In astronomy, a polarizing view is good

By Phil Plait | August 18, 2011 6:49 am

One of the basic principles of modern science is that the physics we understand here, on Earth, work everywhere. This turns out to be a pretty good assumption, because we see it coming true time and again. That knowledge can then be used to figure out things that are happening at very large distances — even well across the Universe.

With that in mind, I present to you LAB-1: a glowing glob of gas as seen by the European Southern Observatory’s Very Large Telescope:

[Click to, um, englobenate.]

This, however, is no ordinary Space Blob: it’s located at the staggering distance of 11.5 billion light years from Earth! Not very many objects have been seen farther away than this, and it’s one of the single biggest discrete structures seen this far away. It’s about 300,000 light years across — three quintillion kilometers, or three times the diameter of our own galaxy. That’s pretty flipping big.

And although it’s faint to our telescopes, at that distance it must actually be tremendously luminous for us to see it at all. Something is making it glow fiercely, but what? One hint is in the cloud’s name: the LAB in LAB-1 stands for Lyman Alpha Blob. Lyman Alpha (written as Lyman-α) is a specific color of light you get from hot hydrogen gas (just so’s you know, it’s emitted when the electron in a hydrogen atom jumps down from the second to the lowest orbital energy state). Normally, Lyman-α is in the ultraviolet, but this blob is so far away the light is shifted by the expanding Universe into the optical region — that’s why it looks green in the image above.

It takes a goodly amount of energy to create Lyman-α light, so something big is going on here. Maybe this gas cloud is collapsing under its own gravity, and is heated up. Or maybe there are big galaxies inside of it, causing it to glow. How can we tell which is the culprit?

It turns out there is a way: polarization.

You may already be familiar with this; many sunglasses are polarized (and it’s used to make 3D movies as well). It’s not hard to understand. Imagine two people standing on opposite sides of a tall picket fence. There are spaces between the pickets, maybe 5 cm wide and two meters tall. One person has a sheet of plywood to hand through to the person on the other side. If they hold the plywood horizontally, it can’t get through. Duh. But if they rotate the sheet so that it’s vertical, it passes between the fence pickets easily.

That’s polarization in a nutshell! Light waves can have a rotation too, like the plywood sheet. A polarized filter only allows light with a specific rotation through it, like the vertical slats of the picket fence. Light is usually not polarized, and can be in any old rotation. But some events can align all the waves of light, polarizing it: reflecting off glass or water, for instance. That’s why polarized sunglasses reduce glare; try looking at a reflection off the windshield of a car and then rotate your glasses 90°; the glare goes away.

Light scattered by gas can be polarized, too, which brings us back to LAB-1. If the light is coming from the gas itself, glowing because it’s warm, then it probably won’t be polarized. But if there are galaxies embedded in it, then the light from those galaxies will scatter off the gas in the cloud, and be polarized.

So astronomers used the VLT to stare at LAB-1 for 15 hours, using a polarimeter (a device for measuring polarized light). And what they saw was that the light coming from the cloud was indeed polarized. Not only that, the polarization was strongest in a ring around the center but not at the center itself, just what you’d expect if there were galaxies inside the cloud. The light from galaxies in the center isn’t scattered; it comes straight through the gas and so is not polarized. But the light farther from the center is scattered toward us (think of it like little bullets being ricocheted off of the gas), and does get polarized.

Tadaaa! The gas cloud is being lit up by galaxies inside it. Since the galaxies are young, they’re probably undergoing furious bouts of star formation, which means they’re pouring out Lyman-α. It’s this that’s getting scattered and polarized by the gas cloud.

This was not an easy observation — 11.5 billion light years is a forbiddingly long way away, and even with an 8-meter telescope it was tough to observe — but it was rather straightforward, and quickly cut through two competing ideas. The physics we understand here on Earth still works, even at a numbing distance of over 100 sextillion kilometers! And because of that, a faint green smudge can reveal entire galaxies hidden from view, betrayed by their own activity.

That’s the power of science, folks. Even more than halfway across the Universe, you can’t hide from it.

Image credit: LAB-1: ESO/M. Hayes; picket fence: Wikipedia.

Related posts:

Found: 90% of the distant Universe (explains Lyman-α light as well)
Most distant object ever seen… maybe
Another record breaker: ultra-deep image reveals ultra-distant galaxy

CATEGORIZED UNDER: Astronomy, Cool stuff, Pretty pictures

Comments (29)

Links to this Post

  1. The Vanishing Rainbow [VIDEO] - The River 100.5 FM | August 21, 2011
  1. Okay, I’ll admit that I don’t know exactly how a polarimeter works. However, if light gets scattered and polarized from something near the center of the gas, wouldn’t the light get polarized into a “ring”? That is, light at the 3 and 9 o’clock positions would be polarized vertically, light at the 6 and 12 o’clock positions horizontally, and different angles in between? Does a polarimeter handle that? (The picket fence analogy obviously would only see light near the 3 and 9 o’clock positions, and the rest of the cloud would be darker the further around the clock you moved.)

    [edit]Well, a quick search seems to say that a polarimeter rotates as it measures the light, so I guess it “sees” all of the angles over time, and can measure which parts are polarized which way.

  2. Bill3

    “Since the galaxies are young, they’re probably undergoing furious bouts of star formation…”

    How can the galaxies be young? It took 11.5 billion years for us to even get to see their furious bouts of star formation. 11.5 billion years ago is old news… :)

    (Using present tense to describe actions that happened light years away, and thus years ago, strikes me as odd – am I the only one?)

  3. Bill3:

    Using present tense to describe actions that happened light years away, and thus years ago, strikes me as odd – am I the only one?

    That topic comes up every now and then. It’s basically been accepted that you talk of things that happened millions, or even billions, of years ago in the present tense, because we are just seeing them now. It would just be overwhelming to use the past tense all the time, and have to add a note “but remember, even though it happened X years ago, since it happened X light years away, we’re just seeing it now”. It’s a lot easier to simply use the present tense.

  4. DeepField

    I have a hard time believing that the blob is a mere 300,000 ly across. The spiral galaxy in the picture, if it is, say, one tenth of the distance away, would be just about 20,000 ly across. Isn’t it too small?

  5. NewEnglandBob

    “…numbing distance of over 100 sextillion kilometers!”

    numbing sex is for young galaxies. 😉

  6. Brandon

    Thats right Ken. I have something to add though. Technically, even when observing objects close to you, like a tree outside, you are seeing the object as it existed in the past, a VERY short time ago, but the past none the less. We still describe it in the present tense though. Expanding this out to galactic scales is the same.

    On another note. Did anyone see the ABC show on Psychics last night. I missed the last 10 mins, but I think they did a good job from a reality standpoint. I hope Phil gives his opinion about it.

  7. a different phil

    Bill3, welcome to the wonderful world of simultaneity, as forced upon us by relativity. See

  8. ScienceJeff

    Is that 300,000 ly using the subtended angle and calculated distance, assuming a static coordinate system, or does it take into account the expansion of the universe? Or is that not a big factor?

  9. Piper

    Something has always puzzled me in this astronomy stuff, and that’s how we measure distances: the more the light is shifted to the red, the farther away the object is. What I don’t get is this: how can you distinguish between a distant gas blob emitting ultraviolet, red-shifted to appear green, and a nearby gas blob emitting green?

    I suspect the answer lies in the known emission wavelengths of hydrogen (and other elements). So, what if it was nearby and emitting green? What would it be then? Or at any of the intermediate distances, emitting something between green and ultraviolet? How are all of those infinite possibilities ruled out?

  10. Chris

    For some polarization fun you can have at home, look at a LCD monitor while wearing polarizing sun glasses. Now rotate your head. Cool huh?

    For even more mindblowing polarization fun, you’ll need 3 pairs of sun glasses (or at least 3 lenses). Hold the two glasses perpendicular to each other and look through them. Theoretically all the light should be gone. Now put the third pair in between the first two at about 45 degrees relative to the others. The light is back! The fun of quantum mechanics.

  11. Charlious

    It looks an old glob of Green gas to me. I have a friend whose a old gas bag but he ain’t green.

  12. CB

    @ Chris:
    Heh, those are fun. I’m fascinated just by the things you can see driving around. Like, most tinted windows appear to have a checker-board pattern in them. Also a lot of automotive paint jobs, especially dark but shiny ones, will show up as a rainbow of colors that changes as you rotate your head.

    Another fun game is to go into a store, find some sunglasses that claim to be polarized, and hold them up together perpendicular and see that they aren’t nearly opaque as they should be, and then inform the clerk on duty that these glasses are not actually polarized and that’s false advertising. The game is to see if they care. Hint: Probably not!

  13. Paul

    DeepField: the expansion of the universe makes objects at extreme distances look larger than they otherwise would be (this also reduces their surface brightness, an effect that can be used to distinguish the cosmological redshift from “tired light” theories.)

  14. eyesoars


    The trick is looking at a detailed spectrum. Various elements in the gas absorb very specific frequencies of light in distinct patterns, and looking at a spectogram will show lots of “drop outs” where frequencies have been completely or nearly completely absorbed. You can then line these drop-outs up, and figure out how much the lines have been shifted, telling you the distance of the intervening gas (and its composition: oxygen has different lines in a different pattern than, say, sulfur or iron or nitrogen).

    Most often the intervening gas is in the envelope of the star. However, for distant galaxies, sometimes several sets of aborption lines can be found, corresponding to clouds of gas between the source and us, each taking out the same set of frequencies, each red-shifted by a different amount.

  15. Piper:

    What I don’t get is this: how can you distinguish between a distant gas blob emitting ultraviolet, red-shifted to appear green, and a nearby gas blob emitting green?

    I’ve often wondered that myself. My understanding now is that they look at the object’s entire spectrum, not just particular wavelengths. It’s not just “Lyman-α has been shifted into the green”, but rather “if we shift the entire spectrum back X, everything lines up with known emissions”, heavily implying that everything was shifted by X in the first place.

  16. Jamey

    What coordinates the polarization across the size of the body? I would presume that since the emitted light is randomly polarized, and the molecules it hits are randomly rotated, moving in generally a random manner, and the photon is hitting the molecule at a random point in the waveform… What provides the aligning factor?

    It’s fairly easy to see that light bouncing from an interface, such as the surface of water, or the hot air/cold air boundary above a road surface, would have an orientation based on that surface – oriented elsewise, it’s going to go elsewhere. But supposedly these galaxies are located within the LAB, and the edges of the blob are going to be too gradual to really make a good interface to bounce off of – and for that matter, the edges are pretty random…

    Other polarized phenomena in space, I can understand – masers depend on one photon exciting the emission of the next, so it’s easy to see why their polarization would be aligned, and synchrotron radiation has the plane in which the radiation is being curved to provide alignment. Pulsars have their magnetic field – likewise sunspots and solar flares…

    So – how do the photons communicate to know which way they’re supposed to line up in this LAB?

  17. Messier Tidy Upper

    So can we expect to have a polarised debate now? 😉 8)

    Remarkable discovery – & great write-up BA thanks – reminds me of Hanny’s Voorweep a bit in appearance. :-)

  18. Messier Tidy Upper

    D’oh! Spelling memory fail. :-(

    Make that Hanny’s Voorweep Voorwerp instead.
    (No wonder it took me a while to find online images & info.)

    See :



    for comparison. :-)

    BTW. Wikipedia is featuring :

    Astronomers announce that TrES-2b [aka Kepler-1b – ed.] has the lowest known albedo of any planet, reflecting less than 1% of the starlight falling upon it.

    on its main front news page currently in case folks are interested. :-)

    I think I remember space-dot-com having a news item on that Hot Jupiter recently too.

  19. Messier Tidy Upper

    The gas cloud is being lit up by galaxies inside it. Since the galaxies are young, they’re probably undergoing furious bouts of star formation, which means they’re pouring out Lyman-α. It’s this that’s getting scattered and polarized by the gas cloud.

    So can we say that these young, distant galaxies are all wearing one huge LAB coat then? 😉

  20. Matt Hayes

    @Jamey – that’s a very insightful question.

    Hi, I’m an author of the paper. I tried to answer a few other questions, but my posts seem not to appear. I will try once more….

    The surface in this case is basically a surface in optical depth; a surface beyond which photons will have passed a certain depth of gas. When light of any kind travels through a gas the photons travel – on average – until the probability of their interaction becomes 1. Or very loosely, until they have crossed an optical depth of 1. What determines this optical depth, and the interaction probability at a given position, is just the density of the medium. That’s your surface – it’s a pretty deep one, but a ‘surface’ nonetheless.

    Then, in the scenario where photons are produced centrally and travel outward radially, in every direction, at some point they traverse a high enough depth of gas that on average they reach the point where their interaction probability has become ~1. Here they scatter. Now, the critical point here is that for Ly-alpha and the hydrogen atom, the polarization fraction is highest when photons scatter at an angle of about 90 degrees. It’s analogous to Rayleigh scattering, and the reason why the blue sky is polarized in a ring around the sun. Thus photons that travel initially in the plane of the observations – the x,y plane only – that are observed after scattering, can *only* have scettered through ~90 degrees; otherwise they would not be observable. Photons that travel in our direction initially and scatter through 90 deg in any direction can never be seen. Similarly photons that star off coming in our direction, scatter, and are subsequently observed can not have scattered through 90 degrees, and will exhibit low polarization fractions. So in the case where photons are produced centrally and scatter at large radii, the prediction is an increase in the fraction of polarization with radius.

  21. IIRC, light reflecting off a non-metallic surface will be polarized to some degree, and the angle at which it reflects can affect the amount of polarization.

    It is also possible to see the polarization of light with the human eye, but it’s difficult.


  22. CB

    @ Matt Hayes:

    Thanks for the explanation, that’s very interesting!

  23. Pardot

    @Messier Tidy Upper: I, too, thought it was Hanny’s Voorwerp at first (and second, and third) glance.

  24. Jamey

    @Matt Hayes – THANK YOU VERY MUCH! Whoot! That does explain it somewhat. Give me a while to wrap my head around it, and I’ll maybe understand it!

  25. Robin

    @ Ken B (#1): Polarimeters come in various types, some that have sensors behind fixed polarizers (for example a horizontal linear polarizer, a linear polarizer rotated to 45°, and a circular polarizer) and some with a sensor or two behind rotating elements or elements that can be automatically moved into the light path and then moved out to be replaced by another). In all cases they determine how much of the light is in a given polarization state (horizontally linearly polarized, vertically linearly polarized, linearly polarized at 45°, linearly polarized at 135°, left circular polarized, or right circular polarized) and thus how much of all the light is polarized. Devices can also be built to not only act as a polarimeter but also to image a scene in a given polarization state.

  26. Robin

    @ Chris (#10):

    The situation you describe (for instance a horizontal linear polarizer followed by a 45° linear polarizer, which is in turn followed by a vertical linear polarizer) doesn’t have to be addressed at all by quantum mechanics and is perfectly described by using the wave nature of light. With all the light being horizontally polarized (after passing thru the first polarizer), it is incident on the 45° linear polarizer. The horizontally polarized light is made of components of 45° polarized light and 135° polarized light, thus the 45° polarized component passes thru the 45° linear polarizer. Similarly, 45° polarized light is made up of components of linearly polarized and vertically polarized light, thus the vertically polarized component passes through the laser filter to the user’s eye. There are several good references for Jones Calculus (which is used to calculate how polarization changes when passing through or reflection off optical components) on the internet.

  27. Matt Hayes

    Oh, very nice post @Robin! The work described here was measuring linear polarization; just Stokes Q and U. So we used 4 angles of a half-wave plate: 0, 22.5, 45, and 67.5 degrees.

  28. Dragonchild

    Pretty much anything viewed from over 10 billion light years out will look “young”.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!


See More

Collapse bottom bar