Hydrogen in Type Ic Supernovae?
ABSTRACT By definition, a Type Ic supernova (SN Ic) does not have conspicuous lines of hydrogen or helium in its optical spectrum. SNe Ic usually are modelled in terms of the gravitational collapse of bare carbon-oxygen cores. We consider the possibility that the spectra of ordinary (SN 1994I-like) SNe Ic have been misinterpreted, and that SNe Ic eject hydrogen. An absorption feature usually attributed to a blend of Si II 6355 and C II 6580 may be produced by H-alpha. If SN 1994I-like SNe Ic eject hydrogen, the possibility that hypernova (SN 1998bw-like) SNe Ic, some of which are associated with gamma-ray bursts, also eject hydrogen should be considered. The implications of hydrogen for SN Ic progenitors and explosion models are briefly discussed. Comment: Accepted by PASP. Several significant changes including one additional figure
- SourceAvailable from: Takeo Minezaki[show abstract] [hide abstract]
ABSTRACT: Observations and modeling for the light curve (LC) and spectra of supernova (SN) 2005bf are reported. This SN showed unique features: the LC had two maxima, and declined rapidly after the second maximum, while the spectra showed strengthening He lines whose velocity increased with time. The double-peaked LC can be reproduced by a double-peaked 56Ni distribution, with most 56Ni at low velocity and a small amount at high velocity. The rapid postmaximum decline requires a large fraction of the γ-rays to escape from the 56Ni-dominated region, possibly because of low-density "holes." The presence of Balmer lines in the spectrum suggests that the He layer of the progenitor was substantially intact. Increasing γ-ray deposition in the He layer due to enhanced γ-ray escape from the 56Ni-dominated region may explain both the delayed strengthening and the increasing velocity of the He lines. The SN has massive ejecta (~6-7 M☉), normal kinetic energy [~(1.0-1.5) × 1051 ergs], a high peak bolometric luminosity (~5 × 1042 ergs s-1) for an epoch as late as ~ 40 days, and a large 56Ni mass (~0.32 M☉). These properties and the presence of a small amount of H suggest that the progenitor was initially massive (M ~ 25-30 M☉) and had lost most of its H envelope, and was possibly a WN star. The double-peaked 56Ni distribution suggests that the explosion may have formed jets that did not reach the He layer. The properties of SN 2005bf resemble those of the explosion of Cassiopeia A.The Astrophysical Journal 12/2008; 633(2):L97. · 6.73 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Synthetic spectra generated with the parameterized supernova synthetic-spectrum code SYNOW are compared to photospheric-phase spectra of Type Ib supernovae (SNe Ib). Although the synthetic spectra are based on many simplifying approximations, including spherical symmetry, they account well for the observed spectra. Our sample of SNe Ib obeys a tight relation between the velocity at the photosphere, as determined from the Fe II features, and the time relative to that of maximum light. From this we infer that the masses and the kinetic energies of the events in this sample were similar. After maximum light the minimum velocity at which the He I features form usually is higher than the velocity at the photosphere, but the minimum velocity of the ejected helium is at least as low as 7000 kms. Previously unpublished spectra of SN 2000H reveal the presence of hydrogen absorption features, and we conclude that hydrogen lines also were present in SNe 1999di and 1954A. Hydrogen appears to be present in SNe Ib in general, although in most events it becomes too weak to identify soon after maximum light. The hydrogen-line optical depths that we use to fit the spectra of SNe 2000H, 1999di, and 1954A are not high, so only a mild reduction in the hydrogen optical depths would be required to make these events look like typical SNe Ib. Similarly, the He I line optical depths are not very high, so a moderate reduction would make SNe Ib look like SNe Ic. Comment: 21 pages and 24 figures, submitted to ApJ06/2001;
- [show abstract] [hide abstract]
ABSTRACT: We present detailed non-LTE (NLTE) synthetic spectra for comparison with a time series of observed optical spectra of the Type Ic supernova SN 1994I which occurred in M51. With the exceptions of Si I and S I, we treat the important species in the formation of the spectrum in full NLTE. We present results for both a hydrodynamic model that has been fitted to the light curve and for an illustrative custom-crafted model that is more massive. Both models give reasonable fits to the overall observed spectra; however, neither is able to reproduce all the observed features. Some conspicuous observed features are absent, and some predicted features are unobserved. No model that we have explored is able to reproduce satisfactorily the observed infrared feature near 1 μm on 1994 April 15 (+7 days), which has been attributed to the triplet He I λ10830 transition. The low-mass hydrodynamic model produces an infrared feature with a blend of He I, C I, O I, and Si I-Si II lines, but it predicts a strong unobserved absorption feature near 6100 Å due to Fe III, and the observed feature just blueward of 6000 Å most likely due to Na D is not reproduced. The more massive model does a better job of reproducing the observed infrared line shape, but also predicts the unobserved feature near 6100 Å. The early-time spectrum of the low-mass model is far too blue; thus, a more massive model may be slightly favored. Since the predicted infrared feature is produced by a blend of so many elements, and there is no overwhelming evidence for other helium features such as λ5876, it may be premature to conclude that SNe Ic unambiguously contain helium. Thus, we conclude that pure C + O cores are still viable progenitors for SNe Ic.The Astrophysical Journal 01/2009; 527(2):739. · 6.73 Impact Factor
arXiv:astro-ph/0604047v2 9 May 2006
HYDROGEN IN TYPE Ic SUPERNOVAE?
David Branch1, David J. Jeffery1, Timothy R. Young2, & E. Baron1
By definition, a Type Ic supernova (SN Ic) does not have conspicuous lines
of hydrogen or helium in its optical spectrum. SNe Ic usually are modelled in
terms of the gravitational collapse of bare carbon–oxygen cores. We consider the
possibility that the spectra of ordinary (SN 1994I–like) SNe Ic have been misinter-
preted, and that SNe Ic eject hydrogen. An absorption feature usually attributed
to a blend of Si II λ6355 and C II λ6580 may be produced by Hα. If SN 1994I–like
SNe Ic eject hydrogen, the possibility that hypernova (SN 1998bw–like) SNe Ic,
some of which are associated with gamma–ray bursts, also eject hydrogen should
be considered. The implications of hydrogen for SN Ic progenitors and explosion
models are briefly discussed.
Subject headings: supernovae: general — supernovae: individual (SN 1994I, SN
A Type II supernova (SN II) has conspicuous hydrogen lines in its optical spectrum. A
Type Ib supernova lacks conspicuous hydrogen lines but does have conspicuous He I lines.
A Type Ic supernova has conspicuous lines of neither hydrogen nor He I. For a review of
supernova spectral classification see Filippenko (1997).
Supernovae of these types are thought to result from core–collapse in massive stars. A
SN Ib is hydrogen–deficient in its outer layers, and displays He I lines owing to nonthermal
excitation by the decay products of radioactive56Ni and56Co (Lucy 1991). In recent years
it has becomes clear that at least some SNe Ib are not completely hydrogen–free; in fact,
it is likely that most or even all SNe Ib eject a small amount of hydrogen at high velocities
(Deng et al. 2000; Branch et al. 2002; Elmhamdi et al. 2006). If some of the events
1Homer L. Dodge Department of Physics and Astronomy, University of Oklahoma, Norman, OK 73019;
2Department of Physics, University of North Dakota, Grand Forks, ND, 58202
– 2 –
classified as Type Ib had been observed earlier, hydrogen lines might have been conspicuous,
in which case they would be classified as Type IIb — the designation used for events such as
SN 1993J that looked like a Type II at early times but later looked like a Type Ib (Filippenko,
Matheson, & Ho 1993).
A SN Ic either is helium–deficient in its outer layers or fails to nonthermally excite its
helium (Woosley & Eastman 1997). It has become common practice to model SNe Ic in
terms of core collapse in bare (or nearly bare) carbon–oxygen cores (Iwamoto et al. 1994;
Foley et al. 2003; Mazzali et al. 2004).
Supernova spectral features are P–Cygni features characterized by line–centered emis-
sion components and blueshifted absorption components, with the absorptions frequently
being more identifiable. Spectral features usually are blended owing to the huge Doppler
broadening. Nevertheless, some identifications of spectral lines are definite. For example,
blends of Fe II lines and features produced by Ca II H&K and the Ca II infrared triplet
appear in all types of supernovae as long as the temperature is sufficiently low. However,
there also are some serious identification ambiguities. The most well known involves the
Na I D–line doublet at mean wavelength λ5892 and the strongest optical line of He I, λ5876,
which are separated by only about 800 km s−1. If the corresponding observed feature is
strong and other He I lines are not present, the feature is produced at least mainly by Na I,
while if other He I lines are clearly present then the feature is at least partly due to He I.
But if the observed feature is not strong, it can be difficult to choose between Na I and He I.
When SNe Ic are interpreted in terms of bare carbon–oxygen cores, an absorption feature
usually near 6200˚ A, which we will refer to as the 6200˚ A absorption, is attributed to the
strongest optical line of Si II, λ6355 (the transition definitely responsible for the similarly
located deep absorption in Type Ia supernovae), perhaps blended with the strongest optical
line of C II, λ6580, forming in higher–velocity ejecta than Si II. Local–thermodynamic–
equilibrium (LTE) calculations of Sobolev line optical depths for a composition dominated
by carbon and oxygen (with hydrogen and helium burned to carbon and oxygen, and solar
mass fractions of heavier elements) show that within certain intervals of temperature Si II
λ6355 and C II λ6580 are expected to have significant optical depths (Hatano et al. 1999).
Nevertheless, the identification of the 6200˚ A absorption is plagued by ambiguities. The
strongest line of Ne I, λ6402, is about 2200 km s−1to the red of Si II λ6355, and Hα
λ6563 is only about 800 km s−1to the blue of C II λ6580. Because these four ions, each of
which could appear in supernova spectra (although Ne I probably would require nonthermal
excitation), have their strongest optical lines to the red, but not too far to the red, of the
6200˚ A absorption, the ambiguity is difficult to resolve when the observed feature is weak
and other lines of these ions do not produce identifiable features.
– 3 –
In this paper we are primarily concerned with the possible presence of hydrogen in
SNe Ic. This issue (and that of helium in SNe Ic) has been addressed in the literature
before; for summaries see Filippenko (1997); Matheson et al. (2001); and Branch (2002).
The presence of hydrogen has been suggested, for example, by Filippenko (1988, 1992);
Filippenko, Porter, & Sargent (1990); Jeffery et al. (1991); and Branch (2002). The currently
prevailing view, however, is that hydrogen is absent and the 6200˚ A absorption is produced
by Si II and/or C II (Wheeler et al. 1994; Millard et al. 1999). In §2, based on a comparison
of spectra of the Type Ib or Type Ib/c SN 1999ex, which we believe to contain hydrogen, and
the Type Ic SN 1994I, we raise the question of whether hydrogen also is present in SN 1994I.
In §3, comparisons of spectra of SN 1999ex and SN 1994I with synthetic spectra generated
with the parameterized resonant–scattering code Synow are presented and discussed. In §4,
the implications of the presence of hydrogen in ordinary (SN 1994I–like), and possibly even
in hypernova (SN 1998bw–like; Foley et al. 2003) SNe Ic, are briefly considered.
2. COMPARISON OF SN 1999ex AND SN 1994I
SN 1999ex was initially classified as Type Ic based on a resemblance of its early spectrum
to SN 1994I (Hamuy & Phillips 1999), but because it later developed He I lines Hamuy et al.
(2002) revised the classification to intermediate Type Ib/c. Although the He I lines were
not as deep as in most events that have been classified as Type Ib, their presence was
definite, so Branch (2002) referred to SN 1999ex as a “shallow–helium” Type Ib. Hamuy
et al. (2002) labelled the 6200˚ A absorption as Si II, but Branch (2002) argued that the
correct identification is Hα, consistent with the presence of Hα in most if not all other
SNe Ib. Recently Elmhamdi et al. (2006) have reinforced the conclusion that SNe Ib,
including SN 1999ex and another transition Type Ib/c event, SN 1996aq, eject hydrogen.
(The overluminous Type Ib SN 1991D (Benetti et al. 2002) could be an exception.)
The spectra of SN 1999ex and the Type Ic SN 1994I shown in Figures 1 to 3 are
from Hamuy et al. (2002) and Filippenko et al. (1995), respectively. The horizontal scale
on all figures in this paper is logarithmic wavelength, which allows the widths of Doppler–
broadened features to be compared on an equal basis across the whole spectrum. The spectra
are “tilted” by multiplying by λα, with α chosen to make the flux peaks near 4600˚ A and
6300˚ A about equally high. The tilting makes it easier to compare spectral features. When
comparing SN 1999ex and SN 1994I spectra we artificially blueshift those of SN 1999ex in
order to compensate for the different photospheric velocities and roughly align the absorption
In Figure 1 a spectrum of SN 1994I obtained 2 days before the time of maximum light
– 4 –
in the B band (day −2) is compared with a day −1 spectrum of SN 1999ex that has been
blueshifted by 4000 km s−1. Figure 2 is like Figure 1, but for day +4 spectra of SN 1994I and
SN 1999ex, with the latter blueshifted by 2000 km s−1. Figure 3 is for a day +10 spectrum of
SN 1994I and a day +13 spectrum of SN 1999ex, with the latter blueshifted by 2000 km s−1.
In all three figures the spectra are similar in many respects, although the features are more
washed out in SN 1994I because of its higher photospheric velocity at each epoch (hence the
necessity to blueshift SN 1999ex to illustrate the similarities). Otherwise the main differences
are that He I λ6678 and λ7065, although weak, are clearly present in SN 1999ex (making it
a Type Ib or at least a Type Ib/c), but they are not clearly present in SN 1994I (making it
a Type Ic). However, it is not possible on the basis of Figures 1 to 3 to exclude the presence
of weak He I features in SN 1994I. Given the similarities in these figures — and they are
rather striking — the question becomes: are the Na I and Si II/C II identifications correct
for SN 1994I? If so, then the resemblance of the SN 1999ex and SN 1994I spectra from about
5500˚ A to 6300˚ A is coincidental.
3. COMPARISONS WITH SYNTHETIC SPECTRA
To further investigate the possibility of hydrogen (and He I) in SN 1994I we have used the
parameterized resonant–scattering Synow code (Branch et al. 2002) to generate synthetic
spectra for comparison with spectra of SN 1999ex and SN 1994I.
3.1. SN 1999ex
In Figure 4 the day +4 spectrum of SN 1999ex is compared with a synthetic spectrum
in which the 6200˚ A absorption is produced by Hα. In the synthetic spectrum the velocity
at the photosphere is 8000 km s−1. Hydrogen is detached from the photosphere at 13,000
km s−1where Hα has an optical depth of 0.5. Hydrogen–line optical depths decrease outward
exponentially but very slowly, with e–folding velocity ve= 20,000 km s−1, to an imposed
maximum velocity of 20,000 km s−1. Given the closeness of the fit to the 6200˚ A absorption,
we know that with fine adjustments of the optical–depth profile we could refine the fit to
be practically perfect. Thus at the Synow level of analysis, Hα is completely adequate and
we would not prefer a different identification without other evidence. For all ions used in
Figure 4, reference–line optical depths and vevalues (in units of 1000 km s−1) are given in
Elmhamdi et al. (2006) show that Si II λ6355 cannot account for the 6200˚ A absorption
– 5 –
of SN 1999ex on its own because even when it is undetached from the photosphere, its
synthetic absorption feature is too blue. It is unlikely that the 6200˚ A absorption is due
mainly to Ne I λ6402 because the observed feature is strong enough that when it is fitted
with Ne I λ6402 other Ne I lines appear in the synthetic spectrum and the fit deteriorates.
This conclusion is based on LTE excitation and must be tested against detailed non-LTE
calculations that take nonthermal excitation into account.
Hamuy et al. (2002) also obtained a day +4 spectrum of SN 1999ex that extends
to nearly 2.5 microns. Synow fitting parameters used to fit the optical He I lines also
give satisfactory fits to features attributed to He I λ10830 and He I λ20851 (Branch 2002;
Elmhamdi et al. 2006), but the infrared spectrum does not provide a useful constraint on
the presence of hydrogen.
3.2. SN 1994I
Synow fits to spectra of SN 1994I in which the 6200˚ A absorption is attributed to
Si II and/or detached high–velocity C II are shown in Millard et al. (1999) and Elmhamdi
et al. (2006). For this paper we began trying to fit the day +4 spectrum of SN 1994I by
varying the input parameters that were used for the synthetic spectrum shown in Figure 4 for
SN 1999ex, i.e., we assumed the presence of hydrogen and He I lines. Figure 5 for SN 1994I
is like Figure 4 for SN 1999ex. The synthetic spectrum has a velocity at the photosphere of
11,000 km s−1(compared with 8000 km s−1for SN 1999ex). Hydrogen lines are detached
from the photosphere at 15,000 km s−1where Hα has an optical depth of 0.4. Hydrogen–line
optical depths decrease outward exponentially with e–folding velocity ve= 20,000 km s−1
to an imposed maximum velocity of 22,000 km s−1. This hydrogen optical–depth profile is
quite similar to that used for SN 1999ex in Figure 4, and in Figure 5 the 6200˚ A absorption is
fit nicely. The 5700˚ A absorption also is fit nicely by He I λ5876. In the synthetic spectrum
He I λ6678 and λ7065 are weak and rather washed out owing to the high photospheric
velocity, but they do more good than harm and their presence in the observed spectrum
cannot be excluded. (However, regarding the strong observed absorption near one micron,
the situation remains as illustrated and discussed in Millard et al. (1999); the synthetic He I
λ10830 absorption is too weak and narrow to account entirely for the observed absorption.)
Reference–line optical depths and vevalues for Figure 5 are given in Table 1.
The synthetic spectrum in Figure 6 is like that of Figure 5 except that it includes
Si II lines instead of hydrogen lines. The figure illustrates that in SN 1994I, as well as in
SN 1999ex, Si II λ6355 is too blue to account for the 6200˚ A absorption on its own (flux
differences between observed and Synow spectra are to be expected, but differences in wave-