Superlong-lasting supernova puzzles scientists

A supernova is considered the terminal explosion of a dying star, but scientists observed what they think were at least three explosions in the case of iPTF14hls

In September 2014, a telescope caught what looked like a normal supernova. But then the exploding star iPTF14hls lit up several times more, and remained bright 600 days – six times longer than comparable supernovae. We spoke with Iair Arcavi at Las Cumbres Obervatory, who published the observation in Nature today.

ResearchGate: When did you realize iPTF14hls was special?

Iair Arcavi: We first realized something strange was going on when an undergraduate student working with me, Andrew Wong, saw that the supernova had gotten faint and then bright again for the first time. He called my attention to it and we immediately started getting more observations. Since then it went up and down in brightness another four times. Now it's been slowly fading for the last year.

RG: What makes this supernova special?

Arcavi: iPTF14hls is special in several ways. Most supernovae get bright and then fade once. 14hls got bright and faded at least five times between 2014 and 2016. This has never been seen before, and we're not sure what the explanation for that behavior is.

Also, every supernova we've seen so far was the final explosion of a star. But we have evidence that 14hls exploded once already in 1954, and maybe also between 2010 and 2014. Then we saw it exploding in 2014, which means it had at least three different explosions! We've never seen that happen before either.

We looked at the spectrum of 14hls by splitting its light into a rainbow – a technique that allows us to measure the composition, velocity and temperature of the material in the supernova. We saw that it looked like a spectrum of the most common type of supernova - Type IIP. Those kinds of supernovae stay bright for 100 days and then fade, whereas 14hls behaved very differently (getting bright and faint at least 5 times in 600 days).

Also, the velocities we measure for 14hls are much higher compared to Type IIP supernovae at the same age. For example, at 600 days after discovery, the spectrum of 14hls looks like that of a normal IIP supernova at 60 days after explosion!

Scientists observed iPTF14hls growing bright and dim again several times over two years. This behavior has never been seen in previous supernovae, which typically remain bright for approximately 100 days and then fade. Adapted from Arcavi et al. 2017, Nature. Credit: LCO/S. Wilkinson

RG: How did you observe this supernova?

Arcavi: In order to study this supernova, we had to observe it for a few hours once every few days for several years. This would have been almost impossible with most observatories which, if you're lucky, give you a night at the telescope every few weeks. Most of our data has come from Las Cumbres Observatory, which is unique in that it operates a network of robotic telescopes around the world. The telescopes are scheduled dynamically by software every 15 minutes. This allows us to get the kind of data we need – an hour or two every few days for years. Without that kind of facility, it would have been very hard to keep monitoring 14hls and collecting the data that is now challenging all our models.

RG: How do you explain this exploding star’s behaviour?

Arcavi: The short answer is, we can't. It just breaks all existing supernova models. The only model that comes close to explaining 14hls is one called "pulsational pair instability". According to that model, which has been only theoretical until (maybe) now, a very massive star (about 100 times the mass of the sun) will get so hot in its core that some of its energy will actually turn into matter and antimatter through Einstein's famous e=mc^2. When this happens, the star becomes unstable and could blow off its outer layers in a giant explosion, but one that leaves the core intact. The star can do this several times before finally exploding in a massive supernova.

RG: Why do you suggest that current models of star evolution and explosion might have to be rewritten?

Arcavi: Even the pulsational pair instability model doesn't predict as much energy as we think we see in iPTF14hls. Also, according to that model, almost all of the hydrogen in the star should have been blown away in the first eruption (i.e. 1954), but we still see a lot of hydrogen (possibly 50 solar masses of it!) now. So either the pulsational pair instability model needs to be revised, or we need a completely new explanation, since none of the other existing models work at all.

An image taken by the Palomar Observatory Sky Survey shows a possible explosion in the year 1954 at the location of iPTF14hls (left), not seen in a later image taken in 1993 (right). Adapted from Arcavi et al. 2017, Nature. Credit: POSS/DSS/LCO/S. Wilkinson

RG: Could there be another explanation for iPTF24hls?

Arcavi: Definitely, but it would be something totally new. For example, if stars can eject material at supernova velocities and energies but not be a supernova, that might explain what we see. But a scenario like that was never proposed before, so it's still to be seen if we can make the physics work.

 RG: You suggest that the star might have been 95-130 times more massive than the sun. Do we know of any other star this massive?

Arcavi: A few, yes. Eta Carina might be that massive, and it had a strong eruption in the 1840's, but it was still much fainter than what iPTF14hls did, though. There could also be stars at those masses in the Magellanic Clouds, small neighboring galaxies to the Milky Way. Such stars are expected to be much more common in the early Universe, but it's quite surprising to find them so nearby, like in the case of 14hls.

Feature image: Artist impression of a supernova. ESA/Hubble