Exploring the Energy Sources Powering the Light Curve of the Type Ibn Supernova PS15dpn and the Mass-Loss History of the SN Progenitor
aa r X i v : . [ a s t r o - ph . H E ] M a y Draft version May 31, 2019
Preprint typeset using L A TEX style AASTeX6 v. 1.0
EXPLORING THE ENERGY SOURCES POWERING THE LIGHT CURVE OF THE TYPE IBN SUPERNOVAPS15DPN AND THE MASS-LOSS HISTORY OF THE SN PROGENITOR
Shan-Qin Wang , and Long Li Guangxi Key Laboratory for Relativistic Astrophysics, School of Physical Science and Technology, Guangxi University, Nanning530004, China; [email protected]
ABSTRACTPS15dpn is a luminous rapidly rising Type Ibn supernova (SN) discovered by Pan-STARRS1 (PS1).Previous study shown that the bolometric light curve (LC) cannot be explained by the Ni cascadedecay model. In this paper, we employ some alternative models, i.e., the magnetar model, and the Ni plus magnetar model, the ejecta-circumstellar medium (CSM) model (the CSI model), and theCSI plus Ni model, to fit the bolometric LC of PS15dpn. We found that all these models canreproduce the theoretical LCs matching the data. By analyzing the parameters of these models andcombining with the spectral features, we conclude that the Ni plus CSI model involving an eruptiveshell expelled by the SN progenitor are the most promising one in explaining the LC of PS15dpn. Inthis scenario, the masses of the ejecta, the CSM, and the Ni are 13 . +4 . − . M ⊙ , 0 . +0 . − . M ⊙ , and0 . +0 . − . M ⊙ , respectively. These parameters are consistent with the conjecture that the progenitorsof SNe Ibn are massive Wolf-Rayet stars. Furthermore, our results indicate that the progenitor ofPS15dpn might expelled a gas shell ∼ Keywords: stars: magnetars – circumstellar matter – supernovae: general – supernovae: individual(PS15dpn) INTRODUCTIONThe interactions of supernova (SN) ejecta and the circumstellar medium (CSM) from the pre-SN outbursts wouldprompt narrow and intermediate-width emission lines and enhance the luminosities of SNe. The SNe showing interac-tion evidence are therefore called “interacting SNe” (Smith 2017). According to the types of the emission lines and thephysical natures, interacting SNe can be divided into at least three classes: type IIn (Schlegel 1990; Filippenko 1997)SNe emitting H α emission lines, type Ibn SNe (Pastorello et al. 2016; Hosseinzadeh et al. 2017) with spectra showingHe emission lines, and type Ia-CSM (Silverman et al. 2013) whose spectra also show H α emission lines.To date, at least 31 SNe Ibn have been discovered and then confirmed (Hosseinzadeh et al. 2019). The shapes of lightcurves (LCs) of SNe Ibn are rather homogeneous. Some SNe Ibn (e.g., SN 2011hw Smith et al. 2012; Pastorello et al.2015a, OGLE-2012-SN-006 Pastorello et al. 2015b and OGLE-2014-SN-131 Karamehmetoglu et al. 2017) have veryslow-declined LCs and most SNe Ibn have very fast-declined LCs.While almost all SNe Ibn were discovered in star-forming regions of the host galaxies and their progenitors havebeen thought to be very massive Wolf-Rayet stars (Pastorello et al. 2016), a SN Ibn (PS1-12sk) was found on a brightelliptical galaxy where the star forming is inactive and its progenitor is elusive (Hosseinzadeh et al. 2019).In this paper, we study a luminous rapidly rising SN PS15dpn discovered by Pan-STARRS1 (PS1) in the area ofthe gravitational-wave GW151226 (Abbott et al. 2016) which was produced by a merger of a black hole-black hole(BH-BH) binary in a luminosity distance ( D L ) of 440 +190 − Mpc. He i z ) of PS15dpn is0 . ± . t peak − t explosion .
10 days) SN and Smartt (2016) demonstratedthat it cannot be explained by the Ni model (Arnett 1982); (2) It is widely believed that the interactions betweenthe ejecta of interacting SNe (including SNe IIn and Ibn) and their CSM would power their unusual LCs; studying theproperties of the ejecta and CSM would provide important information, e.g., the masses of the ejecta and the CSM,the mass-loss rate and mass-loss history of the SN progenitors, and so on (see Smith 2014 for a review and referencestherein).The aim of this paper is exploring the possible energy sources of PS15dpn, the properties of the CSM surroundingthis SN, as well as the mass-loss history of its progenitor. In Section 2, we use five models to fit the bolometric LCof PS15dpn and derive the best-fitting parameters. Our discussion and conclusions can be found in Sections 3 and 4,respectively. MODELING THE BOLOMETRIC LC OF PS15DPNIn this Section, we employ several models to fit the bolometric LC of PS15dpn and adopt the Markov Chain MonteCarlo (MCMC) method to obtain the best-fit parameters. Throughout this paper, the optical opacity κ is fixed to be0.1 cm g − (e.g.,Wheeler et al. 2014). 2.1. The Ni Model
As mentioned above, Smartt (2016) has shown that the Ni model cannot fit the bolometric LC of PS15dpn. Here,we employ the Ni model to model the LC to test the result obtained by Smartt (2016). The parameters of the Ni model employed here are the ejecta mass M ej , the initial scale velocity of the ejecta v sc0 , the Ni mass M Ni , thegamma-ray opacity of Ni decay photons κ γ, Ni , and the moment of explosion t expl .The theoretical Ni-powered LC is shown in Figure 1 and the corresponding parameters are listed in Table 1. It canbe found that the theoretical LC can match the observations. However, the Ni mass (1 . +0 . − . M ⊙ ) is larger thanthe ejecta mass (0 . +0 . − . M ⊙ ), indicating that the Ni model is disfavored and alternative models must be employed.2.2.
The Magnetar Model and the Magnetar Plus Ni Model
Then we use the magnetar model developed by Inserra et al. (2013), Wang et al. (2015a), and Wang et al. (2016)as well as the magnetar plus Ni model (Wang et al. 2015b, 2016) to fit the LC. The parameters of these two modelsare listed in Table 1 and the B p and P are the magnetic strength and initial rotational period of the magnetar,respectively. The theoretical LCs reproduced by these two models are also shown in Figure 1 and their parametersare also listed in Table 1.It is shown that the LC can be powered by a magnetar with B p = 14 . +0 . − . × G and P = 13 . +0 . − . mswithout Ni contribution or a magnetar with B p = 25 . +2 . − . × G and P = 16 . +6 . − . ms aided by 0 . +0 . − . M ⊙ of Ni. Nomoto et al. (2013) shown that an energetic SN explosion can synthesize . . ⊙ of Ni, so the inferred Ni mass is well consistent with this upper limit. The CSI Model and the CSI Plus Ni Model
It can be expected that the SN ejecta with dense He-rich CSM (winds or shells) surrounding the progenitorswould provide at least a fraction of energy to power the LCs of SNe Ibn. Therefore, scenarios taking the ejecta-CSM interaction into account might be promising models to account for the LC of PS15dpn. Therefore, weuse the ejecta-circumstellar medium (CSM) model (the CSI model) (Chevalier 1982; Chevalier & Fransson 1994;Chugai & Danziger 1994; Chatzopoulos et al. 2012; Ginzburg & Balberg 2012; Liu et al. 2018) and the CSI plus Nimodel (Chatzopoulos et al. 2012) to fit the bolometric LC of PS15dpn.The ejecta can be divided into two parts, the inner parts that can be described by ρ ej ∝ r − δ and the outer partthat can be described by ρ ej ∝ r − n . We assume that the density profile of CSM can be descried by a power-law ρ CSM ∝ r − s , s = 2 corresponding to winds and s = 0 corresponding to CSM shells.Letting δ = 1 and n = 10, the parameters of the CSI model we adopt are the energy of the SN ( E SN ), the ejecta mass( M ej ), the CSM mass ( M CSM ), the density of the innermost part of the CSM ( ρ CSM , in ), the radius of the innermost partof the CSM ( R CSM , in ), the efficiency of conversion from the kinetic energy to radiation ( ǫ ), the dimensionless positionparameter ( x ), and the explosion date with respect to the date of the peak ( t expl ). Two additional parameters ( M Ni and κ γ, Ni ) are added to constitute the CSI plus Ni model. However, systematic study (e.g., Lyman et al. 2016) for SNe Ibc shown that the LCs of a fraction of SNe Ibc that are not very luminousmust be powered by 0 . − . ⊙ of Ni (see, e.g., Table 5 of Lyman et al. 2016). The large inferred Ni masses indicate that core-collapseSNe can synthesize more than 0.2 M ⊙ of Ni or some ordinary-luminosity SNe Ibc might be powered by hybrid energy sources (for instance, Ni plus magnetars). Both these two possibilities cannot be excluded. In other words, the upper limit of the Ni yield of core-collapseSNe can be larger than 0.2 M ⊙ . Letting x = r ( t ) /R ( t ), the regions x < x and x > x are the inner part ( ρ ej ∝ r − δ ) and outer part ( ρ ej ∝ r − n ), respectively. The LCs reproduced by the CSI model and the CSI plus Ni model can be found in Fig. 2 and the best-fitparameters can be found in Table 2. The LCs powered by the forward shocks, reverse shocks, as well as Ni decayare also plotted using different lines.In the shell ( s = 0) CSI model, M ej = 17 . +1 . − . M ⊙ , M CSM = 4 . +0 . − . M ⊙ . In the wind ( s = 2) CSI model, M ej = 16 . +2 . − . M ⊙ , M CSM = 0 . +0 . − . M ⊙ . According to the equation ˙ M = 4 πv w q ( q = ρ CSM , in R , in ) andassuming that the velocity of the wind ( v w ) is 100–1000 km s − , we can obtain the mass-loss rate of the wind ( ˙ M ):1.116–11.16 M ⊙ yr − .In the shell ( s = 0) CSI plus Ni model, the values of the ejecta mass, the CSM shell mass, and the Ni mass are13 . +4 . − . M ⊙ , 0 . +0 . − . M ⊙ , and 0 . +0 . − . M ⊙ , respectively. In the wind ( s = 2) CSI plus Ni model, the values ofthese parameters are M ej = 14 . +3 . − . M ⊙ , M CSM = 0 . +0 . − . M ⊙ . In additional, the mass of Ni is 0 . +0 . − . M ⊙ .The corresponding mass-loss rate is 0.784–7.84 M ⊙ yr − .The ejecta masses in these four models are reasonable if the progenitor was a very massive Wolf-Rayet star whosemasses are believed to be about 25 M ⊙ if their metalicity ( Z ) is equal to that of solar ( Z ⊙ ) (Crowther 2007). Therange of Ni mass is consistent with the rough upper limit ( . . ⊙ ) that can be synthesized by the explosion of amassive star. DISCUSSION3.1.
Which Is the Most Reasonable Model?
In Section 2, we demonstrate that the magnetar model, the magnetar plus Ni model, the CSI model, as well asthe Ni plus CSI model can explain the bolometric LC of PS15dpn. It should be pointed out here that he magnetarmodel and the magnetar plus Ni model are not the best models explaining the data because these two models neglectCSI while the He i emission lines indicate that CSI cannot be omitted.The CSI model overcame the disadvantages mentioned above by considering the energy from CSI and the inferredejecta mass is 16 . +2 . − . M ⊙ or 17 . +1 . − . M ⊙ which is favored by the massive Wolf-Rayet progenitor scenario.However, this model neglect the contribution from Ni. The CSI plus Ni model took both CSI and Ni intoaccount. This hybrid model gave rather reasonable parameters: M ej = 14 . +3 . − . M ⊙ , M CSM = 0 . +0 . − . M ⊙ , M Ni =0 . +0 . − . M ⊙ .In principle, the magnetar plus CSI model or the magnetar plus CSI plus Ni model should be considered. However,the CSI model and the CSI plus Ni model can yield good fit and the models involving CSI do not need an additionalmagnetar.In summary, the CSI plus Ni model is the best model to explain the LC of PS15dpn.3.2.
The Properties of the CSM and the Possible Pre-SN Outburst
Adopting the wind CSM model or wind CSM plus Ni model and assuming that the wind speed is ∼ − , the inferred mass-loss rates are ∼ −
10 M ⊙ yr − , at least 10 times that the upper limit of line-driving stellarwind ( ∼ − M ⊙ yr − , Wellons et al. 2012).If the progenitor of PS15dpn is a Wolf-Rayet star, the wind speed must be ∼ − , then the mass-loss rateof the progenitor just prior to the explosion is ∼
10 M ⊙ yr − , significantly larger than the mass-loss rate of iPTF13z( ∼ . − ⊙ yr − , Nyholm et al. 2017).In the shell CSM model and shell CSM plus Ni model, the mass of the shell expelled from the progenitor ofPS15dpn before the SN explosion is ∼ ⊙ . We can estimate the date of the eruption expelling the shell: the radiusof the innermost part of the CSM ( R CSM , in ) is ∼ − × cm (see Table 2) and the speed of the shell can besupposed to be 100–1000 km s − (10 –10 cm s − ), then the time before the SN explosion is ∼ –2 × s ∼ CONCLUSIONSPS15dpn is a luminous rapidly rising SN Ibn discovered by Pan-STARRS1 (PS1). Smartt (2016) demonstrated thatPS15dpn cannot be explained by Ni model. In this paper, we investigate the possible energy sources that can powerthe bolometric LC of PS15dpn.We found that the Ni model is disfavored since the inferred Ni mass is larger than the ejecta mass, supportingthe conclusion of Smartt (2016). Alternatively, we used the magnetar model and the magnetar plus Ni model and The mass-loss rate dominated by line-driven winds is proportional to Z . (Vink et al. 2001) and the mass-loss rate can be higherand the eventual mass of the SN progenitor can therefore be lower if Z > Z ⊙ . Moreover, the mass transfer in binary system would yield awider range of the masses of the aged massive stars. found that a magnetar with B p = 14 . +0 . − . × G and P = 13 . +0 . − . ms can reproduce a LC fitting the datawithout Ni contribution. By taking the Ni into account, the values of B p and P are 25 . +2 . − . × G and16 . +6 . − . ms, respectively; the mass of Ni is 0 . +0 . − . M ⊙ .The caveat of the magnetar model and the magnetar plus Ni model is that they omit the contribution from theinteraction between the ejecta and the pre-existing CSM whild the confirmed He i emission lines indicative of theinteraction between the ejecta and the dense He-rich CSI provide evidence that the ejecta-CSM interaction wouldprovide a fraction of energy to power the LCs of SNe Ibn (including PS15dpn).Therefore, we employed the CSI model and the CSI plus Ni model to fit the LC of PS15dpn. We found that boththese two models did a good job of fitting the data and the inferred values of ejecta mass are about 13–17 M ⊙ whichis consistent with the scenario that the progenitors of SNe Ibn are very massive Wolf-Rayet stars surrounded by thedense He-rich CSM erupted from the progenitors. The CSM mass required by these two models is ∼ ⊙ .The models containing Ni contribution is more reasonable since PS15dpn with a peak bolometric luminosity ∼ × erg s − is not superluminous SNe Ib like ASASSN-14ms whose peak bolometric luminosity is ∼ . × erg s − (Vallely et al. 2018) and the luminosity provided by a moderate amount of Ni cannot be neglected. Hence,the CSI plus Ni model is more favorable. The inferred Ni mass of the CSI plus Ni model is 0 . +0 . − . M ⊙ (CSMis a wind) or 0 . +0 . − . M ⊙ (CSM is a shell), consistent with the rough upper limit of the Ni yield that can besynthesized by core-collapse SNe.The required mass-loss rate in the wind CSI plus Ni model is extremely high ( ∼
10 M ⊙ yr − ). The shell CSI plus Ni model can avoid this tricky problem if we suppose that the progenitor expelled a gas shell ∼ Abbott, B. P., Abbott, R., Abbott, T. D., et al. 2016b, PhRvL,116, 241103Arnett, W. D. 1982, ApJ, 253, 785Chatzopoulos, E., Wheeler, J. C., & Vinko, J. 2012, ApJ, 746,121Chevalier, R. A. 1982, ApJ, 258, 790Chevalier, R. A., & Fransson, C. 1994, ApJ, 420, 268Chugai, N. N., & Danziger, I. J. 1994, MNRAS, 268, 173Crowther, P. A. 2007, ARA&A, 45, 177Filippenko, A. V. 1997, ARA&A, 35, 309Foley, R. J., Smith, N., Ganeshalingam, M., et al. 2007, ApJL,657, L105Ginzburg, S., & Balberg, S. 2012, ApJ, 757, 178Hosseinzadeh, G., Arcavi, I., Valenti, S., et al. 2017, ApJ, 836,158Hosseinzadeh, G., McCully, C., Zabludoff, A. I., et al. 2019,ApJL, 871, L9Inserra, C., Smartt, S. J., Jerkstrand, A., et al. 2013, ApJ, 770,128Karamehmetoglu, E., Taddia, F., Sollerman, J., et al. 2017,A&A, 602, A93Liu, L. D., Wang, L. J., Wang, S. Q., & Dai, Z. G. 2018, ApJ,856, 59 Lyman, J. D., Bersier, D., James, P. A., Mazzali, P. A., Eldridge,J. J., Fraser, M. & Pian, E. 2016, MNRAS, 457, 328Nomoto, K., Kobayashi, C., & Tominaga, N. 2013, ARA&A, 51,457Nyholm, A., Sollerman, J., Taddia, F., et al. 2017, A&A, 605, 6Pastorello, A., Benetti, S., Brown, P. J., et al. 2015a, MNRAS,449, 1921Pastorello, A., Smartt, S. J., Mattila, S., et al. 2007, Nature, 447,829Pastorello, A., Wang, X.-F., Ciabattari, F., et al. 2016, MNRAS,456, 853Pastorello, A., Wyrzykowski, Valenti, S., et al. 2015b, MNRAS,449, 1941Palazzi, P., Cappellaro, E., et al. 2016, GCN, 19145Schlegel, E. M. 1990, MNRAS, 244, 269Silverman, J. M., Nugent, P. E., Gal-Yam, A., et al. 2013, ApJS,207, 3Smartt, S. J. 2016, ApJL, 827, L40Smith, N. 2014, ARA&A, 52, 487Smith, N. 2017,
Observational and Physical Classification ofSupernovae . In: Alsabti A., Murdin P. (eds)
Handbook ofSupernovae . Springer, 2017, p. 403 (arXiv:1612.02006)
Smith, N., Mauerhan, J. C., Silverman, J. M., et al. 2012,MNRAS, 426, 1905Vallely, P. J., Prieto, J. L., Stanek, K. Z., et al. 2018, MNRAS,475, 2344Vink, J., de Koter, A., & Lamers, H. J. G. L. M. 2001, A&A,369, 574Wang, L. J., Wang, S. Q., Dai, Z. G., Xu, D., Han, Y. H., Wu,X. F., & Wei, J. Y. 2016, ApJ, 821, 22Wang, S. Q., Wang, L. J., Dai, Z. G., & Wu, X. F. 2015a, ApJ,799, 107 Wang, S. Q., Wang, L. J., Dai, Z. G., & Wu, X. F. 2015b, ApJ,807, 147Wellons, S., Soderberg, A. M., & Chevalier, R. A. 2012, ApJ,752, 17Wheeler, J. C., Johnson, V., & Clocchiatti, A. 2015, MNRAS,450, 1295
Table 1 . Parameters of the Ni model, the magnetar model, and the magnetar plus Ni model. The uncertainties are 1 σ . M ej M Ni B p P v sc0 κ γ, Ni κ γ, mag t expl ⋆ χ / dof(M ⊙ ) (M ⊙ ) (10 G) (ms) (10 cm s − ) (cm g − ) (cm g − ) (days) Ni 0 . +0 . − . . +0 . − . - - 2 . +0 . − . . +0 . − . - − . +0 . − . . / . +0 . − . . +0 . − . . +0 . − . . +0 . − . - 1 . +23 . − . − . +0 . − . . / Ni 0 . +0 . − . . +0 . − . . +2 . − . . +6 . − . . +0 . − . . +0 . − . . +89 . − . − . +0 . − . . / ⋆ The value of t expl is with respect to the date of the SN peak. Table 2 . Parameters of the CSI model and the CSI plus Ni model. The uncertainties are 1 σ . s E SN M ej M CSM M Ni ρ CSM , in R CSM , in ǫ x κ γ, Ni t expl ⋆ χ / dof(10 erg) (M ⊙ ) (M ⊙ ) (M ⊙ ) (10 − g cm − ) (10 cm) (cm g − ) (days)CSI 0 0 . +0 . − . . +1 . − . . +0 . − . . +0 . − . . +4 . − . . +0 . − . . +0 . − . - − . +0 . − . . / . +0 . − . . +2 . − . . +0 . − . . +6 . − . . +1 . − . . +0 . − . . +0 . − . - − . +0 . − . . / Ni 0 1 . +0 . − . . +4 . − . . +0 . − . . +0 . − . . +6 . − . . +0 . − . . +0 . − . . +0 . − . . +17 . − . − . +0 . − . . / Ni 2 1 . +0 . − . . +3 . − . . +0 . − . . +0 . − . . +7 . − . . +1 . − . . +0 . − . . +0 . − . . +11 . − . − . +0 . − . . / ⋆ The value of t expl is with respect to the date of the SN peak. Figure 1 . The bolometric LCs reproduced by the Ni model (the top left panel), the magnetar model (the top right panel)and the magnetar + Ni model (the bottom panel). Data are taken from Smartt (2016). The abscissa represents time sincethe explosion in the rest frame.
Figure 2 . The bolometric LCs reproduced by the CSI model ( s = 0, 2; the top panels) and the Ni plus CSI model ( s = 0,2; the bottom panels). The LCs powered by the forward shocks, reverse shocks, as well as56