Peremennye Zvezdy (Variable Stars) 34, No. 4, 2014 Received 1 December; accepted 10 December.

Article in PDF

Explosion of an LBV star in the galaxy UGC 8246

E.A. Barsukova¹, V.P. Goranskij², A.F. Valeev^1,3, and S.S. Kaisin¹

Special Astrophysical Observatory, Russian Academy of Science, Nizhnij Arkhyz, Karachai-Cherkesia, 369167, Russia; e-mail: bars@sao.ru

Sternberg Astronomical Institute, Lomonosov Moscow University, Universitetski pr., 13, Moscow, 199992, Russia

Kazan Federal University, Kremlevskaya 18, Kazan, 420008, Russia

1. INTRODUCTION

Discoveries of a large number of supernovae using networks of automatic robotic telescopes and preliminary classification of their spectra fill the maximum-luminosity gap between classical novae and supernovae. The optical transients reaching in their outbursts the absolute visual magnitude from 8 to 17 and evolving their spectra to cooler subtypes during outbursts are called Intermediate-Luminosity Red Transients (ILRTs). This family of astrophysical objects is heterogeneous and contains luminous blue variable (LBV) stars with giant eruptions ( Car, low-luminosity SNe 1961V and 1954J with spectra resembling SNe IIn in outbursts), calcium transients like SN 2008S in NGC 6496 or NGC 300 OT having super-AGB stars as progenitors, and luminous red novae like V1006/7 in M31, V838 Mon in the Galaxy, and M85 OT2006-1. Van Dyk et al. (2000) called lower-luminosity SNe IIn "supernova impostors", but later this term was extended to other transients with peak luminosities in the gap (Berger et al. 2009; Smith et al. 2009).

Some LBVs and Galactic red novae turned out to be binaries or multiple systems. The most promising scenario for luminous red novae as merging events in binary systems was suggested by Tylenda & Soker (2006). Tylenda et al. (2011) observed a merging event in the contact binary before the outburst of V1309 Sco as a red nova. On the other hand, Martini et al. (1999) suggested that a red nova phenomenon is due to a nuclear event in a single star, in which a slow shock drives the photosphere outward. Barsukova et al. (2014) specify that this is the phenomenon connected with the adiabatic expansion of the star envelope after an episodic energy release in the star center. The release may be due to merging of stellar nuclei inside a massive common envelope or to instability of a single stellar nucleus of a young star.  Car, the LBV which experienced a Great Eruption in the mid-1800s accompanied by an episodic mass-loss event, is known as a binary (Corcoran & Ishibashi 2012). The companion is not seen in the spectrum, but it is found to be an O4-O6 giant having mass of 40-50  based on indirect evidence (Mehner et al., 2010). RXTE X-ray observations show cyclic variability with the 2024-day period that may be the signature of the star's motion in an elliptic orbit with and a semi-major axis  AU (Ishibashi et al. 1999). However, the role of this companion in the evolution of  Car and in the Great Eruption scenario is still unknown. Kashi & Soker (2010) and Kashi (2010) suggest that most outbursts of LBV-type systems are powered by gravitational energy released from either a vigorous mass transfer process from the evolved primary star to a main sequence secondary star or a merger of two stars. So, major LBV eruptions can be triggered by stellar companions, and in extreme cases, a short-duration event with a huge mass transfer rate can lead to a bright transient event on time scales of weeks to months. Alternative hypotheses explain LBV eruptions either with enhanced stellar wind episodes when a massive star luminosity sometimes exceeds the Eddington limit, or with interior explosions due to instabilities in stellar nuclei or envelopes that may be pulsational instabilities (Owocki & Shaviv 2012).

Note that one of LBVs in the galaxy NGC 7259, SN 2009ip, previously went through two typical impostor outbursts that reached in maxima and then became a true core-collapse supernova of -18th absolute magnitude in the peak of brightness (Mauerhan et al. 2013; Margutti et al. 2014).

A recent discovery of an LBV-type transient as a possible supernova in UGC 8246 on 2013 December 20.93 UT was reported by Wang and Gao (2013). The object had the  magnitude of approximately 17.75. UGC 8246 is an SB(s)cd type galaxy, its redshift is and distance,  Mpc; the Galactic interstellar absorption is (NED). With these parameters, the absolute magnitude of the transient at the time of discovery was . Additional photometry acquired in 2013 December was published by Elenin, Wang & Gao, Luppi & Buzzi (CBAT) and by Brimacombe (2014). -band observations by Wang & Gao, continued during 10 days, show no essential light decay, which strongly suggests that their observations cover the brightness maximum. However, this suggestion may be incorrect in the case of plateau shape of the light curve. Tartaglia et al. (2014a) carried out spectroscopic observations and classified the object as a supernova impostor similar to the prototypical SN 1997bs. SN 1997bs was classified by Van Dyk et al. (2000) as a superoutburst of a very massive luminous blue variable star, analogous to  Car. In the spectrum of the UGC 8246 transient taken on January 8, 2014, Tartaglia et al. (2014a) found H emission with an unresolved narrow component superimposed on broader wings (FWHM of about 1800 km s). Taking into account scientific interest to SN impostors, we continued spectroscopic and photometric observations of this object.

2. OBSERVATIONS AND DATA REDUCTION

We obtained medium-resolution optical spectra of the transient PSN J13100734+3410514 in UGC 8246 with the SCORPIO focal reducer (Afanasiev & Moiseev 2005) mounted on the 6-m BTA telescope of the Special Astrophysical Observatory (SAO) on 2014 February 8 and June 7. Seeing was measured as on February 8 and on June 7. Additional photometry was performed using the SAO 1-m Zeiss telescope and a CCD photometer with an EEV42-40 chip on 2014 April 3, November 14 and 22. November observations were carried out with the and filters at good weather conditions with seeing, and the total exposure times were 1200 s or longer.

In the photometric mode with the SCORPIO, we obtained CCD frames with standard , and Cousins filters. The color image generated using these frames is shown in Fig. 1. It demonstrates that the transient occurred in a spiral arm rich in star-forming regions.

To perform relative photometry, we used the comparison star marked "std" in Fig. 1, its SDSS magnitudes being . We transferred SDSS magnitudes to magnitudes by interpolation using and AB magnitudes of Lyr as described in our previous paper (Barsukova et al. 2012). As a result, the magnitudes of the standard star were derived as (18.95, 17.97, 17.25, 16.49). All currently available photometry is presented in Table 1. The light curves plotted using our data and all published observations in the and filters are shown in Fig. 2.

On February 8, 2014, the spectroscopic camera functioned in the long-slit mode. First it was equipped with the VPHG 1200G grism (nominal spectral range  Å, resolution 5 Å, dispersion 0.88 Å pixel), and later it was replaced with the VPHG 550G grism (spectral range  Å, resolution 10 Å, dispersion 2.1 Å pixel). Actual resolution measured using emission lines of night sky was 5.4 Å for the VPHG 1200G grism and 14.6 Å for VPHG 550G. In the first spectrum, we found an essential contribution of a nearby galactic H II region in the brightest lines of the stellar spectrum, so we took spectra with different position angles of the slit. Slit locations at different telescope pointings are shown as straight lines in Fig. 1 and they are listed in Table 2 for individual spectra. On June 7, 2014, we used only the VPHG 550G grism. Spectra were reduced using standard ESO MIDAS procedures for the long-slit mode. Basic parameters of our spectra are collected in Table 2. The blue and green spectral regions are displayed in Fig. 3, and the whole optical spectra are presented in Fig. 4. The wavelengths in these figures have been corrected with respect to bright emission components of Balmer lines.

3. RESULTS

Analyzing published early non-homogeneous observations presented in Table 1, one can suggest that the star had a color index about zero at the outburst maximum. We do not know the interstellar reddening in the galaxy UGC 8246 and therefore can determine only the upper limit to the bolometric magnitude ( ) and the lower limit to the effective temperature ( ). Our February photometry taken 50 days after the maximum gave the following magnitudes and color indices: , , . Reddening continued in June; on the 169th day after the light maximum, was 086. In November, on the 336th day at the level of , the star became even redder, with . The star had a primary decay of in 50 days and a nonlinear secondary decay both in the and bands. By the 169th day after maximum, the speed of the decay decreased to about 07  in 100 days.

The spectrum of the UGC 8246 transient taken in February was rich in emission lines. It resembled spectra of Fe II-class novae in the fireball stage (e.g., see Williams 2012; we used the finding list of optical emission lines in his Table 2 for identification). The lines we have identified are listed in Table 3. The strongest emission lines in our spectra are Balmer lines, [O II], and [O III]. Oxygen lines are evidently radiated in the nearby H II regions in the spiral arm of the galaxy. As seen in Fig. 5, the regions of [O III] emission are superimposed on the stellar spectrum only partly. Otherwise, this figure shows that the main component of H emission mostly has a stellar origin. But oxygen lines are located at the same radial velocity, +920 km s, as the strongest H component. Direct images taken in the filter in November confirm the association of the transient with a compact H II region, the brightest part of which is located in 1 0 east and 0 5 south of the transient. These spectra and the direct images definitely provide evidence that the explosion of this transient has happened in the active star forming region rich in young massive stars.

The equivalent widths of Balmer lines on February 8 were the following:  Å;  Å;  Å;  Å. H, in a blend with Ca II H, form a deep absorption line. On June 7, the equivalent widths changed to -350 Å (H), -41 Å (H), and -7 Å (H).

With the best resolution of 5.4 Å, the H line profile looks complex (Fig. 6). The strongest narrow component has  km s after correction for the spectral resolution. The profile extends over  km s. Additionally, there is another weak blue narrow emission component in the profile, displaced by  km s with respect to the main component and having approximately the same half width, 320 km s. A dip between these components is at the velocity of -380 km s and looks like an absorption. Such profiles with an absorption component are rare in spectra of SN impostors but do exist. With wings, an absorption dip and a narrow component, the H line profile of UGC 8246 transient resembles Balmer line profiles of the type IIP supernova 1994W in NGC 4041 (Sollerman et al. 1998), but that was a true and extremely luminous supernova. A similar shape of Balmer line profiles was observed by Humphreys et al. (2012) in the spectrum of the peculiar type IIn SN or impostor 2011ht in UGC 5460. That star had reached M in maximum. However, the wing span in the profiles of these two supernovae was 4 times larger than that of the UGC 8246 transient. Also, wings of the H line profile are twice narrower than wings of the same profile of the "calcium transient" in NGC 5775 observed by us (Barsukova et al. 2012). There was no absorption component in the profile of that transient. The narrow component in H profile of the UGC 8246 transient was poorly resolved with the VPHG 550G grism. The line displayed a blue shoulder that confirmed the presence of a similar blue narrow component at the velocity about -660 km s. Explaining profiles of similar shape in Balmer lines of SN 1994W, Chugai et al. (2004) identified their components with a photoionized expanding circumstellar envelope, shocked cool gas in the forward post-shocked region, and multiple Thompson scattering in the envelope. Probably cool gas is responsible for the absorption component in the profiles of Balmer lines. The models gave a rather deep absorption component in a case of homologous expansion of the ejecta without post-explosion radiative acceleration but did not take into consideration light echo in the circumstellar medium.

One can expect that the light echo of an LBV explosion expanding through the circumstellar envelope formed by stellar wind of the progenitor should influence the Balmer line profiles. An external observer can see only those illuminated dust particles or emissive excited atoms that are located on the surface of the ellipsoid with the flare source in one focus, and the observer in the other one. In this case, the total light path from the source to a particle and of the reflected light from the particle to the observer is the same for all visible particles at the time after the flare, namely . During the first weeks after the outburst, the ellipsoid is narrow and elongated along the line of sight. Then the photoionization front in the expanding gaseous envelope visible for the observer in the first weeks includes only the part of the envelope that is located along the line of sight and moving to the observer. The light from the excited gas at the opposite side of the envelope and from other parts of it has not yet reached the observer. Thus, in the first weeks, we expect to have more negative velocity of the main emission component relative to its velocity during the last stages. To test the assumption on the light echo, we measured velocities of peaks in the H and H profiles in the spectrum by Tartaglia et al. (2014a) taken on January 8, 2014 and found 820 and 850 km s, respectively, against 970 and 920 km s in our February 8 spectra. So the velocity difference is really present, and is equal to  km s. Later, the visible front of ionization will get to the opposite side of the wind envelope, with velocities directed away from the observer, and the central-peak velocity will increase. Our spectrum taken on June 7 gives larger velocities, 1060 km s for H and 960 km s for H (these values were corrected with respect to galactic [O III] lines). Note that the presence of a negative excess in the velocity of the main emission component of on January 8, 2014 confirms the hypothesis that the transient was discovered at an early stage of the outburst and testifies against a plateau stage. Thus, the effects of light echo in Balmer-line profiles in the spectra of SN impostors are strong and should be taken into account in the spectral line modeling.

We have tried to separate components of the H line profile (Fig. 6) using the MIDAS procedure DEBLEND/LINE. These components were fitted with Gaussian functions. The fit with a broad component along with two narrow ones represents the total profile badly, especially to the red side. The observed curve runs evidently above the model one. An attempt to fit the profile including one absorption component represents the data better but results in a very broad main component and a broad absorption which eats up half of main emission width. Such a model is hard to explain physically. But the fit without any broad component and with three narrow ones is the best. In short, the red excess in the profile can be successfully fitted with an additional Gaussian. Parameters of the components in this model are the following: for the blue one,  km s,  km s,  Å; for the central one,  km s (adopted),  km s,  Å; for the red one,  km s,  km s,  Å. Note that the FWHM values in this model include the instrumental profile. The fitting and its three Gaussian components are shown in Fig. 6.

We have an alternative explanation of the H line profile based on this fitting. A separate narrow blue component may be the radiation of a frontal lobe of bipolar ejecta thrown out in the direction of the observer, while the opposite receding lobe is partly covered by the volume of the approaching lobe. The non-covered radiation of the receding lobe is visible in the red part of the profile. Note that a contribution from the H II region may be partly added to the red component (see Fig. 5). Such a bipolar eruption formed the "Homunculus Nebula" of  Car in the mid-1800s explosion. The lobes in the Homunculus are known to be expanding radially with velocities in the range of 650-700 km s (Davidson & Humphreys 1997) which is in agreement with 670 km s for the UGC 8246 transient. However, the UGC 8246 transient outburst was not so long in duration, only 200 days vs. 20 years for Car. In this case, the dip between two narrow components in the profile of H cannot be treated as an absorption.

The rich weak-line spectrum indicates a dense environment containing evolved material ejected from the progenitor by the stellar wind in the super-Eddington regime. There are many Fe II emission lines in the spectrum, the strongest ones have P Cyg profiles. The absorption component of the strongest Fe II 5169 Å line is located at the velocity of  km s. Forbidden emissions are absent in the stellar spectrum, but they are present in the spectrum of the galaxy environment. The stellar spectrum indicates low excitation and absence of a hot ionizing source. He I 5876 Å is very strong, it has a P Cyg profile; its absorption component is centered at the velocity of -1500 km s and expands to -1850 km s with respect to the center of the emission component. The center of the emission component is displaced by -400 km s against main components of the Balmer lines. The profile looks strange and unlike any other lines, and there is an assumption that it is being formed entirely in the ejecta, not in the wind. He I 4471 Å is also strong, so we have to identify emission lines at 4922 and 5016 Å as He I lines, though they can contain some contribution from close Fe II components. The Na I DD blend has a P Cyg profile and is as strong as He I 5876 Å. The absorption component of Na I is displaced to the blue side from the emission component's center by 700 km s.

The Ca II H & K lines are in absorption. In the spectra, we also identify the Mg I 4571 Å intercombination line that may be formed in massive and rarefied nebulae, and sometimes occurs in ejecta of red novae (Goranskij & Barsukova 2007). Recently Tartaglia et al. (2014b) found absorption lines of Ba II and Sc II in the spectra of the SN impostor 2007sv that exploded in the galaxy UGC 5929; these species are not met in the spectra of classical novae. The presence of Sc II, Mg II, and Si II was noted in the spectra of SN 1994W (Chugai et al. 2004). We tried to find these lines in our spectra of the UGC 8246 transient, and the attempt seems to be successful only for Sc II. There is an obvious depression at 5640-5684 Å that cannot be explained by absorption of other chemical elements except Sc II. The line of Sc II 5526 Å can contribute to the absorption component of the Fe II 5535 Å P Cyg-type profile. We also identified a strong absorption feature located towards short wavelengths from H at 4294-4315 Å, like the G band in early G-type stars. This molecular CH band was found in the spectrum of the  Car-type transient SN 2011ht by Humphreys et al. (2012).

In the low-resolution noisy spectrum taken on June 7 (on the 169th day after maximum), when the object's brightness was at the level of 208 in the band, only the strongest lines of the transient and the galaxy are visible. The Balmer lines definitely belong to the transient: the star's continuum can be distinguished in the spectrum, and the stellar image is visible in the direct frames. Figure 7 compares the H line profiles in the February 8 and June 7 spectra. The narrow H line became noticeably broader on June 7 than on February 8. In spite of low resolution, the red and blue components stay visible in the profile as shoulders. Note that these components may be the radiation of bipolar ejecta. The half width of the profile increased from 300 to 700 km s. These estimates were corrected for the spectral resolution, and their accuracy is about 50 km s. Both profiles are plotted in intensity units relative to the stellar continuum. The star faded between these dates by 10 in the band. The contribution of continuum in the band changed between these dates from 82 to 78 per cent. As the H flux declined slower than the continuum, the equivalent width of the H line increased its absolute value from 180 to 350 Å. We explain the broadening of the H line by two reasons. First, it is the radiative acceleration of the stellar wind. The second reason is the expansion of visible volume of the photoionized region of stellar wind due to expanding light echo.

The observations of LBV-type transient remnants with the Hubble space telescope or Spitzer infrared space telescope show that the LBV stars survive in their explosions. Some of them undergo repeated explosions. Thus, future observations of the UGC 8246 transient are encouraged.

4. CONCLUSIONS

We present the results of our spectroscopy and CCD photometry of the intermediate-luminosity red transient PSN J13100734+3410514 in UGC 8246 performed in February and June, 2014 with the Russian 6-m BTA telescope and SCORPIO spectral camera and follow-up CCD photometry with the SAO 1-m telescope.

The transient showed a number of characteristics typical of LBV stars with giant eruptions. Namely, the amplitude of the outburst exceeded three magnitudes; the bolometric magnitude at maximum was evidently above the Humphreys-Davidson limit (the upper boundary of the supergiant luminosities in the diagram; see Humphreys & Davidson 1979); a significant reddening of the spectrum occurred during the outburst decline; the spectrum is typical of LBV stars; moderate wind velocities are characteristic of LBV stars; a possible bipolar ejecta signature was seen in the Balmer line profiles. There is an evidence for a light echo of the outburst expanding in the surrounding gaseous medium after the outburst. We assume, like other researchers, that this transient was an explosion of a high-mass LBV star.

Acknowledgments.

This work was supported by the Russian Foundation for Basic Research (RFBR) with the grants 13-02-00885 and 14-02-00759. E.A.B. thanks the President of the Russian Federation for the grant supporting the Leading scientific school NSh-4308.2012.2. A.F.V. thanks the President for grants for young PhD researchers MK-6686.2013.2 and MK-1699.2014.2 and the RFBR for the grant 12-02-31548. We thank Drs. D.Yu  Tsvetkov and O.V. Maryeva for discussion. Observations with the Russian 6-m telescope are financially supported by the Ministry of Education and Science of Russian Federation (agreement No. 14.619.21.004, project identifier RFMEFI61914X0004). This research has made use of the Sloan Digital Sky Survey and of the NASA/IPAC Extragalactic Database (NED).

References:

Afanasiev, V.L., Moiseev, A.V., 2005, Astronomy Letters, 31, 194

Brimacombe, J., 2014,
http://spacefinalfrontier.blogspot.ru/2014/02/luminous-blue-variable-psn.html

Barsukova, E.A., Fabrika, S.N., Goranskij, V.P., Valeev, A.F., 2012, Variable Stars, 32, No. 2

Barsukova, E.A., Goranskij, V.P., Valeev, A.F., Zharova, A.V., 2014, Astrophys. Bull., 69, 67

Berger, E., Soderberg, A.M., Chevalier, R.A. et al., 2009, Astrophys. J., 699, 1850

Chugai, N.N., Blinnikov, S.I., Cumming, R.J., et al., 2004, MNRAS, 352, 1213

Corcoran, M.F., Ishibashi, K., 2012, in Eta Carinae and the Supernova Impostors, R. Davidson & R.M. Humphreys (eds.), Springer, New York, p.195

Davidson, K., Humphreys, R.M., 1997, Ann. Rev. Astron. & Astrophys., 35, 1

Goranskij, V.P., Barsukova, E.A., 2007, Astron. Reports, 51, 126

Humphreys, R.M., Davidson, K., 1979, Astrophys. J., 232, 409

Humphreys, R.M., Davidson, K., Jones, T.J., et al., 2012, Astrophys. J., 760, 93

Ishibashi, K., Corcoran, M.F., Davidson, K., et al., 1999, Astrophys. J., 524, 983

Kashi, A., Soker, N., 2010, arXiv:1011.1222

Kashi, A., 2010, AIP Conference Proc., 1314, 55

Margutti, R., Milisavljevic, D., Soderberg, A.M. et al., 2014, Astrophys. J., 780, 21

Martini, P., Wagner, R.M., Tomaney, A. et al., 1999, Astron. J., 118, 1034

Mauerhan, J.C., Smith, N., Filippenko, A.V. et al., 2013, MNRAS, 430, 1801

Mehner, A., Davidson, K., Ferland, G.J. et al., 2010, Astrophys. J., 710, 729

Owocki, S.P., Shaviv, N.J. 2012, in Eta Carinae and the Supernova Impostors, K. Davidson & R.M. Humphreys (eds.), Springer, New York, p.275

Smith, N., Ganeshalingam, M., Chornock, R. et al., 2009, Astrophys. J., 697, L49

Sollerman, J., Cumming, R.J., Lundquist, P., 1998, Astrophys. J., 493, 933

Tartaglia, L., Pastorello, A., Benetti, S. et al., 2014a, Astronomer's Telegram, No. 5737

Tartaglia, L., Pastorello, A., Taubenberger, S. et al., 2014b, arXiv:1406.2120

Tylenda, R., Soker, N. 2006, Astron. & Astrophys., 451, 223

Tylenda, R., Hajduk, M., Kaminski, T. et al., 2011, Astron. & Astrophys., 528, 114

Van Dyk, S.D., Schuyler, D., Peng, C.Y. et al, 2000, Publ. Astron. Soc. Pacific, 112, 1532

Wang, B., Gao, X., 2013, CBAT
(http://www.cbat.eps.harvard.edu/unconf/followups/J13100734+3410514.html)

Williams, R., 2012, Astron. J., 144, 98

---