T h e     S W I R T

P r o j e c t







SWIRT (Supernova Watchdogging InfraRed Telescope) is a project involving the Teramo , Rome and Pulkovo (in St. Petersburg) Observatories (view here the staff), which has two main goals:

  1. the near-infrared (NIR) search of Supernova (SN) events in a sample of nearby galaxies
  2. the NIR follow-up of Supernovae
The NIR spectral range has been preferred to the visible one since the drastic decrease of dust extinction at these wavelengths, which allows the observation of sources well beyond the dust clouds in the Galaxy and in external galaxies. In fact, according to Rieke & Lebovsky, ApJ, 288, 618, (1985), the extinction in K-band amounts to just one tenth of the one in V-band, which means, e.g., that a type II SN in the Virgo Cluster is no longer detectable in the optical bands when AV=7 mag, whereas it could be still observed in K-band when AV=20 mag.

As a consequence the SWIRT Project could improve the evaluation of the explosion frequencies for different SN types and the understanding of the physical processes which are at work in the explosive event.

The observations are carried out with the near-infrared camera SWIRCAM, mounted at the focal plane of the AZT-24 Telescope.



Supernovae

The explosion of a Supernova (SN) is one of the most spectacular events in nature: a new bright source suddenly appears in a previous dark region of the sky and, in few days, it can become "more luminous than the host galaxy" (Lundmark, 1920).

Supernovae represent the final explosive outcomes of the evolution of stars which are able to synthesize heavy elements with atomic number greatest than 24.

These phenomena are interesting in their own right since some of the most complicated physical processes, "from the explosion mechanisms to nucleosynthesis, radiation transport and shock physics" (Leibundgut B., A&A Rv., 10, 179 (2000)) are simultaneously at work.

Moreover SNe have a great importance both on the astrophysical point of view and more in general in physics. In fact:

  1. the explosion causes the ejection of stellar matter nuclearly burned during the quiescent and explosive phases of the exploding star in such a way as to pollute the interstellar medium. Keeping in mind that stars originate from the interstellar medium, it is quite evident that young stellar populations are enriched in Helium and metal contents with respect to old stellar populations due to the fact that a greater number of SNe has exploded in the considered galaxy, causing a huger pollution. In addition, owing to the explosion, a pressure wave propagates in the interstellar medium inducing an increase of the local density; in some cases, if the density contrast η=Δρ/ρ is of the order of or greater than
    10-3 a self-gravitating gaseous sphere forms and a proto-star is produced. As a consequence SNe take part to the cyclic "death/birth" process of stars which contains the chemical and nuclear history of galaxies. The marks of this evolution are imprinted in the properties of low mass stars which, evolving very slowly, maintain unaltered the chemical composition at the time of the star formation in such way that they represent the main source of information about the chemical evolution of galaxies;
  2. in a Supernova explosion a huge amount of kinetic energy is delivered in the interstellar medium. In this way SNe can play a key role in the ejection of matter from the galaxy where the phenomenon occurs;
  3. the analysis of a SN explosion represents an useful tool to verify the accuracy of stellar models for stars in advanced evolutionary phases ( Asymptotic Giant Branch - AGB - and successive phases);
  4. in the explosion of a specific class of Supernovae (SNe II) very high energy neutrino beams are emitted (Eν ~ 1053 erg). If the SN event occurs in the Galaxy these beams can be detected with the current generation of neutrino detectors, obtaining information both on the physical mechanism which determines the explosion and on the neutrino properties;
  5. SNe which emit neutrinos produce also gravitational waves since the mass distribution is asymmetric in the exploding object; if the phenomenon occurs in the Galaxy or in a nearby galaxy these signals can be recorded with the detectors currently available, whereas the next generation of gravitational wave detectors would have to be able to distinguish above the noise signals from galaxies up to about 1 Mpc;
  6. another class of Supernovae (SNe Ia) exhibits a peculiar homogeneity in the profile of the light curve and in the spectra at maximum of brightness, so that it is common opinion that they can be used as standard candles on very large distance scale (~ 300 Mpc) where the ordinary distance indicators can not be further used;
  7. it is common opinion that SNe could produce cosmic rays (emission of high energy particles) and that they could be the main location where cosmic rays themselves are accelerated;
  8. in a Supernova explosion radioactive elements with long decay time are produced; the X photons delivered in the decays participate in some way to the observed background X radiation so that SNe represent the ideal target of X detector satellites to study the properties of the unstable nuclei and their abundances in stars.
In spite of their great importance many uncertainties there still exist about both the physics of the explosion and the evolution of the stellar progenitor system of these events since the extreme physical conditions of an exploding star are known very badly (in some respects unknown at all); in addition only in recent years observational searches have provided a huge amount of data and, in turn, a critical and more detailed investigation has become possible.

Observational data provide for each event light curves in different photometric bands and spectra. By an analysis of the spectra it is possible to determine both the chemical composition of the ejected matter and also the velocity of the ejected matter. In this way it is possible to have a posteriori a direct information about both the produced nucleosynthesis and the energy delivered during the explosion and, in turn, to derive information about the explosion mechanism of a given SN event.

On the other hand, the light curve represents the time evolution of the magnitude in a specific band; by performing an integration over the observed fluxes in different photometric bands, it is possible to compute the bolometric light curve, i.e. the time evolution of the total luminosity.
The light curves associated with different types of Supernova are all characterized by the same behaviour, a very fast rise until a maximum and then a slow fading down, though the exact shape is determined by the explosion mechanism and the produced nucleosynthesis.



According to the observational evidence, Minkowsky (1940) first suggested to distinguish two different classes of SNe
(see figure):

Increasing observational data, it became clear that this scheme was unable to account for the large differences existing between distinct events and, hence, trying to prevent an enormous proliferation of classes, a further subdivision was proposed. In particular it was suggested to distinguish SNe II in two subclasses on the base of the light curve profile:

Recently a new subclass of type II has been identified, namely type IIn supernovae, which exhibit narrow lines in their spectra. In any case, according to our knowledge of the SN phenomenon, it seems quite reasonable that SNe IIn are type IIL supernovae embedded in a dense circumstellar medium.

On the other hand, it was suggested to distinguish SNe I in two subclasses basing on the different properties of spectra at maximum. In recent years, SNe I have been distinguished according on the presence of the strong SiII line (λ=6150 A); in addition SNe I which do not exhibit the SiII line have been further subdivided according on the presence or absence of the HeI line (λ=5876 A). As a result, type I Supernovae have been distinguished in:
Nowadays it is common opinion that SNe Ib and Ic represents a quite similar phenomena, the observed differences being due to a different chemical composition of the photospheric layers at the time of the explosion.
It is worth noticing that the previous classification scheme has been obtained referring only to the properties of spectra at the maximum phase and to the shape of light curves. Moreover this phenomenological classification reflects more deep differences about both the stellar progenitor and the explosion mechanism. In particular it is common opinion that SNe Ia (or "classic" SNe I or briefly SNe I) originate by the thermal disruption of a low mass star, whereas SNe II and SNe Ib/c are due to the iron core collapse of an initial massive star.


To date, more than one thousand SNe have been observed in the external galaxies. From these data, it has been inferred that type-Ia supernovae are the only ones occurring in elliptical galaxies, and that the SN rates are higher in late-type galaxies (late spirals and irregulars). However, the estimates of the SN rates are largely uncertain and it is common to find in literature SN rates ranging from 0.1 or less to several SNU (remember that 1 SNU is the unit of measure of the SN rate and corresponds to the explosion of 1 supernova per century in a galaxy with luminosity in B band equal to 1010 solar luminosities). This uncertainty is mainly due to the fact that the current surveys, aimed to discover extragalactic SNe, are performed only in the visible: therefore, a large number of SNe can be not detected, namely those occurring beyond the clouds of dust which produce a large extinction of light.

An infrared survey should produce a significant improvement in the determination of the SN rates, making possible to observe also highly obscured SN explosions.

Of course, the most spectacular supernovae should be those exploding in our galaxy. Such an event has occurred several times in the past centuries, and more or less accurate recordings can be found in the history. The most famous is the supernova of 1054 AD, recorded by Chinese astronomers and exploded in Taurus. Probably this is not the first supernova observed: historical chronicles report a "new star" observed in Lupus by Arabic astronomers in 1006 AD, plus other recordings by Chinese astronomers before the year 1000 AD. The "absolutely first" recording should be attributed again to Chinese astronomers in 352 BC.

Also famous are the Ticho’s supernova in Cassiopeia (1572), the Kepler’s supernova in Ophiucus (1604) and a suspected supernova in Cassiopeia observed by Flamsteed in 1667. In the twentieth century only a supernova exploded near our galaxy, in the Large Magellanic Cloud: it has been named Supernova 1987A and has been very extensively studied, leading to important improvements in this field.

The supernovae exploded nearby are still visible via the remnants they left in the site of the explosion. Recently many other SN remnants have been discovered and with the advent of the Hubble Space Telescope the number of such objects has increased hugely. Beyond the very spectacular aspect of these remnants, their properties (such as luminosity, color indexes, radio emission) are very useful to the astronomers to test the models about the physical processes occurring before, during and after a supernova explosion and to date the explosion itself, thus refining the estimates on the SN rate in our galaxy.






The SWIRT sample of galaxies

A first sample for the SWIRT survey was compiled by the RC3 Catalogue (De Vacouleurs et al., Third Reference Catalogue of Bright Galaxies, Springer-Verlag, 1991) and included 1448 nearby galaxies (Battinelli et al., Mem. Soc. Astron. It., 65, 863). It was aimed mainly to test some crucial characteristics related to the feasibility of the project, such as the infrared visibility of the galaxies and their spatial and temporal sampling: this test was necessary because of the lack of reference images at near-infrared wavelengths (the 2MASS data had not been released yet). After a first year of commissioning, the catalogue was revised and the number of galaxies was reduced to a maximum of 600.

The current sample of galaxies consists of 514 spiral galaxies, plus 48 elliptical and 44 irregular ones which were kept because they were already observed in the first year. The routine observations, however, are performed on the 514 spirals, whereas the other galaxies are assumed as a test for a future extension of the project to galaxies of any morphological type. The data concerning the SWIRT sample are summarized in the following table.

SWIRT galaxy sample data

Maximum number of galaxies evaluated accordingly to the average weather conditions at Campo Imperatore and human resources

~600

Total selected galaxies

606

Spirals

514

Elliptical

48

Irregulars

44

Distribution over the sky

No constraints on RA
DEC > -10°
Most of the galaxies come from Virgo and Coma clusters

Inclination

60° ≤ i ≤ 90° (edge on) for the 67% of the sample
0° ≤ i ≤ 30° (face on) for the 33% of the sample

B absolute magnitude

&le -17

Angular diameter

≥ 1.5 arcmin to avoid spatial undersampling
≤ 4.0 arcmin to fit the SWIRCAM field-of-view

> 5.0 arcmin for 25 galaxies in order to test a better
sampling of the nuclear regions

Distance modulus

≤ 33.5 for face-on galaxies
≤ 32 for edge-on galaxies

Control time

20d ≤TC≤ 90d

<TC> = 40d

Other

No stars-poor fields
No very bright stars (in the infrared) close to the galaxies

As it is apparent from the table, the sample is not homogeneous as to the inclination: high inclination (edge-on) galaxies are preferred. This is because the main aim of the SWIRT survey is to exploit the potential of the infrared observations with respect to the visible ones: this difference is expected to be strengthened just in the edge-on galaxies, where the extinction by dust is maximum. Nevertheless, a "reference" subsample of low-inclination (face-on) galaxies has been included in the galaxies, in order to compare the any possible SWIRT detection of SNe in these galaxies with the optical observations.

The constraints put on the diameters and distance moduli made not possible to select the galaxies uniformly distributed over the sky. This fact is visible in the map on the right, where the Virgo and Coma clusters can be recognized aroud RA=12h.

Despite its non-uniformity, this distribution has an advantage for the SWIRT observations. It comes out, indeed, that the maximum number of galaxies visible in the sky occurs during the late winter, when the night length is maximum (see figure on the left): that means, there is a positive correlation between the number of galaxies to observe and the length of the night, which makes the time available for each galaxy (roughly the ratio between these two quantities) nearly constant over the year.

The map reported above shows also that about 40 galaxies are located at high declinations, and hence can be always observed during the year. Moreover, a statistical count performed on the sample shows that 60 galaxies can be observed for 11 months per year and the most part of the sample is visible above an elevation of 30° for at least 8 months.

A SWIRT model for the galaxies in the sample has been developed (Caratti , Dolci & Li Causi, OACT Internal Report, 12/2001) in order to compute the control time for each galaxy: the control times range between 20 days (for some distant edge-on galaxies) and 90 days (for some nearby face-on galaxies). On average, the sample has to be surveyed every 40 days, a number consistent with the time available for the observations and the total number of galaxies.

Finally, it is worth of note that some fields have been discarded because of the presence of very bright stars near the galaxies which could give problems with the automatic routines of SN search.
Other fields were discarded because of the lack of stars bright enough, which made not possible to recombine the dithered frames in the image preprocessing (Di Paola, OAR Internal Report OAR/00/IR1 jan. 00).



Outcomes: the discoveries of SN2000E and SN2002cv

The current SWIRT archive includes 1091 frames from the old catalogue, acquired between Oct 5, 1999 and Jun 26, 2001 and concerning 404 out of 1448 galaxies, plus 710 frames from the new catalogue, acquired between Jun 26, 2001 and Nov 24, 2001 and concerning 299 out of 514 galaxies.

The images have been processed and analyzed, and results similar to those shown in the figure have been obtained.

Even if many possible candidates have been monitored, no unambiguous detection of SNe has been recorded yet. However, at least 5 active regions not visible in the optical range have been discovered and 4 known asteroids have been serendipitously observed: moreover, a flux variation has been recorded in more than 80 sources (Caratti, Degree Thesis, Univ. di Roma "La Sapienza").

However, a more detailed data analysis is still in progress. A broad sample of the observed SWIRT galaxies can be found in the AZT24 image gallery. The complete set of data is available in the SWIRT public archive.

Curiously, the two supernovae discovered by the SWIRT team are in galaxies not belonging to the catalogue.

SUPERNOVA SN2000E has been discovered on Jan 26, 2000 by G. Valentini and co-workers in NGC6951 during the follow-up of SN1999el (Valentini et al., 2000, IAU Circ. No. 7351). It has been classified as a Type-Ia supernova. The parent galaxy, NGC6951, is located in the Cepheus constellation and its distance is about 24 Mpc (80 millions light-years).
This event is particularly important because, to date, the possibility that two almost contemporary supernovae occur in the same galaxy seems to be rather low. Visit here also the TNT page with further information and images of SN2000E.



SUPERNOVA SN2002cv has been discovered on May 13, 2002 by V. Larionov and co-workers in NGC3190 during the follow-up of SN2002bo (Larionov et al., 2000, IAU Circ. No. 7901). UGC3190 is a spiral galaxy in the Leo constellation, at a distance of about 20 Mpc (66 millions light-years).
Its classification has been rather difficult, because the optical spectrum was heavily reddened and no features were detectable below 600 nm. Basing on the interpretation of near-infrared spectra taken at UKIRT telescope, SN2002cv is probably a Type-Ia supernova.
The discovery of this supernova is probably of fundamental importance, for at least two reasons.
First of all, it has been discovered by infrared observations, and revealed itself as the most extincted supernova ever discovered: only other two supernovae were discovered with infrared techniques (Maiolino et al., 2002, Astron. Astrophys., 389, 84), but their extinction was lower.
Second, the very high extinction of this supernova made it invisible in the optical range (no signal detected above the noise at limiting magnitude of 24) and this fact could imply that the SN rate is remarkably higher than expected, and also that quasi-simultaneous SN events in the same galaxy are not so rare as previously believed.



Other surveys

Currently, at least eleven SN surveys are operating around the world (e. g. LOTOSS (Traffers, 1997), BAO (Li, 1996), Nearby Galaxy SN Search (Strolger, 1998), etc...), performed by observatories or other scientific institutions. Moreover, at least other nine SN surveys are being performed by amateur astronomers (e. g. Puckett observatory (Puckett, 1998), Tenagra (Schwartz, 1997), etc...).

In general, detailed information about each survey cannot be found in the literature: only in few cases the limiting magnitude and the control time can be known from IAU Circulars. We know, however, that almost all of these surveys do not make use of filters. We can deduce therefore a mean limit magnitude of about 19 for them and really short control times (less than 20 days).

Some data are available on the web for LOTOSS: this is essentially a network of three telescopes, located in California (0.8 meter reflector telescope (Richmond, Traffers & Filippenko, 1993)), in Oregon (0.4 meter Schmidt-Cassegrain telescope) and Patagonia (0.5 meter Ricthey-Cretien telecope). LOTOSS catalogue is made up of 14140 galaxies and 503 of them are present in our catalog too (click here to download the complete gzipped list). The exposure time for each galaxy is between 30 and 40 seconds and every night up to 1200 images are realized. In the last two years, LOTOSS has discovered 95 (out of 331) SNe, resulting one of the most prolific survey in the world.

Since we do not know how many SWIRT galaxies are included in other catalogues, we can only overestimate the number of SWIRT SNe that could be previously discovered by other surveys, assuming that all our sample is regularly observed by the other surveys.

Recomputing the SWIRT galaxy model for the visible bands and a survey limiting magnitude 19, we obtained that at least 2 ± 1 SNe could be detected by other surveys. Since the same result was 3 ± 1 SNe for the SWIRT survey, we conclude that at least one SWIRT SN (about 25% of the total) should not be detected by any other visual survey.



Near-infrared follow-up of Supernovae

The most important result of the follow-up program has been the determination of the lightcurves for the supernova SN1999el (discovered by Cao et al., IAU Circ. No. 7288 (1999)). This supernova has been observed at the AZT24 telescope in the J, H, and K bands until 150 days after the B maximum and then at the TNG telescope, over the 200th day, as a target of opportunity in the J and K bands. Moreover, it has been observed in the optical B, V, R, I bands at the TNT 72-cm telescope of Teramo Observatory.

SN1999el is a type-IIn supernova and only another similar object (SN1998S) has been extensively studied in the past (see, for example, Fassia et al., Mon. Not. R. Astron. Soc., 318, 1093 (2000)). Therefore, this work represents a cornerstone in this field and is going to be published on Astrophysical Journal (Di Carlo et al., submitted to Astroph. J. (2001)).

Another important result is the follow-up of SN2000E whose infrared and optical lightcurves have been also determined (Valentini et al., Mem. Soc. Astron. It., in press).

Other supernovae have been observed, mostly in the infrared.
In 2000, SN2000ev (type IIn).
In 2001, SN2001Q (type II), SN2001V (type Ia) and SN2001cy (type II), shown in the figure.
In 2002, SN2002bo (type Ia) has been extensively followed in the near-infrared, and some qualitative NIR spectra have been acquired. Also SN2002cv (type Ia) has been followed, but it was too faint for attempting any spectroscopic observation.

The analysis of these data is currently in progress. Some images are shown in the AZT24 image gallery.

Last update Dec 12, 2002

Mauro Dolci, INAF - OACT