Astronomical Science DOI: doi.org/10.18727/0722-6691/5005

Supernova 1987A at 30

Jason Spyromilio1 ranged across astroparticles and all made famous by HST, Chandra and Bruno Leibundgut1 wavelengths of the electromagnetic ATCA images (see Figure 1, left), is readily Claes Fransson2 spectrum, then one might consider one visible in images from NAOS-CONICA Josefin Larsson2 had found the perfect source. (NACO), and the Spectrograph for INtegral Katia Migoto2 Field Observations in the Near-Infrared Julien Girard1 SN 1987A is just such a heavenly object! (SINFONI) as well (Larsson et al., 2016). That it is circumpolar for the more south- Figure 1 right shows a very recent NACO ern astronomical sites adds to its opti- image at 2.15 μm (Ks-band). The flux in 1 ESO mality. Located in the Large Magellanic the Ks-band is dominated by Brackett-γ 2 Department of Astronomy, The Oskar Cloud (LMC), it is near enough to be emission at 2.165 µm illuminated from the Klein Centre, Stockholm University, resolved, yet far enough to not be a outside. The illumination of the ring at Sweden threat, along an almost unextinguished the earliest times, by the ultraviolet (UV) line of sight, in a relatively uncluttered light emerging as the shock broke the part of its host galaxy. It is evolving on surface of the progenitor star, was used Thirty years on, SN 1987A continues to a human timescale. Many articles on to determine a geometric distance to the develop and, over the last decade in 1987A have appeared in the pages of LMC and remains one of the anchors of particular, has: revealed the presence of The Messenger over the years and there- the extra-galactic distance ladder. The a large centrally concentrated reservoir fore we will not dwell here on the past ring illumination also provided the first of dust; shown the presence of molecu- but rather focus on the current state and evidence of the radiation from the shock lar species within the ejecta; expanded ponder an exciting future. A comprehen- break-out, lasting only a few minutes but such that the ejecta structure is angu- sive review (McCray & Fransson, 2016) with a temperature of about a million larly resolved; begun the destruction of appeared recently and covers SN 1987A degrees. the circumstellar ring and transitioned over the past 10 000 days. to being dominated by energy sources Later, when the fastest ejecta, moving external to the ejecta. We are partici- Despite fading by seven orders of magni- at ~10 % of the speed of light, reached pating in a live experiment in the creation tude from its peak, the supernova and its the ring, the shocked gas emitted brightly of a supernova remnant and here the surroundings remain readily observable. at wavelengths from radio to X-ray. recent progress is briefly overviewed. Lengthy exposure times are still necessary Recently, observations from HST have Exciting developments can be expected to study the details and, critically, the time- been used to show that the ring is begin- as the ejecta and the reverse shock scales over which the supernova changes ning to suffer from the effects of the ejecta continue their interaction, the X-rays remain of order half a year (approximately colliding into it (Fransson et al., 2015). It penetrate into the cold molecular core the light travel time of the ejecta at this will take a while, but the ring is currently and we observe the return of the mate- epoch); therefore continued vigilance is being destroyed. A simple extrapolation rial into the interstellar medium. We needed. ESO, together with the Hubble estimates this process will be complete by anticipate that the nature of the remnant Space Telescope (HST) and the Australia ~ 2025. However, new spots of emission of the leptonisation event in the centre Telescope Compact Array (ATCA), are the outside the ring have appeared and con- will also be revealed. sole observatories that, thanks to the evo- tinued observations may yet provide sur- lution of their observing capabilities, have prises about the ­surrounding structure. provided the necessary continuous ultra- We get to watch in real time as a shock In the preface to the first SN 1987A con- violet/optical/near-infrared/radio coverage wave with well-defined characteristics ference thirty years ago (Danziger, 1987), of the supernova, and fortunately the fire- impinges on a well-understood structure Lodewijk Woltjer, then Director General of works are still continuing. Together with (both in density and composition); a text ESO, welcomed the participants with the similar monitoring by the Chandra and book illustration of shock theory. prescient statement “It is very well possi- X-ray Multi-Mirror Mission (XMM-Newton) ble that […] SN 1987A will remain observ- space telescopes in X-rays, these facili- able for thousands of years to come.” ties have provided a nearly complete Radioactivities multi-wavelength coverage of the devel- opment of the supernova. These tele- Inside the ejecta, radioactive species Introduction scopes are providing a legacy dataset for freshly synthesised in the explosion pro- this object that cannot be repeated. vide gamma rays and energetic positrons If an observational astronomer was that deposit their energy into the ejecta, allowed to pick the parameters of the The progenitor star (Sanduleak –69°202) provided they do not escape. Which iso- object of study, then ideally the angular of SN 1987A cleared a volume around topes are present, and how much of size would be matched to the resolution itself, sweeping-up material blown off in each, are critical to our understanding of of the telescope, the dimensions of the earlier evolutionary stages (about the emerging spectrum. The isotope mix physical processes would be matched 8000 years ago) into an hour-glass struc- also places limits on the mass cut (the to the angular size and the variability ture dominated by an equatorial ring. This mass coordinate in the proto-neutron star matched to the proposal cycles for tele- structure, first observed with the New where the ejection starts and the collapse scope time. If, in addition, the physics Technology Telescope (NTT) in 1989, and ends) and provides a measure of the

26 The Messenger 167 – March 2017 Figure 1. Left: combined HST (green), Chandra (blue) and ALMA (red) image of SN 1987A. Right: NACO data taken in January 2017 in the Ks-band. The ejecta emission in the centre of the ring is well resolved in both images; the short axis of the ring projects to 1 arcsecond on the sky. The west side of the ring can now be seen to be signifi- cantly brighter than the east side.

nucleosynthetic yield of the supernova. ARray (NuSTAR) detection (Boggs et al., core. This distribution is one of the main The presence of 56Ni (source of 56Fe) had 2015) and the INTErnational Gamma-Ray diagnostics of the explosion dynamics long been confirmed in SN 1987A, as Astrophysics Laboratory (INTEGRAL) during the first seconds. had 57Co. Theory predicted that 44Ti detection by Grebenev et al. (2012) of should also be made in the explosion of hard X-ray lines from 44Ti. Observations supernovae. Combining Very Large Tele- with SINFONI (Kjaer et al., 2010; Larsson Shocks scope (VLT) and HST spectra with time- et al., 2016) and HST (Larsson et al., dependent non-local thermodynamic 2011) have revealed a complex structure While the forward shock moves through equilibrium (LTE) radiative transfer calcu- of emission from atomic species that, in the ring, slowing down the ejecta and lations, it was determined that 44Ti had some cases, are collocated with the radi- accelerating the ring material, the reverse taken the role of key energy supplier to oactive species and in others are illumi- shock is formed by the supernova ejecta the supernova eight years after the explo- nated by external sources. In particular, hitting the decelerated medium behind sion (Jerkstrand et al., 2011). the SINFONI observation of the 1.644 µm the forward shock. The reverse shock is [Si I] + [Fe II] line gives a three-dimen- formed in successfully slower and denser It was consequently exciting to see both sional view of the 44Ti distribution in the regions in the supernova ejecta. the Nuclear Spectroscopic Telescope ejecta, responsible for powering the inner The supernova ejecta are being exposed from the outside to X-rays from the ring 120 Figure 2. The light interaction and at some point during the curve of the super- supernova’s teenage years the dominant nova over the past 5000 days. Different source of energy became the conversion 100 components and of kinetic energy from the supernova wavelengths are iden- ejecta with the surroundings. Quite ele- tified. The dimming of gantly, just as the radioactive elements ) 80 the ring in optical –1 whose decay had powered the emission s emission and the

g brightening of the of the junior supernova exponentially er

emission from the 3 decreased, the X-rays from the interaction –1 60 Ring: 0.5–8 keV ejecta, consequent 0 Ejecta: R-band × 150 with the inner ring provided the energy for

(1 on the input of energy

from the reverse an outside-in look at the ejecta. These

Flux shock, are evident. X-rays are mainly absorbed in the hydro- 40 Ring: [O III] × 300 gen rich envelope and the metal core is still powered by decay of 44Ti.

20 The different emission sites (ionised but Ring: R-band × 5 unshocked ring material, shocked ring gas, reverse shock, inner supernova 0 4000 6000 8000 10 000 ejecta) account for the vastly different Days since explosion velocities, and are easily separated in the

The Messenger 167 – March 2017 27 Astronomical Science Spyromilio J. et al., Supernova 1987A at 30

–13 The ejecta continue to be observed –13 Hα unshocked ring across the optical and near-infrared spectral region. SINFONI with adaptive Hα [N II] optics has been critical in providing a view of the near-infrared emission at a –13.5 –14 [N II] spatial resolution comparable to that of [O I] HST. This work has established the

)] geometry of the emission in the lower –1

Å –14 ionisation lines of [Fe II] and [Si I]. These

–1

s “core” elements play a critical role in

–2 –15 cooling the ejecta while at the same time –14.5 Shocked ring tracing the radioactive energy deposition. rg cm However, the SINFONI data also provided (e –1000 0 1000 x [S II] us with the ability to find previously unde- He I [Flu tected molecular gas. –15 lo g Before construction of the VLT had even been approved, the discovery of hot –15.5 (2000 K) molecular emission (CO and Ejecta Hα SiO) from within the ejecta of 1987A initi- ated the discussion of supernova chem- –16 Reverse shock istry. Chemical models also predicted Ejecta [O I] ­formation of molecular hydrogen (Culhane –15 –10–50 51015 & McCray, 1995). H2 is notoriously shy Velocity (1000 km s–1) and challenging to detect in the part of the spectrum approaching the thermal Figure 3. The UVES spectrum of SN 1987A around The dominant feature of the velocity infrared, and therefore it was a pleasant the Hα line in different zooms; spectrum taken in ­profile from the whole supernova shown confirmation to detect the emission January 2012. The UVES slit covers the whole –1 supernova, sampling the emission from all compo- in Figure 4 is the ~ 500 km s emission (­Fransson et al., 2016) in SINFONI data nents; see text for details. from the shocked ring, seen in Hα some 20 years after the explosion. and the [N II] lines either side of it. The Except for its very presence, the fact that optical and infrared spectrum of the narrow feature on top of the broader the observed distribution reaches almost supernova. The UV-Visual Echelle ­profile is emission from the ionised ring to the centre means that hydrogen was Spectro­graph (UVES) spectroscopy gas at the systemic velocity of the super- mixed by the Rayleigh-Taylor instabilities ­complements the HST imaging by provid- nova, + 287 km s–1. Over a range of shortly after the explosion, as predicted ing separation of the different velocity 10 000 km s–1 and at a flux level less than by explosion models (for example, components, and is perfectly suited both 1 % of the peak of the ring emission, we ­Wongwathanarat et al., 2015). in wavelength coverage and resolution see a strong box-like emission of Hα. (see Figure 4 for the different velocity This is emission from the supernova components in the Hα region). Spectra ejecta passing through the reverse Dust and molecules such as this have allowed us to follow shock. The emission from the inner core the evolution of both the ring and ejecta can be seen as the rising blue emission As new facilities come online, SN 1987A emission continuously. for velocities less than about 2500 km s–1. is not only a natural target but also a

Figure 4. Left: SINFONI spectral image in the molecular hydrogen line H2 2.41 μm at 2.40 µm (from Fransson et al., 2016). The bright ring is the contin- uum emission from the shocks and is not related to the molecular hydro- gen in the inner ejecta. Right: 3D dis- Y tribution of the 1.64 µm [Si I] + [Fe II] line from SINFONI. The ring is shown for reference and the tick marks on the box are at steps of 1000 km s–1.

X Z

28 The Messenger 167 – March 2017 Wavelength remnant before the reverse shock dis- 1 cm 3 mm 1 mm 300 μm 100 μm turbs the environment. We will need to remain vigilant to see whether the dust and molecules survive or are destroyed. Supernovae are certain to be prolific 102 ­polluters of the early Universe in metals. Whether they also contribute dust and ALMA: central ejecta molecules to the mix, we will see in the experiment taking place before our tele- scopes. ) Jy ALMA: torus The future (m ν

F 1 10 What can we expect for the future? The holy grail is certainly the nature of the compact object in the centre. The length of the neutrino burst indicated the forma- tion of a neutron star, but with later fall- back it is also possible that a black hole could have been formed. ATCA has been giving us a detailed view of the material around the supernova (Potter et al., 2009) 10 0 30 GHz100 GHz300 GHz 1 THz3 THz and combining these radio data with Frequency ALMA mm/sub-mm data provides an interesting constraint set for the presence Figure 5. ALMA's angular resolution permits the the cold dust with the inner ejecta of of a plerion, a region ionised by the separation of the continuum dust emission from the 1987A (Indebetouw et al., 2014; see Fig- strong magnetic field of the neutron star, synchrotron power law radiation resulting from shocks in the ring. The measurements shortward of ure 5). The mass of the dust, however, in the centre of the remnant (Zanardo et 400 µm are from Herschel (Matsuura et al., 2011) remains “stubbornly high”, above 0.5 M⊙ al., 2014). Higher angular resolution and APEX (Lakicevic et al., 2012) At low frequencies (Matsuura et al., 2015). If the dust sur- observations with ALMA will help greatly. data come from ATCA (Potter et al., 2009; Zanardo vives travelling through the reverse et al., 2014). From Indebetouw et al. (2014). shock, then core-collapse supernovae From X-ray observations there is a strong may be a significant contributor to the upper limit to the luminosity in the source of new understanding. The dust budget in the early Universe. 3–10 keV band of 3 × 1033 erg s–1 (Frank Spitzer Space Telescope observed the et al., 2016). The absorption of the X-rays supernova out to 30 µm and detected the The ALMA spectra provided further by the ejecta may, however, still be large presence of warm dust. Observations excitement in that, in addition to the dust (Fransson & Chevalier, 1987; Orlando et with the Gemini South telescope and the observations, they revealed strong emis- al., 2015), although this decreases rapidly VLT in the 10 µm band showed that the sion from CO 2–1 and SiO 5–4 transi- with the expansion of the ejecta. The thermal infrared emission comes from the tions. Combined with a detection by clumpiness of the ejecta, as revealed by equatorial ring. In 2010 the Herschel sat- ­Herschel of the CO 6–5 and 7–6 transi- SINFONI and ALMA, is the main uncer- ellite observed the supernova in the far tions, the temperature and mass determi- tainty. The optical/infrared may also pro- infrared and detected an enormous nation could be refined by Kamenetzky vide an interesting window, since any excess of emission longward of 100 µm et al. (2013). Approximately 0.01 M⊙ of absorbed X-rays may be reprocessed (Matsuura et al., 2011). Contemporane- CO emits at a temperature of ~ 20 K and into this wavelength range. The infrared ously, Lakićević et al. (2012) used APEX at an expansion velocity of 2000 km s–1. sensitivity of the James Webb Space Tel- to detect emission from the supernova at It remains to be seen whether this is the escope (JWST) and the superior spatial 300 and 870 µm. Combining the flux with same CO that was detected at 2000 K resolution of the Extremely Large Tele- models of dust, Matsuura et al. (2011) and 2000 km s–1 during the supernova’s scope (ELT), will be especially exciting concluded that between 0.1 and 1 M⊙ of first year or represents new formation. here. With the new and old facilities, we dust formed in the supernova. will, hopefully, during the coming decade Fascinating new ALMA observations at reveal whether the leptonisation event Given the angular resolution of Spitzer high angular resolution are in the process that was responsible for the neutrinos and Herschel, the location of the cold of being published (Abellan et al., 2017, ended in a neutron star or a black hole. dust remained uncertain. Observations in submitted). The cold CO and SiO in the 2013 with ALMA at 1 mm and 450 µm at centre of the supernova are seen to be As for the ejecta and circumstellar an angular resolution of ~ 0.5 arcseconds separated into clumps. These are unique medium, we are already seeing the proved the unambiguous association of observations of the birth of a supernova decaying ring emission. The shock wave,

The Messenger 167 – March 2017 29 Astronomical Science Spyromilio J. et al., Supernova 1987A at 30

however, continues out into the circum- ring have also been found. The neutron the sites. They have all contributed through their stellar medium beyond the ring, and will star has so far defied detection. expertise and enthusiasm to generating this beautiful and unique data set. We also want to thank the time hopefully reveal more of the several solar allocation committees over the decades who have masses of material thought to have been There have indeed also been great sur- recognised the uniqueness of the supernova and lost by the progenitor star. The mass of prises. Tracing the explosion mechanism have supported this research. the ring is only ~ 0.06 M . The X-ray by directly observing the geometry of the ⊙ The authors thank the current editor, himself a emission is expected to decay more element distribution in the inner ejecta naked eye observer of SN 1987A, for improving slowly than the optical ring emission. It confirms the non-spherical explosion the article and wish the next editor a Galactic will therefore continue to illuminate more models. The illumination of the inner ­supernova for herself. and more of the ejecta, and thus also ejecta by X-rays from the ring is a new give a new view of the abundance and feature in the development of SN 1987A. References hydrodynamic structure of the ejecta. The It provides a novel and unexpected win- supernova will then gradually transform dow on parts of the ejecta that have so Boggs, S. E. et al. 2015, Science, 348, 670 into a supernova remnant similar to other far been unobservable. The molecules in Culhane, M. & McCray, R. 1995, ApJ, 455, 335 Danziger, I. J. 1987, Proc. ESO Workshop on young remnants. We can follow this in the inner ejecta were predicted early on SN 1987, ESO, Garching real time for hundreds of years, perhaps and have finally been found. ALMA, with Frank, K. A. et al. 2016, ApJ, 829, 40 longer, as suggested by Lo Woltjer. In all its first observations, together with Fransson, C. & Chevalier, R. A. 1987, ApJ, 322, 15 these aspects ESO can continue to play Spitzer and Herschel, have told us more Fransson, C. et al. 2007, The Messenger, 127, 44 Fransson, C. et al. 2015, ApJ, 806, L19 a leading role. about the dust in SN 1987A than we Fransson, C. et al. 2016, ApJ, 821, 5 could have guessed two decades ago. Grebenev, S. A. et al. 2012, Nature, 490, 373 Ten years ago we were already musing The reverse shock has been firmly Indebetouw, R. et al. 2014, ApJ, 782, L2 about the future development of SN 1987A observed and adds another important Jerkstrand, A. et al. 2011, A&A, 530, A45 Kamenetzky, J. et al. 2013, ApJL, 773, L34 (Fransson et al., 2007). We predicted aspect to the evolution of the supernova. Kjaer, K. et al. 2010, A&A, 517, 51 exciting events — the destruction of the Lakićević, M. et al. 2012, A&A, 541, L1 inner ring and the illumination of material Larsson, J. et al. 2011, Nature, 474, 484 outside the ring. We were also hoping Larsson, J. et al. 2016, ApJ, 833, 147 Matsuura, M. et al. 2011, Science, 333, 6047 Acknowledgements to find the compact remnant inside Matsuura, M. et al. 2015, ApJ, 800, 50 SN 1987A and hopefully other surprises. McCray, R. & Fransson, C. 2016, ARAA, 54, 19 It is a pleasure to thank all the staff at the various Orlando, S. et al. 2015, ApJ, 810, 168 The ring has started to fade, indicating observatories and in particular those involved in the Potter, T. M. et al. 2009, ApJ, 705, 261 that it will be destroyed in the near future, Paranal and ALMA operations, both in preparing the Zanardo, G. et al. 2014, ApJ, 796, 82 and the first traces of material beyond the observations in Garching and in executing them on

A wide-field image (about 4 degrees in extent) of the Large Magellanic Cloud and 30 Doradus taken by the ESO 1-metre Schmidt telescope in 1986, before the appearance of SN 1987A.

30 The Messenger 167 – March 2017