論文の公開元へTheoretical study on the origin of supernova remnants associated with magnetars
概要
When massive stars (M > 8 M⊙) use up their nuclear fuels, they cannot be supported by their own pressure gradients and explode as core-collapse supernovae (CCSNe). SNe synthesize heavy elements at the inside of stars and the explosion phases and eject them to interstellar space. Our universe has become rich with various elements by repeating the process.
In the center of SN explosions, cores of stars become compact objects such as neutron stars (NSs) or black holes, and ejected materials make supernova remnants (SNRs). Actually, various compact objects including radio pulsars, magnetars, and central compact objects have been discovered in SNRs. Especially, magnetars have attracted many researchers as test sites for strong magnetic fields which cannot be reached on the ground because they have a few hundred times stronger magnetic fields (B ≥ 1014 G) than that of a typical NS. Furthermore, recent observations have discovered magnetar-like NSs having intermediate magnetic fields and burst activities (e.g., DAi et al. 2016; Rea et al. 2016). The border between each species becomes ambiguous. While their characters as magnetized objects have been revealed as increasing the number of discoveries, the origin of diversity of NSs remains still unexplained. In order to investigate the origin of diversity, we focus on the origin of magnetars.
The α − Ω dynamo scenario is a prospective candidate of the mechanism forming magnetars with strong magnetic fields (Duncan & Thompson, 1992; Thompson & Duncan, 1993). Rapidly rotating NSs are formed as results of core collapses while preserving the angular momenta. The rapidly-rotating NSs amplify the own magnetic fields by the α−Ω dynamo effect. In order to obtain such strong magnetic fields (B ≥ 1014 G), magnetars have to rotate with very short spin periods (P0 < 3 ms). When a spin period is P0 = 1 ms, the corresponding rotational energy is 1052 erg.
The enormous energy is thought to be emitted through magnetic dipole radiation by the strong magnetic field, and the energy is injected to the surrounding SN ejecta. Thus, magnetars have been thought to be engines of energetic SNe such as superluminous SNe (SLSNe) or hypernovae (HNe) associated with gamma-ray bursts (GRBs). This model can represent the light curves of SLSNe (e.g., Kasen & Bildsten 2010). Nicholl et al. (2017) analyzed light curves of 38 SLSNe by magnetar models and reported that some SLSNe have explosion energies of 1052 erg. These energetic SLSNe will be referred to as eSLSNe below. There are some speculations that SNe harboring magnetars have a lot of energy. However, there is few direct evidence that magnetars exist inside such energetic SNe, and the explosion energies of SNRs associated with magnetars have been estimated as approximately 1051 erg which is significantly less than that expected from the magnetar model on energetic SNe (e.g., Vink & Kuiper, 2006).
Optically thick SN ejecta prevent us from confirming the initial properties of magnetars. However, since rotational energies of magnetars are radiated by their magnetic dipole radiation and interact with the surrounding ejecta, we can infer the initial properties of magnetars from the properties of the ejecta. Since magnetars are relatively young, some of them are associated with SNRs. So far, 23 magnetars have been discovered, and eight of them have been reported to be associated with SNRs (Kaspi & Beloborodov, 2017). The associated SNRs give some information of progenitors and initial properties of magnetars. Progenitors mass can be inferred by combining mass densities and emitting volumes estimated from observations of the SNRs. By using this method, the progenitor masses of N 49 and CTB 109 were inferred to be more than 30 − 40 M⊙ (Uchida et al. 2015; Nakano et al. 2017). Martin et al. (2014) compared magnetar SNRs with other pulsar SNRs and found no difference between them in elements, ionization states, and luminosities.
Since only using observational results is not enough to follow the evolutions of the SNe for about ten thousands years, theoretical approaches bridging between SNe and SNRs are also necessary. The missing links between energetic SNe supposedly powered by magnetars and SNRs harboring magnetars have not been understood because they have been studied separately. In order to investigate the origin of magnetars, we need to consistently investigate the evolutions from the birth of magnetars to the phases they have been observed.
First, we tried to describe the evolutions of SNRs with central energy sources by using a selfsimilar solution in chapter 4. In many astronomical events, a transient flow of gas often exhibits self-similarity. In fact, self-similar solutions are known for specific phenomena (e.g., Sedov,1959; Taylor, 1950; Chevalier, 1982; Yahil, 1983). Blandford & McKee (1976) presents a self-similar solution for ultra-relativistic flow as a result of point explosion with a central energy source whose energy injection rate follows a power law with respect to time. In this study, we discussed the corresponding non-relativistic flow using a self-similar solution. We constructed a model in which a blast wave is driven by the central energy source embedded in an ultra-relativistic gas or photon gas pushing the outer region through the contact discontinuity and derive the self-similar solution. We also investigated the allowed ranges for which self-similar solutions exist. Furthermore, we also calculated the same model numerically to confirm our self-similar solutions and investigated what occurs outside the applicable range. We found that the fluid dynamics cannot be described by our self-similar solution when the energy was injected by magnetic dipole radiation from a magnetar into the envelope of a massive star.
Thus, we investigated the dynamical evolutions of SNe having magnetars by using numerical simulations. In order to constrain the initial properties of magnetars, we discussed the properties of two objects, CTB 109 (chapter 5) and SN 2012au (chapter 6).
In chapter 5, we performed hydrodynamical simulations in order to link SNe forming magnetars and SNRs associated with magnetars. As already mentioned, while some of millisecond magnetars are thought to be engines of energetic SNe with explosion energies of 1052 erg, the explosion energies estimated from the SNRs associated with magnetars are only 1051 erg. In order to confirm the relation, we need to consistently investigate the evolutions from the birth of magnetars to the phases they have been observed.
The explosion energies of SNRs have been estimated by the Sedov solution using the electron temperatures of SNRs. The model assumes that the SNR is in an adiabatic phase and the radiative cooling is negligible. If the ion temperatures are significantly larger than electron temperatures by one order of magnitude or the radiative cooling is not negligible, the explosion energies might be underestimated. There remains a possibility that SNRs associated with magnetars are observed as normal SNRs even though the original SN are energetic with the explosion energy of 1052 erg. Thus, we perform hydrodynamical simulations considering non-thermal equilibriums and radiative cooling and investigate evolutions of SNRs originating from HNe of massive stars.
Comparing the observation of the SNR with our simulation result, we discuss the discrepancy between the energy estimated from the SNR associated with the magnetar and the energy of the SN supposedly powered by the magnetar. CTB 109 is an SNR associated with the magnetar 1E 2259+586. The SNR has been well observed by X-ray observatory, and the progenitor mass is estimated as M ∼ 40 M⊙ (e.g., Sasaki et al. 2004, 2013; Nakano et al. 2015, 2017). The progenitors of other magnetars are also thought to be massive stars (e.g., Figer et al. 2005; Uchida et al. 2015). Thus, in order to investigate evolutions of SNRs associated with magnetars originating from massive stars (M = 40 M⊙), we performed hydrodynamical simulations and discussed the possibility that the origin of CTB 109 is an energetic SN with explosion energy of 1052 erg by comparing our simulation results with the observation results.
Our simulation results indicated that SNe with a large amount of energies of 1052 erg evolve while maintaining most of the energies, and they are observed as SNRs which are completely different from that of a typical explosion energy of 1051 erg. Comparing the radius, the electron temperature, and the luminosity of CTB 109 with those of our simulation results, the observables of CTB 109 can be explained by simulation results of a typical explosion energy of 1051 erg rather than that of an HN. Thus, we concluded that the explosion energy of CTB 109 was not 1052 erg but 1051 erg.
In chapter 6, we performed 1D hydrodynamical simulations and constrained the way of the injection of the explosion energy and the property of the pulsar of SN 2012au. A type Ib SN 2012au has been focused to provide some clues to link normal SNe with energetic SNe associated with such as GRBs since the property of SN 2012au such as the peak luminosity (6.7 × 1042 erg/s), the synthesized mass of 56Ni (0.3 M⊙), and the explosion energy (1052 erg) were similar to those of HNe (Takaki et al. 2013; Milisavljevic et al. 2013).
This HN has been proposed to harbor a pulsar wind nebula because the HN was still visible in optical bands even at 6.2 yrs after the explosion. Milisavljevic et al. (2018) detected forbidden transition lines of oxygen and sulfur with expansion velocities of 2, 300 km/s on the spectrum at 6.2 yrs. Although no emission line of [OIII] was detected in the spectrum in 1 yr after the explosion, the emission line of [OIII] was detected clearly, and the luminosity was estimated to be 1038 erg/s. The HN needs a heating source to be visible and obtain highly ionized oxygens in 6.2 yrs. Milisavljevic et al. (2018) proposed that a pulsar wind nebula existed in the center of the HN as the power source.
The explosion mechanism and the pulsar property have not been completely explained yet. The expansion velocity of 2, 300 km/s might be too slow to be expected in the ejecta of an HN explosion. We suspected that this velocity corresponds to the minimum velocity due to the gravity of the central object. The matter with velocities lower than this minimum velocity cannot escape from the gravity of the central object. Satisfying the condition leaving enough ejecta mass with velocities slower than 2, 300 km/s before the ejecta are accelerated by the pulsar, we can constrain the explosion mechanism. Thus, we performed 1D hydrodynamical simulations in order to constrain the possible explosion mechanism.
In order to synthesize 56Ni with 0.3 M⊙ estimated from the peak luminosity and keep enough ejecta mass with velocities slower than 2, 300 km/s, we found that the energy injection time for the explosion needs to be a very short time with 0.01 sec from our simulation results. This is much shorter than time scales of explosions by a neutrino heating mechanism or a jet by magnetohydrodynamics. We also found that the allowed dipole magnetic field is B ≤ 1014 G. This result means it is possible that there is a magnetar-like NS inside SN 2012au. If the pulsar has B = 1014 G, the initial spin period is derived as 40 ms to account for the luminosities of forbidden lines at 6.2 yrs after the explosion. Thus, the pulsar has a too little rotational energy of the order of 1049 erg to power this HN.
From the above studies, we found no evidence that the origin of SNRs associated with magnetars is energetic SNe with 1052 erg. The engine of SN 2012au which is one of HNe is also not a millisecond magnetar. These results do not support the scenario that millisecond newborn-magnetars are engines of eSLSNe/HNe.
