SN 2006X is a bright Type Ia supernova presently visible in the Virgo cluster galaxy M 100 (NGC 4321). Here is a V-band image obtained with the CTIO 1.3-m telescope at 06:50 UT on 21 February 2006 (UT), with SN 2006X and various field stars labeled. North is up and east is to the left.

B- and V-band templates obtained with the Apache Point Observatory 3.5-m telescope on 24 April 2000 (UT), can be obtained by clicking here .

Based on calibration from 5 photometric nights using observations of the Landolt (1992) fields Ru 149 and PG1047, given below are the mean magnitudes and colors for the field stars. The infrared magnitudes of star 1 were measured on 14 nights with respect to the Persson et al. (1998) standards P9144 and LHS2397a. The measured J and K magnitudes are 0.07 of a magnitude brighter than the values from the 2MASS survery, which obtained J = 14.793 (0.025) and K = 13.995 (0.040). 


   *     V              B-V            V-R            V-I              J               K

   1   17.053  0.012   1.180  0.017   0.744  0.012   1.440  0.014    14.721 0.006    13.925  0.009
   2   14.115  0.004   0.649  0.003   0.379  0.006   0.745  0.007        
   3   15.501  0.004   0.880  0.006   0.538  0.004   1.043  0.012        
   4   15.726  0.006   0.626  0.007   0.365  0.004   0.739  0.008       
   5   16.214  0.004   0.953  0.004   0.573  0.005   1.086  0.006         
   6   14.168  0.008   0.644  0.004   0.355  0.005   0.714  0.006
As one can see from the graphs below, the V-R and V-I colors of the field stars are well correlated with the B-V colors. SN 2006X, however, was redder than the reddest of the field stars while at maximum, and 30 days after maximum it will become much redder than the field stars. So knowing one's color terms will be important.

Below we see BVRI light curves of SN 2006X through May 26 UT. The cyan-colored squares and the magenta-colored squares are data obtained using image subtraction of the B- and V-band templates, plus PSF magnitudes from dophot. The rest of the BVRI photometry was obtained from aperture magnitudes using IRAF. We do not show three B-band points obtained when it was faint and when the image subtractions failed (due to moonlight washing out the galaxy light, I think). We have corrected the B- and V-band data for time dilation, applied K-corrections, and S-corrections to Bessell filters. Such a fit to the B-band data gives an observed decline rate parameter is dm15(B) = 1.28 mag. Given the reddened nature of this object, the true decline rate would be more like 1.42 (Phillips et al. 1999, AJ , 118 , 1766, Eqn. 6). We have included a B-band fit based on the B-band template of Goldhaber et al. (2001, ApJ , 585 , 359). We used a stretch factor of 0.90 to scale the time axis of the Goldhaber et al. template.

Note how red this object was, even at the time of maximum light. (The unreddened color at t = 0 would be B-V ~ -0.04.) From the maximum B and V magnitudes, plus the B-V color at 35 days after the time of B-band maximum, we find E(B-V) = 1.336 +/- 0.028. We show for reference the Lira (1995) unreddened line offset by 1.34 mag to the red. The tail of the distribution clearly has a steeper slope than the Lira line. Some SNe just do. SN 1999ac had a slope there which was considerably more shallow than the Lira line.

In the next graph we show the J- and K-band light curves, calibrated using the derived magnitudes of the field star 7 arcsec north of the SN. The data are K-corrected, S-corrected to the system of Persson et al. (1998) and corrected for time dilation. The black lines are based on the J- and K-band templates of Krisciunas et al. (2004, AJ , 127 , 1664, Table 12 and Fig. 9), multiplied by a stretch factor of 0.90. Note the marginal evidence for a 3rd maximum (or shoulder) at ~70 days after T(Bmax). Kasen (2006, ApJ, in press) has come up with an explanation for the secondary maximum and has predicted that some Type Ia SNe might show a third maximum at ~80 days.

If we fit polynomials to the data we find the following maximum magnitudes and times of maximum. They are probably good to +/- 0.5 to 1.0 days. 


   Filter      max mag     JD of max           dt
                        [2,450,000 + ...]   of maximum
                                            wrt Bmax

     B          15.36       3786.0
     V          14.07       3787.9           +1.9 d
     R          13.52       3787.9           +1.9 d
     I          13.30      [3783.2]         [-2.2 d]    
     J          12.88       3782.1           -3.9 d
     K          12.82       3782.8           -3.2 d
We have used our unreddened V minus near-infrared loci for mid-range decliners to derive approximate color excesses (Krisciunas et al. 2000. ApJ , 539 , 658). Actually, the loci given in our original paper are first or second order. Below we show the fourth order fits to our compactified data, offset by the indicated color excesses. In any case, SN 2006X is clearly a highly reddened object, and it seems to be reddened by unusual dust, with R_V ~ 1.5 (somewhat like SN 1999cl), and A_V(host) = 1.99 +/- 0.05.

If we adopt the observed maximum infrared magnitudes given above, total J-band extinction of 0.33 mag and total K-band extinction of 0.14 mag, we get infrared absolute magnitudes at maximum of M_J = -18.49 and M_K = -18.36, with uncertainties of about 0.18 mag. In the graph below the triangles represent SNe whose distances are known via Cepheids or the method of Surface Brightness Fluctuations (SBF distances). SN 2006X is shown as orange triangles. Altogether we now have 27 J-band absolute magnitudes at or near maximum, 26 H-band points, and 25 K-band points.

Excepting SNe 1991bg and 1999by, the mean absolute magnitudes are as follows: M_J (max) = -18.61 +/- 0.030, M_H (max) = -18.30 +/- 0.035, and M_K (max) = -18.44 +/- 0.030.

Perhaps the most robust way to determine the host galaxy extinction and the value of R_V is to take the observed maximum apparent magnitudes and first correct them for Milky Way extinction. E(B-V)_Gal = 0.026 mag from Schlegel et al. (1998). For R_V (Gal) = 3.1, we obtain Galactic extinction corrections ranging from 0.106 mag for B to 0.009 mag for K. Next, we can extimate the BVRI absolute magnitudes at maximum using the linear relations in Table 4 of Prieto, Rest, and Suntzeff (2006, astro-ph/0603407), obtaining M_B -19.12, M_V = -19.055, M_R = -19.07, M_I = -18.81. We used a decline rate of Delta m_15(B) = 1.42. From the data shown in the 3 panel graph above we estimate that in the near-infrared M_J = -18.63 and M_K = -18.40 at maximum light.

Next we take the differences of the apparent magnitudes (corrected for Galactic extinction) minus the estimated absolute magnitudes. In the graph below one can see that as the wavelength tends to infinity (inverse wavelength tends to zero), the effect of host galaxy extinction diminishes. In the limit we obtain the distance modulus of SN 2006X.

We can do a multi-dimensional chi-2 minimization. In the contour plot below the X-axis represents A_V (host) and ranges from 1.60 to 2.28. The Y-axis represents R_V, ranging from 1.1 to 2.6. Shown are the 1, 2, and 3 sigma contours.

We obtain a host-galaxy extinction A_V(host) = 1.976 +/- 0.21 mag. The total V-band extinction toward SN 2006X is 2.06 mag. We obtain R_V = 1.56 (+0.32, -0.28) and a distance modulus of 31.134 mag, corresponding to a distance of 16.9 Mpc. This is to be compared to the Cepheid-based distance modulus of m-M = 31.04 +/- 0.17 (Ferarrese et al. 1997, ApJ , 475 , 853), or a distance of 16.1 +/- 1.3 Mpc.

I think perhaps the reason this BVRIJK solution gives such large uncertainties for A_V and R_V is that we don't really know the absolute magnitudes just from the decline rate. In the case of SNe 2001el and 2004S, which were essentially clones of each other, we made templates from SN 2001el and adjusted them to the light curves of SN 2004S. Since SN 2001el had roughly an order of magnitude more dust extinction than SN 2001el, the difference of dust extinction really gave us a measure of the host galaxy extinction of SN 2001el. In the case of SN 2006X our relative uncertainty of A_V was 11 percent. For SN 2001el/2004S the relative error on the difference of the respective A_V's was 5 percent.

Kevin Krisciunas

Univ. of Notre Dame

[and Cerro Tololo Obs.]

Go back to Kevin Krisciunas home page by clicking here.