An Einstein supernova in the sky

March 5, 2015 3:12 PM

Astronomers using the Hubble space telescope have discovered four images of the same supernova arranged in an Einstein Cross. They've released pictures and a video to explain what we're looking at.

The team talks about their discovery of supernova Refsdal, named for Norwegian astrophysicist Sjur Refsdal.

posted by Athanassiel (39 comments total)

Galaxy field images like that never cease to humble and awe me.
posted by stinkfoot at 3:20 PM on March 5, 2015 [2 favorites]

For anyone like me who didn't tumble right away to what is depicted: the configuration is known as an Einstein Cross. It's a well-known but rarely seen effect of gravitational lensing, which is in line with Einstein's assertion that a massive object warps the fabric of space-time — and thus warps the path taken by light rays around the object.

For once, "awesome" is not an overstatement.
posted by bearwife at 3:20 PM on March 5, 2015 [8 favorites]

It's a crime that they can't refurb Hubble and keep it alive for a couple more decades.
posted by Thorzdad at 3:29 PM on March 5, 2015 [2 favorites]

I always thought gravitational lensing was a pretty cool concept, but I found this even more mind-blowing:

Computer models of the cluster predict that another image of the stellar blast will appear within five years. Astronomers may have missed an earlier appearance of the supernova in 1995. These multiple appearances of the exploding star are due to the various paths its divided light is taking through the maze of clumpy dark matter in the galactic grouping.

So you can see the exact same event not only in different points in the sky, but at different times (years apart, at that) as well. Crazy.
posted by photo guy at 3:36 PM on March 5, 2015 [6 favorites]

so very cool
posted by djseafood at 3:38 PM on March 5, 2015

So I had this strange feeling of deja vu with this press release, because I thought we'd discussed this extensively at our astro-ph coffee sessions last year. And it turns out, we did!

Here's the preprint, posted to the arXiv last November. It's back in the news now because it's been accepted by Science (the journal), but several follow-up papers have already been published in the last few months (e.g.) - a major victory for the rapid dissemination of knowledge.
posted by RedOrGreen at 3:40 PM on March 5, 2015 [1 favorite]

From the main link:
> These multiple appearances of the exploding star are due to the various paths its divided light is taking through the maze of clumpy dark matter in the galactic grouping. Each image takes a different route through the cluster and arrives at a different time, due, in part, to differences in the length of the pathways the light follows to reach Earth.

Or, in other words: "Space is big. Really big. You just won't believe how vastly, hugely, mind-bogglingly big it is."
posted by ardgedee at 3:42 PM on March 5, 2015 [4 favorites]

Apparently there's a version of this, a quasar behind a galaxy in this case, that can be seen by (extremely well equipped) amateur astronomers. Which is astounding to me.
posted by tavella at 3:55 PM on March 5, 2015 [1 favorite]

So I had this strange feeling of deja vu with this press release . . .

Yeah it was up a few years ago, but you're seeing it again because it had to route around a bunch of clumpy dark matter.
posted by The Bellman at 3:56 PM on March 5, 2015 [9 favorites]

It's amazing that that much mass can warp an image around itself to appear on the other side. And that we have a theory for that warping. And we have built a camera to collect that image. And that part of the image is SIMULTANEOUSLY 5 billion (the galaxy) AND 9.3 billion (the supernova) years old and doesn't look like that any more and deserves the word 'awesome'.

Also.
"Your momma so fat, you stand behind her an' you look like quadruplets."
posted by Zack_Replica at 3:58 PM on March 5, 2015 [15 favorites]

It's a crime that they can't refurb Hubble and keep it alive for a couple more decades.

First: The fact that it's still on orbit and productive after almost 25 years is basically amazing. Also, it's 53 days until exactly 25 years on orbit.

Second: It would probably have been cheaper to just throw up a new sat every 5 years rather than send a manned mission to fix/upgrade it. Yes, the Hubble Service Missions were amazing, but they were expensive. We could have bought many telescopes with the cost of those missions.

Third: HST is constrained by two big factors. One, it's warm, which makes IR hard to do, two, the primary mirror is 1m, which limits how much light it can collect.

The JWST, with a 6.5m mirror, a huge sunshade, and an orbit at L2, will be able to see far deeper than HST can. But, at L2, there won't be any service missions.

The other three of the four Great Observatories, CGRO, AXAF, and SIRTIF, were all one shots.

There may be a time where the repairable/upgradeable orbital scope will make more sense. But Hubble? The HST was expensive to build and even more expensive to fly and the service missions were expensive squared.

I'm glad we have the HST, and it has done amazing things. But there's a real argument that maybe we could have done so much more with a series of one shots, rather than using the incredibly expensive manned space program to upgrade a frame in space.

And this has nothing to do with the mirror. Yes, the mirror was misfigured. But it was misfigured in an extremely accurate way. Once the instruments were built for the mirror as flown, not the mirror as designed, they, and the mirror, performed spectacularly.
posted by eriko at 4:10 PM on March 5, 2015 [13 favorites]

An "Einstein Cross" sounds like something out of an anime show.
posted by Sangermaine at 4:19 PM on March 5, 2015 [4 favorites]

Thank you for posting this! First I had heard of it.
Mind-boggling.
posted by spock at 4:34 PM on March 5, 2015 [1 favorite]

When you see the Einstein Cross for the first time
You understand now why you came this way
'Cause the image you might be runnin' from is so small
But it's as big as the promise, the promise of a comin' day
posted by ZenMasterThis at 4:37 PM on March 5, 2015 [4 favorites]

Given the title, if we sing this post, shouldn't it be to Oasis?
posted by librosegretti at 4:38 PM on March 5, 2015

No. Oasis still sucks.
posted by ZenMasterThis at 4:49 PM on March 5, 2015 [4 favorites]

the primary mirror is 1m

Hubble's primary mirror is 2.4m and neither grinding it nor getting it and an assembly capable of using it into orbit was trivial. Hubble was designed under two major assumptions, one of which was brilliantly forward-looking and the other touchingly naive: Things will break, but manned space travel is going to only become cheaper and more common as time passes, so designing for field service is a sensible thing to do.

Suggesting that a succession of new telescopes would have been cheaper than the excursions to fix and upgrade Hubble is kind of silly, since the STS has been the only launch system capable of putting something that size in orbit since roughly 1976. In fact, the USAF's insistence that STS be capable of lofting modules that size is one reason the STS had so many problems.

The JWST will be wonderful if it works, and if any little thing goes wrong all we will be able to do is shrug helplessly from the ground. With Hubble we were able to correct both design mistakes and equipment failures without starting from scratch with a blank sheet of paper each time, a thing NASA does with expensive and depressing regularity.
posted by localroger at 4:51 PM on March 5, 2015 [1 favorite]

I hope there's a chance for a sixth image, too. Einstein deserves a magen david.
posted by Joe in Australia at 4:57 PM on March 5, 2015 [3 favorites]

You know what would be really cool? If gravitational lensing prevented us from seeing some amazing phenomenon behind us, but gravitational lensing in front of us sent it back our way so we could see it. Objects in the gravitational mirror are — by definition — closer than they appear.
posted by Riki tiki at 5:09 PM on March 5, 2015 [1 favorite]

Sorry, ZenMasterThis. (And also astronomers - I have taken poetic liberties; I know stars don't really explode.)

How many supernovas flare?
How would we know that they are there?
When will we ever get so high?
Slowly exploding into sight
Just as fast as the speed of light
When will we ever get so high?

Some day we will find it
Caught by Hubble's telescope
Einstein cross is shining in the sky
Some day we will find it
Warped by dark matter
It's the Refsdal supernova
The Refsdal supernova in the sky

Thanks to Sjur we now know why
Gravity makes it multiply
It looks that way to the human eye
Slowly exploding into sight
Just as fast as the speed of light
When will we ever get so high?

Some day we will find it
Caught by Hubble's telescope
Einstein cross is shining in the sky
Some day we will find it
Warped by dark matter
It's the Refsdal supernova
The Refsdal supernova in the sky

Science can reveal
So many things we never would have seen
And you and I, we live and die
The supernova's still exploding
And we know why
Why, why, why, why
posted by Athanassiel at 5:28 PM on March 5, 2015 [4 favorites]

Joe in Australia: actually there's certainly a fifth image at least, and definitely not only six. Gravitational lenses always produce an odd number of images. If you imagine the outgoing wavefront as a sphere, strong lending can fold it and produce an even number of extra images from folding, thus an odd number overall. However one image is usually very demagnified and too faint to detect.
posted by edd at 5:41 PM on March 5, 2015 [1 favorite]

Edd: wow, that's a great explanation - should "lending" be "bending", though?

One other thing I've been wondering about: could a gravitational lens endanger us by focusing radiation from a supernova? I presume there's some formula relating the mass and distance of the lens to the distance of the area it focuses; this implies that not all lenses can focus enough radiation to be dangerous to us. But that doesn't mean that no lens can do it.
posted by Joe in Australia at 5:51 PM on March 5, 2015

HST is constrained by two big factors.

One being that he's dead...
posted by Greg_Ace at 5:54 PM on March 5, 2015 [1 favorite]

Time means nothing. Space means less. Does speed count for anything?
posted by mule98J at 5:58 PM on March 5, 2015

the STS has been the only launch system capable of putting something that size in orbit since roughly 1976

Hubble is almost identical in mass and similar in size to KH-9 Hexagon, which has been happily launched on Titan IIID's since 1971. The NRO didn't need the shuttle, but if they were going to use it, it had to fit Hexagon, so it's no coincidence that these sizes are similar. You would probably have to add a fairing to launch HST as it was designed for the shuttle. It could have been done.
posted by kiltedtaco at 6:11 PM on March 5, 2015

HST is constrained by two big factors.

One being that he's dead...
posted by Greg_Ace

You messed up your link, although your point stands.
posted by the phlegmatic king at 6:11 PM on March 5, 2015 [2 favorites]

So you can see the exact same event not only in different points in the sky, but at different times (years apart, at that) as well. Crazy.

That is amazing. I always had a crackpot theory that if you assembled the right kind of matter in the right kind of configuration, you could use it to look out into the universe and see anything that ever happened. Perhaps that is actually possible.
posted by charlie don't surf at 6:23 PM on March 5, 2015 [1 favorite]

It is kind of cool to contemplate that you're seeing these flares at different times because, while the original event occurred nine billion years ago, there is a four or five light-year difference in the path the light takes along the various bent trajectories. That fraction of the circumference of the Earth is a bit under one inch. Yet on the scale of the Universe you have enough time to go through high school or get a college degree in the time it takes light to cross that inch.
posted by localroger at 7:07 PM on March 5, 2015 [1 favorite]

Since our TV broadcasts and such fade into the background noise of the universe much too soon, the universe had to find some way to syndicate its own old reruns somehow, right?
posted by sparkletone at 7:26 PM on March 5, 2015

Feynman: Hey Einstein, show Schrodinger your cross!
Einstein: Fuck you, Schrodinger.
posted by w0mbat at 8:02 PM on March 5, 2015 [1 favorite]

It's a crime that they can't refurb Hubble and keep it alive for a couple more decades

I've always thought the U.S. government should see if the Chinese would like to have it. It would save the de-orbiting costs, give them a good reason to goose their manned space program if they wanted to upgrade it, maybe even create a revenue stream for replacement hardware from the U.S. Certainly it would be an excellent zero-cost goodwill gesture. The in-flight hardware is old enough that's there's likely no secrets to steal.
posted by CynicalKnight at 8:04 PM on March 5, 2015

I've always thought the U.S. government should see if the Chinese would like to have it.

That seems unlikely considering how much tech it (probably) shares with KH-11 spy satellites.
posted by ROU_Xenophobe at 9:07 PM on March 5, 2015 [1 favorite]

This page explains: " Sometimes, the distant light source, lensing galaxy and the observer line up precisely, and we can see an "Einstein ring" – a perfect loop of light from the source encircling the lensing mass. But if there is any misalignment along the way, we observe partial arcs or spots. Depending on the relative positions of the bodies, four such spots can be seen, forming an Einstein cross. The lensing effect serves as a "natural telescope" for astronomers, who can determine the mass of the lensing galaxy and its dark-matter content based on the amount of distortion observed."
posted by spock at 11:03 PM on March 5, 2015 [4 favorites]

I particularly like the proposal to use gravitational lensing for interstellar communication, even if there are a few small practical issues to iron out first.
posted by Devonian at 3:25 AM on March 6, 2015

Joe in Australia: sorry strong lensing. Strong is the sort that heavily distorts shapes into arcs and multiple images. Weak involves distortions of images that are slight enough you need to detect it statistically - averaging over the shapes of lots of background galaxies in an area as the distortions are less than typical ellipticities of galaxies. Micro lensing is when a lensing object passes in front of another over time and just causes a change in the background object's brightness.

And yes, a supernova has to be pretty much in our backyard to be any threat. In our galaxy there aren't objects big enough to do an awful lot of magnification in most circumstances. It's hard to come up with very good assumptions to figure this out, and this post will ramble as I do calculations as I go. In principle a source can be magnified infinitely but this requires perfect alignment of point sources and lenses and neither are in reality point like, and the alignment is always changing somewhat on galactic scales as things move about (the galaxy is where such a circumstances as you're wondering about are most likely even though it's very unlikely I think). Typical micro lensing like this can magnify up to hundreds of times, but only rarely and lasting maybe days before the alignment drifts out.

An unlensed supernova is safe out to maybe a hundred light years (and that's quite a pessimistic estimate perhaps), and if we allowed a hundred times magnification it would extend that distance to a thousand light years. For best conditions for lensing the lens needs to be half way between so we are talking about lenses within 500ly or so, and there's a few million like that. Most of those will be relatively low mass stars and they obviously lens less, so the chances of them causing a problem are lower. Lensing is proportional to the square root of the lens mass.

For say a ten solar mass lens in optimal conditions the radius of an Einstein ring is 0.016 arcseconds (HST resolution is 0.05). To get a magnification of 100 you need the alignment between source and lens to be 100 times better than that, so it'd have to be spending a fair bit of time lined up just right to within about 0.2 light years, and at exactly the right time for when the supernova went off. Which obviously isn't very likely in the first place.

Or to put it another way that ~0.0001 arcsecond lensing 'footprint' to get a hundred times magnification is about 10^-20 of the whole sky, so even with maybe tens of millions of stars around in that volume your alignment fraction for that is tiny, multiplied by the low chance of a supernova going off in that volume.
posted by edd at 4:14 AM on March 6, 2015 [1 favorite]

Hubble's primary mirror is 2.4m

(swings at ball, ball hits shank, goes 50 feet to the right. "oooh, you hate to see that.")

Yeah, I shanked that. You are correct. In my defense, I was in a tiki bar, with tiki drinks, getting a bit tiki buzzed.

the STS has been the only launch system capable of putting something that size in orbit since roughly 1976

Nope. As mentioned, the Titan IIID was built to fly KH-9 and KH-11 sats, and it could put 12 metric tonnes into LEO, HST masses about 11.1 metric tonnes. The Titan 4 nearly doubled that to 21 metric tonnes. Indeed, the HST looks a lot like a KH-11 with different solar arrays -- and the reason the mirror is 2.4m is that that was the largest mirror a KH-11, I mean, the HST could fly and fit into a 4m payload shroud or the Shuttle cargo bay.

Heck, in 2012, NRO gave two (believed to be) KH-11s to NASA, one of them is going to be used for WFIRST, which the combination of STS, STS wind-down, and the JWST was put on the back burner for lack of budget. To NRO, they're outdated and not worth flying. To NASA, they're wide-field versions of the HST optics.

STS was fantastically expensive. If it wasn't flying, we could have used that budget for a dozen space telescopes and more -- indeed, arguably, the right answer, budget wise, wasn't flying the HST service missions, it was to built HST2 and launch it on a Titan IV or Delta IV.

That makes me sad, because the HST service missions were some of the great moments in manned spaceflight. But costwise? Between the extra costs to make HST serviceable, the training for the service missions, and the massive costs of the Shuttle flights themselves, we made the wrong call there. This is why no other space telescope, flying or planned, is built to be serviceable. Unless and until manned spaceflight becomes much cheaper, it'll always be cheaper to fly another instrument than to go update the one on orbit.

Then, of course, we get results like this, or things like the HDF/UDF images, or the legions of images that HST has given us over the last 25 years, and I go back to "Yeah, it was as expensive as hell, but look what we got." That picture of the Eagle Nebula launched a thousand scientific careers. We can't lose that capability.

Then I go right back, and say "No, we can't -- and the right answer is to loft a new one on a Delta IV, not try to update HST again."

Then I wonder: Making HST serviceable was a big part in why it was so expensive to build, and when it went up with the primary having the wrong figure, it was a public relations disaster. It may be that if we'd flown a non-serviceable version up, we'd never get another one after that debacle. Thank all that is clever that while the mirror was ground wrong, it was ground very precisely wrong, so COSTAR and later instruments could fix the image.

But, regardless. We have HST now, and it's still pulling science like this. I hope JWST and WFIRST continue the legacy, and I'd love to see another optical observatory on orbit when HST finally retires.
posted by eriko at 4:45 AM on March 6, 2015

The Hubble Wars by Eric J. Chaisson goes into quite a lot of detail on Hubble's history, including bits on its relation to Keyhole satellites.
posted by edd at 7:10 AM on March 6, 2015 [1 favorite]

Thanks for that really detailed answer, edd. So, to put in British tabloid terms:

ASTRONOMER SAYS
it's not very likely that
ROGUE STARS
which may have been one of the factors
THAT KILLED THE DINOSAURS
MAY SLAY US ALL.

COSMIC MAGNIFYING GLASS WILL
probably not be positioned accurately enough to
BURN US ALL IN OUR BEDS.
posted by Joe in Australia at 5:03 AM on March 7, 2015 [2 favorites]

eriko: Indeed, the HST looks a lot like a KH-11 with different solar arrays ...

Yeah, I vaguely remember this from my grad school days, but one of the bigger problems with HST - bigger than the incorrect optics, which people were well on their way to living with using deconvolution algorithms[*] - was the pointing stability. Basically the telescope used to go into this oscillation mode where it would bob up and down while its solar arrays flexed in the opposite direction, like a giant space bird. Apparently the folks at Space Telescope asked the NRO about this, and got a couple of embarrassed coughs back as the only response.

Not sure how much of this is urban legend, and how much is well documented in the Hubble Wars book referenced above by edd, but it boggled my mind that (a) the NRO knew about this potential problem but let the Hubble design get launched anyway, and (b) they must have found out about this problem the hard way - and it must have been harder to compensate when you're looking down rather than up? Who knows how much money those guys are wasting behind the curtain of black ops.

[*] There were some furious arguments when COSTAR was being planned about whether or not deconvolution was going to be "good enough" to live with. Basically, knowing the response to a point source (a star, for example), you could iteratively subtract scaled copies of that response from each bright area in your blurry image and put back scaled copies of the corrected response when you were done, a process called "Clean" that radio astronomers pioneered and that we still use every day. The images wouldn't have looked as good, the spectra for science would have been unaffected, and it would have been fixed later on by installing next-generation instruments with internal optics that accounted for the exact mirror shape, as ACS and WFC3 do now, for example. But the "something must be done" train left the station with "eyeglasses for the Hubble" and "even better than planned", launching the giant publicity juggernaut that is now the NASA/STScI machine, and the rest is history.
posted by RedOrGreen at 3:18 PM on March 9, 2015