EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 638 (June 2020) A&A, 638 (2020) A83 Full HTML
Open Access
Issue
A&A
Volume 638, June 2020
Article Number A83
Number of page(s) 15
Section Stellar structure and evolution
DOI https://doi.org/10.1051/0004-6361/202037536
Published online 17 June 2020

© C. Weinberger et al. 2020

Licence Creative CommonsOpen Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Open Access funding provided by Max Planck Society.

1. Introduction

Supernova explosions play a crucial role in the chemical and kinematic evolution of the Universe. Self-consistent detailed models for the explosion mechanism and the ensuing kinematics of the ejected material are still lacking. Despite the frequent occurrence of supernova explosions (1.9 ± 1.1 per century in the Galaxy; Diehl et al. 2006), observational constraints on the explosion mechanisms are still sparse, due to the large variety of progenitor systems and large parameter space of their models.

For modeling core-collapse supernova explosions, reviving the stalled shock and triggering an explosion presents a major challenge (Janka 2012; Burrows et al. 2018), as it has long been understood that the prompt explosion mechanism following core bounce cannot explode the star. Energy deposition by neutrinos in a gain region close to the stalled shock is considered as the driving force of the shock revival. Heating-induced creation of hydrodynamic effects, observed in 2D and 3D models, such as neutrino-driven convection and the standing accretion shock instability (SASI; Blondin & Shaw 2007; Marek & Janka 2009; Bruenn et al. 2013, 2016; Hix et al. 2016) enhance the explodability, however, different implementation schemes favor the dominance of either neutrino-driven convection or the SASI (Pan et al. 2016; Summa et al. 2016; Skinner et al. 2016). Inclusion of additional, microphysical effects, for example, strangeness corrections (Melson et al. 2015) and rotation from the progenitor (Müller et al. 2017; Summa et al. 2018; Takiwaki et al. 2016; Iwakami et al. 2014) affect the explodability. Model calculations of supernova explosions performed by various groups lead to successful explosions in a mass range 8 ≲ M/M⊙ ≲ 25. Asymmetries evolved from hydrodynamic effects become frozen in the explosion and are reflected in the kinematics and mass distribution of the ejecta (Nomoto et al. 1995; Buras et al. 2006; Fryer & Kusenko 2006; Takiwaki et al. 2012; Wongwathanarat et al. 2015; Orlando et al. 2016; Wongwathanarat et al. 2017).

The variety of observed type Ia supernova luminosity and temporal behavior of light curves in the first few hundred days (Phillips et al. 1999) cannot be reproduced by a single progenitor type (Wang & Han 2012; Hillebrandt et al. 2013). Mergers of binary white dwarfs (van Kerkwijk et al. 2010; Pakmor et al. 2010, 2013; Ruiter et al. 2012; Kashyap et al. 2018) and mass accretion on single white dwarf stars (single degenerate scenario) are expected to lead to a central thermonuclear runaway, disrupting the white dwarf star in the process. On single degenerates, both stable mass accretion on white dwarfs towards the Chandresekhar mass limit (Nomoto et al. 1984; Parthasarathy et al. 2007; Maeda et al. 2010; Woosley & Kasen 2011; Hachisu et al. 2011; Seitenzahl et al. 2013; Chen et al. 2014; Fink et al. 2014) and surface helium detonation (Livne 1990; Fink et al. 2010; Shen & Bildsten 2014; Leung & Nomoto 2020) can increase the central density sufficiently to ignite nuclear fusion.

The high opacity of the ejected material does not allow direct observations of the first stages of the explosions. One of the most promising methods of deducing physical constraints for the explosion mechanisms is to directly observe the decay of radioactive isotopes produced during explosive nucleosynthesis. The best candidates for observing supernova interior physics through their nucleosynthesis imprints are the isotopes 56Ni and 44Ti due to their high abundances (≈10−2 M⊙ for core-collapse supernovae, 0.5 M⊙ for SNe Ia in 56Ni, 10−6 − 10−4 M⊙ in 44Ti, see also below). Both isotopes have, for gamma-ray observations, ideal radioactive decay times: 6 d and 77 d for 56Ni and its daughter nucleus 56Co, respectively, allow only a very short post-explosion observation window in which the majority of emitted gamma-rays are expected to be absorbed by the dense stellar material. With a half life of 58.9 ± 0.3 yr (Ahmad et al. 2006), a considerable amount of 44Ti is still decaying even after centuries when the supernova remnant has long become optically thin to X- and gamma-ray emission. In contrast to the decay of synthesized 26Al and 60Fe with half lives of 700 kyr and 2 Myr, producing a diffuse emission throughout the Milky Way (Plüschke et al. 2001; Bouchet et al. 2015; Siegert & Diehl 2017), the emission signature of 44Ti decay is expected to be the one of individual point sources.

Spatially co-produced in core-collapse supernovae, 44Ti and 56Ni are mainly synthesized during alpha-rich freeze out (Woosley et al. 1973) deep in the central region of the supernova, where nucleosynthesis is strongly dependent on the thermodynamic conditions of the inner ejecta (Magkotsios et al. 2010). Due to the hydrodynamic instabilities required for an effective explosion, mixing and asymmetric expansion of the burning volume invalidates the 1D model characteristics of a mass cut, an idealized radius (Woosley & Weaver 1995), separating the gravitationally bound material from the ejecta. Since the nuclear burning occurs close to such a mass cut, the final amount of ejected 44Ti is subject to uncertainty and ranges between 10−5 − 10−4 M⊙ (Timmes et al. 1996; Limongi & Chieffi 2018).

Type Ia supernovae typically produce in centrally ignited pure deflagration models, and up to a few times 10−5 M⊙ in delayed detonation models (Maeda et al. 2010; Seitenzahl et al. 2013; Fink et al. 2014). Double detonation and surface He deflagration of sub-Chandrasekhar mass white dwarfs produce yields up to 10−3 M⊙ of 44Ti (Fink et al. 2010; Woosley & Kasen 2011; Moll & Woosley 2013). The subclass of helium surface explosions synthesizes 56Ni very little, potentially producing subluminous type Ia supernova events. Simulations (Waldman et al. 2011) and observations (Perets et al. 2010) suggest special configurations of binary white dwarfs, which, when exploding as peculiar type Ia supernovae, can produce of the order of 10−2 M⊙ of 44Ti.

Observations of late optical spectra allow for the determination of mass ratios of synthesized elements, constraining the burning mechanisms (Eriksen et al. 2009; Pakmor et al. 2010; Jerkstrand et al. 2015; Maguire et al. 2018; Mori et al. 2018). However, the total ejected mass remains a free parameter. Nucleosynthesis yields can be estimated from the bolometric light curves of explosion, however, inferring the ejected mass of radioactive material is highly model dependent (Seitenzahl et al. 2014). Comparing optical and infrared spectra to the late-time light curve, an ejected 44Ti mass of (1.5 ± 0.5) × 10−4 M⊙ is obtained for SN 1987A (Jerkstrand et al. 2015) in agreement with NuSTAR findings of (1.5 ± 0.3) × 10−4 M⊙ (Boggs et al. 2015) based on the 68 and 78 keV line. This is, however, in disagreement with the results obtained from a multicomponent long-time light-curve model of (0.6 ± 0.2) × 10−4 M⊙ (Seitenzahl et al. 2013) and direct detection of 44Ti decay with INTEGRAL/IBIS at 68 and 78 keV of (3.1 ± 0.8) × 10−4 M⊙ (Grebenev et al. 2012).

Direct observational evidence for the production of 44Ti can be obtained through the decay chain of 44Ti→44Sc→44Ca, where the prominent decay lines arise at energies of 68 keV and 78 keV for the decay of 44Ti, and at 1157 keV for the 44Sc decay. These lines have a probability of 93.0%, 96.4%, and 99.9% per decay, respectively (Chen et al. 2011). In addition, a fluorescence photon is emitted from shell transitions in 44Sc with a probability of 16.7% at 4.1 keV. With a significant difference of half life times of 60 yr in the first decay and 4 h (Audi et al. 2003) in the subsequent decay, the activity of all three decay channels can be safely assumed to be identical after correcting for the branching ratios. In this work, we utilized INTEGRAL/SPI data to search for emission of 44Ti in all three decay lines simultaneously in the young close by supernova remnants Cassiopeia A, Tycho, Kepler, G1.9+0.3, Vela Junior, and the extragalactic (but very young) SN 1987A. We aim to constrain both 44Ti ejecta yields and explosion kinematics for remnants with ages of less than a few centuries where a substantial amount of 44Ti may still be present. We re-evaluated previous analyses of these remnants concerning the decay of 44Ti (see Sect. 2) in the 68 and 78 keV lines and improved on them by including decay signature of the 44Sc daughter nucleus at 1157 keV, which can only be seen in SPI. The 1157 keV line was studied for the remnant Cassiopeia A and Vela Jr.

This paper is structured as follows: Sect. 2 provides an overview of the six target remnants. Section 3 describes SPI data and spectral analysis, followed by Sect. 4, which includes results for the six remnants. Finally, in Sect. 5, we discuss our results and give a summary of astrophysical implications.

2. Young supernova remnants

The astrophysical parameters of the six most promising candidates to coherently observe the 44Ti decay chain are listed in Table 1. Evidence for the signature of the 44Ti decay has been claimed in the majority of these historic supernova explosions. Within SPI’s narrow-line sensitivity verification of the 44Ti signal and detection of the ensuing decay of 44Sc is feasible within a limited age-distance volume including these remnants. We searched for signatures of the 44Ti decay at 68 and 78 keV due to the de-excitation of 44Sc*, and at 1157 keV due to the subsequent de-excitation of 44Ca. We utilized data from the INTEGRAL mission from 2003 to 2019, meaning INTEGRAL revolutions 43 to 2047. For our study, we applied an average date of 01.01.2011 AD as an observation date for all calculations and included age uncertainties in our mass derivation.

Table 1.

Astrophysical parameters for the analyzed objects including the distances of supernova remnants used in the mass estimate for 44Ti.

2.1. Cassiopeia A

Cassiopeia A is one of the best studied supernova remnants in the Milky Way (Vink 2004). With an approximate age of 340 years, it is the youngest known Galactic supernova remnant, attributed to a core-collapse explosion. From the detection of hydrogen and weak helium lines in a supernova light echo, attributed to Cassiopeia A, the explosion is characterized as a type IIb supernova (Krause et al. 2008a). Explosions of this type are typically produced by 15–25 M⊙ stars (Young et al. 2006). The decay of 44Ti has been consistently measured (Siegert et al. 2015; Grefenstette et al. 2014; Iyudin et al. 1994; Vink et al. 2001) from Cassiopeia A with an average inferred 44Ti ejecta mass of (1.37 ± 0.19) × 10−4 M⊙.

2.2. SN 1987A

The explosion of SN 1987A occurred in the Large Magellanic Cloud (LMC) on February 24 1987, giving rise to a peculiar light curve containing a plateau phase (type II-P supernova). Detection of 56Co lines (Matz et al. 1988; Tueller et al. 1990) for the first time directly has confirmed that supernova light is indeed powered by this isotope, produced in the inner regions at the time of core-collapse. Recent refinements of astronomical precision yield a distance of 49.6 ± 0.5 kpc (Pietrzyński et al. 2019) to the LMC. Direct proof for the decay of 44Ti has been reported from measurements with IBIS/INTEGRAL and NuSTAR in the hard X-ray lines of 44Sc. However, the two measurements show a discrepancy in the ejecta mass, with (3.1 ± 0.8) × 10−4 M⊙ and (1.5 ± 0.3) × 10−4 M⊙, respectively (Grebenev et al. 2012; Boggs et al. 2015). NuSTAR constrains the ejecta kinematics to an expansion velocity of less than 4100 km s−1, and a bulk Doppler shift suggests an asymmetric explosion.

2.3. Vela Jr.

COMPTEL discovered a significant gamma-ray emission in the energy range centered at the 1157 keV 44Ca line. The signal was located in the direction of the Vela region with a flux of (3.8 ± 0.7) × 10−5 ph cm−2 s−1 (Iyudin et al. 1998). This emission was attributed to the decay of 44Ti in a previously unknown type II supernova remnant, which had been identified through detailed analysis of the X-ray emission of this region, and was called RX J0852.0-4622 or “Vela Jr.” (Aschenbach 1998). An estimated age of ≈680 yr and a distance of ≈200 pc is derived (Aschenbach et al. 1999). The remnant has an apparent diameter of 2° (Aharonian et al. 2007). Follow-up observation with ASCA and XMM-Newton report a line at 4.4 keV (Tsunemi et al. 2000; Iyudin et al. 2005, respectively) possibly from 44Sc fluorescence. However, the more likely remnant age of 2.4–5.1 kyr derived from the expansion rate of the supernova (Allen et al. 2014) would exclude detectability of 44Ti decay emission. This would then also be in agreement with the upper limits determined with IBIS/INTEGRAL (Tsygankov et al. 2016).

2.4. G1.9+0.3

G1.9+0.3 is presumably the youngest supernova remnant seen in the Galaxy so far. It has been identified by Reynolds et al. (2008) in the years 1985 and 2008 in the radio and X-ray regime. It is identified as a type Ia explosion. Radio observations suggest a distance of 8.5 kpc, placing the remnant in the Galactic center (Reynolds et al. 2008). From the apparent increase in size, an age of ≈100 yr is deduced, which suggests a very high expansion velocity of 14 000 km s−1 for the shock front. Using Chandra data, a detection of a soft X-ray component at 4.1 keV from the fluorescence line of 44Sc has been reported (Borkowski et al. 2010) inferring a 44Ti ejecta mass of (1 − 7) × 10−5 M⊙. Extrapolating the detected flux in the 44Sc fluorescence line to the hard X-ray lines provides a line flux estimate for the 68 keV 44Ti line, which is below the upper limits determined from NuSTAR and IBIS/INTEGRAL instruments ((0.7 − 1.5) × 10−5 ph cm−2 s−1, 9 × 10−6 ph cm−2 s−1; Zoglauer et al. 2015; Tsygankov et al. 2016, respectively).

2.5. Tycho

The Tycho supernova remnant is attributed to an explosion of type Ia, from measurements of the light echo and a comparison of model light curves with X-ray spectra (Badenes et al. 2006; Krause et al. 2008a). Tycho exploded in 1572 AD. Detection of synchrotron emission from a thin shell supports the idea of particle acceleration in young supernova remnants (e.g., Slane et al. 2014). X-ray line emission from intermediate to iron group elements in the interior of the supernova remnant has been found to be clumped (XMM-NewtonMiceli et al. 2015), and the hard X-ray lines of the 44Ti decay chain were also detected (Swift/BAT Troja et al. 2014). These measurements suggest the presence of titanium both in the shocked shell and the interior region of the remnants. However, upper limits obtained from NuSTAR measurements exclude the presence of 44Ti within a 2′ remnant radius at the Swift/BAT detection level over a large range of expansion velocities (Lopez et al. 2015). The distance to the remnant is somewhat uncertain, with estimates ranging from 1.7 to 5.1 kpc (Hayato et al. 2010; Slane et al. 2014; Albinson et al. 1986; Völk et al. 2008). We adopted a distance of 4.1 ± 1 kpc (Hayato et al. 2010).

2.6. Kepler

Johannes Kepler detected this supernova in 1604 AD. Also named G4.5+6.8, this is the youngest Galactic supernova for which an optical transient has been observed. This supernova occurred at a distance of 5.1 ± 0.8 kpc (Sankrit et al. 2016), and is located 460 − 600 pc above the Galactic plane. Due to the detection of strong iron lines in the ejecta, it is attributed to a type Ia explosion (Reynolds et al. 2007), also supported by the absence of a central compact object. 44Ti decay lines have not been found, with upper limits of 1.8 × 10−5 ph cm−2 s−1 in the 1157 keV line (Dupraz et al. 1997) determined from COMPTEL data and 6.3 × 10−6 ph cm−2 s−1 determined from INTEGRAL/IBIS for the 68 and 78 keV line (Tsygankov et al. 2016).

3. Data and analysis method

3.1. Instrument and analysis method

ESA’s gamma-ray space observatory INTEGRAL (Winkler et al. 2003) carries two main instruments on board, the imager IBIS and the spectrometer SPI. The SPI camera (Vedrenne et al. 2003) is a germanium detector array consisting of 19 hexagonally shaped detectors, optimized for high-resolution spectroscopy in the energy range between 18 keV and 8 MeV, with a spectral resolution of ≈2.3 keV (full width at half maximum, FWHM) at 1 MeV. SPI electronics records 16384 energy channels in the range of 18 keV − 2 MeV, which is called SPI’s “low-energy range”. We analyzed SPI data for the signatures of the 44Ti decay lines, which have centroid energies in the laboratory at 68, 78 keV and 1157 keV. Our analysis is focused on the energy bands 30 − 100 keV and 1090 − 1210 keV in order to constrain potential underlying continuum emission at lower energies and account for potential large line broadening at higher energies.

SPI data after initial energy calibration and pre-processing comprises spectra in 0.5 keV bins for each of the 19 detectors, accumulated over exposures of typically 2000 s, called pointings. The orientation of the satellite is shifted by ≈2° between each consecutive exposure in a rectangular-shaped dithering pattern consisting of 5 × 5 sets of coordinates around the observation target. We include data in our analysis in which the celestial objects of interest are within the partially coded field of view of SPI of 34° ×34°. In general, celestial photons entering the aperture of SPI are partially blocked by a coded mask placed 171 cm above the detector plane, imprinting shadow grams on the camera. For an idealized source at long integration times, this creates relative detector intensity distributions (detector patterns), since the absolute number of measured photons per detector is governed by the visibility of the source through the mask (Fig. 1). Sources within a field of view of 16° ×16° are fully coded by the tungsten mask, with decreasing coding fraction towards the coding limit at 34° ×34°.

thumbnail Fig. 1.

Intensity distribution of a celestial source located at the center of the plot (blue dot) (Siegert et al. 2019). The orientation of the SPI instrument is changed in steps of two degrees to visit the locations marked on the map with black dots. The instrument remains centered on each celestial position for ≈2000s, which is called “one pointing”. The source located at the blue position is folded through the instrumental response function to determine the intensity distribution in the detector plane.

The main challenge of SPI data analysis consists of extracting the sparsely populated celestial detector pattern above a large, underlying, instrumental background. The latter is introduced by interaction and activation of satellite and instrument material by cosmic ray bombardment. In our spectroscopic analysis method, we compare the combination of the celestial detector patterns and the background detector patterns to the time series of measured patterns for the 19 detectors by fitting (time dependent) scaling parameters for both contributors. The celestial detector intensity distribution is calculated by applying the energy-dependent image response function (IRF) specific to SPI’s tungsten mask and the re-orientations during the dithering exposures. The model of the background consists of two separate components, one for continuum emission and one for nuclear de-excitation lines at specific energies. Both components are determined over a broad energy range and from multiple years of data. Degradation effects of detectors and time dependent variance of background level are taken into account by modeling the background per detector on an orbital timescale. The mathematical description of our modeling method is given by

(1)

which means that the data and model per energy bin k are represented by the sum of the celestial components i of the total number of Nl celestial sources convolved through the image response function R per detector j and the sum of all background components Nb of detector j. No prior knowledge concerning the energy spectrum of the celestial sources is assumed in our analysis. In general, the model is fit to the data by minimizing the Cash Statistic (Cash 1979), adjusting the scaling parameters θi in Eq. (1), where different timescales for the scaling of the components are allowed. We use the spimodfit analysis tool (Strong et al. 2005; Halloin 2009), which applies a Levenberg-Marquardt algorithm to determine the maximum likelihood solution for all intensity parameters θi. The software is based on the ISDC software spiros (Dubath et al. 2005), however optimized for high spectroscopic resolution of low signal to noise sources. Unless otherwise stated, uncertainties are given as 1σ. We use the Pearson χ2 as an absolute goodness-of-fit criterion. We note that the chosen absolute goodness-of-fit criterion () itself carries an uncertainty (Andrae et al. 2010). The smallest possible timescale we utilized is a single pointing. To minimize the contamination of our data set by known periods of increased background, we excluded orbit phases below 0.10 and above 0.88, during which the satellite passes through the Van Allen radiation belts.

3.2. Background modeling

SPI instrumental background mainly originates from the bombardment of the satellite by cosmic ray particles. Interaction of these cosmic rays can induce nuclear reactions in the satellite materials. Subsequent decays from excited nuclear levels lead to the emission of nuclear de-excitation lines, which fall into the energy range of the SPI detectors. Among others, bremsstrahlung is a second dominant contribution to the background, forming an underlying continuum.

To determine the temporal and spectral behavior of the background, we used the knowledge gained from 17 yr of integrated mission data. Long-term temporal variation is introduced by the degradation of the lattice structure of the Germanium detectors and the absolute production rate of cosmic rays, which is anticorrelated with the solar cycle. Short-term variations are introduced by solar flares.

The detailed spectral shapes of the continuum emission and nuclear de-excitation lines are determined separately. Since a physically based model is difficult to construct and calibrate at the required precision, we used an empirical description of the background. This is based on previous attempts to model the highly variable instrumental background in SPI, as, for example, in Knödlseder et al. (2004) and Jean et al. (2003), and supersedes the standard on-off methods as presented in Dubath et al. (2005). All background components are determined as a linear superposition of an underlying continuum normalized to a central pivot energy, superimposed by emission lines. The line shapes are represented by Gaussian functions convolved with a degradation function, which accounts for the degradation of the germanium charge collection efficiency (Kretschmer 2011). We determine the spectral shape on a 3 d (one orbit) period separately for each detector to trace the time-dependent degradation of the detectors. This timescale is chosen as the best compromise between accumulating sufficient statistics and appropriate determination of temporal variations of the spectral shape. Secondary contributions to the background are smeared out in our coded-mask analysis by accumulating data over multiple pointings (Siegert et al. 2019). The consistency of this high-resolution, time-dependent background modeling approach is demonstrated, for example, in Siegert et al. (2016), Siegert (2017), and Diehl et al. (2018).

3.3. Spectral analysis

With the modest spatial resolution of SPI of ≈3°, supernova remnants cannot be resolved in separate clumps of ejecta. To enhance sensitivity for the relatively low-intensity total celestial signal in a line, a model for the line shape has to be adopted. We describe the emission produced by radioactive decay with Gaussian shaped lines, plus a power-law-shaped continuum.

(2)

where F0 is the measured line flux, E0 is the energy of the Doppler shifted-line centroid, and σ is line width. We interpret any broadening of the line, which would be additional to the detector resolution, as Doppler broadening caused by the expansion velocity of the ejecta. We determine the line parameters separately for each line when possible. No changes in the kinematics of the ejecta are expected, allowing for a combined three-line fit, assuming identical Doppler parameters for all three lines. We further allow for an underlying celestial continuum accounting for bremsstrahlung processes, with normalization parameter A0 and power-law index α. The mass of the ejected 44Ti per line is determined by

(3)

where FL is the flux of the specific line L = (68, 78, 1157) keV, d is the distance to the source, N = 44 is the number of nuclei in 44Ti, u is the atomic mass number, τ = 86.6 yr is the decay constant of 44Ti, and t is the age of the supernova remnant. All line fluxes F0 are normalized with the branching ratio bL1 of the specific line L. For comparison, all flux values are stated as the normalized flux FL = F0/bL. In a multiline fit, the branching ratios determine the relative line intensities, and the normalized flux FL is fit as the parameter of interest. Velocities corresponding to the Doppler broadening of the lines are calculated for the ejecta from the FWHM of the line, assuming a uniformly expanding sphere. While this model might not adequately describe asymmetries as seen in the supernova remnant Cassiopeia A (Grefenstette et al. 2017), it provides a reasonable first-order approach for determining fluxes from the remaining unresolved sources. Confidence intervals for our results of the spectral fits are estimated from the 68th percentile interval of a Metropolis-Hastings algorithm, minimizing the Pearson χ2 as test statistics. Upper limits are given at 2σ (i.e., Δχ2 = 4, for one degree of freedom, d.o.f.). We derived our upper limits by varying only the integrated flux of the respective line, assuming values for Doppler broadening and shift, and utilizing the best fit values for the underlying continuum.

4. SPI results

4.1. Cassiopeia A

We used all available data for Cassiopeia A up to 2019 AD, containing a total exposure of 11.2 Ms for our analysis. Figure 2 shows the spectrum of Cassiopeia A in the energy ranges of interest. The average reduced χ2 per fit energy bin is 1.001 (χ2/d.o.f. = 92 772/92 658). We adopted a uniform power law underlying the line emission across the entire energy range between 30 keV and 1200 keV with a fit power-law index of α = −2.6 ± 0.4. The line signal with the highest significance for a single Gaussian shaped line is found for the 78 keV line, with a significance of 3.6σ. The strong background lines of germanium lead to relatively large flux variations in the energy range between ≈50 − 65 keV, to such an extent that the 68 keV decay line is only marginally detectable.

thumbnail Fig. 2.

Spectra for Cassiopeia A in the energy region 50–100 keV in 1 keV binning and 1090–1210 keV in 10 keV binning containing all potential decay lines from the decay chain of 44Ti. We fit the spectrum with a power law accounting for underlying continuum most likely produced by synchrotron emission at the shock front and Gaussian shaped line profiles in each region. The lines are Doppler shifted and broadened, which cannot solely be explained by the instrumental resolution at the energies, respectively. The lines are determined from a combined fit. This means that we fit one uniform power law over the entire energy range and identical Doppler parameters and integrated flux values for all three lines simultaneously. Flux values are corrected for the respective branching ratios of the lines.

In order to validate our findings and to avoid spurious detection, we searched for an emission that could mimic 44Ti decay from any celestial point source in a 20° ×20° area centered at Cassiopeia A. Possible emission was determined at locations in a spherical square-shaped grid, where each point is separated by 2° from the adjacent point, yielding a total of 121 grid points. Figure 3 shows the map of source significances for a 78+1157 keV signal. Detection at 4.9σ is only found at the location consistent with Cassiopeia A, while other test points do not show significant signals. We additionally show (Fig. 3) the 1σ uncertainty band determined form these 121 source locations, which, if interpreted as statistical signal fluctuations of a zero signal, bracket the spectrum of Cassiopeia A. A significant excess above the uncertainty band is clearly visible at the 78 keV line, while the majority of the 68 keV line lies within the statistical background-uncertainty band.

thumbnail Fig. 3.

Upper panel: significance map of 44Ti decay evaluated for the 78 and 1157 keV line with color-coding in significance levels. Excess at the 4.9σ level is only found at the location of Cassiopeia A (center). The second red cross marks the location of Tycho’s supernova remnant. Lower panel: spectrum of Cassiopeia A is shown in black. The red shaded area contains the uncertainty band from the spectra obtained in a 20° ×20° area centered at Cassiopeia A. The gray spectrum shows the mission integrated background. It is evident that the uncertainty scales with the strength of the background leading to an increased uncertainty in the region 50–65 keV. Arrows indicate the laboratory-determined centroid energies of both 44Ti decay lines. Clustered excess over the uncertainty band is found in the vicinity of the 78 keV line. Unfortunately, the 68 keV emission is located at the edge of a strong background line complex allowing for only a marginal detection probability of the line.

The 78 keV line, if represented by a Gaussian-shaped line, represents an integrated flux of (3.3  ±  0.9)  ×  10−5 ph cm−2 s−1 and a 3.6σ detection level. With Eq. (3), this flux corresponds to an ejected 44Ti mass of (2.1  ±  0.6)  ×  10−4 M⊙, for a remnant age of 330 yr and a distance of 3300 ± 100 pc. The uncertainty is mainly due to the uncertainty in the determined flux, however, the uncertainty in the distance estimate is also incorporated in the result. The line is centered at (77.7  ±  0.5) keV, which is slightly red-shifted with respect to the laboratory-determined decay energy of 78.3 keV (Firestone et al. 2003). This Doppler shift translates into a bulk motion of (2400 ± 1500) km s−1 away from the observer. The FWHM of the line is (2.3 ± 0.8) keV. This is broadened with respect to the instrumental resolution of 1.6 keV FWHM at 78 keV. We interpret this broadening of the line as Doppler broadening due to the expansion of the supernova remnant; this translates the FWHM of the line into an expansion velocity of (5500 ± 2700) km s−1.

Representing the 1157 keV line with a simple Gaussian on top of the uniform power law, our best fit values suggest a significantly higher flux of (9.5 ± 3.0) × 10−5 ph cm−2 s−1. This flux can be overestimated due to the assumption of a single underlying continuum fit across the broad energy range. The continuum flux from our best fit power law (largely determined at energies below 100 keV) in the energy region between 1125–1175 keV is 1.0 ± 7.0 × 10−7 ph cm−2 s−1, consistent with zero. We estimated the potential offset in the fit continuum flux density by allowing a separate, constant offset in the energy range 1090–1210 keV, which accounts for a flux of 3.1 ± 1.5 × 10−5 ph cm−2 s−1 in the energy region between 1125–1175 keV. Therefore, the line flux at 1157 keV can be reduced to (6.4 ± 3.4) × 10−5 ph cm−2 s−1. The Gaussian is centered at (1151 ± 7.9) keV, which is red shifted but compatible with the laboratory-determined energy of 1157 keV. The line width is (40.0 ± 6.7) keV FWHM. This corresponds to (8900 ± 1500) km s−1 expansion velocity.

We find a combined signal with a significance of 4.9σ by simultaneously fitting two lines with identical Doppler shift, Doppler broadening, and integrated flux. The overall Doppler shift of the lines corresponds to a bulk ejecta velocity of (1800 ± 800) km s−1. The Doppler broadening for both lines translates to (6400 ± 1900) km s−1 expansion velocity, in agreement with values determined for the 78 keV line alone. Due to the higher relative spectral resolving power of SPI at higher energies, the expansion velocity can be better constrained including the 1157 keV line. The combined fit contains a flux of (4.2 ± 1.0) × 10−5 ph cm−2 s−1 per line. This higher flux corresponds to a 44Ti mass of (2.6 ± 0.6) × 10−4 M⊙. Table 2 contains measured line parameters and derived physical quantities.

Table 2.

Line shape parameters and derived quantities for the 44Sc and 44Ti decay in Cassiopeia A.

Even though the uncertainties are high, we also include the 68 keV line in our analysis. However, we adopted the kinematic values determined from the combined 78 and 1157 keV line fit for this line, as the strong fluctuations induced by the strong background lines might lead to an artificial broadening of the line. The 68 keV line is then observed with a single-line significance of 2.2σ. When linked to a common origin, the total significance for the three fit lines is then increased to 5.4σ.

4.2. SN 1987A

SPI was pointed towards the LMC including SN 1987A for a total of 7 Ms. With SPI’s angular resolution of ≈2.7° , we cannot distinguish between SN 1987A and other potential or known sources of high energy emission. In particular, the pulsar PSR B0540-69 and the high-mass X-ray binary LMC-X1 are located less than 1° apart from SN 1987A. We obtain an average reduced χ2 of 1.000 (χ2/d.o.f. = 51 716/51 731) per energy bin. In Fig. A.4, we show the spectrum obtained from our data for a source located at the position of SN 1987A. We find no significant flux excess in either energy region that could be attributed to the decay chain emission of 44Ti. Tueller et al. (1990) determined an expansion velocity of 3100 km s−1 from the line profiles of measured radioactive 56Co. Assuming co-moving 44Ti ejecta, we adopted this value to determine 2σ upper limits on the flux. In our analysis, we searched for the combined signal of all three lines simultaneously, for which we determined a value of 1.8 × 10−5 ph cm−2 s−1 per line, corresponding to an upper mass limit of ejected 44Ti of 6.9 × 10−4 M⊙ for a distance of 49.6 kpc, and a remnant age of 24 yr. No systematically increased flux is observed in the 1157 keV line with respect to the lines at 68 and 78 keV.

4.3. Vela Junior

We modeled the supernova remnant as a source of extended emission with a 2D Gaussian emission profile and a diameter of 0.6° for the width of the remnant. This means that the 2° diameter contains ≈90% of the expected 44Ti signal. The obtained fit is satisfactory with a reduced χ2 of 0.998 (χ2/d.o.f. = 65 423/65 535) per energy bin. We find no signal for the decay of 44Ti (Fig. A.3). We determined a 2σ upper limit of 2.1 × 10−5 ph cm−2 s−1 for the combined signal of all three lines, assuming no bulk motion and an expansion velocity of 3000 km s−1. This corresponds to an upper limit for the ejected 44Ti mass of 3.3 × 10−5 M⊙ for the remnant age and distance of 690 yr and 200 pc, respectively. Considering the updated age and distance estimates of 2.4–5.1 kyr and 700 ± 200 pc (Allen et al. 2014), the ejecta mass limits determined from our results significantly increase to a value ≤2.2 × 10−1 M⊙ for the lower age limit of 2.4 kyr.

4.4. Tycho’s supernova remnant

SPI was pointed towards the region containing Tycho’s supernova remnant for a total of 10 Ms. Figures A.1 and A.2 show the spectra obtained for Tycho in both energy regions relevant for our 44Ti search. The average reduced χ2 is 0.997 (χ2/d.o.f. = 79 982/80 196) per energy bin. We find no significant excess for the emission in the three lines of the 44Ti decay chain. To determine our upper limits, we adopted an expansion velocity of 5000 km s−1. This value is in agreement with the expansion velocities found in the central ejecta Sato & Hughes (2017). We determined a 2σ upper limit of 1.4 × 10−5 ph cm−2 s−1 for each line in the 44Ti decay, corresponding to an 44Ti ejecta-mass limit of 4.8 × 10−4 M⊙ for a distance of 4.1 kpc and a remnant age of 438 yr.

4.5. G1.9+0.3

Due to its location close to the Galactic center, several hard X-ray sources (Bird et al. 2016) are present within the 2.7° PSF of SPI around the position of G1.9+0.3. No signature for the decay of 44Ti is expected for these other sources, so it appears a safe assumption that potential flux excess in the 68 and 78 keV regime can be attributed to the emission from G1.9+0.3. The average reduced χ2 is 1.011 (χ2/d.o.f. = 318 356/315 005). We interpret this as due to the possible presence of unresolved sources in SPI’s field of view in the Galactic central region. The resulting spectra (Figs. A.1 and A.2) show an underlying continuum from the spatially coincident sources, upon which we search for the imprints of the three decay lines.

We find no significant excess in both energy ranges, determining a corresponding upper limit of 1.0 × 10−5 ph cm−2 s−1 for an assumed expansion velocity of 5000 km s−1 for 44Ti containing ejecta. This translates into a 44Ti yield of 0.3 × 10−4 M⊙ for a remnant age of 120 yr and a distance of 8.5 kpc. Even though this velocity is lower than the expansion of the remnant’s blast wave, we believe our assumption is plausible, as the distribution of the ejecta containing radioactive 44Ti is uncertain anyway and may consist of clumps as seen for Cassiopeia A.

We determined velocity-dependent limits, which depend on the expected line width for expansion velocities between 0 and 15 000 km s−1. We obtain limits in the range from 0.7 to 1.5 × 10−5 ph cm−2 s−1, assuming the same Doppler velocities for all three lines.

4.6. Kepler’s supernova remnant

In our analysis, we do not find emission from Kepler’s supernova remnant in the two energy bands (Figs. A.1 and A.2). The (2σ) upper limit is 1.1 × 10−5 ph cm−2 s−1. For a remnant age of 406 yr and distance of 5.1 kpc, the flux limit corresponds to a 44Ti ejecta mass limit of 4.0 × 10−4 M⊙.

5. Discussion

5.1. Cassiopeia A

For six analyzed supernova remnants, we find significant detection only for Cassiopeia A, with an integrated flux of (4.2 ± 1.0) × 10−5 ph cm−2 s−1 corresponding to an 44Ti ejecta mass of (2.6 ± 0.6) × 10−4 M⊙.

Conventional models of core-collapse supernova explosions (Timmes et al. 1996; Magkotsios et al. 2010), including models specific to the progenitor evolution of Cassiopeia A (Young et al. 2006), suggest 44Ti ejecta of less than 1.0 × 10−4 M⊙, significantly lower than the amount that we determine for Cassiopeia A. This underproduction of 44Ti in models is also supported by measurements from other instruments, consistently showing higher ejecta mass for Cassiopeia A ((2.4 ± 0.9) × 10−4 M⊙Siegert et al. 2015) and ((1.5 ± 0.2) × 10−4 M⊙; Grefenstette et al. 2017) than the values from models of ≤1.0 × 10−4 M⊙ obtained for a 30 M⊙ star (Timmes et al. 1996; Limongi & Chieffi 2018). Figure 4 shows the expected yield of 44Ti for various supernova scenarios. The green shaded area represents standard, mostly piston-driven2 explosion models, in which nucleosynthesis is calculated by post-processing from the modeled thermodynamic evolution of the remnant. Harris et al. (2017) point out that including nucleosynthesis networks into the simulation, rather than post processing yields, can change the production of alpha nuclei, especially at intermediate mass range A = 36–52, by an order of magnitude. The high 44Ti mass seen in Cassiopeia A shows that a more detailed treatment of explosive nucleosynthesis appears necessary. The 3D model of Wongwathanarat et al. (2017), representing the special case of Cassiopeia A, suggests higher 44Ti masses, especially considering the clumpy and asymmetric distribution in the supernova remnant. The measured expansion velocity (6400 ± 1900) km s−1 is compatible with models including Rayleigh-Taylor instabilities that lead to large-scale mixing of the inner ejecta with overlying stellar shells in type II-b supernova models (Nomoto et al. 1995). We determined a bulk motion of (2200 ± 1300) km s−1. This suggests that the bulk of the ejecta is receding from the observer. Both the kinematics and the ejected mass of 44Ti support the interpretation that Cassiopeia A is an asymmetric supernova explosion. While we determined kinematic constraints from a spectral analysis, other evidence for an asymmetric explosion is provided from the spatially resolved analysis of the remnant with the NuSTAR telescope. Grefenstette et al. (2017) have found that the majority of the 44Ti -containing ejecta is expelled in a large solid angle, where the bulk of the ejecta moves away from the observer. Our measurements and the resulting velocity spread, determined from the Doppler broadening, is consistent with NuSTAR findings, which suggest a clumped nature of 44Ti -containing ejecta (Grefenstette et al. 2017). Despite concurring kinematic constraints, we determine a higher integrated flux in the 44Ti decay lines. Due to the different angular resolutions of the NuSTAR telescope (18″ FWHM; Harrison et al. 2013) and the SPI spectrometer (2.7° FWHM; Vedrenne et al. 2003), different spherical surface areas for the integration of the flux are considered in both analyses. Grefenstette et al. (2017) considered the flux of an integrated emission from a region of 120″ radius centered on Cassiopeia A, containing all spatial points in which emission from 44Ti decay is detected in their analysis, and they gave upper limits on regions outside the 120″ radius. X-ray measurements both suggest a forward shock radius of 153″ (Gotthelf et al. 2001) and the presence of iron at radii between 110″–170″ (Willingale et al. 2002). Co-moving 44Ti can be present at large radii extending as far outwards as the observed iron distribution. Flux from these unresolved regions contributes to the total flux of the 44Ti emission. Within the SPI’s angular resolution, the entire surface area of the remnant is included, constituting the increased integrated flux we measured in our analysis. The increased line-of-sight beam width in SPI of 2.7° FWHM also includes areas outside of the supernova remnant. This also means that unresolved or previously unknown sources can contribute to the total flux in SPI measurements.

thumbnail Fig. 4.

Left panel: comparison of upper limits and detection of Cassiopeia A with predictions made from core-collapse supernova models. Right panel: predicted 44Ti ejecta yields for different type Ia supernova models. Upper limits obtained for the three youngest type Ia supernovae exclude the double detonation model and faint SN2005E-like scenarios. We include the detections of 44Ti in Cassiopeia A and SN1987 A with the NuSTAR and INTEGRAL/IBIS telescopes in green and blue, respectively. Upper limits from further NuSTAR and IBIS observation are additionally included in green and blue (see Table 3 for references). Masses are determined for the distances in Table 1 and remnant age at the average observation date in the cited publications.

For the first time, we also identified a very broad decay signature in the high-energy decay line at 1157 keV, which also reveals the kinematic evolution of the supernova remnant. We determine an expansion velocity of (8900  ±  1500) km s−1 and a line that is not significantly redshifted with (1600  ±  2000) km s−1. This line contains an integrated flux of (9.5  ±  3.0) × 10−5 ph cm−2 s−1 ((6.4  ±  3.4) × 10−5 ph cm−2 s−1). A systematic offset of the high-energy line has been also observed with COMPTEL (Iyudin 1999) and with earlier SPI data (Siegert et al. 2015), however with less exposure on Cassiopeia A. In contrast to our updated values, Siegert et al. (2015) provided a more constraining expansion velocity and an overall different kinematic behavior of 44Sc decay in comparison to the 78 keV line for a line centered at (1158.0 ± 3.6) keV. As shown by Grefenstette et al. (2017)44Ti is ejected in clumps in Cassiopeia A. Each clump translates into a separate peak in the energy range between 1130–1180 keV, which blends into a broadened line. We described the entire emission by one Gaussian, which captures the overall expansion of the entire remnant. The line measured by Siegert et al. (2015) in the narrow energy window around 1157 keV only captures parts, or one separate ejecta clump, of the total emission in the 1157 keV regime. We find an increased flux in the 1157 keV line, that cannot be explained by systematic effects alone. We speculate that the flux included in this line could be enhanced for different reasons: (1) Excitation of the nuclear transition in 44Ca, in addition to the decay of 44Sc. Interaction with ambient material can lead to excitation of the nucleus. An excitation of the stable 44Ca nucleus by cosmic rays in the shock region of the supernova envelope might thus contribute to the flux in the 1157 keV line. This mechanism would only influence the flux of the 44Ca∗ line, as the half life of 44Sc is too short for efficient cosmic-ray-induced excitation. More analysis, in particular of other candidate nuclear de-excitation lines, is required to support this hypothesis. The most promising approach would be the detection of the de-excitation lines at 4.4 MeV and 6.1 MeV, which are the most prominent de-excitation lines caused by cosmic ray interaction in the shock front (Summa et al. 2011). A first search for these lines shows that flux values as high as those suggested by Summa et al. (2011) can be excluded. (2) McKinnon et al. (2016) pointed out that two thirds of dust in the Milky Way-like galaxies can be produced by type II supernova events. The presence of dust grains composed of ejecta material in the vicinity of the supernova remnant or in the line of sight towards the supernova remnant can alter the observed flux ratios beyond the branching ratios. Attenuation coefficients for 68 and 78 keV photons are higher than for the 1157 keV line (Iyudin et al. 2019) for common dust grain compositions. Including correction for branching ratios, we derived the following flux ratio:

(4)

This suggests that 16–80% of the emission in the 78 keV line could be absorbed by dust, located between INTEGRAL and Cassiopeia A. (3) The assumption of a Gaussian-shaped line for the total emission does not correctly represent the ejecta kinematics as found in Cassiopeia A. This can artificially lead to an increased flux in the 1157 keV line.

5.2. SN 1987A

We determine an upper limit of 1.8 × 10−5 ph cm−2 s−1 per decay line, assuming all lines share identical Doppler characteristics. We attribute this flux to the decay of 44Ti, corresponding to an upper ejecta mass limit of 6.9 × 10−4 M⊙ for a remnant age of 24 yr and a distance of 49.6 kpc. We find no evidence for a systematically increased flux in the high energy line at 1157 keV. Assuming that the 44Ti ejecta are contained in the central region of the supernova, the expansion velocity of the 44Ti ejecta should be lower than 1800 km s−1 (McCray 2017). In addition to an upper limit determined from the combined three lines, we give upper limits only for the 1157 keV line for expansion velocities corresponding to the interior of the supernova core (vexp ≤ 1800 km s−1). For this velocity range, we determine flux limits that range between (1.7 − 3.2) × 10−5 ph cm−2 s−1. The flux limit we derived for the 44Ti decay chain combined fit is compatible with direct detection of 44Ti, as found by IBIS/INTEGRAL and NuSTAR (Grebenev et al. 2012; Boggs et al. 2015). Both analyses suggest narrow line broadening, which is compatible with slowly expanding 44Ti ejecta. Given the NuSTAR and IBIS/INTEGRAL fluxes, a significant offset in flux of the 1157 keV line, either from less efficient absorption at higher energies or from an additional excitation, as seen in Cassiopeia A, should be detectable, albeit with small significance, within the SPI’s sensitivity.

5.3. Vela Junior

Vela Jr. still poses a mystery 20 yr after the serendipitous detection of gamma-ray emission in the 1157 keV 44Ca line by COMPTEL (Iyudin et al. 1998). Our upper limit of 2.1 × 10−5 ph cm−2 s−1 assumes an extended source for the combined signal in all three decay lines, excluding a signal at the level reported by Iyudin et al. (1998) from COMPTEL data. Emission at the COMPTEL level was also excluded from the nondetection of the scandium fluorescence line Slane et al. (2001). More recent studies with IBIS/INTEGRAL (Tsygankov et al. 2016) found no excess in the energy bands of the 68 and 78 keV lines, with an upper limit for the 44Ti flux of 1.8 × 10−5 ph cm−2 s−1, which excludes the COMPTEL detection. Tsygankov et al. (2016) point out that they considered the remnant as a point-like source neglecting the apparent 2° diameter of the remnant. Measurements of the radial displacement in the northern rim of Vela Jr. suggest that the remnant age is 2.4 − 5.1 kyr (Allen et al. 2014) at a distance of 0.5 − 1.0 kpc, which is in contrast with the 0.7 kyr age and 200 pc distance estimates discussed by Iyudin et al. (1998). Our results substantiate the higher age and larger distance.

5.4. Tycho’s supernova remnant

Our upper limit of 1.4 × 10−5 ph cm−2 s−1 for the Tycho supernova remnant is in agreement with the upper limits determined with NuSTAR and INTEGRAL/IBIS (Wang & Li 2014; Lopez et al. 2015) of ≥10−5 ph cm−2 s−1 for moderate expansion velocities, with 44Ti spatially distributed over the entire remnant. However, detection has been claimed from observations with Swift/BAT (Troja et al. 2014) at a flux of 1.3 × 10−5 ph cm−2 s−1 and 1.4 × 10−5 ph cm−2 s−1 for the 68 and 78 keV lines, respectively. For an assumed distance of 4.1 kpc, the upper limits correspond to an ejecta mass of the order of 10−4 M⊙ of 44Ti. Lopez et al. (2015) provided an ejecta-mass upper limit of 2.4 × 10−4 M⊙ of 44Ti for a distance of 2.3 kpc. In all cases, we can exclude the double detonation and Ca-rich models as explosion scenario for this type Ia explosion remnant. The results are, however, in agreement with delayed detonation models. This model may be favored as it best reproduces the measured X-ray spectra from Tycho (Badenes et al. 2006).

5.5. G1.9+0.3

We find no significant excess of 44Ti line emission at the position of the supernova remnant G1.9+0.3. We determine an upper flux limit of 1.0 × 10−5 ph cm−2 s−1. This translates into a mass limit of 0.3 × 10−4 M⊙ at a distance of 8.5 kpc and a remnant age of 120 yr, which excludes both double detonation and Ca-rich models for this candidate type Ia explosion. Predictions for classical delayed detonation models (Maeda et al. 2010) suggest less than 10−5 M⊙ of 44Ti, which would be in agreement with our results. We compare our results with the signal detected by Borkowski et al. (2010). Using Chandra data, they reported a line at 4.1 keV, which they attribute to a fluorescence transition in 44Sc following the electron capture decay on 44Ti. Their inferred mass of (1 − 7) × 10−5 M⊙ of ejected 44Ti translates into a flux of (0.3 − 1.9) × 10−5 ph cm−2 s−1 in the decay lines at 68, 78, and 1157 keV. Our upper limits (0.7 to 1.5 × 10−5 ph cm−2 s−1) are in conflict with the extrapolated fluxes ((0.3 − 1.9) × 10−5 ph cm−2 s−1) expected from the 44Sc fluorescence line for expansion velocity below ∼15 000 km s−1. The discrepancy between the 4.1 keV fluorescence line and the decay emission at hard X-ray energies suggests that the fluorescence line is not necessarily produced from the decay of 44Ti alone. Since the line is produced by the emission of a Kα photon, which is independent of the scandium isotope, other stable isotopes (e.g., 45 Sc ) co-produced in the supernova explosion can contribute to the fluorescence emission at 4.1 keV.

Zoglauer et al. (2015) determined a 2σ upper limit for the flux in the 68 keV line of 1.5 × 10−5 ph cm−2 s−1 for a non-shifted line with a 4 keV width (1σ), using measurements of the NuSTAR telescope. Results from the IBIS telescope (Tsygankov et al. 2016) yield a 3σ upper limit of 9 × 10−6 ph cm−2 s−1.

5.6. Kepler’s supernova remnant

We determined a flux limit for Kepler of 1.1 × 10−5 ph cm−2 s−1. The derived mass limit disagrees with the double detonation and Ca-rich scenario; however, due to the uncertain distance to the remnant, ranging between 4.4 kpc and 5.9 kpc, the double detonation scenario cannot be explicitly excluded. Our result is in agreement with the nondetection of 44Ti emission with the COMPTEL and IBIS instrument (Iyudin 1999; Dupraz et al. 1997; Tsygankov et al. 2016).

5.7. Implications for supernova models

The detection of 44Ti in only one Galactic supernova remnant provides a striking and significant conclusion regarding supernovae in the Galaxy. With an average core-collapse supernova rate of ≈1 − 3 century−1 (Diehl et al. 2006; van den Bergh & Tammann 1991), and current understanding of nucleosynthesis in supernovae, five supernova remnants with a 44Ti -decay line flux of more than 10−5 ph cm−2 s−1 are expected to be visible in the Galaxy. Detection of a single remnant at the position of Cassiopeia A is unlikely, with a probability of less than 2.7% (The et al. 2006; Dufour & Kaspi 2013). With the high 44Ti mass measured in Cassiopeia A (and also SN 1987A), it is possible that the 44Ca content in the Galaxy is produced by a few, rare, 44Ti producing supernovae. It is possible that both Cassiopeia A and SN 1987A are the prototypes for asymmetric explosions producing a high ejecta 44Ti mass, whereas the majority of core-collapse supernovae explode in a more symmetric scenario, producing less 44Ti ejecta.

We use the solar abundance value [44Ca/56Fe]⊙ = 1.2 × 10−3 (Anders & Grevesse 1989) of the 44Ca to 56Fe ratio, which are the end products of the 44Ti and 56Ni decay chain respectively, as another criterion to judge supernova model types, assuming that each of those supernova types would be the sole source of 44Ca and 56Fe as found in the Sun. Figure 5 shows the 44Ti to 56Ni ratio of the candidate sources in our analysis (dot symbols), together with modeled values for several supernova types (star symbols), and the solar 44Ca to 56Fe ratio (red line).

thumbnail Fig. 5.

[45Ti/56Ni] ratio of our mass estimates and several supernova explosion models. The red line marks the solar [44Ca/56Fe] ratio (1.2 × 10−3; Anders & Grevesse 1989), which is used as a reference criterion to judge supernova model subtypes. Ejected 56Ni masses for Cassiopeia A, SN 1987A, Tycho, Kepler and G1.9+0.3 are taken from (Eriksen et al. 2009; Woosley et al. 1987; Badenes et al. 2006; Patnaude et al. 2012; Borkowski et al. 2013), respectively. Vela Jr. 56Ni ejecta mass is modeled from a 25 M⊙ star (Maeda & Nomoto 2003). Model yields are from Wongwathanarat et al. (2017), Maeda & Nomoto (2003), Seitenzahl et al. (2013), Fink et al. (2014, 2010).

For Vela Jr., we adopted a 56Ni ejecta-mass estimate from the explosion model of a 25 M⊙ (Maeda & Nomoto 2003). We note that the values for Vela Jr. may deviate from the solar ratio due to the assumed distance and age (200 pc; 690 yr), which underestimates the updated values determined by Allen et al. (2014). The ejected 56Ni mass for Cassiopeia A is inferred from near-infrared spectral analysis (Eriksen et al. 2009). For SN 1987A, we used the 56Ni ejecta-mass estimate based on the early bolometric light curve by Woosley et al. (1987). For type Ia supernova remnant candidates, 56Ni masses are obtained from Badenes et al. (2006, Tycho), Patnaude et al. (2012, Kepler), and (2013, G1.9+0.3), who inferred ejecta masses by comparing measured X-ray spectra with long-term simulated remnant models.

We infer from Fig. 5 that the measured 44Ti-to-56Ni ratios of Cassiopeia A and SN 1987A are plausibly consistent with asymmetric core-collapse supernova models being responsible for the solar abundance ratio.

In contrast, the majority of type Ia supernova scenarios do not plausibly reproduce the measured solar 44Ca to 56Fe ratio. Type Ia supernovae appear to consistently produce a 56Ni ejecta mass between 10−1 − 100 M⊙ (Dhawan et al. 2016; Wang et al. 2008; Stritzinger et al. 2006). In model calculations, nucleosynthesis yields of intermediate mass elements (such as 44Ti) are highly dependent on physical conditions during nuclear burning. In centrally ignited, pure-deflagration, and delayed-detonation scenarios, nucleosynthesis occurs mainly in a high-density regime, producing only a small amount of 44Ti, while the majority of nuclear fuel is completely burned to iron group elements (Seitenzahl et al. 2013; Fink et al. 2014). On the other hand, nucleosynthesis in double-detonation supernova scenarios also occurs during the initial burning of the surface helium layer at lower densities, allowing for the production of large amounts of 44Ti on the surface (Fink et al. 2010; Sim et al. 2012). Pure-deflagration and delayed-detonation type Ia supernovae (green and red star symbols, Fig. 5) therefore could not be major contributors to solar 44Ca, while double detonation supernovae (blue star symbols in Fig. 5) could. Our constraints for all three candidate remnants of type Ia supernovae show that they cannot be sole sources of the solar abundance of 44Ca and 56Fe together.

Under the assumption that core-collapse supernovae are, in general, responsible for the solar [44Ca/56Fe]⊙ ratio, we can use our 44Ti limits and the average simulated 44Ti and 56Ni yields to constrain type Ia sub-type rates: We obtain a ratio ≥2.6 : 1 of delayed detonation/deflagration to double detonation events, necessary to maintain and not violate the solar [44Ca/56Fe]⊙ ratio.

5.8. Summary

In this work, we searched for the signature of gamma rays produced in the decay chain of 44Ti in the six young nearby supernova remnants Cassiopeia A, SN 1987A, Vela Jr., Tycho’s supernova, Kepler’s supernova, and G1.9+0.3. In Table 3, we list the mass estimates we derived from SPI/INTEGRAL data acquired over the entire mission duration of 17 yr.

Table 3.

Values for the masses and fluxes of the six young supernova remnants.

We only detect emission in the supernova remnant Cassiopeia A. Inferred masses of more than 2 − 3 × 10−4 M⊙ ejected 44Ti exceed theoretical predictions. Upper limits determined for Vela Jr. exclude the detection of the 44Sc decay line found with COMPTEL by Iyudin (1999). The detection of 44Ti hard X-ray lines in SN 1987A (Grebenev et al. 2012; Boggs et al. 2015) cannot be confirmed. We exclude models predicting high yields of 44Ti such as the double-detonation models and Ca-rich explosion (Waldman et al. 2011; Perets et al. 2010) model for the three Galactic thermonuclear supernova remnants G1.9+0.3, Tycho, and Kepler.

1

This is equal to the probability of photon emission per decay (93.0%, 96.4%, 99.9%, respectively; see Sect. 1).

2

Models that artificially inject the energy necessary for explosion.

Acknowledgments

The INTEGRAL/SPI project has been completed under the responsibility and leadership of CNESS; we are grateful to ASI,CEA, CNES, DLR (Nos. 50OG 1101 and 1601), ESA, INTA, NASA and OSTC for support of this ESA space science mission. Thomas Siegert is supported by the German Research Society (DFG-Forschungsstipendium SI 2502/1-1).

References

  1. Aharonian, F., Akhperjanian, A. G., Bazer-Bachi, A. R., et al. 2007, ApJ, 661, 236 [NASA ADS] [CrossRef] [Google Scholar]
  2. Ahmad, I., Greene, J. P., Moore, E. F., et al. 2006, Phys. Rev. C, 74 [CrossRef] [Google Scholar]
  3. Alarie, A., Bilodeau, A., & Drissen, L. 2014, MNRAS, 441, 2996 [Google Scholar]
  4. Albinson, J. S., Tuffs, R. J., Swinbank, E., & Gull, S. F. 1986, MNRAS, 219, 427 [NASA ADS] [CrossRef] [Google Scholar]
  5. Allen, G. E., Chow, K., DeLaney, T., et al. 2014, ApJ, 798, 82 [NASA ADS] [CrossRef] [Google Scholar]
  6. Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197 [Google Scholar]
  7. Andrae, R., Schulze-Hartung, T., & Melchior, P. 2010, ArXiv e-prints [arXiv:1012.3754] [Google Scholar]
  8. Aschenbach, B. 1998, Nature, 396, 141 [NASA ADS] [CrossRef] [Google Scholar]
  9. Aschenbach, B., Iyudin, A. F., & Schönfelder, V. 1999, A&A, 350, 997 [Google Scholar]
  10. Audi, G., Bersillon, O., Blachot, J., & Wapstra, A. H. 2003, Nucl. Phys. A, 729, 3 [NASA ADS] [CrossRef] [Google Scholar]
  11. Badenes, C., Borkowski, K. J., Hughes, J. P., Hwang, U., & Bravo, E. 2006, ApJ, 645, 1373 [NASA ADS] [CrossRef] [Google Scholar]
  12. Beckmann, V., Gehrels, N., Favre, P., et al. 2004, ApJ, 614, 641 [NASA ADS] [CrossRef] [Google Scholar]
  13. Bird, A. J., Bazzano, A., Malizia, A., et al. 2016, VizieR Online Data Catalog, J/ApJS/223/15 [Google Scholar]
  14. Blondin, J. M., & Shaw, S. 2007, ApJ, 656, 366 [NASA ADS] [CrossRef] [Google Scholar]
  15. Boggs, S. E., Harrison, F. A., Miyasaka, H., et al. 2015, Science, 348, 670 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  16. Borkowski, K. J., Reynolds, S. P., Green, D. A., et al. 2010, ApJ, 724, L161 [NASA ADS] [CrossRef] [Google Scholar]
  17. Borkowski, K. J., Reynolds, S. P., Hwang, U., et al. 2013, ApJ, 771, L9 [NASA ADS] [CrossRef] [Google Scholar]
  18. Bouchet, L., Jourdain, E., & Roques, J.-P. 2015, ApJ, 801, 142 [NASA ADS] [CrossRef] [Google Scholar]
  19. Bruenn, S. W., Mezzacappa, A., Hix, W. R., et al. 2013, ApJ, 767, L6 [NASA ADS] [CrossRef] [Google Scholar]
  20. Bruenn, S. W., Lentz, E. J., Hix, W. R., et al. 2016, ApJ, 818, 123 [Google Scholar]
  21. Buras, R., Rampp, M., Janka, H. T., & Kifonidis, K. 2006, A&A, 447, 1049 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  22. Burrows, A., Vartanyan, D., Dolence, J. C., Skinner, M. A., & Radice, D. 2018, Space Sci. Rev., 214 [Google Scholar]
  23. Cash, W. 1979, ApJ, 228, 939 [NASA ADS] [CrossRef] [Google Scholar]
  24. Chen, J., Singh, B., & Cameron, J. A. 2011, Nucl. Data Sheets, 112, 2357 [NASA ADS] [CrossRef] [Google Scholar]
  25. Chen, M. C., Herwig, F., Denissenkov, P. A., & Paxton, B. 2014, MNRAS, 440, 1274 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  26. Dhawan, S., Leibundgut, B., Spyromilio, J., & Blondin, S. 2016, A&A, 588, A84 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  27. Diehl, R., Halloin, H., Kretschmer, K., et al. 2006, Nature, 439, 45 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  28. Diehl, R., Siegert, T., Greiner, J., et al. 2018, A&A, 611, A12 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  29. Dubath, P., Knödlseder, J., Skinner, G. K., et al. 2005, MNRAS, 357, 420 [NASA ADS] [CrossRef] [Google Scholar]
  30. Dufour, F., & Kaspi, V. M. 2013, ApJ, 775, 52 [NASA ADS] [CrossRef] [Google Scholar]
  31. Dupraz, C., Bloemen, H., Bennett, K., et al. 1997, A&A, 324, 683 [NASA ADS] [Google Scholar]
  32. Eriksen, K. A., Arnett, D., McCarthy, D. W., & Young, P. 2009, ApJ, 697, 29 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  33. Fink, M., Röpke, F. K., Hillebrandt, W., et al. 2010, A&A, 514, A53 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  34. Fink, M., Kromer, M., Seitenzahl, I. R., et al. 2014, MNRAS, 438, 1762 [NASA ADS] [CrossRef] [Google Scholar]
  35. Firestone, R. B., & Revay, Z., & Molnar, G. L. 2003, Capture Gamma-Ray Spectroscopy and Related Topics (World Scientific) [Google Scholar]
  36. Fryer, C. L., & Kusenko, A. 2006, ApJS, 163, 335 [NASA ADS] [CrossRef] [Google Scholar]
  37. Gall, E. E. E., Polshaw, J., Kotak, R., et al. 2015, A&A, 582, A3 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  38. Gotthelf, E. V., Koralesky, B., Rudnick, L., et al. 2001, ApJ, 552, L39 [NASA ADS] [CrossRef] [Google Scholar]
  39. Grebenev, S. A., Lutovinov, A. A., Tsygankov, S. S., & Winkler, C. 2012, Nature, 490, 373 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  40. Grefenstette, B. W., Harrison, F. A., Boggs, S. E., et al. 2014, Nature, 506, 339 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  41. Grefenstette, B. W., Fryer, C. L., Harrison, F. A., et al. 2017, ApJ, 834, 19 [NASA ADS] [CrossRef] [Google Scholar]
  42. Hachisu, I., Kato, M., Saio, H., & Nomoto, K. 2011, ApJ, 744, 69 [Google Scholar]
  43. Halloin, H. 2009, spimodfit Explanatory Guide and Users Manual, version 2.9 (Giessenbachstraße 1, 85748 Garching, Germany: Max Planck Institut für extraterrestrische Physik) [Google Scholar]
  44. Harris, J. A., Hix, W. R., Chertkow, M. A., et al. 2017, ApJ, 843, 2 [NASA ADS] [CrossRef] [Google Scholar]
  45. Harrison, F. A., Craig, W. W., Christensen, F. E., et al. 2013, ApJ, 770, 103 [NASA ADS] [CrossRef] [Google Scholar]
  46. Hayato, A., Yamaguchi, H., Tamagawa, T., et al. 2010, ApJ, 725, 894 [NASA ADS] [CrossRef] [Google Scholar]
  47. Hillebrandt, W., Kromer, M., Röpke, F. K., & Ruiter, A. J. 2013, Front. Phys., 8, 116 [Google Scholar]
  48. Hix, W., Lentz, E., Bruenn, S., et al. 2016, Acta Phys. Pol. B, 47, 645 [NASA ADS] [CrossRef] [Google Scholar]
  49. Iwakami, W., Nagakura, H., & Yamada, S. 2014, ApJ, 793, 5 [NASA ADS] [CrossRef] [Google Scholar]
  50. Iyudin, A. 1999, Nucl. Phys. A, 654, 900c [NASA ADS] [CrossRef] [Google Scholar]
  51. Iyudin, A. F., Diehl, R., Bloemen, H., et al. 1994, A&A, 284, L1 [NASA ADS] [Google Scholar]
  52. Iyudin, A. F., Schönfelder, V., Bennett, K., et al. 1998, Nature, 396, 142 [NASA ADS] [CrossRef] [Google Scholar]
  53. Iyudin, A. F., Aschenbach, B., Becker, W., Dennerl, K., & Haberl, F. 2005, A&A, 429, 225 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  54. Iyudin, A. F., Müller, E., & Obergaulinger, M. 2019, MNRAS, 485, 3288 [NASA ADS] [CrossRef] [Google Scholar]
  55. Janka, H.-T. 2012, Annu. Rev. Nucl. Part. Sci., 62, 407 [NASA ADS] [CrossRef] [Google Scholar]
  56. Jean, P., Vedrenne, G., Roques, J. P., et al. 2003, A&A, 411, L107 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  57. Jerkstrand, A., Timmes, F. X., Magkotsios, G., et al. 2015, ApJ, 807, 110 [NASA ADS] [CrossRef] [Google Scholar]
  58. Kashyap, R., Haque, T., Lorén-Aguilar, P., García-Berro, E., & Fisher, R. 2018, ApJ, 869, 140 [NASA ADS] [CrossRef] [Google Scholar]
  59. Knödlseder, J., Valsesia, M., Allain, M., et al. 2004, in Proc. 5th INTEGRAL Workshop on the INTEGRAL Universe (ESA SP-552), 16-20 February 2004, Munich, Germany, eds. V. Schönfelder, G. Lichti & C. Winkler, 33 [Google Scholar]
  60. Krause, O., Birkmann, S. M., Usuda, T., et al. 2008a, Science, 320, 1195 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  61. Krause, O., Tanaka, M., Usuda, T., et al. 2008b, Nature, 456, 617 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  62. Kretschmer, K. 2011, Ph.D. Thesis, Technische Universität München, Garching bei München, Online available at: https://mediatum.ub.tum.de/?id=1074263 [Google Scholar]
  63. Leung, S.-C., & Nomoto, K. 2020, ApJ, 888, 80 [Google Scholar]
  64. Limongi, M., & Chieffi, A. 2018, ApJS, 237, 13 [NASA ADS] [CrossRef] [Google Scholar]
  65. Livne, E. 1990, ApJ, 354, L53 [NASA ADS] [CrossRef] [Google Scholar]
  66. Lopez, L. A., Grefenstette, B. W., Reynolds, S. P., et al. 2015, ApJ, 814, 132 [CrossRef] [Google Scholar]
  67. Maeda, K., & Nomoto, K. 2003, ApJ, 598, 1163 [NASA ADS] [CrossRef] [Google Scholar]
  68. Maeda, K., Röpke, F. K., Fink, M., et al. 2010, ApJ, 712, 624 [NASA ADS] [CrossRef] [Google Scholar]
  69. Magkotsios, G., Timmes, F. X., Hungerford, A. L., et al. 2010, ApJS, 191, 66 [NASA ADS] [CrossRef] [Google Scholar]
  70. Maguire, K., Sim, S. A., Shingles, L., et al. 2018, MNRAS, 477, 3567 [NASA ADS] [CrossRef] [Google Scholar]
  71. Marek, A., & Janka, H.-T. 2009, ApJ, 694, 664 [NASA ADS] [CrossRef] [Google Scholar]
  72. Matz, S., Share, G., Leising, M., et al. 1988, Nature, 331, 416 [NASA ADS] [CrossRef] [Google Scholar]
  73. McCray, R. 2017, Handbook of Supernovae (Springer International Publishing), 1 [Google Scholar]
  74. McKinnon, R., Torrey, P., & Vogelsberger, M. 2016, MNRAS, 457, 3775 [NASA ADS] [CrossRef] [Google Scholar]
  75. Melson, T., Janka, H.-T., Bollig, R., et al. 2015, ApJ, 808, L42 [NASA ADS] [CrossRef] [Google Scholar]
  76. Miceli, M., Sciortino, S., Troja, E., & Orlando, S. 2015, ApJ, 805, 120 [NASA ADS] [CrossRef] [Google Scholar]
  77. Müller, B., Melson, T., Heger, A., & Janka, H.-T. 2017, MNRAS, 472, 491 [NASA ADS] [CrossRef] [Google Scholar]
  78. Moll, R., & Woosley, S. E. 2013, ApJ, 774, 137 [NASA ADS] [CrossRef] [Google Scholar]
  79. Mori, K., Famiano, M. A., Kajino, T., et al. 2018, ApJ, 863, 176 [Google Scholar]
  80. Nomoto, K., Thielemann, F.-K., & Yokoi, K. 1984, ApJ, 286, 644 [Google Scholar]
  81. Nomoto, K. I., Iwamoto, K., & Suzuki, T. 1995, Phys. Rep., 256, 173 [NASA ADS] [CrossRef] [Google Scholar]
  82. Orlando, S., Miceli, M., Pumo, M. L., & Bocchino, F. 2016, ApJ, 822, 22 [NASA ADS] [CrossRef] [Google Scholar]
  83. Pakmor, R., Kromer, M., Röpke, F. K., et al. 2010, Nature, 463, 61 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  84. Pakmor, R., Kromer, M., Taubenberger, S., & Springel, V. 2013, ApJ, 770, L8 [NASA ADS] [CrossRef] [Google Scholar]
  85. Pan, K.-C., Liebendörfer, M., Hempel, M., & Thielemann, F.-K. 2016, ApJ, 817, 72 [NASA ADS] [CrossRef] [Google Scholar]
  86. Parthasarathy, M., Branch, D., Jeffery, D. J., & Baron, E. 2007, New Astron. Rev., 51, 524 [NASA ADS] [CrossRef] [Google Scholar]
  87. Patnaude, D. J., Badenes, C., Park, S., & Laming, J. M. 2012, ApJ, 756, 6 [NASA ADS] [CrossRef] [Google Scholar]
  88. Perets, H. B., Gal-Yam, A., Mazzali, P. A., et al. 2010, Nature, 465, 322 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  89. Phillips, M. M., Lira, P., Suntzeff, N. B., et al. 1999, ApJ, 118, 1766 [Google Scholar]
  90. Pietrzyński, G., Graczyk, D., Gallenne, A., et al. 2019, Nature, 567, 200 [NASA ADS] [CrossRef] [Google Scholar]
  91. Plüschke, S., Diehl, R., Schönfelder, V., et al. 2001, Exploring the Gamma-Ray Universe, eds. A. Gimenez, V. Reglero, & C. Winkler, ESA Spec. Publ., 459, 55 [Google Scholar]
  92. Reed, J. E., Hester, J. J., Fabian, A. C., & Winkler, P. F. 1995, ApJ, 440, 706 [NASA ADS] [CrossRef] [Google Scholar]
  93. Reynolds, S. P., Borkowski, K. J., Hwang, U., et al. 2007, ApJ, 668, L135 [NASA ADS] [CrossRef] [Google Scholar]
  94. Reynolds, S. P., Borkowski, K. J., Green, D. A., et al. 2008, ApJ, 680, L41 [NASA ADS] [CrossRef] [Google Scholar]
  95. Ruiter, A. J., Sim, S. A., Pakmor, R., et al. 2012, MNRAS, 429, 1425 [Google Scholar]
  96. Sankrit, R., Raymond, J. C., Blair, W. P., et al. 2016, ApJ, 817, 36 [NASA ADS] [CrossRef] [Google Scholar]
  97. Sato, T., & Hughes, J. P. 2017, ApJ, 840, 112 [NASA ADS] [CrossRef] [Google Scholar]
  98. Seitenzahl, I. R., Ciaraldi-Schoolmann, F., Röpke, F. K., et al. 2013, MNRAS, 429, 1156 [Google Scholar]
  99. Seitenzahl, I. R., Timmes, F. X., & Magkotsios, G. 2014, ApJ, 792, 10 [NASA ADS] [CrossRef] [Google Scholar]
  100. Shen, K. J., & Bildsten, L. 2014, ApJ, 785, 61 [NASA ADS] [CrossRef] [Google Scholar]
  101. Siegert, T. 2017, Ph.D. Thesis, Technische Universität München, Garching bei München, Online available at: https://mediatum.ub.tum.de/1340342 [Google Scholar]
  102. Siegert, T., &Diehl, R. 2017, in 14th International Symposium on Nuclei in the Cosmos (NIC2016), eds. S. Kubono, T. Kajino, S. Nishimura, et al., 020305 [Google Scholar]
  103. Siegert, T., Diehl, R., Krause, M. G. H., & Greiner, J. 2015, A&A, 579, A124 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  104. Siegert, T., Diehl, R., Khachatryan, G., et al. 2016, A&A, 586, A84 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  105. Siegert, T., Diehl, R., Weinberger, C., et al. 2019, A&A, 626, A73 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  106. Sim, S. A., Fink, M., Kromer, M., et al. 2012, MNRAS, 420, 3003 [NASA ADS] [CrossRef] [Google Scholar]
  107. Skinner, M. A., Burrows, A., & Dolence, J. C. 2016, ApJ, 831, 81 [NASA ADS] [CrossRef] [Google Scholar]
  108. Slane, P., Hughes, J. P., Edgar, R. J., et al. 2001, ApJ, 548, 814 [NASA ADS] [CrossRef] [Google Scholar]
  109. Slane, P., Lee, S.-H., Ellison, D. C., et al. 2014, ApJ, 783, 33 [NASA ADS] [CrossRef] [Google Scholar]
  110. Stritzinger, M., Mazzali, P. A., Sollerman, J., & Benetti, S. 2006, A&A, 460, 793 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  111. Strong, A. W., Diehl, R., Halloin, H., et al. 2005, A&A, 444, 495 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  112. Summa, A., Elsässer, D., & Mannheim, K. 2011, A&A, 533, A13 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  113. Summa, A., Hanke, F., Janka, H.-T., et al. 2016, ApJ, 825, 6 [NASA ADS] [CrossRef] [Google Scholar]
  114. Summa, A., Janka, H.-T., Melson, T., & Marek, A. 2018, ApJ, 852, 28 [Google Scholar]
  115. Takiwaki, T., Kotake, K., & Suwa, Y. 2012, ApJ, 749, 98 [NASA ADS] [CrossRef] [Google Scholar]
  116. Takiwaki, T., Kotake, K., & Suwa, Y. 2016, MNRAS, 461, L112 [NASA ADS] [CrossRef] [Google Scholar]
  117. The, L.-S., Clayton, D. D., Diehl, R., et al. 2006, A&A, 450, 1037 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  118. Timmes, F. X., Woosley, S. E., Hartmann, D. H., & Hoffman, R. D. 1996, ApJ, 464, 332 [NASA ADS] [CrossRef] [Google Scholar]
  119. Troja, E., Segreto, A., La Parola, V., et al. 2014, ApJ, 797, L6 [NASA ADS] [CrossRef] [Google Scholar]
  120. Tsunemi, H., Miyata, E., Aschenbach, B., Hiraga, J., & Akutsu, D. 2000, PASJ, 52, 887 [NASA ADS] [Google Scholar]
  121. Tsygankov, S. S., Krivonos, R. A., Lutovinov, A. A., et al. 2016, MNRAS, 458, 3411 [NASA ADS] [CrossRef] [Google Scholar]
  122. Tueller, J., Barthelmy, S., Gehrels, N., et al. 1990, ApJ, 351, L41 [CrossRef] [Google Scholar]
  123. van den Bergh, S., & Tammann, G. A. 1991, ARA&A, 29, 363 [NASA ADS] [CrossRef] [Google Scholar]
  124. van Kerkwijk, M. H., Chang, P., & Justham, S. 2010, ApJ, 722, L157 [NASA ADS] [CrossRef] [Google Scholar]
  125. Vedrenne, G., Roques, J.-P., Schönfelder, V., et al. 2003, A&A, 411, L63 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  126. Vink, J. 2004, New A Rev., 48, 61 [NASA ADS] [CrossRef] [Google Scholar]
  127. Vink, J., Laming, J. M., Kaastra, J. S., et al. 2001, ApJ, 560, L79 [CrossRef] [Google Scholar]
  128. Völk, H. J., Berezhko, E. G., & Ksenofontov, L. T. 2008, A&A, 483, 529 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  129. Waldman, R., Sauer, D., Livne, E., et al. 2011, ApJ, 738, 21 [NASA ADS] [CrossRef] [Google Scholar]
  130. Wang, B., & Han, Z. 2012, New A Rev., 56, 122 [Google Scholar]
  131. Wang, B., Meng, X.-C., Wang, X.-F., & Han, Z.-W. 2008, Chin. J. Astron. Astrophys., 8, 71 [NASA ADS] [CrossRef] [Google Scholar]
  132. Wang, W., & Li, Z. 2014, ApJ, 789, 123 [NASA ADS] [CrossRef] [Google Scholar]
  133. Willingale, R., Bleeker, J. A. M., van der Heyden, K. J., Kaastra, J. S., & Vink, J. 2002, A&A, 381, 1039 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  134. Winkler, C., Courvoisier, T. J.-L., Di Cocco, G., et al. 2003, A&A, 411, L1 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  135. Wongwathanarat, A., Müller, E., & Janka, H.-T. 2015, A&A, 577, A48 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  136. Wongwathanarat, A., Janka, H.-T., Müller, E., Pllumbi, E., & Wanajo, S. 2017, ApJ, 842, 13 [NASA ADS] [CrossRef] [Google Scholar]
  137. Woosley, S. E., & Kasen, D. 2011, ApJ, 734, 38 [NASA ADS] [CrossRef] [Google Scholar]
  138. Woosley, S. E., & Weaver, T. A. 1995, ApJS, 101, 181 [NASA ADS] [CrossRef] [Google Scholar]
  139. Woosley, S. E., Arnett, W. D., & Clayton, D. D. 1973, ApJS, 26, 231 [NASA ADS] [CrossRef] [Google Scholar]
  140. Woosley, S. E., Pinto, P. A., Martin, P. G., & Weaver, T. A. 1987, ApJ, 318, 664 [NASA ADS] [CrossRef] [Google Scholar]
  141. Young, P. A., Fryer, C. L., Hungerford, A., et al. 2006, ApJ, 640, 891 [NASA ADS] [CrossRef] [Google Scholar]
  142. Zoglauer, A., Reynolds, S. P., An, H., et al. 2015, ApJ, 798, 98 [NASA ADS] [CrossRef] [Google Scholar]

Appendix A: Additional supernovae spectra

Figures A.1–A.4 contain the spectra of the supernova remnants G1.9+0.3, Tycho, Kepler, Vela Jr., and SN 1987A not shown in the main text.

thumbnail Fig. A.1.

Spectra containing the 2σ upper limits determined for the Type Ia supernovae. From top to bottom: G1.9+0.3, Tycho and Kepler.

thumbnail Fig. A.2.

Same as Fig. A.1 for the energy range containing the 44Sc decay line.

thumbnail Fig. A.3.

Spectrum of Vela Jr. in both energy regions. No significant excess in can be found in either of the two regions for all three lines. The remnant was modeled with a Gaussian-shaped emission region with 0.6° Gaussian width. Upper limits are shown in green. Red dashed line is the flux determined with COMPTEL data (Iyudin 1999).

thumbnail Fig. A.4.

Upper panel: spectrum in the energy range from 50–100 keV for a source at celestial position of SN 1987A. The spectrum is fit with a power law accounting for the emission of LMC-X1 and PSR B0540-69, which are spatially indistinguishable from SN 1987A. The green line shows the power-law continuum and the 2σ line limits. Lower panel: same as upper panel but for energy range 1090–1210 keV. We allow for a constant offset to account for possible diffuse emission sources located within the 2.7° angular resolution of SPI. We do not find significant flux excess for any of the 44Ti decay lines. The 2σ upper-limit flux determined for an expansion velocity of 3000 km s−1 is 1.8 × 10−5 ph cm−2 s−1 corresponding to 7.0 × 10−4 M⊙ synthesized 44Ti (green line). Red dashed line corresponds to the IBIS flux (Grebenev et al. 2012) as it would be seen in SPI data.

Appendix B: Investigating systematics of analysis approach

We tested our analysis method for consistency in two scenarios. We used data taken in the high latitude region close to the Galactic north pole and analyzed the two point sources NGC 4388, for which we expected no emission in the 44Ti decay chain and an “empty” region with no known celestial source. Both “sources” are close to the Galactic north pole with latitude 60° <  b <  80° and longitude −320° < l <  −270° containing a total of 11.5 Ms exposure for the entire INTEGRAL mission time. This region contains only a few sources, making it ideal for testing the analysis approach in an “empty” celestial region. The sources can be outside the fully coded field of view for various observation times, reducing the exposure on specific locations. The image response function was not calculated for angles between the pointing direction and source location of more than 25 degree offset, and we assumed no source contribution in these cases. Figures B.1 and B.2 show the spectra for the galaxy cluster NGC 4388 and “empty” space. With our background modeling approach, the strong background lines in the region between 50 and 68 keV, 90 keV, as well as 1115–1125 keV are adequately suppressed. The photon index of the continuum emission (γ = 1.740 ± 0.074) of NGC 4388 (l = 279.1°; b = 74.3°) is consistent in the spectral index with the measurements of 1.72 ± 0.05 (Beckmann et al. 2004). We found no spurious 44Ti signatures in either location. In both regions, fluctuations around the expected continuum and baseline are present, however, these fluctuations are on scales smaller than the instrumental energy resolution at the respective energies and compatible with statistical fluctuations.

thumbnail Fig. B.1.

Spectrum of galaxy NGC 4388 in black. The gray spectrum shows the average background count rate. Clearly visible are the strong background lines between 50 and 68 keV and around 92 keV. Our background modeling approach efficiently suppresses contribution from strong lines.

thumbnail Fig. B.2.

Spectra of an arbitrarily chosen celestial location with no known sources present. The gray line shows the average count rate for the background. Influence from strong background lines is suppressed by our modeling approach. Larger errors on the fit results are found in accordance with statistical analysis for energy bins with strong background. Dead time corrected exposure at the positions is 9 Ms.

All Tables

Table 1.

Astrophysical parameters for the analyzed objects including the distances of supernova remnants used in the mass estimate for 44Ti.

In the text
Table 2.

Line shape parameters and derived quantities for the 44Sc and 44Ti decay in Cassiopeia A.

In the text
Table 3.

Values for the masses and fluxes of the six young supernova remnants.

In the text

All Figures

thumbnail Fig. 1.

Intensity distribution of a celestial source located at the center of the plot (blue dot) (Siegert et al. 2019). The orientation of the SPI instrument is changed in steps of two degrees to visit the locations marked on the map with black dots. The instrument remains centered on each celestial position for ≈2000s, which is called “one pointing”. The source located at the blue position is folded through the instrumental response function to determine the intensity distribution in the detector plane.
In the text
thumbnail Fig. 2.

Spectra for Cassiopeia A in the energy region 50–100 keV in 1 keV binning and 1090–1210 keV in 10 keV binning containing all potential decay lines from the decay chain of 44Ti. We fit the spectrum with a power law accounting for underlying continuum most likely produced by synchrotron emission at the shock front and Gaussian shaped line profiles in each region. The lines are Doppler shifted and broadened, which cannot solely be explained by the instrumental resolution at the energies, respectively. The lines are determined from a combined fit. This means that we fit one uniform power law over the entire energy range and identical Doppler parameters and integrated flux values for all three lines simultaneously. Flux values are corrected for the respective branching ratios of the lines.
In the text
thumbnail Fig. 3.

Upper panel: significance map of 44Ti decay evaluated for the 78 and 1157 keV line with color-coding in significance levels. Excess at the 4.9σ level is only found at the location of Cassiopeia A (center). The second red cross marks the location of Tycho’s supernova remnant. Lower panel: spectrum of Cassiopeia A is shown in black. The red shaded area contains the uncertainty band from the spectra obtained in a 20° ×20° area centered at Cassiopeia A. The gray spectrum shows the mission integrated background. It is evident that the uncertainty scales with the strength of the background leading to an increased uncertainty in the region 50–65 keV. Arrows indicate the laboratory-determined centroid energies of both 44Ti decay lines. Clustered excess over the uncertainty band is found in the vicinity of the 78 keV line. Unfortunately, the 68 keV emission is located at the edge of a strong background line complex allowing for only a marginal detection probability of the line.
In the text
thumbnail Fig. 4.

Left panel: comparison of upper limits and detection of Cassiopeia A with predictions made from core-collapse supernova models. Right panel: predicted 44Ti ejecta yields for different type Ia supernova models. Upper limits obtained for the three youngest type Ia supernovae exclude the double detonation model and faint SN2005E-like scenarios. We include the detections of 44Ti in Cassiopeia A and SN1987 A with the NuSTAR and INTEGRAL/IBIS telescopes in green and blue, respectively. Upper limits from further NuSTAR and IBIS observation are additionally included in green and blue (see Table 3 for references). Masses are determined for the distances in Table 1 and remnant age at the average observation date in the cited publications.
In the text
thumbnail Fig. 5.

[45Ti/56Ni] ratio of our mass estimates and several supernova explosion models. The red line marks the solar [44Ca/56Fe] ratio (1.2 × 10−3; Anders & Grevesse 1989), which is used as a reference criterion to judge supernova model subtypes. Ejected 56Ni masses for Cassiopeia A, SN 1987A, Tycho, Kepler and G1.9+0.3 are taken from (Eriksen et al. 2009; Woosley et al. 1987; Badenes et al. 2006; Patnaude et al. 2012; Borkowski et al. 2013), respectively. Vela Jr. 56Ni ejecta mass is modeled from a 25 M⊙ star (Maeda & Nomoto 2003). Model yields are from Wongwathanarat et al. (2017), Maeda & Nomoto (2003), Seitenzahl et al. (2013), Fink et al. (2014, 2010).
In the text
thumbnail Fig. A.1.

Spectra containing the 2σ upper limits determined for the Type Ia supernovae. From top to bottom: G1.9+0.3, Tycho and Kepler.
In the text
thumbnail Fig. A.2.

Same as Fig. A.1 for the energy range containing the 44Sc decay line.
In the text
thumbnail Fig. A.3.

Spectrum of Vela Jr. in both energy regions. No significant excess in can be found in either of the two regions for all three lines. The remnant was modeled with a Gaussian-shaped emission region with 0.6° Gaussian width. Upper limits are shown in green. Red dashed line is the flux determined with COMPTEL data (Iyudin 1999).
In the text
thumbnail Fig. A.4.

Upper panel: spectrum in the energy range from 50–100 keV for a source at celestial position of SN 1987A. The spectrum is fit with a power law accounting for the emission of LMC-X1 and PSR B0540-69, which are spatially indistinguishable from SN 1987A. The green line shows the power-law continuum and the 2σ line limits. Lower panel: same as upper panel but for energy range 1090–1210 keV. We allow for a constant offset to account for possible diffuse emission sources located within the 2.7° angular resolution of SPI. We do not find significant flux excess for any of the 44Ti decay lines. The 2σ upper-limit flux determined for an expansion velocity of 3000 km s−1 is 1.8 × 10−5 ph cm−2 s−1 corresponding to 7.0 × 10−4 M⊙ synthesized 44Ti (green line). Red dashed line corresponds to the IBIS flux (Grebenev et al. 2012) as it would be seen in SPI data.
In the text
thumbnail Fig. B.1.

Spectrum of galaxy NGC 4388 in black. The gray spectrum shows the average background count rate. Clearly visible are the strong background lines between 50 and 68 keV and around 92 keV. Our background modeling approach efficiently suppresses contribution from strong lines.
In the text
thumbnail Fig. B.2.

Spectra of an arbitrarily chosen celestial location with no known sources present. The gray line shows the average count rate for the background. Influence from strong background lines is suppressed by our modeling approach. Larger errors on the fit results are found in accordance with statistical analysis for energy bins with strong background. Dead time corrected exposure at the positions is 9 Ms.
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.