Found: Heart of darkness

Aug 01, 2011
This is the portion of sky in which astronomers found the Segue 1 dwarf galaxy. Can you see it? Credit: Marla Geha

Astronomers using the 10-meter Keck II telescope in Hawaii have confirmed in a new paper that a troupe of about 1,000 small, dim stars just outside the Milky Way comprise the darkest known galaxy, as well as something else: a treasure trove of ancient stars.

By “dark” astronomers are not referring to how much light the galaxy, called Segue 1, puts out, but the fact that the dwarf galaxy appears to have 3,400 times more mass than can be accounted for by its visible . In other words, Segue 1 is mostly an enormous cloud of decorated with a sprinkling of stars.

Using the DEIMOS instrument on the Keck II telescope, astronomers could identify which stars were moving together as a group. They are circled here in green. Credit: Marla Geha

The initial announcement of the “Darkest Galaxy” was made two years ago by Marla Geha, a Yale University astronomer, Joshua Simon from the Carnegie Institution of Washington, and their colleagues. This original claim was based on data from the Sloan Digital Sky Survey and the Keck II telescope. Those observations indicated the stars were all moving together and were a diverse group, rather than simply a cluster of similar stars that had been ripped out of the nearby and more star-rich Sagittarius dwarf galaxy. A competing group of astronomers at Cambridge University were, however, not convinced.

So Simon, Geha and their group returned to Keck and went to work with the telescope’s Deep Extragalactic Imaging Multi-Object Spectrograph (DEIMOS) to measure how the stars move not just in relation to the , but also in relation to each other.

If the 1,000 or so stars were all there was to Segue 1, with just a smidgeon of dark matter, the stars would all move at about the same speed, said Simon. But the Keck data show they do not. Instead of moving at a steady 209 km/sec relative to the Milky Way, some of the Segue 1 stars are moving at rates as slow as 194 kilometers per second while others are going as fast as 224 kilometers per second.

“That tells you Segue 1 must have much more mass to accelerate the stars to those velocities,” Geha explained. The paper confirming Segue 1’s dark nature appeared in the May 2011 issue of The Astrophysical Journal.

By subtracting out all the other objects in the image and leaving the Segue I member stars, the “darkest galaxy” emerges. Credit: Marla Geha

The mass required to cause the different star velocities seen in Segue 1 has been calculated at 600,000 solar masses. But there are only about 1,000 stars in Segue 1, and they are all close to the mass of our Sun, Simon said. Virtually all of the remainder of the mass must be dark matter.

Stellar Old Folks Home
Equally exciting news from Segue 1 is its unusual collection of nearly primordial stars. One way to tell how long ago a star formed is by its heavy element content, which can be gleaned from the characteristic absorption features in the star’s spectrum. Very old or primitive stars come from a time when the universe was young and few large stars had yet grown old enough to fuse lightweight atoms like hydrogen and helium into heavier elements like iron and oxygen. Early, and therefore ancient, stars that formed from early gas clouds are therefore very low in heavy elements. 

The researchers managed to gather iron data on six stars in Segue 1 with the Keck II telescope, and a seventh Segue 1 star was measured by an Australian team using the Very Large Telescope. Of those seven, three proved to have less than one 2,500th as much iron as our own Sun.

“That suggests these are some of the oldest and least evolved stars that are known,” said Simon. 

All three images above are combined in this captioned mosaic. Credit: Marla Geha, Keck Observatory

Searches for such primitive stars among the Milky Way’s billions have yielded less than 30.

“In Segue 1 we already have 10 percent of the total in the Milky Way,” Geha said. “For studying these most primitive stars, dwarf galaxies are going to be very important.”

Dark Matter Demolition Derby
The confirmation of the large concentration of dark matter in Segue 1 underscores the importance of other research that has focused on Segue 1. In particular, some researchers have been looking with the space-based Fermi Gamma Ray Telescope in hopes of catching sight of a faint glimmer of which could be created, theoretically, by the collision and annihilation of pairs of dark matter particles.

So far the Fermi telescope has not detected anything of the sort, which isn’t entirely surprising and doesn’t mean the dark matter isn’t there, said Simon.

“The current predictions are that the Fermi telescope is just barely strong enough or perhaps not quite strong enough to see these gamma rays from Segue 1,” Simon explained. So there are hopes that Fermi will detect at least the hint of a collision. 
“A detection would be spectacular,” said Simon. “People have been trying to learn about dark matter for 35 years and not made much progress. Even a faint glow of the predicted gamma rays would be a powerful confirmation of theoretical predictions about the nature of dark matter.”

In the meantime, astronomers suspect there are other, perhaps even darker dwarf hovering around the Milky Way, waiting to be discovered. “We’d like to find more objects like Segue 1,” Simon said.

Explore further: Astronomer confirms a new "Super-Earth" planet

Provided by W. M. Keck Observatory

3.8 /5 (12 votes)

Related Stories

Studying the physics of galaxies

Nov 03, 2014

Assistant Professor of Astronomy Evan Kirby arrived on campus in August. Born and raised in New Orleans, Kirby earned his BS in 2004 from Stanford University; his undergraduate thesis involved trips to Pasadena ...

How small can galaxies be?

Sep 29, 2014

Yesterday I talked about just how small a star can be, so today let's explore just how small a galaxy can be. Our Milky Way galaxy is about 100,000 light years across, and contains about 200 billion stars. Th ...

'Ridiculously' dim bevy of stars found beyond Milky Way

Apr 27, 2012

(Phys.org) -- A team of American, Canadian and Chilean astronomers have stumbled onto a remarkably faint cluster of stars orbiting the Milky Way that puts out as much light as only 120 modest Sun-like stars. ...

A Slimmer Milky Way Revealed by New Measurements

Jun 17, 2008

The Milky Way Galaxy has lost weight. A lot of weight. About a trillion Suns' worth, according to an international team of scientists from the Sloan Digital Sky Survey (SDSS-II), whose discovery has broad ...

Recommended for you

Kepler proves it can still find planets

Dec 18, 2014

To paraphrase Mark Twain, the report of the Kepler spacecraft's death was greatly exaggerated. Despite a malfunction that ended its primary mission in May 2013, Kepler is still alive and working. The evidence ...

User comments : 47

Adjust slider to filter visible comments by rank

Display comments: newest first

gxander
1 / 5 (1) Aug 01, 2011
Any indication of gravitational lensing caused by a quiet black hole holding the group together?
hard2grep
not rated yet Aug 01, 2011
Galaxy? what galaxy?
Pyle
4.2 / 5 (5) Aug 01, 2011
Is anybody trying to find a proposed distribution of DM for this galaxy?

"The stars are moving differently than if they were the only thing there, must be lots of DM." Is a very unsatisfactory answer in my book.

Because this galaxy is so close to us I am looking forward to much more research here.

@gxander: Seems to me that lensing from a black hole would be substantially different than lensing from a "halo" of DM causing the star movement. This could be a great opportunity to put some nails in the coffin of several competing theories.

This is a small enough galaxy, that theories such as MOG should be able to be tested without the effort required for larger systems. Throw in the Milky Way, 1000 stars and a predicted BH and see whether it comes close to the observations. I sure wish I could do the math.
RobertKarlStonjek
4.7 / 5 (15) Aug 01, 2011
The stars in the top right and bottom right are identical to about half way across the image (note the large bright star with the smaller star to the immediate left, the two smaller stars further to the left and so on).

This is one of those 'please explain' moments ~ whose been fiddling with the telescope??
Pyle
3 / 5 (2) Aug 01, 2011
My guess would be that those are stars in the foreground that moved, but it sure looks really fishy to me too.

Good eye RKS. Anybody have a "real" answer?
krwhite
5 / 5 (1) Aug 02, 2011
I too see the obvious image repeating there. What on earth is this?
Bobamus_Prime
5 / 5 (1) Aug 02, 2011
Very keen eye. Would be nice to know what caused the double image.
Umbersaw
5 / 5 (1) Aug 02, 2011
There's a bug in the matrix!

Well spotted.
Mayday
3 / 5 (4) Aug 02, 2011
So a little bad photoshop work is all it takes these days to claim scientific discovery and get this kind of press? I do not think this fake galaxy needs investigation. I think that the "astronomers" perpetrating this hoax are the ones who need investigating. I hope this site does some follow-up. These kind of stories showing up here are annoying. Please stop wasting our time.
Mayday
2.5 / 5 (4) Aug 02, 2011
I had to look again. It's like being unable to avert your eyes from a car crash. Or a bad comedian. This story really does belong in The Onion. I particularly love the way they point out the stars in one portion of the image, but not THE SAME STARS in another portion of the image. I'd like to challenge Marla Geha to retract this story and publish an explanation. If she even exists. I'm savin' this one; it's hilarious.
dzifo
not rated yet Aug 02, 2011
It should be gravitational lensing of background stars caused by dark matter.
What else it can be?

If you have looked at pictures of Abell 1689 it seems somewhat similar.
But of course the scale is dramatically smaller. And this lensing is incredibly smooth. And formation of dark matter can't be spherical but rather should be filamentary with cut off around these stars.

Anyways it's amazing.
Mayday
3.7 / 5 (3) Aug 02, 2011
It's not gravitational lensing, just someone playing tricks. If you turn the image sideways and merge the two duplicated star fields with your eyes, you can clearly see the cut-off across the center of the image. Upright, the right half of the image is two duplicated images -- every star is exactly the same, in exactly the same spot -- even the grain is the same. The left half of the image is of a different star field. Doing "science" in Hawaii is a lot of laughs, I guess. I hope someone won a bet, getting this published. It would be interesting to know the story behind it.
david13579
5 / 5 (2) Aug 02, 2011
The repeating stars are actually not part of what they claim is the dark galaxy so they really gain nothing by faking it, something else is probably causing it. I compared it using an image editor and saw that it really us half the stars repeating but not the stars that concern the article.
Mayday
2.5 / 5 (4) Aug 02, 2011
With all due respect; I beg to differ. Several of the green-circled stars also are repeated, but are not circled again. I believe the whole thing is a fake. There is no way that this composited image shows, proves or demonstrates any real astronomy.
david13579
5 / 5 (2) Aug 02, 2011
With all due respect; I beg to differ. Several of the green-circled stars also are repeated, but are not circled again. I believe the whole thing is a fake. There is no way that this composited image shows, proves or demonstrates any real astronomy.


You are right, I check again and there are 12 stars that are circled and also are repeated ( and their repetitions are not circled by the way)
Look here http://i.imgur.com/9nTQD.jpg
dzifo
5 / 5 (2) Aug 02, 2011
You are right Mayday.
Alignment with picture orientation is too perfect and the border is too sharp.

You could imagine that research was done using other data and this picture was made only for presentation purposes. But if they have circled some random stars after picture was made ... this looks like a bad joke.
TheRedComet
5 / 5 (3) Aug 06, 2011
This was taken by W. M. Keck Observatory which is a two telescope observatory. Which results in two images intended to be turned in to a composite image. I looks like they used the pre composite image instead of the post composite image.
TheRedComet
not rated yet Aug 06, 2011
@Mayday
The green circles represent different star groups moving together using DEIMOS. DEIMOS uses a narrow-band filter. so the stars should still be their they are just not moving with the others.
Mayday
3 / 5 (2) Aug 07, 2011
If this is science, and not a fake, then this team should be able to pass the same old test with flying colors: repeat the observation and publish.

I call fake. And I predict that we won't hear a peep from Hawaii. Case closed, IMHO.
david13579
5 / 5 (3) Aug 07, 2011
If it is indeed a double picture in need of superimposing to create a composite,then why is it only one side? the other side is not repeated.
TheRedComet
not rated yet Aug 07, 2011
@ Mayday
The green circles are not repeated because they are the ones that are moving at different rates. All the stars would have been subjected to DEIMOS narrow-band filter in order analysis their red shift. This data would not have been shown on this image because it would clutter the image. If the rest of the math was shown it would look something like this keep in mind Im that good at math. C = 299,792,458 m/s is the speed of light in vacuum and h = 6.62606896(33)×1034 J s = 4.13566733(10)×1015 eV s is Planck's constant. Then you would have the specific red shift given in GHZ. For some reason I think people would be saying WTF Keck do we really need all that text I cant even see what your talking about.
ubavontuba
1 / 5 (1) Aug 07, 2011
I simply can't believe what I'm seeing here. I verified the article is staight from the W. M. Keck Observatory journal (they use these same images, and the images are credited to the paper's author.):

http://keckobserv...arkness/

And I verified the paper was published in the May 2011 issue of The Astrophysical Journal:

http://iopscience...33/1/46/

And here's an accessible copy of the orignal paper:

http://authors.li...ys_J.pdf

But it's not clear from the graphs and charts in the paper that these Keck images are the actual images used for this study.

So I can definitely conclude that the W. M. Keck Observatory article is using these images, but I'm uncertain as to whether these images and the study's images are one and the same.

Can anyone clarify further?

TheRedComet
not rated yet Aug 07, 2011
@ david13579
I was mistaken the image was not intended to become a composite. I dont know why both side are not identical. I speculate that it has something to do with DEIMOS filtering out interfering intense light sources. Confusing but perhaps need to make a more accurate reading.
TheRedComet
not rated yet Aug 07, 2011
The third image down shows Segue I member stars that is the picture you need to focus on. If one of those stars are not repeated in the four pictures above that it then you have something.
ubavontuba
1 / 5 (1) Aug 07, 2011
The third image down shows Segue I member stars that is the picture you need to focus on. If one of those stars are not repeated in the four pictures above that it then you have something.
It's already been established that some of the selected Segue 1 stars are repeated in the double image, but strangely aren't selected twice.

david13579 even plotted them and highlighted them here:

http://i.imgur.com/9nTQD.jpg

I can't wait to hear the explanation for this.

Good job to everyone, particularly RobertKarlStonjek (for being first to point out the discrepancy)!

TheRedComet
not rated yet Aug 07, 2011
It's already been established that some of the selected Segue 1 stars are repeated in the double image, but strangely aren't selected twice.

david13579 even plotted them and highlighted them here:


No he didnt those stars are still their they are just not moving at the same rate. That was the propose of the image to show the different speeds that the stars are moving.
ubavontuba
1 / 5 (1) Aug 07, 2011
No he didnt those stars are still their they are just not moving at the same rate. That was the propose of the image to show the different speeds that the stars are moving.
You are mistaken. If the stars david13579 highlighted are the same as those above, they can't logically be included in one portion of the total image as "moving together" but not the other.

Maybe you think they suddenly "put the brakes on" from one image segment to the other?
TheRedComet
not rated yet Aug 07, 2011
I didnt get the same results as david13579 there is not a single circle repeated from the top image to the bottom one. Maybe you should try it and see if we get three different results.
Isaacsname
not rated yet Aug 07, 2011
Did you know if you connect the dots in the pic it makes an image of a facepalm ?
TheGhostofOtto1923
not rated yet Aug 07, 2011
Aw thats not dark. THIS is dark.
http://www.youtub...G0AqCZUQ

-Note the visual reference to the flood of humanity upon the earth-
TheRedComet
5 / 5 (1) Aug 07, 2011
As a kid I loved falling to sleep with Burzum. But IMO Sun o))) better represents the darkness www.youtube.com/w...related.
TheGhostofOtto1923
not rated yet Aug 07, 2011
I know them.

But otto is Purist. No death/doom/strange punctuation names, only Blackness from the Pit:
http://www.youtub...syOzpSJU
ubavontuba
1 / 5 (1) Aug 07, 2011
I didnt get the same results as david13579 there is not a single circle repeated from the top image to the bottom one. Maybe you should try it and see if we get three different results.
Uh, what don't you get about: "It's already been established that some of the selected Segue 1 stars are repeated in the double image, but strangely aren't selected twice."

TheRedComet
not rated yet Aug 07, 2011
"If the 1,000 or so stars were all there was to Segue 1, with just a smidgeon of dark matter, the stars would all move at about the same speed, said Simon. But the Keck data show they do not. Instead of moving at a steady 209 km/sec relative to the Milky Way, some of the Segue 1 stars are moving at rates as slow as 194 kilometers per second while others are going as fast as 224 kilometers per second."

david13579 observation was flawed. None of the stars that are circled in the top image are circled in the lower. Simply because he says that some are repeated and others are not doesnt make it true.
yyz
5 / 5 (3) Aug 08, 2011
Wow, there seems to be quite a bit of confusion with the presented images here, much of it well founded. Let me offer a possible explanation:

Looking at the first two images, it appears that the field in the upper-right quadrant is the same as in the lower-right quad. The upper and lower left quads appear to be adjacent, not duplicates. Some CCD imagers, as I suspect w-Keck's DEIMOS, are comprised of CCD arrays that cover larger areas of the sky than a single CCD (in this case 4 CCDs).

IMO, the upper and lower left and lower right CCD images are all are covering adjacent fields (L-shaped). The upper right quad contains a copy of the lower-right quad. Regardless, the green circles, denoting the observed stellar distribution of Segue 1 stars has been overlain on this erroneous 4-CCD representation of the galaxy as noted above, thus accounting for the discrepancies in the right side of the image. Note there are no stellar discrepancies in the left portion of the image.

con't

david13579
5 / 5 (3) Aug 08, 2011


david13579 observation was flawed. None of the stars that are circled in the top image are circled in the lower. Simply because he says that some are repeated and others are not doesnt make it true.


And THAT is the problem. The same stars appear on top but are not circled, so, they form part of Serge 1 on the bottom and not on the top?.

The 12 stars I marked appear twice while they are circled once.
If you don't believe they are repeated, cut the image in have and superimpose both halves with one half not 100% opaque and then slowly move on the images up and down and you will notice those 12 do repeat.
yyz
5 / 5 (3) Aug 08, 2011
con't

In reality, the main problem here is in the images presented to accompany the PR, as noted above by david13579. Ignoring them (misprints are not uncommon in PR material you know), the science presented in the paper seems quite sound. Faking these PR pics makes absolutely no sense.

I have been following the published work on Segue 1 since its discovery in 2006. Dr Geha and her colleagues are quite well regarded in this field and with Segue 1 in particular. Another paper regarding these same Keck observations of Segue 1 can be found here: http://arxiv.org/...85v1.pdf

The original 2008 paper by Geha et al is here: http://arxiv.org/...81v1.pdf

Discovery of Segue 1: http://arxiv.org/...48v1.pdf

A wiki page on the galaxy with links to other relevant papers: http://en.wikiped.../Segue_1

con't

yyz
5 / 5 (4) Aug 08, 2011
con't

"Is anybody trying to find a proposed distribution of DM for this galaxy?"

The modelled distribution of DM in all galaxies includes several models largely inconsistent with each other. However, today's arXiv had a paper concerning how dwarf galaxies may (or not) act as accurate tracers for DM distribution in these systems: http://arxiv.org/...95v1.pdf

The paper also offers some interesting insights into how DM haloes of satellite galaxies may combine over time to form the massive DM haloes around average galaxies found today. Most interestingly, "Our results therefore suggest that the processes that govern the spatial distribution of galaxies, once they have merged into larger halos, must be luminosity dependent, such that luminous galaxies become poor tracers of the underlying dark matter."

This of course has implications wrt models of DM distribution in (most) average-size galaxies formed by accretion of smaller satellites.
yyz
5 / 5 (3) Aug 08, 2011
Looking at Keck info on the DEIMOS instrument, imaging is provided by 2x4 arrays of 2Kx4K CCDs: http://www2.keck....date.pdf [p. 24]

I seem to recall from one of the earlier Keck papers on Segue 1 linked above that due to the size of the galaxy at the plane of the telescope the galaxy could be completely detected on just a single 4 CCD array. This would jive with my earlier supposition that using a 4 CCD image with the top right quad image being a copy of the bottom right quad, would form an L-shaped contiguous field of Segue 1.

I've emailed Dr Geha for her response regarding this apparent illustration mixup and will post a reply if received.
ubavontuba
1 / 5 (1) Aug 08, 2011
@yyz
This would jive with my earlier supposition that using a 4 CCD image with the top right quad image being a copy of the bottom right quad, would form an L-shaped contiguous field of Segue 1.
I don't know about that yyz. I've been over these images in detail and there isn't a single seem or split image anywhere. Therefore, the lower right is just as likely to be the duplicate (making the upper right contiguous to the left half of the image). Or, perhaps only two of them are contiguous, or perhaps none of them are contiguous. How can we trust this image wasn't simply randomly assembled?

It seems unlikely that Marla Geha could make such a mistake, even going so far as to circle different stars in the same star-field, without some form of serious incompetence (or even fraud) being evident. This makes me question not only the validity of this work, but any related work as well.

yyz
5 / 5 (2) Aug 08, 2011
"It seems unlikely that Marla Geha could make such a mistake..."

I agree it seems unlikely. She has extensive experience w-Keck & DEIMOS, using the combination for earlier observations of Segue 1, Andromeda & MW dwarf galaxies and stars near Sag-A* at the center of our galaxy. Do note that in this (and previous) studies, individual CCDs are used to precisely locate potential candidates in the field, not mosaic images of the larger field. This is in part to assure precision placing of the fiber guides to the spec side of DEIMOS for RV determinations of individual stars. This was done for both sets of observations of Segue performed w-Keck('08 & this paper).

Also, I'm not sure if you're aware that most images (*especially* those not appearing in the original paper) prepared for press releases are done so by students or others not affiliated with the paper. I believe this is the case here. Team members may or may not have even seen the PR images before release.

con't
yyz
5 / 5 (2) Aug 08, 2011
con't

Simple errors in IRAF routines can produce erroneous images such as that seen here. In fact, I have seen previous examples of the same type of error in PR images from Keck, VLT, and SDSS, so it's not out of the question that's what happened here.

In any case, given Segue 1's potential for future DM studies (GR annihilation signatures for one), new papers should be out soon with data on stars in this galaxy, using 8-10m class scopes needed for RV work on stars this faint. Hopefully before the end of the year. Why not wait for confirmation of these admittedly difficult observations before passing judgment on this or prior work by Geha.

What (besides the bonky ills. in the PR) do you see as a problem in the *paper* by Geha et al?
Pyle
4 / 5 (4) Aug 08, 2011
Hey all.

I beat yyz to the punch.

Hi-

Thanks for your email and eagle eyes. You are completely right-- I create the PR image of Segue 1 by putting together 4 separate images. In my code, I mistakenly repeated one of these images and appreciate your pointing this out. We will be updating the website later today.

The new version can also be found at:

http://www.astro....1_pr.pdf

best,
-Marla

All you analyzers get to work. Hopefully somebody can match this new image to the star field as well as the consistency of the image.

yyz - the paper was fine for me, but admittedly I'm in over my head as usual. Very interesting all the same.
yyz
not rated yet Aug 08, 2011
"I beat yyz to the punch."

Scooped, I tell ya. ;)

(Thanks for the update Pyle......and thanks Marla.)

ubavontuba
1 / 5 (1) Aug 08, 2011
What (besides the bonky ills. in the PR) do you see as a problem in the *paper* by Geha et al?
This error is on par with that of a surgeon amputating the wrong leg of a patient. But as I stated above, I didn't see that the pictures were necessarily tied to the paper. However, as a lead researcher, that Marla Geha made such an obvious blunder makes me wonder if she's prone to making mistakes in the research, that might not be so obvious.

Think about it. Would you ever fully trust the wayward surgeon again?

Good work to Pyle and yyz.

Pyle
3.7 / 5 (3) Aug 09, 2011
Think about it. Would you ever fully trust the wayward surgeon again?

I think you misstate the importance of the photo for the PR piece. This is more like doing the surgery on the correct leg and then photographing the wrong leg for the after image. This is a PR image and really has nothing to do with the research except to "sell" it. Because of the potential implications of this research with respect to our understanding of DM I think the PR is a great thing.

Remember most of the work is done with raw numbers generated and not the pictures. I am looking forward to more of Prof. Geha's work with Segue 1.

The modelled distribution of DM in all galaxies includes several models largely inconsistent with each other.
Hopefully the work on these local systems can shed some light on the validity of our different DM models as well as providing constraints on modified gravity theories.
ubavontuba
3 / 5 (4) Aug 09, 2011
I think you misstate the importance of the photo for the PR piece.
Perhaps. But I think this is a golden opportunity for Marla Geha to "humanize" science with some self-deprecating PR.

I think if I were in her place, I'd be all over the media, tastefully musing at my misfortune (while promoting my work).

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.