNe given their FIR luminosities, the time span of the
observations and limit magnitudes. Despite of these expectations, no SN was
detected. Also in this case we attribute this lack of detection to the presence
of high extinctions, but the limits are less stringent.
4 Incoming HST and VLA observations
In order to obtain more useful data, recently an HST proposal by our group
was approved for cycle 12. The aim of the proposal is to obtain second epoch
images of a sample of 37 nearby starburst galaxies already observed by NIC-
MOS in the F160W lter. This \snapshot" program is already active and at
the moment of writing the rst galaxy was already observed (see the gure).
If all the galaxies will be observed, we expect to detect up to 50 SNe, value
corresponding to no extinction. Even if all SNe suer an extinction of A V = 30
we would still expect to detect 8 SNe. Therefore we are looking forward these
observations.
Fig. 1. Left panel: NICMOS image of the nucleus of NGC3690 in the F160W lter
taken in Aug 2003; center panel: residuals with an image taken in 1997 with the
same instrument setting; a fraction of less then 0.3% of the ux in the central arcsec
(H 13:6) remains in the residual image. right panel: a simulated SN with magnitude
H=19.5 is added to show the residual noise level and the NICMOS detection power.
The circle is 1 arcsec of diameter.
All the detected objects will be observed spectroscopically. Any event de-
tected in galaxies within 100 Mpc will be observed by VLA: in these starburst
galaxies we expect to nd SNe with peculiar properties, as the radio emission

The Infrared Supernova Rate 5
of core-collapse SN is dominated by the interaction of the ejecta with the cir-
cumstellar medium. The radio properties of these SNe can be used to derive
the density and the structure of the circumstellar medium and the details of
the late stages of the presupernova stellar evolution. Comparison with the ex-
isting radio SN models will also test their validity in a wider range of physical
conditions than previously available.
References
1. Cappellaro, E., Evans, R., & Turatto, M., 1999, A&A, 351, 459
2. Richmond, M. W., Filippenko, A. V., Galisky, J., 1998, PASP, 110, 553
3. Navasardyan, H., Petrosian, A. R., Turatto, M., Cappellaro, E., & Boulesteix,
J., 2001, MNRAS, 328, 1181.
4. Mannucci, F., et al., 2003, A&A, 401, 519
5. Mattila, S., & Meikle, W. P. S., 2001, MNRAS, 324, 325
6. van Buren, D., Jarrett, T., Terebey, S., Beichman, C.,
7. Grossan, B., Spillar, E., Tripp, R., Pirzkal, N., Sutin, B. M., Johnson, P., Barn-
aby, D., 1999, AJ, 118, 705
8. SWIRT project web site,
www.te.astro.it/attivitascientica/telescopi/azt24/swirt eng.html
9. Maiolino, R., Della Valle, M., Vanzi, L., & Mannucci, F., 2001, IAU Circ., 7661,
2
10. Maiolino, R., Vanzi, L., Mannucci, F., Cresci, G., Ghinassi, F., & Della Valle,
M., 2002, A&A, 389, 84