Science keeps advancing, in fits and starts. It was a good week for intriguing results from experiments.
The first bit of news, which has been the subject of the most internet buzz, is a new paper by Chilean astronomers C. Moni Bidin, G. Carraro, R. A. Mendez, and R. Smith, which claims that there’s no evidence for dark matter in the dynamics of stars near the Sun. If this were true, it would imply something funny going on with the distribution of nearby dark matter, which could have significant implications for direct searches here on Earth (see below). It wouldn’t really be much of a threat to the idea of dark matter itself, since there’s plenty of evidence for dark matter elsewhere. But it might mean that the distribution in the Milky Way was very different from the kinds of models we like to use, for example by being much lumpier.
We just heard a great physics colloquium here at Caltech by Katie Freese, who talked about this result very briefly. Her opinion matched those of the skeptics in Ron Cowen’s article linked above: this paper makes a lot of assumptions, some of the a bit dubious, and we would need to see something much more solid before we become convinced. The biggest issue is that they don’t actually measure the DM distribution near the Sun; they try to measure it in a region between 1500 and 4000 parsecs below the galactic plane (which is actually pretty far away), and then fit to a model and extrapolate to what we should have nearby. This kind of procedure relies on our understanding of the vertical structure of the galactic disk, which isn’t all that great. So it’s definitely an intriguing result, one that should be taken seriously and followed up by other surveys, but nothing to lose sleep over just yet.
The second bit of news is another puzzling absence: a lack of neutrinos that were predicted to be produced by gamma-ray bursts. These bursts are among the most energetic events in the post-Big-Bang universe, and for a long time were a major mystery to astrophysicists. More recently, a consensus had grown up that GRB’s (as they are called) are associated with intense beams of particles created by newly-born supernovae. That’s a model that seems to fit most of the data, anyway, and it also makes a pretty good prediction for the production of associated neutrinos. But a new paper by the marvelous IceCube experiment has thrown a spanner into the works: they should have been able to see those neutrinos, and they don’t.
IceCube consists of a series of thousands of detectors arrayed within a cubic kilometer of Antarctic ice, and looks for flashes of light associated with high-energy particles passing through. Recently they have been keeping an eye out for signs of neutrinos that should be associated with GRB’s that have been detected by the Swift and Fermi satellites — but no luck. It’s a puzzle that will send GRB theorists back to the drawing board. One of the funny aspects of the story is that particles from GRB’s are a leading candidate to serve as the origin of high-energy cosmic rays, but that seems to be out the window now. It’s still possible that the cosmic rays come from active galactic nuclei, but there’s another group of theorists who have something new to chew on.
The final bit of news is even dicier, and hasn’t received any internet buzz at all yet — I only heard about it through Katie Freese’s talk. Maybe because there was no press release and the shocking claim is hidden within the guts of a technical paper with a boring title. The paper is by our friend Juan Collar and Nicole Fields. Recall that the DAMA dark matter experiment looks for an annual modulation due to the fact that the Earth moves through the dark-matter “wind” at different velocities during different times of the year. And they see a signal — very strongly — but many people have questioned whether what they are seeing is really due to dark matter. Juan has been leading another experiment, CoGeNT, which has been trying to check DAMA’s results — and has found a very tentative signal that seems to agree with them (which wasn’t what most people were expecting).
One of the reasons for the skepticism is that there are other experiments, which aren’t tuned specifically to look for annual modulations, but nevertheless should be sensitive to dark matter at the level implied by DAMA and CoGeNT — and they see nothing. More recently, some of these experiments have started looking for annual modulations — and they see nothing. Here for example is a recent paper by the CDMS experiment that says exactly that.
But the new paper by Collar and Fields claims that CDMS have analyzed their own data incorrectly. They argue that (1) CDMS isn’t really sensitive to the kind of annual modulation purportedly seen by CoGeNT, and (2) if you look carefully there is actually a statistically significant (more than 5 sigma!) bump at low energies, consistent with the kind of low-mass dark matter particle you would need to explain the annual modulations.
My impression is that the CDMS folks are unmoved by this argument. It’s certainly always very hard to analyze the data from somebody else’s experiment. This kind of controversy comes down to very particular aspects of data collection, analysis, and sources of systematic error. It’s way over my head, so I have no professional opinion about who is right. But at the very least it’s a reminder (as if we needed one) that the dark-matter-detection game is heating up, and big news might be creeping up on us. The universe loves puzzles.
Also, if light moves much slower than c, more than making up the difference in speed, doesn’t that in itself screw up the expectation that photons and neutrinos arrive simultaneously?
You can calculate the effective speed of light in a medium as c/n, where n is the index of refraction. The exact value of n is dependent on both the wavelength of light and the physical properties of the interstellar medium. The interstellar medium is going to vary quite a bit over the course of a photon’s trip from a supernova to Earth, so I don’t think we could meaningfully specify a single speed.
However, there is another wrinkle about which I forgot earlier: neutrinos are actually produced earlier in the process of a supernova explosion than visible light. Neutrinos are produced right away, when the core of the star collapses. That collapse creates a massive amount of pressure, which leads to a shock wave. Visible light is produced only when the shock wave exits the star, which does not happen instantaneously. (The shock wave will travel at some speed above the speed of sound in the star, which varies based on density and material properties.) It’s this effect that actually accounts for most of the time lag between photons and neutrinos from a supernova like SN1987A.
The expectation is not that they will arrive simultaneously, for all of these reasons. Photons should arrive only a short amount of time after neutrinos based on current knowledge, though, if GRBs come from supernovae. That behavior is what IceCube did not observe.
@ Kevin,
My point is that the the delay for neutrinos with mass over typical distances for GRBs (billions of LY) could amount to years. I had no idea intergalactic space could slow down light that much, but surely there’s no reason to believe that would match the neutrino delay so closely?
I’m actually starting to believe the slowdown of light in space you’re talking about is too little to be in any way relevant. Many GRBs last less than a minute and contain a fairly wide range of frequencies that should travel at different speeds. This doesn’t seem to happen, or the burst would be smeared out when it arrives here. A minute over a billion light years is one part in 10^14 -10^15.
When you say the delay could amount to years, that entirely depends on the mass and energy of the neutrinos. There’s a good discussion of actual measurements of neutrino speed on Wiki, and you can see the deviation from light speed is very small for typical neutrino energies.
You are probably right that the slowdown from the interstellar medium is usually negligible. Most of the time lag comes from the fact that light is generated by the shock wave which is slowed within the exploding star itself.
However, photons above a certain energy (~10^14 eV) will interact with cosmic background radiation and pair produce electrons and positrons, meaning we actually can’t detect those signals. Such photons are, of course, well outside the visible light range. This opacity of space to high-energy photons is another reason to use neutrinos for astronomy. (This has little direct relevance to the above discussion, but I think an it’s interesting point, and it does affect future investigation of GRBs.)
Regarding the time delay of neutrinos relative to photons, there are probably more careful ways to derive a precise number, but a nice hand-wavy way to think about it is this:
The (muon) neutrino mass is somewhere in the neighborhood of 0.3eV, give or take a few orders of magnitude (my guess is it’s probably lower, actually), while IceCube is looking for neutrinos in the neighborhood of 10^14 eV. Thus the rest mass energy is about 3 parts in 10^15 of the total energy, and a particle with these characteristics would experience only a negligible a time delay relative to a particle traveling at exactly c. The time window that IceCube uses is enough to compensate for this potential difference.
And in the usual (“fireball”) theories, neutrinos come from the interactions of jetted material, so the neutrinos would be in the same (or narrower) jets than the photons. If they’re emitted isotropically, then you’re pretty much looking at a whole new theory of high energy neutrinos associated with GRBs, which is why the IceCube result warrants a “back to the drawing board” conclusion.
Pingback: Read/Write Daily: Cosmic Puzzles
@24. And at least three authors in that previous post explain why the claim against MOND is unfounded. Milgrom has written beautiful stuff explaining by the Bullet cluster results have not falsified MOND, in despite of many misconceptions in the science news.
Moreover, the same Bullet cluster is problematic for the dark matter hypothesis. See for instance the Astrophysical Journal article Bullet Cluster: A Challenge to ΛCDM Cosmology
And it has been also shown that MOND can explain the observed shock velocity in the Bullet cluster, whereas dark matter cannot. Check MNRAS 2007, 383, 417. A figure comparing MOND vs the dark matter prediction can be found online here
@ KWK,
Ok, thanks. The energy difference between the neutrinos from SN1987A and what IceCube is sensitive to explains why the time window is small enough. (10^10 vs 10^14 eV) Assuming my earlier calculations were correct, that means the average neutrino should be >3 eV to arrive even 3 hours later.
@ Kevin,
Thanks for the Wiki link, have to read.
Very informative post Sean, but I’m puzzled by the sentence
“Recently they have been keeping an eye out for signs of neutrinos that should be associated with GRB’s that have been detected by the Swift and Fermi satellites — but no luck”
Taking this sentence at face value, if GRB neutrinos have already been detected by SWIFT and FERMI , doesn’t that suggest a shortcoming of the IceCube experiment, rather than any great mystery about GRBs?
Regards, Cormac
Oops, I see it now – did you mean the GRBs have been detected by SWIFT and FERMI, not the neutrinos? In that case, do we know IceCube can deliver, has it detected other neutrino events successfully? I’m always wary of non-results with new experiments…
Yes, Fermi and Swift detected the gamma-rays, not the neutrinos. IceCube has detected plenty of atmospheric neutrinos.
@cormac–that’s right, the GRBs have been detected by Swift and Fermi because of the light they emit, not the neutrinos. And IceCube does see neutrinos all the time (usually those generated by other cosmic rays striking the atmosphere), but what it hasn’t seen–what no detector has seen yet, and part of what makes the whole project interesting–is neutrinos from specific astrophysical sources like Gamma-Ray Bursts or Active Galactic Nuclei.
Sean, can you report on Verlinde’s colloquium at Caltech?
Thanks
shantanu
Thanks Sean and KWK. In which case this is a very interesting result! I take it IceCube is sensitive to all types of neutrinos…(it’d be funny if it turned out to be similar to the missing solar neutrinos but I’m sure the boffins have that covered)
An alternative answer to this puzzle:
http://www.ioannisxydous.gr/
Best Wishes
Ioannis Xydous
Electronic Engineer
Sociological Experimental Challenge: Visit almost any physics dept.’s faculty page, & eyeball the pix. Nearly every faculty member displays a pic ~10 yrs younger than they actually are ! Its easy to ballpark their age, since on avg. most get their PhD ~ 25. Simply subtract the dates.
Dr.Freeze’s appears ~ 35; Her PhD is in `84 => 25+28 = 53. Dare I say we have a pic ~ 18 yrs out of date ? Male faculty pix exhibit ~ the same intentional time dilation.
Strikes me this is a blatant recruitment ploy, as bright young kids prefer bright young faculty, in this youth-obsessed culture we have. We need to project a dept. image of Who we Really are, not some air-brushed marketing of who we wish we were, or used to be.
Methodology wars between rival direct DM search teams, continued negative results both coming from both indirect and direct searches, no support for Supersymmetry– the leading particle physics theory that predicts the existence of extra massive weakly interacting particles. All of the above points to the fact the CDM theory is in serious crisis. On the one hand it’s good that scientists are doggedly testing the theory; on the other hand, the continued negative results aren’t very frustrating. After all, despite all of the advances in modern science, we still do not know what accounts for 95% of the mass-energy of Universe!
Jimbo #41:
Interesting comment (though of course completely off-topic). I’d never seen Katie Freese’s homepage, so went to verify your assertion. However I discovered a photo on there which is clearly from June 2011, and looks much the same as the main mugshot. So it would appear that it isn’t an elaborate outdated-photo-recruitment-ploy, just that you aren’t particularly good at guessing a person’s age from their photograph (neither am I).
Somewhat more disconcerting though is the link on her webpage offering me the opportunity to “download the entire zip file of photos of Dr. Freese”.
Sesh #43; You can argue interpretation all you want. I can get w/in 5-7 yrs at worst, & NoWay that pic is an accurate rendering of a 50-something woman. I suspect the zip file will serve up any dilation you desire. In addition, my observation is true of male faculty 40-60 as well. Ex.#2: PhD in `81 => ~ 56 yrs of age. Pic ~ 37: http://physics.berkeley.edu/index.php?option=com_dept_management&act=people&Itemid=299&task=view&id=361
:
@Jimbo: it is simpler than that, really: who has the time to update those pages? Typically, these are taken when a faculty member joins a department, then refreshed every ~20 years by the Chairperson, if he or she gives a hoot.
(but you are correct about KF’s age, yet she looks pretty much as in that picture: good genes in her case)
Recent observations of our Galaxy challenge the existence of dark matter:
http://www.ras.org.uk/news-and-press/219-news-2012/2118-do-the-milky-ways-companions-spell-trouble-for-dark-matter
Pingback: Någorlunda aktuellt: mörk materia och så | Stjärnstoft och kugghjul
Sean et al., not sure whether this thread is still read but just in case, I wanted to point out that the main error of Moni Bidin et al has now been spotted: to make a long story short, they neglected what is known as the “asymmetric drift”. Their error is that they assumed that the mean azimuthal velocity of stars is constant as a function of radius at all heights, but this is actually not true. What is constant at z=0 is the circular velocity for a “cold” population i.e. one that has very low velocity dispersion. But in reality, stars are not on perfectly circular orbits, and the more the velocity dispersion is high, the more you have elliptical orbits, the lower the mean azimuthal velocity. The difference between circular velocity and mean azimuthal velocity is known as the asymmetric drift, because it makes the local stellar velocity distribution asymmetric with a kink towards low azimuthal velocities. Now, the crucial point is that the velocity dispersion is *not* constant with radius, and its radial variation also depends on height. So the mean azimuthal velocity is not constant with radius, and its radial variation also depends on height. In reality, the gradient of the mean azimuthal velocity is not zero but about 7km/s/kpc in the plane, growing to 11km/s/kpc at a height of 1kpc, and 40km/s/kpc at a height of 3.5kpc. Moni Bidin et al. had claimed that a gradient of 17 km/s/kpc on the mean azimuthal velocity was needed to be compatible with the expected amount of DM in the solar neighbourhood, but dismissed the possibility by confusing the mean azimuthal veklocity with the circular velocity. But, at their probed heights, this gradient is perfectly compatible with the numbers above. QED This is all explained in more details in today’s rebuttal: http://arxiv.org/abs/1205.4033
Ben, I hadn’t seen that, thanks for pointing it out.
Pingback: Dark Matter: Still Existing (One in a Continuing Series) | Cosmic Variance | Discover Magazine