Report

44Ti gamma-ray emission lines from SN1987A reveal an asymmetric explosion

See allHide authors and affiliations

Science  08 May 2015:
Vol. 348, Issue 6235, pp. 670-671
DOI: 10.1126/science.aaa2259

Stellar metals shine toward our eyes only

Taking a different look at a familiar star may still yield surprises. Boggs et al. trained the x-ray vision of the NuSTAR observatory on the well-studied supernova 1987A. Core-collapse explosions such as SN 1987A produce a titanium isotope, 44Ti, whose radioactive decay yields hard x-ray emission lines. All the emission associated with 44Ti appears to be from material moving toward us, with none moving away. This implies that the explosion was not symmetric. These findings help to explain the mechanics of SN 1987A and of core-collapse supernovae in general.

Science, this issue p. 670

Abstract

In core-collapse supernovae, titanium-44 (44Ti) is produced in the innermost ejecta, in the layer of material directly on top of the newly formed compact object. As such, it provides a direct probe of the supernova engine. Observations of supernova 1987A (SN1987A) have resolved the 67.87- and 78.32–kilo–electron volt emission lines from decay of 44Ti produced in the supernova explosion. These lines are narrow and redshifted with a Doppler velocity of ~700 kilometers per second, direct evidence of large-scale asymmetry in the explosion.

Supernova 1987A (SN1987A) in the Large Magellanic Cloud provides a unique opportunity to study a nearby (50 kpc) core collapse supernova (CCSN) explosion and its subsequent evolution into a supernova remnant. SN1987A has validated some of our most basic assumptions about CCSN. The burst of neutrinos observed on Earth that preceded the visible emission confirmed that the overall explosion is driven by the collapse of the central core to a neutron star (1, 2). The direct gamma-ray detections of 0.07 solar mass (M) of 56Co (3, 4) and 0.003 M of 57Co (5), and the correlation between the exponential decay of the optical light curve and lifetime of these isotopes (6, 7), confirmed that the light curve is driven by radioactive decay of these nuclei.

SN1987A, however, held a few surprises. Models and light curves supported red supergiant progenitors for CCSN, but the SN1987A progenitor was blue (8). The 56Co gamma-ray lines emerged from the explosion several months before expected, implying that some of the innermost heavy nuclear products had been mixed into the outer envelope (3). The 56Co gamma-ray line spectroscopy showed mixing to velocities of ∼3000 km s−1, several times higher than expected from spherically symmetric models, and also a net redshift of ∼500 km s−1, which indicated large-scale asymmetry (9, 10). Although subsequent x-ray observations have revealed expanding, brightening ejecta, there has been no evidence of the compact object created in the explosion (11, 12).

Here we present observations of 44Ti emission from SN1987A with the Nuclear Spectroscopic Telescope Array (NuSTAR) focusing high-energy x-ray telescope (13). 44Ti production occurs deep inside the supernova near the dividing line between ejecta and material that falls back on the compact object, and is sensitive to the CCSN explosion energy and asymmetries (14, 15). The decay of 44Ti (85.0 years) results in the production of photons at 67.87 keV (93.0% of decays) and 78.32 keV (96.4% of decays) (16). NuSTAR observed SN1987A for multiple epochs between September 2012 and July 2014 with a total exposure of 2596 ks, at an average time of ~26.6 years after explosion. A full list of the observations is provided in table S1. Both the 67.87-keV line and the 78.32-keV lines of 44Ti are clearly seen in the NuSTAR spectrum (Fig. 1). The lines are well fit with a simple Gaussian line shape plus the underlying continuum, demonstrated through the chi-square test. All uncertainties are quoted here at the 90% confidence level unless otherwise stated. In the combined analysis, we measure a 67.87-keV line flux of Embedded Imagephotons cm−2 s−1, corresponding to a 44Ti yield of Embedded ImageM. This derived yield assumes the decay rate measured in the laboratory and nearly neutral 44Ti. The NuSTAR optics response cuts off at 78.39 keV, potentially cutting off any blueshifted emission. However, examination of the 67.87-keV line reveals a net redshift, allowing us to combine the lines for optimal analysis and use the individual lines to check for consistency (17). The signal-to-noise ratio for the detection of these combined lines is ∼8.5σ. The line widths are consistent with the NuSTAR spectral resolution, and the corresponding upper limit on any Doppler broadening is 4100 km s−1 full width at half-maximum (FWHM). The 67.87-keV line centroid is redshifted by Embedded Image keV. Corrections due to recession velocity and the look-back effect (17) are ≤ 340 km s−1 combined. Taking these corrections into account, the redshift velocity of the 44Ti lines in the rest frame of SN1987A is Embedded Image km s−1.

Fig. 1 SN1987A 55- to 80-keV background-subtracted spectrum measured with NuSTAR.

Data from both telescopes are combined (for presentation only) and shown with 1σ error bars. Both of the 44Ti lines are clearly measured. The vertical green lines are the rest-frame energies of the 44Ti lines (67.87 and 78.32 keV). The redshift is evident in both lines, indicating the asymmetry of the explosion. Also shown is the best-fit model, convolved through the NuSTAR instrument response, for case (1), where the fitting parameters for the two lines are tied together (supplementary materials).

The NuSTAR yield can be compared to previous measurements, indirect estimates, and theoretical predictions. The International Gamma Ray Astrophysics Laboratory (INTEGRAL) observatory reported first detection of the 44Ti 67.87-keV and 78.32-keV lines from SN1987A (18). INTEGRAL could not spectrally resolve these lines, but measured a combined flux for both lines of (1.7 ± 0.4) × 10−5 photons cm−2 s−1, corresponding to an initial mass of synthesized 44Ti of (3.1 ± 0.8) × 10−4 M. The yield measured by NuSTAR is inconsistent with the high yield found by INTEGRAL. Chandra X-ray Observatory Advanced CCD Imaging Spectrometer (ACIS) limits on Sc Kα emission have been used to set an upper limit on the 44Ti mass of <2 × 10−4 M (19), consistent with the NuSTAR yield. Between 1994 and 2001, the ultraviolet, optical, and infrared (UVOIR) emission was dominated by the radioactive decay of 44Ti, which led to a number of estimates of the 44Ti yield based on detailed modeling of the UVOIR spectra (2022) and bolometric luminosity (23, 24). These observational studies do not all agree within their uncertainties, but generally fall in the range of (0.5 to 2) × 10−4 M of 44Ti produced in the explosion (24). The 44Ti yield measured by NuSTAR is in good agreement with most estimates based on the UVOIR bolometric light curves and spectroscopic modeling, providing support for the detailed models underlying these estimates; however, our measured redshift reveals a more complicated explosion structure than assumed in these models. Finally, theoretical predictions of the 44Ti yield for SN1987A fall roughly within this same range of (0.5 to 2) × 10−4 M, with lower yields generally corresponding to spherically symmetric models, and yields increasing with larger asymmetries. See reference (24) for a compilation of theory references.

NuSTAR observations have set an upper limit on the Doppler broadening of <4100 km s−1 FWHM, consistent with the widths of ∼3000 km s−1 originally measured for the 56Co lines (9, 10). Although the limit on broadening itself is not surprising, the measured redshift is both statistically significant and large compared to the upper limit on Doppler broadening, indicative of an asymmetric ejection of 44Ti in the initial explosion. The 56Co gamma-ray lines also showed redshifts (∼500 km s−1), but the significance was marginal. The 56Co detection also stands contrary to predictions of spherically symmetric explosion models that would produce blueshifted gamma-ray lines due to increased absorption of the receding redshifted emission. The redshifted 56Co lines suggest large-scale asymmetry in the explosion.

There has been growing evidence for asymmetries in supernovae explosions over the past decades (25). In SN1987A itself, asymmetry was initially supported by extensive evidence for mixing and polarized optical emission as reviewed in (26, 27), and later by spatially resolved images of the ejecta (27, 28). NuSTAR observations of the spatial distribution of 44Ti in the Cas A supernova remnant shows direct evidence of asymmetry (29). Our results here suggest an even higher level of asymmetry for SN1987A. For comparison, NuSTAR measured a redshift for the integrated Cas A spectrum of (2100 ± 900) km s−1 and a line broadening corresponding to a fastest ejection velocity of ∼5000 km s−1. Given that ejection velocity and the age of the remnant (340 years), the estimated “look-back” redshift velocity for Cas A is ∼1400 km s−1, consistent with the measured redshift. From the spatially integrated 44Ti spectrum alone, Cas A would not appear to have a statistically significant asymmetry: The spatial brightness distribution in Cas A revealed the asymmetries.

In the 44Ti-powered phase, the dominant energy input to the ejecta comes through the subsequent positron emission of 44Sc, when most of the gamma rays escape the ejecta without interacting. These positrons are produced deep in the ejecta, and both simple estimates and detailed models suggest that they are locally absorbed and instantaneously thermalized (20, 22). The implication is that the UVOIR emission of SN1987A during the 44Ti-powered phase should be dominated by the ejecta spatially coincident with the 44Ti ejecta. In principle, UVOIR spectral imaging in the 44Ti-powered phase can yield direct evidence for asymmetries. Hubble Space Telescope (HST) obtained resolved spectral images of the SN1987A ejecta (28) from June 2000 (4857 days after explosion), near the end of the phase when the UVOIR emission was truly dominated by 44Ti decay (21). They reveal a bipolar structure elongated along the north-south direction. There is a clear gradient in velocity across ejecta, with the northern component showing a redshift of about 500 km s−1 in the [Ca II] λ7300 emission line, whereas the southern component showed a larger redshift of about 1700 km s−1. The ejecta exhibit an overall redshift of ~1000 km s−1. At the time, this asymmetry and overall redshift were noted but not emphasized, as they could be the result of blending of the [Ca II] λ7300 line with a [O II] λ7320 line. This shifted velocity distribution is consistent with our measured redshift of the 44Ti lines. On the basis of our 44Ti line profile, we might naïvely imagine the picture of a bright, redshifted clump or jet of 44Ti, with the UVOIR emission tracing the spatial and velocity distribution of this clump. However, the spatial distribution of the ejecta in this HST observation does not immediately reveal such a large spatial asymmetry.

A single-lobe (i.e., very asymmetric) explosion model for SN1987A (30) could explain the observed evidence that 56Ni was mixed to speeds exceeding 3000 km s−1 and redshifted, as evidenced by both the gamma-ray emission and the infrared forbidden line profiles of [Fe II] (mainly produced through 56Ni decay) around 400 days after the explosion (31, 32). In this model, the single lobe is oriented at an angle pointing away from us, producing the redshifted lines (30). The NuSTAR observations appear consistent with these single-lobe models. One consequence of such a highly asymmetric explosion is that the compact object produced by SN1987A would, presumably, receive a kick velocity opposite the direction of the ejecta (33).

Supplementary Materials

www.sciencemag.org/content/348/6235/670/suppl/DC1

Materials and Methods

Figs. S1 to S3

Tables S1 and S2

References (3436)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: This work was supported under NASA Contract no. NNG08FD60C and made use of data from the NuSTAR mission, a project led by the California Institute of Technology, managed by the Jet Propulsion Laboratory, and funded by the National Aeronautics and Space Administration. We thank the NuSTAR Operations, Software, and Calibration teams for support with the execution and analysis of these observations. This research has made use of the NuSTAR Data Analysis Software (NuSTARDAS) jointly developed by the ASI Science Data Center (ASDC, Italy) and the California Institute of Technology (USA). D.B. acknowledges funding from the French Space Agency (CNES). T.K. was supported by Japan Society for the Promotion of Science Grant-in-Aid for Young Scientists (B) (no. 24740185). N.J.W. acknowledges funding from the Technical University of Denmark. NuSTAR data are accessible from NASA’s High Energy Astrophysics Science Archive Research Center (HEASARC, http://heasarc.gsfc.nasa.gov/).
View Abstract

Subjects

Navigate This Article