Photometric and Spectroscopic study of a highly reddened type Ia supernova SN 2003hx in NGC 2076

We present $UBVRI$ CCD photometry and optical spectra of the type Ia supernova SN 2003hx which appeared in the galaxy NGC 2076, obtained till $\sim$ 146 days after the epoch of $B$ band maximum. The supernova reached at maximum brightness in $B$ band on JD 245 2893 $\pm$ 1.0 with an apparent magnitude of 14.92 $\pm$ 0.01 mag which was estimated by making template fits to the light curves. SN 2003hx is an example of a highly reddened supernova with $E(B-V)$ = 0.56 $\pm$ 0.23. We estimate $R_v$ = 1.97 $\pm$ 0.54 which indicates the small size of dust particles as compared to their galactic counterparts. The luminosity decline rate is $\Delta m_{15}(B)$ = 1.17 $\pm$ 0.12 mag and the absolute $B$ band magnitude obtained from the luminosity versus decline rate relation (Phillips et al. 1999) is $M^B_{max}$ = -19.20 $\pm$ 0.18 mag. The peak bolometric luminosity indicates that $\sim$ 0.66 $M_\odot$ mass of $^{56}$ Ni was ejected by the supernova. The spectral evolution indicates the supernova to be a normal type Ia event.


INTRODUCTION
Type Ia supernovae, being one of the most luminous stellar outburst, form a fairly homogeneous class of objects and are considered as standard candles for determining extragalactic distances and cosmological parameters. Type Ia supernovae are produced by thermonuclear explosions of white dwarfs (Hoyle & Fowler 1960) primarily composed of Carbon and Oxygen nuclei. The probable explosion scenario involves a binary system in which a white dwarf accretes matter from its companion star until it reaches the Chandrasekhar mass limit of 1.4 M⊙. Though type Ia supernovae show homogeneity in both their photometric as well as spectroscopic properties (Höflich et al. 1996) but many SNe show significant deviations. Li et al. (2001) studied a sample of type Ia supernovae and conclude that ∼ 64 percent of them belong to the normal group, almost 20 percent belong to the over luminous group of events such as SN 1991T whereas about 16 percent belong to the sub luminous type such as SN 1991bg.
The last few decades have witnessed a large number of diverse data sets for SNe Ia. Nevertheless, type Ia supernovae seem to follow a few common patterns. One of these is the correlation between the peak luminosity and the linear decline rate (Phillips 1993). The color evolution, spectral appearance and the host galaxy morphology are the other correlations. The peak absolute magnitude of type Ia SNe is correlated with the Hubble type of the parent galaxy. SNe Ia hosted by elliptical galaxies are comparatively fainter than SNe Ia in spirals (Della Valle & Panagia 1992, Howell 2001. Even amongst the normal group of type Ia SNe significant photometric and spectroscopic uncertainties exist. Nugent et al. (1995) saw that the spectral variations in type Ia SNe correlate with the expansion velocity, the effective temperature and the peak luminosity. Thus, it is not sufficient to describe SNe Ia by a single parameter such as the early light curve decline rather the diversity noticed is multi-dimensional (Hatano et al. 2000, Benetti et al. 2004. It is therefore, necessary to study the photometric and spectroscopic evolution of individual type Ia SNe. The integrated flux in optical bands provides a meaningful estimate of the bolometric luminosity which is directly related to the amount of radioactive 56 Ni synthesised and ejected in the explosion (Arnett 1982, Höflich et al. 1996, Pinto & Eastman 2000a) which can later be used to test various explosion models of type Ia supernovae.
We present in this paper the optical photometric and spectroscopic observations of highly reddened type Ia supernova SN 2003hx. SN 2003hx was discovered on unfiltered KAIT images on 2003 September 12.5 UT (magnitude 14.3) and 13.5 UT (magnitude 14.4) by Burket, Papenkova & Li (2003). A KAIT image of the same region taken on 2003 March 7.2 UT shows nothing at the location of the supernova to a limiting magnitude of ∼ 18.5. SN 2003hx is located at α = 05 h 46 m 46 s .97, δ = −16 • 47 ′ 00 ′′ .6 (J2000) which is 5 ′′ .2 West and 2 ′′ .6 South of the nucleus of the galaxy NGC 2076. A spectrum of SN 2003hx taken on 2003 September 13.78 UT at the Australian National University (ANU) 2.3-m telescope shows it to be a type Ia supernova around maximum light (Salvo, Norris & Schmidt, 2003). The Si II 635.5 nm line gives an expansion velocity of ∼ 12000 km/sec if the NED recession velocity for the host is adopted as 2142 km/sec (Salvo, Norris & Schmidt, 2003). It was confirmed to be a type Ia supernova around 10 days past maximum light with the spectropolarimetric observations using the ESO very large telescope on 2003 September 15.4 UT (Wang & Baade, 2003). The interstellar Na I D line has an equivalent width of 4.86Åwhich indicates significant dust extinction. The observed degree of polarization is ∼ 2 percent (Wang & Baade, 2003). If this polarization is due to the dust in the host galaxy, then it implies that the dust particles are significantly smaller in size than their Galactic counterparts. These observations yield a value of 2.2 for the ratio of total to selective extinction, Rv = Av/E(B − V ) (Wang & Baade, 2003).
We have carried out the optical photometric and spectroscopic observations of the type Ia supernova SN 2003hx. A brief description of the observations and data analysis is given in section 2, whereas the development of the light curves and color curves are presented in section 3. Section 4 discusses about the reddening estimate. The description about the absolute magnitude, the bolometric luminosity and the estimation of 56 Ni ejected are discussed in section 5. Spectral evolution has been studied with a comparison to other type Ia supernovae in section 6. Conclusion forms section 7 of the paper.

OBSERVATIONS AND DATA REDUCTION
The observations of SN 2003hx were carried out with the 2-m Himalayan Chandra Telescope (HCT) at Indian Astronomical Observatory (IAO), Hanle during 2003 September 18 to 2004 February 03 which was six days after the discovery on 2003 September 12. The Himalayan Faint Object Spectrograph Camera (HFOSC) equipped with the SITe 2 K × 4 K pixel CCD was used. The central 2 K × 2 K region was used for imaging covering a field of 10 ′ × 10 ′ on the sky corresponding to a plate scale 0.296 arcsec pixel −1 . The gain and read out noise of the CCD camera are 1.22 e − /ADU and 4.87 e − respectively.

Photometry
The broad band U BV RI photometric observations of SN 2003hx were carried out at 16 epochs during 18 September 2003 to 03 February 2003. All the images were bias subtracted, flat fielded and cosmic ray removed in the standard fashion using various tasks in IRAF. Landolt (1992) standard region PG 0231+051 was imaged along with the supernova field in U BV RI filters on 26 October 2005 under good photometric sky conditions. The values of atmospheric extinction on the night of 26/27 October 2005 determined from the observations of PG 0231+051 bright stars are 0.30 ± 0.01, 0.20 ± 0.009, 0.12 ± 0.007, 0.08 ± 0.004 and 0.04 ± 0.003 in U, B, V, R and I filters respectively. The observations of PG 0231+051 were used to generate secondary standards in the supernova field. The U BV RI magnitudes of 10 secondary stars using the transformation equations obtained are listed in Table 1 and are marked in Figure 1. These magnitudes were used to calibrate the data obtained on other nights. The sequence photometry of SN 2003hx field was carried out by the American Association of Variable Star Observers (AAVSO). We compared our field calibration with that of AAVSO and found it to be consistent. Thus, our calibration of SN 2003hx field is secure.
We performed aperture photometry on the local standards using an aperture of 3 -4 times the F W HM of the seeing profile that was determined on the basis of an aperture growth curve. Accurate estimate of sky and its subtraction plays a crucial role in the photometry when the object falls on the varying background. The supernova was just next to the nucleus of the galaxy, we therefore had to perform template subtraction to get a better estimate of the underlying background. The template observations were taken with the same instrumental setup, on 26 October 2005, nearly two years after the supernova discovery, when the supernova had faded. These templates were subtracted from the images of SN obtained during 2003 -2004. One such subtracted image, showing only the supernova SN 2003hx, is shown in Figure  2. When the seeing was not very good, the template subtraction was not perfect and showed signature of the galaxy. Here, the supernova magnitudes were determined by using the profile-fitting method with a fitting radius equal to that of the F W HM of the seeing profile. The difference between the aperture and profile-fitting magnitudes was obtained us- ing standards and was applied to the supernova magnitudes. Thus, the supernova magnitudes were obtained by differentially calibrating with respect to the secondary standards listed in Table 1. The supernova magnitudes derived in this way are given in Table 2.

Spectroscopy
Spectroscopic observations of SN 2003hx were carried out on four nights starting 19 September 2003 which was 7 days after the discovery on 12 September 2003. The log of spectroscopic observations is given in Table 3. All the spectra were obtained at a resolution of ∼ 7Å in the wavelength range 3300-6000Å, 3500-7000Å, 5200-9200Åand 5200-10300Å. Spectroscopic data were reduced using the standard routines within IRAF. The data were bias corrected, flat-fielded and the one dimensional spectra were extracted using the optimal extraction method. FeAr and FeNe sources were used for wavelength calibration. The wavelength calibrated spectra were corrected for instrumental response using spectra of spectrophotometric standards observed on the same night and brought to the same flux scale. The final spectrum on a relative flux scale were obtained by combining the flux calibrated spectra in the two different regions scaled to a weighted mean.

LIGHT CURVES AND COLOR CURVES
In this section we study the multi band light curve and color curve evolution of SN 2003hx and compare it with other type Ia supernovae.

U BV RI light curves
We present the U BV RI light curves of SN 2003hx in Figure  3. Since our observations started ∼ 6 days after the discovery, we do not have observations around the peak brightness. The peak brightness and the JD corresponding to the peak brightness in different bands was estimated by making template fits to the observations. Our observations span a period of ∼ 150 days. The frequency distribution of our data is N (U, B, V, R, I) = (3,11,16,14,13). In order to determine peak brightness in different bands we have adopted the template fitting method. We attempted to fit the different template sets in BV I bands given by Hamuy et al. (1996). We adopted a χ 2 minimizing technique which solved simultaneously for the peak magnitude and the peak time in different bands. We see that the B band fits best with the template of SN 1992al whereas the lowest value of χ 2 is attained in the case of SN 1992A for V and I bands. Since Hamuy et al. (1996) does not present the R band template, we have taken the R band template for type Ia supernova from Schlegel (1995). The R band template was similarly fit to the observations to determine the peak magnitude and the peak time in R band. Figure 4 shows the light curves in BV RI bands including the template fits to the data. The main parameters of SN 2003hx as estimated from template fits are listed in Table 4. Leibundgut (1988) showed   Hamuy et al. (1996) in the BV I bands and Schlegel (1995)  that for a normal type Ia supernova, a 2 day difference is seen between the time of maxima in the B and V bands. We see from the template fits that SN 2003hx reached maximum brightness in the V , R and I bands earlier than in the B band. An excellent match of the SN 1992al template with that of SN 2003hx indicates that the peak in the B band occurred at JD 245 2893 ± 1.0. These peak magnitudes obtained using template fitting are further used to calculate the peak luminosity in section 5. In Figure 4 we see a pronounced secondary maximum for SN 2003hx in the I band. This secondary maximum was seen ∼ 21 days after the B maximum and was 0.32 magnitude fainter than the first maximum. The magnitude of the secondary I maxima is listed in Table 4. R band also shows a noticeable rise similar to the I band at similar epochs. This secondary maxima in the I band is a remarkable feature of type Ia SNe and becomes more pronounced in the near-IR band. Such behaviour has been seen for many type Ia SNe. Elias et al. (1981) and Pinto & Eastman (2000b) pointed out that this secondary maxima is due to a temporary increase in absorption, which reduces with the fall in the degree of ionization several weeks after maximum light. ∆m15(B), the number of magnitudes in the B band by which the SN declines in the first 15 days after maximum, is a characteristic feature of the type Ia SNe. The fitted template of SN 1992al has a ∆m15(B) = 1.11. We calculate ∆m15(B) = 1.17 ± 0.12 by taking the B band peak magnitude obtained by the template fit and the observed B band magnitude ∼ 15 days after the B maximum. The average decline rate in different bands is also estimated, and listed in Table 4, from our observations using a time baseline of 20 days.  Table 4 and discussed in section 4. For a comparison, we show here the color curves of other type Ia supernovae which were reddening corrected using Cardelli extinction law (Cardelli, Clayton & Mathis 1989)   We also compare the colors of SN 2003hx with those obtained using the intrinsic color curves of the SNe Ia population given by Nobili et al. (2003). Nobili et al. (2003) present the intrinsic color curves of 48 type Ia SNe till 40 days after B band maximum. The total selective extinction was estimated,taking it as a fit parameter,by comparing the observed colors of SN 2003hx with the intrinsic colors given by Nobili et al. (2003). The overall shape of the observed (V − R) and (R − I) color curves are similar to the intrinsic  Table 4.

Comparison of the light curves
In Figures 9, 10 Table 5.

REDDENING ESTIMATE
The estimated reddening in the direction of SN 2003hx due to our own galaxy from Schlegel, Finkbeiner & Davis (1998) is E(B-V) = 0.084 mag. The supernova occurred very close to the nucleus of the galaxy NGC 2076, we therefore expect substantial reddening due to the host galaxy. The optical spectra demonstrates that SN 2003hx is a highly reddened supernova (refer Section 6). Figure 13 shows the strong interstellar NaID lines at the rest wavelength of the host galaxy   Table 4. We also estimate the total extinction following the photometric methods of Phillips et al. (1999) and the Lira's method (1995). However, the best estimate of reddening comes from a good measurement of ∆m15 from where we deduce the intrinsic luminosity. This relationship between ∆m15 and intrinsic luminosity is much more exhaustively tested than the extinction law. We obtain, using this method, our independent measure of Rv towards SN 2003hx. The value of Rv is listed is listed in Table 4. The values of total selective extinction obtained from different methods is listed in Table 4. Wang & Baade (2003) based References: 1- Lira et al. 1998, 2-Contardo, Leibundgut & Vacca 2000, 3-Lira et al. 1998, 4-Richmond et al. 1995, 5-Suntzeff et al. 1999, 6-Strolger et al. 2002, 7-Valentini et al. 2003, 8-Sahu et al. 2006, 9-Anupama et al. 2005, 10-Misra et al. 2006, 11-Present work Figure 13. Spectra of SN 2003hx around NaID absorption line. The two components due to the Milky Way and the host galaxy is clearly seen.  SN 2003du, SN 2002hu, SN 2000E, SN 1999aw, SN 1998bu, SN 1994D, SN 1991T and SN 1990N. All the light curves are shifted to match the time of B maximum and peak magnitude in B band.  2004S, SN 2003du, SN 2002hu, SN 2000E, SN 1999aw, SN 1998bu, SN 1994D, SN 1991T and SN 1990N. All the light curves are shifted to match the time of B maximum and peak magnitude in V band.
on the spectropolarimetric observations inferred an equivalent width of 4.86Åfor the interstellar Na I D line, indicating significant dust extinction and a polarization of ∼ 2 percent. If this polarization is due to dust in the host galaxy, it implies that the dust particles are smaller in size than their galactic counterparts. A similar conclusion about the dust particle size was arrived at by Sahu et al. (1998), in their  SN 2003du, SN 2002hu, SN 2000E, SN 1999aw, SN 1998bu, SN 1994D, SN 1991T and SN 1990N. All the light curves are shifted to match the time of B maximum and peak magnitude in R band.  2004S, SN 2003du, SN 2002hu, SN 2000E, SN 1999aw, SN 1998bu, SN 1994D, SN 1991T and SN 1990N. All the light curves are shifted to match the time of B maximum and peak magnitude in I band.
study of dust property of the host galaxy NGC 2076. Our independent estimates of Rv also indicates the small size of dust particles as compared to their galactic counterparts.
Wand & Baade (2003) based on these observations, found the ratio of total to selective extinction Rv = Av/E(B − V ) to be 2.2. We adopt E(B − V ) = 0.56 ± 0.23 (using the value obtained from the relation between ∆m15 and intrin-sic luminosity) and Rv = 1.97 ± 0.54 to estimate the total extinction in different filters, the values of which are listed in Table 4. We see that the reddening due to our own galaxy is very small in comparison to the total reddening. Thus, a large amount of extinction could arise in the host galaxy of SN 2003hx.

ABSOLUTE LUMINOSITY AND BOLOMETRIC LIGHT CURVE
Assuming H0 = 70 km sec −1 Mpc−1 and the radial velocity of NGC 2076 as vr = 2142 ± 5 km sec −1 , we find a distance modulus of 32.456 mag. The total extinction estimated is mentioned in Table 4. The bolometric light curve of SN 2003hx is estimated using the optical observations presented here. During the early phase most of the flux emerges in the optical from a type Ia supernova (Suntzeff 1996), thus the integrated flux in U BV RI bands gives a good estimate of the bolometric luminosity. The peak bolometric luminosity is directly related to the radioactive Nickel ejected in the explosion. The dereddened magnitudes were converted to flux using calibrations by Fukugita, Shimasaku & Ichikawa (1995). Since we have very few U band observations, we corrected for the missing passband flux in the optical for a contribution of 10 percent as shown by Contardo et al. (2000). Also, to construct the full U V OIR bolometric light curve we should combine the ultraviolet and the near-IR data with the optical data. Suntzeff (1996) constructed the full U V OIR bolometric light curve for SN 1992A by integrating the flux in wavelength range of 2000Å to 2.2 µm. This shows that both the ultraviolet and the near-IR contribution is ∼ 10 percent each of the total U V OIR luminosity for ∼ 80 days since the time of B band maximum. Thus, a correction for the missing flux is required and has to be applied to the bolometric luminosity which is estimated using the optical data alone. We correct for a total contribution of 20 percent from both ultraviolet and near-IR regions. In Figure 14 we show the U V OIR bolometric light curve as dots. The dash line in Figure 14 shows the contribution derived from the BV I bands alone, as obtained from fitted templates from -5 to 80 days with reference to the time of B band maximum. The U V OIR luminosity estimated indicates a peak luminosity of logL = 43.01. The light curves are powered by radioactivity and the amount of 56 Ni mass ejected may be estimated using the peak luminosity (Arnett 1982). Assuming a rise-time of ∼ 18 days for SN 2003hx and the peak U BV RI bolometric luminosity determined, the amount of 56 Ni is estimated to be MNi = 0.50 M⊙. Using the corrected bolometric luminosity, the 56 Ni mass estimate is 0.66 M⊙.
We see that the peak bolometric luminosity and the ejected 56 Ni mass of SN 2003hx is comparable to that of SN 1998bu and SN 1999aw though the amount of 56 Ni ejected in the other two cases is slightly higher than SN 2003hx. Table  5 lists the parameters of different type Ia supernovae including SN 2003hx. We present plots of two simple parameters: absolute magnitude versus ∆m15 ( Figure 15) and log(L bol ) versus 56 Ni ( Figure 16). We see that MB and ∆m15 as well as log(L bol ) and 56 Ni for type Ia supernovae follow a linear relation and the location of SN 2003hx in both these plots agrees well with the rest of the sample.

SPECTRAL EVOLUTION
Optical spectra of SN 2003hx were obtained on four different epochs during +10 days -+53 days past the B maximum. Figures 17 and 18 show the spectral evolution of SN 2003hx. All the spectra show a deep absorption around 6100Å due to SiII, indicating the supernova to be a normal type Ia event.
The NaI D lines are clearly visible and strong, suggesting a high value of reddening, consistent with the estimates in Section 4. A comparison of the spectra of SN 2003hx with those of other normal type Ia events, namely, SN 2003cg (Elias-Rosa et al. 2006), SN 2003du (Anupama et al. 2005, and SN 1996X (Salvo et al. 2001) indicates an overall similarity in the spectra. However, there are a few discrepancies. The +13d spectrum (Figure 19 indicates very weak or no Ca II IR triplet in SN 2003hx, while it is present, but weak compared to the other SNe at later phases ( Figure 20). Further, the S II features at 5300-5600Å appear to be stronger in SN 2003hx.
The expansion velocity is measured for the first three epochs, based on the absorption minimum of Si II 6355Å line, and is found to be ∼ 10, 400 km sec −1 . As our spectroscopic coverage is rather sparse, not much can be inferred   about the velocity evolution in SN 2003hx. However, as seen in Figure 21, the velocity estimated for the first three epochs compares well with the estimates for other type Ia events, SN 2003cg (Elias-Rosa et al. 2006), SN 2003du (Anupama et al. 2005, SN 2002er (Kotak et al. 2005), SN 2002bo (Benetti et al. 2004), SN 1996X (Salvo et al. 2001) and SN 1994D (Patat et al. 1996).

CONCLUSIONS
We present here the U BV RI photometric observations of SN 2003hx over a period of ∼ 146 days after B maximum obtained using the 2.01-m Himalayan Chandra Telescope (HCT). We also present the optical spectra of SN 2003hx at four epochs. We study the light curve evolution in BV RI bands and estimate the peak magnitudes and the time of maximum in different bands using the template fitting method.  Figure 21. Expansion velocity derived from the absorption minima of SiII 6355Å line for SN 2003hx, SN 2003cg, SN 2003du, SN 2002er, SN 2002bo, SN 1996X and SN 1994D (see text for references).