In a galaxy far, far away lies a puzzling star, one that seems to be dying multiple deaths, leaving astronomers to rethink the way these beacons of the sky end their lives.

Stars can die in a few ways, but the most spectacular is as a supernova, an explosion that can outshine even its host galaxy. And while astronomers believe they have a relatively good understanding about how and why these stellar explosions occur, supernova iPTF14hls — discovered in 2014 — just doesn't fit.

This star died as a Type II-P supernova in a galaxy 500 million light years away. These types of supernovas occur when a star eight to 15 times the mass of our own sun exhausts its energy and explodes. Typically, they continue to shine for roughly 100 days. Over time they begin to fade.

But not iPTF14hls. Instead, it lasted 600 days. 

In a new study published in Nature, the team of international researchers discuss what could possibly account for the extremely long brightening period. 

But nothing fit.

The mystery deepens

One method of studying stars is by analyzing their spectrum, where visible light is separated into its various colours. This can provide a wealth of information about a star, including data on its composition.

Initially, it looked like a run-of-the-mill supernova. But when lead author of the study Iair Arcavi, from the department of physics at the University of California, took another look, he was confounded.

"I was really, really not expecting what we saw," he told CBC News. "What we saw was the spectrum of the most typical supernova observed. They always get bright for 100 days; this one didn't do that. It got faint and then bright," he said. 

Four months after the initial brightening, it still only looked like a supernova that was just one month old.

"Six hundred days later, the spectrum looked like that of a 60-day-old supernova," Arcavi said. "It was evolving, just in slow motion."

Supernova brightness

This graph illustrates how iPTF14hls ​grew ​bright ​and ​dim ​again ​at ​least ​five ​times ​over ​two ​years. ​This ​behaviour ​has ​never ​been seen ​in ​previous ​supernovae, ​which ​typically ​remain ​bright ​for ​approximately ​100 ​days ​and ​then ​fade. (LCO/S. ​Wilkinson)

There seemed to be no rhyme or reason to it. It was like it just kept exploding over and over again.

As well, as supernovae expand and the shells of gas are blown off, astronomers normally see slower-moving material, but that didn't happen either.

And then came the temperature. If something expands but keeps the same brightness, it should cool. But once again iPTF14hls would have none of that: it remained at the same temperature.

The researchers ruled out all the possible explanations. There was one final possibility: something called a pulsational pair instability supernova.

In this kind of stellar death, a star that has about 105 solar masses (one solar mass is equal to the mass of our sun) dies in an explosion, but instabilities produce several outbursts over an extended period of time, which would fit with iPTF14hls. 

But again, it wasn't so simple.

Archival data suggests that there was an explosion in the exact same location — in 1954, more than 60 years ago.

Supernova

These two images show the location of supernova iPTF14hls in 1954 and the absence of the supernova in 1993. (Carnegie Science)

But a subsequent image of the galaxy in 1993 showed no sign of the supernova. 

While it's not entirely certain that it is the same star exploding, Arcavi said that there's just a one- to five-per-cent chance that it was a different star.

'Incredibly puzzling'

But the pulsational pair instability model isn't a 100- per-cent fit, either.

"A pulsational pair could give outbursts at very different time scales — could be years, could be decades," he said. "But it predicts that the star should lose all of hydrogen in the first explosion. And we still see a lot of hydrogen in the 2014 one ... which the model can't really explain."

And there's more.

"The other problem is the energy we deduced for the explosion that happened now in 2014, is so high, it's more than all the total energy predicted by that model for all the explosions together," Arcavi said. "So, that's a problem."

Weird supernova

The galaxy and location of iPTF14hls, in crosshairs. (Iair Arcavi)

"It's crazy," said Maria Drout of the University of Toronto's Dunlap Institute for Astronomy and Astrophysics. Though not involved with the study, Drout most recently was part of the team that was the first to image the source of a gravitational wave earlier this year. "It's incredibly puzzling."

"I don't think this changes the picture for your plain old vanilla Type II-P supernovae. But it points out a new and unusual way that massive stars can explode or die," she said.

Arcavi said the ability to monitor the sky continuously — something that wasn't possible in the past — helped them discover this strange supernova. 

"It begs the question: how many have we missed in the past?" he said.

For Nick Konidaris, a co-author and an one of the inventors of the SED Machine, a tool that allows telescopes to image the sky much faster than conventional methods, the discovery is what he always hoped would be the result. 

"When we started designing and conceiving this SED Machine instrument, the thing we were really hoping for was this kind of discovery," he told CBC News. "Where it was something new and exotic and could teach us something new about our universe."

And it worked.

Arcavi is excited to continue monitoring the strange star. As to whether or not it could brighten again, he says, "I'm not making any bets."

He's also looking forward to discovering if astronomers will produce new models and thus a new category for this type of supernova.

"It has these two aspects, a Dr. Jekyll and Mr. Hyde supernova that on the one hand is so weird and unpredictable, and on the other hand just has this textbook vanilla supernova spectrum, and we really don't know how to reconcile those two sides of it."