2 SN 2001db

2.1 Observations

2.1.1 Imaging

The infrared imaging observations were obtained with SOFI, the near-IR camera (Lidman et al. 2000) at the NTT ESO telescope, in the Ks band (see Table 1). NGC 3256 was observed for the first time on January 9, 2001, and subsequently three more times: February 8, March 18, and April 1. Each observation consisted of a series of 15 images, 60 s each, alternated with 15 images sampling the near sky for an optimal removal of the background. The total on-source integration time was therefore 15 min per epoch, obtained under sub-arcsec seeing conditions. The data reduction was carried out with a ``home'' package developed by our team (Hunt et al. 1994). The images obtained during the monitoring in 2001 were compared with an archival image obtained with SOFI on November 28, 1999.

In Fig. 1 we show the archival SOFI image of NGC 3256 (left) and the first image obtained in 2001 (right) where SN 2001db was detected. The magnitude of the SN at the discovery epoch was $K{\rm s}=16.03\pm 0.20$ . The SN is located at RA(J2000)  $=10^{\rm h}$27$^{\rm m}$504, Dec (J2000) =-43 54' 21'', i.e. offset 57 to the West and 57 to the South of the Ks nucleus of the galaxy (uncertainty on the relative position about 04). In Table 2 and in Fig. 2 (upper panel) we give the Ks-band light curve, both in terms of relative and absolute magnitude (assuming a distance modulus of 33.06).

Photometric measurements performed on ESO archival images (Feb. 1993) and FORS images (Apr. 2001, May 2001) imply m_V >22 for the SN (Fig. 2 lower panel).

   

Table 2: Broad lines detected and ascribed to the SN. Pa$\beta$ was observed on April 21, 2001, while all the optical lines were observed on May 16, 2001. Quoted errors are at 1$\sigma$.

Line	$\lambda _{\rm rest}$	Flux	FWHMa	$V_{\rm peak}^{b}$
		$\rm (10^{-15}~erg~s^{-1}~cm^{-2})$	$\rm (km~s^{-1})$	$\rm (km~s^{-1})$
NaIc	5893 Å	$0.34 \pm 0.07$	$6540\pm1000$	$340 \pm 800$
[OI]c	6300 Å	$0.10 \pm 0.03$	$9500\pm2000$	$320 \pm 900$
H$\alpha$	6562 Å	$4.40 \pm 0.40$	$5200\pm600$	$-2140 \pm 250$
FeII	7155 Å	$0.17 \pm 0.05$	-d	$-2290 \pm 500$
[CaII]	7324 Å	$1.05 \pm 0.1$	$8416\pm500$	$-1820 \pm 500$
CaII	8542 Å	$0.94 \pm 0.1$	$3000 \pm 800$	$-2310 \pm 300$
CaII	8662 Å	$1.50 \pm 0.15$	$6500 \pm 700$	$-370 \pm 300$
Pa$\beta$	12 818 Å	$7.61 \pm 0.7$	$4900 \pm 700$	$-2570 \pm 300$

Notes: ^a Full width half maximum in km s^-1. ^b Velocity shift of the line peak with respect to the rest frame velocity of the galaxy,
in km s^-1. ^c Marginal detection. ^d Strongly blend with CaII.

2.2 Optical spectroscopy

The spectroscopic follow-up was performed both in the optical and in the infrared.

The optical spectrum was obtained with FORS1 (Szeifert 2002) at the ESO VLT-UT1 on May 16, 2001 (i.e. when the SN was at least 4 months old) with the grism GRIS_300V and a spectral resolution of 500. The 15 m exposure was split in two different integrations, to allow an optimum subtraction of cosmic rays. Background was removed by interpolating the spectrum above and below the galaxy (after standard flat-fielding and bias subtraction).

Figure 3 (upper panel) shows the optical spectrum of the SN 2001db. There are several narrow emission lines, most of which are probably from HII regions and SNRs in the environment of SN 2001db. As mentioned above, the continuum is not due to the SN 2001db but is dominated by background emission from the host galaxy. The spectrum clearly shows a broad component of H$\alpha$ (FWHM $~~ \sim5000$ km s^-1) which is a clear signature of the SN. The signatures of the SN are better observed if the contribution of the the underlying background continuum is removed (by using as a template the contiguous regions of the galaxy intercepted by the slit and properly re-scaled). The resulting spectrum is shown in Fig. 3 (lower panel). The H$\alpha$ has a strongly asymmetric profile, whose peak is blueshifted by about 2000 km s^-1 with respect to the parent galaxy (see Schlegel 1990 for a discussion on the blueshift), but it also has a prominent red tail, a profile similar to that observed in type IIL and IIP SNe (Filippenko 1997). No other SN signatures are found in the blue part of the spectrum. A summary of the properties of all the broad lines associated with the SN is given in Table 2. We note that the profile of the broad component of [Ca II] 7324 Å and CaII 8662 Å is not as asymmetric as the other broad lines. At the same redshift of the peak of the broad H$\alpha$ we have also detected the FeII 7155 Å line.

2.3 Infrared spectroscopy

The infrared spectrum was obtained in the J band with ISAAC (Cuby et al. 2002) at the ESO VLT-UT1 on April 21, 2001. Observations were performed with the 1'' slit and with the low resolution grating (set at the 4th order), giving a spectral resolution of 500. The observations were obtained with a set of single integrations of 30 s each (10 times 3 s), by moving the object along the slit to obtain an optimal sky subtraction, for a total of 45 min of integration. Finally, the spectra obtained at different positions along the slit were subtracted from each other to remove the background, then flat fielded, aligned, co-added and calibrated. The atmospheric trasmission features were corrected as described in Maiolino et al. (1996).

Figure 3: Upper panel: optical spectrum of the SN obtained on May 16th, 2001. Only the narrow emission lines and absorption lines are marked. Lower panel: optical spectrum after removing the background galaxy light. n and b refer to narrow and broad components, respectively; the former most likely associated to the background HII regions while the latter are associated to the SN.

Figure 4: Infrared spectrum obtained on April 21st, 2001. n and b refer to narrow and broad components, respectively; the former most likely associated to the background HII regions and SN remnants, while the latter are associated with the SN 2001db.

Figure 4 shows the infrared ISAAC spectrum extracted from an aperture of 1''. Here the broad component of Pa$\beta$ is more prominent, relative to the narrow component, with respect to H$\alpha$. The profile of the broad component of Pa$\beta$ is nearly identical to that of H$\alpha$. There are also indications of a broad component of Pa$\gamma$ and HeI 1.0083 $\mu$m, but assessing the reality of these features is difficult because they are at the edge of the spectrum. Pa$\beta$ might be characterized by an absorption on the blue side (i.e. a P-Cygni profile) which might be filled by the [FeII] line. However, the flux of the latter line is a factor of $\sim$3 higher than the companion [FeII]1.3206 $\mu$m, as expected by atomic constants, thus it is unlikely that [FeII]1.2567 $\mu$m is affected by a significant absorption feature beneath its profile. Most likely Pa$\beta$ does not have a P-Cygni signature, similarly to H$\alpha$. It is worth noting that the strong [FeII] emission relative to Pa$\beta$ (narrow) indicates that the underlying emission is not simply due to HII regions, but must also be contributed significantly by SN remnants.

2.4 Extinction

The Balmer decrement for the narrow components of H$\alpha$ and H$\beta$ is $\sim$12. This implies, for a Galactic extinction curve, an equivalent screen extinction A_V = 4.2 mag. If dust is mixed with the emitting gas the real optical depth is even higher. The Na I interstellar absorption doublet at $\sim$5890 Å has an equivalent width of 5.87 Å. However we note that due to the blending with the near HeI $\lambda$5876 Å emission line, the equivalent width of Na D has been underestimated. According to the relations of Barbon et al. (1990) and Benetti (priv. comm.), and after assuming a Galactic extinction curve, EW(Na D)$~\geq~$5.87 Å implies an equivalent screen extinction $A_V\geq3{-}4.9$ mag.

We can exploit the ratio of the broad component of Pa$\beta$ and H$\alpha$ to constrain directly the reddening affecting the SN. We will assume that during 25 days elapsed between the infrared and the optical spectrum the line flux did not change significantly (e.g. Danziger et al. 1991; Xu et al. 1992; Benetti et al. 1998); afterwards we will discuss the case of rapid line variability. The observed ratio between the broad components of Pa$\beta$ and H$\alpha$is 1.73. The case B recombination gives an intrinsic ratio Pa$\beta$/H $\alpha = 0.06$, which would imply an extinction of A_V = 6.7 mag. Yet, in type II SNe the intrinsic ratio Pa$\beta$/H$\alpha$ appears to be lower than in the case B, although only a few simultaneous observations of Pa$\beta$ and H$\alpha$ are available for type II SNe. The Pa$\beta$/H$\alpha$ ratio observed in 1987A was about 0.12 (Xu et al. 1992). By interpolating the data obtained by Fassia et al. (2000) at different epochs we obtain a similar ratio Pa$\beta$/H $\alpha = 0.09$, once corrected for foreground reddening. If we conservatively assume Pa$\beta$/H $\alpha =0.12$, the observed value in SN 2001db gives an extinction toward the SN of A_V = 5.3 mag.

To discuss the rapid variability case we assume that the hydrogen line flux has decreased at the same rate of the continuum. In this case the line flux cannot have changed by more than 40%, which gives a lower limit on the inferred optical extinction of A_V > 4.6 mag or A_V > 5.7 mag for an intrinsic ratio of 0.12 and 0.06 (case B) respectively.

Summarizing, we derive that the optical extinction toward the SN is larger than A_V > 3 (lower limit inferred above from the Na D absorption and narrow Balmer decrement) and most likely in the range 4.6 to 6.7 mag. In the following we assume $A_V \approx 5.6$ mag, with an uncertainty of about 1 mag.

Note that the upper limit on the V magnitude of the SN also provides a constraint on the extinction when compared with the light curve of SN templates (Fig. 2). With an extinction A_V < 2 mag the SN would have been detected in the optical image of April 17. This lower limit on the extinction A_V > 2 is fully consistent with the value inferred above.

In this section we have adopted a Galactic "standard'' extinction curve, but there are indications that starbust galaxies might have different extinction curves. Calzetti et al. (2000) estimated the extinction curves for a sample of starburst galaxies. These extinction curves apply mostly for the "mixed" case and, therefore, are probably inappropriate for our SN. However, aware of this caveat, we have also tried to estimate the extinction by using the extinction curves given in Calzetti et al. (2000) and obtained (accounting also the uncertainties in the reddening curves) a visual extinction ranging from 4.1 to 8.4 mag.

In the compilation given in Mattila & Meikle (2000) and in Schmidt et al. (1994) the extinction A_V of core collapsed SNe is generally lower than $\sim$1.5 mag, with the exception of SN 1973R for which Schmidt et al. estimate A_V = 2.7 mag. The distribution of extinctions in these two compilations is shown in Fig. 5. If these compilations are representative, then the extinction inferred for the IR SN 2001db is probably the highest among the SNe so far discovered (Fig. 5).

Figure 5: Distribution of extinction for the optically discovered SNe reported in the literature (see text), compared with the extinction measured for the infrared SN 2001db.

2.5 Supernova type and age

The lack of any absorption blueward of the line peak (i.e. the typical P-Cygni profile) may dis-favor a type IIP. However, since the continuum is dominated by the background emission of the parent galaxy, the lack of P-Cygni absorption is more difficult to assess. Based in Fig. 3 (lower panel), we cannot exclude the presence of an absorption feature blueward of the H$\alpha$ with an EW less than $\sim$2 Å. The shape of the broad hydrogen lines (strongly asymmetric and with a blueshifted peak) does not favor a type IIn. The available photometric points also do not discriminate between type IIL and IIP. In Fig. 2 we report the light curves of SN 1980k and SN 1999em taken as representative of type IIL and type IIP, respectively (data from Dwek et al. 1983; Barbon et al. 1982; Hamuy et al. 2001). Both the optical and infrared light curves have been extinguished by an A_V =5.6 (i.e. A_K=0.56), as obtained in the former section. The infrared observations can be roughly fit both with the light curve of SN 1980k offset by 65 days, and with the light curve of the SN 1999em offset by 90 days and 1.0 mag. The average K-band light curve for type II SNe obtained by Mattila & Meikle (2000) is nearly identical to the SN 1980k light curve in Fig. 2, but offset by 55 days.

It is interesting to note that, in any case, the V-band magnitude of the SN was most likely fainter than 20 at its maximum (see Fig. 2) and, therefore, it would have been missed by most of the optical SN search programs (e.g. Richmond et al. 1998).

Figure 6: Left: Ks-band image of UGC 4881 obtained with ARNICA in December 1999. Right: difference between the ARNICA image of December 1999 and the NICS image of February 2001, after matching of the two PSF. Both images are $35''\times 35''$ in size.