“Nothing travels faster than light, with the possible exception of bad news, which follows its own rules.” -Douglas Adams

My inbox is on fire today with messages about this story about neutrinos breaking the speed of light:

What’s going on here? A group (a large group, mind you) of physicists known as the OPERA collaboration have made a neutrino beam, and have been studying it for the past few years.

Making a neutrino beam is the easiest type of beam to make, by the way. All you do is shoot a bunch of high-energy particles into the Earth, like so.

Image credit: CERN Neutrinos to Gran Sasso.

You shoot a high-energy beam of protons into a fixed target, and you make all sorts of unstable particles — things like pions, kaons and other mesons — which have a lifetime of at most a paltry few nanoseconds.

You focus this beam very tightly, so that the decay products you get out travel in a narrowly collimated beam as well. What are these decay products?

Among other things, you get a bunch of high-energy muon neutrinos. And if you fire it through the Earth, everything that isn’t a neutrino gets wiped out in short order by the intervening atomic material.

But the muon neutrinos, for the most part, pass straight through the Earth uninhibited. Why? Because neutrinos hardly interact with anything at all! We’ve built neutrino beams like this before: from Fermilab (in Batavia, Illinois) to Minnesota, from KEK (in Japan) to Super-Kamiokande, and others.

And what we’d expect, based on measurements of neutrino mass, is that these particles should be traveling at almost, but just a hair under the speed of light!

And then you go and detect your neutrino.

But I just said they don’t interact with anything! So how do you do this?

Image credit: Super-Kamiokande.

You build a giant tank of something liquid for neutrinos to interact with. And although nearly all of your neutrinos pass right through it, every once in a while, one neutrino undergoes an interaction (through the weak force) with one of the atoms in your detector!

And when it does, because of how hugely energetic these neutrinos are, you produce either a muon (for a mu-neutrino) or an electron (for an electron-neutrino) that’s moving close to the speed of light in vacuum, and faster than the speed of light in your liquid!

Image credit: Georgia State University.

When you move faster than the speed of light in a medium, you give off a special type of light known as Čerenkov radiation. If you line the outer rim of your neutrino detector tank with photomultiplier tubes, you can not only detect this radiation, you can use the information from it to reconstruct exactly where and when, in your tank, this neutrino interacted with one of your atoms!

Image credit: Tomasz Barszczak.

Now, in the past, we’ve found that these neutrinos move, more or less, at the speed of light in vacuum (c), as expected. One experiment based out of Chicago, a few years ago, found marginal evidence that neutrinos might move just a tiny bit faster than the speed of light, at 1.000051 (+/- 0.000029) c.

Of course, this result is consistent with neutrinos moving at or slower than the speed of light; the errors are not significantly smaller than the measured difference from c. So OPERA, whose detector is shown below, performed this measurement with great care, and announced their results today.

The 730 kilometer trip should have taken these neutrinos 2.43 milliseconds, were they traveling at the speed of light. But according to the OPERA collaboration, the neutrinos arrived 60 nanoseconds earlier than expected, with a claimed uncertainty of only ten nanoseconds!

Translating that into a measurement for the speed of neutrinos, that means they are traveling at 1.0000247 (+/- 0.0000041) c.

Now, measurement at this level of precision is not easy, and I am certainly not going to be the first person to come out and say I don’t believe, based on this, that neutrinos move faster than the speed of light. (But, as one of many, I don’t.)

Because there’s a much better constraint out there on the speed of high-energy neutrinos from some time ago. Above is a Hubble Space Telescope time-sequenced image of the closest supernova in my lifetime: Supernova 1987A, which took place in the Large Magellanic Cloud 168,000 light-years away.

This supernova was discovered, optically, on February 24, 1987. About three hours earlier, 23 neutrinos were detected over a timespan of less than 13 seconds. The reason for the 3 hour delay? When the core of a star collapses (in a type II supernova; see here), most of the energy is radiated away in the form of neutrinos, which pass freely through the outer material of the star, while the emission of visible light occurs only after the shock wave reaches the stellar surface.

Image credit: TeraScale Supernova Initiative.

However!

Even if you assume that the light and neutrinos were created at the same time, but the visible light moved at c and the neutrinos moved faster than light, which is why they got here first, know what value you’d get for the speed of these neutrinos?

1.0000000020 c, which is inconsistent with the results from the OPERA collaboration.

Now, something fishy and possibly very interesting is going on, and there will certainly be scientists weighing in with new analysis in the coming weeks. But in all the excitement of this group declaring that they observe neutrinos moving faster than the speed of light, don’t forget what we’ve already observed to much greater precision! And be skeptical of this result, and of the interpretation that neutrinos are moving faster than light, until we know more.

Comments

  1. #1 JeffW
    September 22, 2011

    Thanks for the reply and the clarity. Oh and… FIRST!

  2. #2 Excited State
    September 22, 2011

    “But the muons, for the most part, pass straight through the Earth uninhibited. Why? Because neutrinos hardly interact with anything at all!”

    Typo, I believe. Muons should be neutrinos in the first mention.

  3. #3 Dave
    September 22, 2011

    Straight away on reading this story in the press, I thought “Better check out what Ethan has to say!” and as usual I’m rewarded with a thorough and clear explanation of why I should remain cautiously skeptical of press headlines on science press releases. Thanks Ethan!

  4. #4 William Russell
    September 22, 2011

    When you say “…I am certainly not going to be the first person to come out and say…” do you mean you aren’t saying that at all, or that you are saying it, but you’re certain others have already done so?

  5. #5 Ethan Siegel
    September 22, 2011

    Excited State @2: fixed; thanks.

    William Russell @4, I am not the first, nor am I the only. I added a link to Lubos’ take, and clarified in the text above.

  6. #6 John Halewood
    September 22, 2011

    What occurs to me is that we might have this the wrong way round. Neutrinos pretty much fail to interact with anything, so (if massless) travel at the speed of light. Photons, however, do interact, otherwise I wouldn’t be able to see to write this. If one is consistently seeing neutrinos from an event arriving before photons, and experimental error can be ruled out, then I would assume that the photons are being slowed by interacting with something along the path. This may be a clue to the dark matter/energy theory (if said presumed stuff interacts with photons more than neutrinos).
    On the other hand, I recall this is also predicted by the Standard Model Extensions, which include Lorentz invariance violations at high energies, but that was never really my field of physics.

  7. #7 James Davis
    September 22, 2011

    I note that the guys who discovered it think it’s probably a systematic error of some kind, and have said that they want help figuring out what the error probably is. This is yet another case of bad media reporting, I think.

  8. #8 Ethan Siegel
    September 22, 2011

    From the humor department:

    If this result were verified, Italian scientists would next attempt to apply these results to trains. Perhaps someday, one of them might actually arrive early, too.

  9. #9 Atlee
    September 22, 2011

    Ha! My friend called me a few minutes before this was up and I quoted Carl Sagan saying that, “extraordinary claims require extraordinary evidence.”
    I then said I fully expect a blog post very soon on Science Blogs. I am glad to see that the title is almost verbatim what I said.
    This post was excellent, especially when explaining the experiment.

  10. #10 Lloyd
    September 22, 2011

    As always, the message is in the medium. Thanks for keeping watch.

  11. #11 Andrew Foland
    September 22, 2011

    Apologies if I missed it, but I didn’t see this in the writeup above: the actual OPERA arxiv submission.

  12. #12 NewEnglandBob
    September 22, 2011

    Since thoughs are way faster than light or neutrinos, I knew this several years ago.

  13. #13 Ethan Siegel
    September 22, 2011

    Andrew, thanks for the link.

    A cursory read-through of their paper suggest that there are many questions to be addressed, such as their claimed accuracy of 20 cm in the accuracy of their baseline distance, their quantification of systematics errors as 7.4 ns (they add them in quadrature, which is standard practice; if they added them linearly, the error would be more like ~18 ns), etc.

    It is worth noting that even scaling the energies from the SN1987a explosion (which has a peak neutrino energy of ~10 MeV) up to OPERA’s energies of ~3 GeV would be insufficient to allow these two pieces of data to be consistent with one another. I would be willing to bet dollars to donuts that there is a mistake somewhere in their accounting for those 60 nanoseconds, but it sounds like that is what their hunch is, too.

  14. #14 Randy Owens
    September 22, 2011

    1.0000000020 c, which is inconsistent with the results from the OPERA collaboration.

    Would there actually be any reason to expect SN1987a neutrinos to travel the same speed as CERN’s neutrinos?

    Added: I see you just mentioned the comparative energies, and that would suggest that, if anything, CERN neutrinos should be even slower than SN1987a neutrinos, under the standard assumption that for superliminal particles, more energy means slower speed. But I’ll let the question stand as a matter of abstract interest.

  15. #15 Andrew Foland
    September 22, 2011
  16. #16 Fred Bacon
    September 22, 2011

    I’ve already had a similar discussion elsewhere. A back of the envelope calculation suggests that if their numbers were correct, then the neutrinos from SN 1987a would have arrived roughly 4 years before the light from the supernova. Their not even close to being right.

  17. #17 Jonathan
    September 22, 2011

    This is why Physics demands 5 sigma certainty.

  18. #18 Raging Bee
    September 22, 2011

    I’m surprised the religious wingnuts aren’t already here claiming this proves that science can never be trusted.

  19. #19 Randy Owens
    September 22, 2011

    @Raging Bee: Didn’t you see? They all fell off the edge of the earth two posts ago. No wonder they’ve been so quiet.

  20. #20 Robert S.
    September 22, 2011

    At what level of precision do gravitational effects start making a difference. Does the path that goes down 11km from the geoid make any difference? If so, how do you calculate the path length/time differential from a straight line, in vacuum with no gravitational disturbance?

  21. #21 Daniel
    September 23, 2011

    I just finished going over the paper myself; what struck me was the direct correlation of particle speed with the degree of superluminality (figure 13). Now, the paper goes on to say that the uncertainty is too high to verify the existence of this energy dependency, and in all fairness, that’s a perfectly fair conservative tack because the error bars are larger than the effect size.

    If, however, this relationship did exist, it would explain the results of previous lower-energy experiments _and_ the timing of the SN1987A pulse. I don’t think the neutrino is a superluminal particle in the same way as a tachyon, and so wouldn’t inherit its inverse relationship between energy and speed. If these data are correct, they suggest that the c barrier simply doesn’t exist for high-energy neutrinos, and that they smoothly cross the c line, though the relationship between energy and speed clearly can’t be linear. (Otherwise, the SN1987A neutrinos would have arrived far later than they did.)

  22. #22 Thomas
    September 23, 2011

    Ok, fair warning, I am a Political Scientist, which is a contradiction in terms at best, and explains any mistakes I make by attempting to comment on this subject.

    So, as an uninformed observer, I’ll add my belief that its much more likely that there is a problem with the Opera observations than there is with a theory as well tested as Special Relativity.

    But given that, I am interested in the Hubble observations of SN1987A and the neutrino burst that was observed approximately 3 hours before that.

    First, am I wrong that when we detect a burst of neutrino’s it does not come with a return address? And, in fact, that we “know” what caused the observed burst of 24 neutrinos, because we had visual observations of a supernova three hours later? From these two observations we then are able to graph the speed of neutrinos to the speed of light with high precision.

    However, because of the huge distances involved, isn’t it also true that ANY observed neutrino burst we might have detected around that time period would graph very closely to the speed of light, regardless of where it’s actual origin was?

    And in fact there was a separate burst which was detected three hours before the 24 neutrino burst which consisted of 5 neutrinos, and which is not generally sourced to SN1987A (according to your wiki article).

    So my question is, how often do we receive bursts of neutrinos which would be consistent with (or close enough for government work) models of stellar collapse similar to what we think happened with SN1987? If we detect such bursts every day or every week, it seems likely that one could just happen to be detected at a time that was convenient for us and cause us to assume that the two events were related.

    If we eliminate the assumption that neutrinos travel at the speed we think they do, how solid is the observation that the 24 neutrino burst we observed coincident with the SN1987A observation?

  23. #23 Thomas
    September 23, 2011

    Ok, fair warning, I am a Political Scientist, which is a contradiction in terms at best, and explains any mistakes I make by attempting to comment on this subject.

    So, as an uninformed observer, I’ll add my belief that its much more likely that there is a problem with the Opera observations than there is with a theory as well tested as Special Relativity.

    But given that, I am interested in the Hubble observations of SN1987A and the neutrino burst that was observed approximately 3 hours before that.

    First, am I wrong that when we detect a burst of neutrino’s it does not come with a return address? And, in fact, that we “know” what caused the observed burst of 24 neutrinos, because we had visual observations of a supernova three hours later? From these two observations we then are able to graph the speed of neutrinos to the speed of light with high precision.

    However, because of the huge distances involved, isn’t it also true that ANY observed neutrino burst we might have detected around that time period would graph very closely to the speed of light, regardless of where it’s actual origin was?

    And in fact there was a separate burst which was detected three hours before the 24 neutrino burst which consisted of 5 neutrinos, and which is not generally sourced to SN1987A (according to your wiki article).

    So my question is, how often do we receive bursts of neutrinos which would be consistent with (or close enough for government work) models of stellar collapse similar to what we think happened with SN1987? If we detect such bursts every day or every week, it seems likely that one could just happen to be detected at a time that was convenient for us and cause us to assume that the two events were related.

    If we eliminate the assumption that neutrinos travel at the speed we think they do, how solid is the observation that the 24 neutrino burst we observed coincident with SN1987A is in fact linked?

    I think I saw one comment that if the Opera results were real, we would have detected the neutrinos 4 years earlier than we did. So did we? :)

  24. #24 Thomas Fisher
    September 23, 2011

    Stupid internets. Didn’t mean for that to be posted (the second one is the one I intended) so sorry for the double post :)

  25. #25 Thomas Fisher
    September 23, 2011

    @Daniel wrote “(Otherwise, the SN1987A neutrinos would have arrived far later than they did.)”

    Do we actually know that the observed neutrinos were related to SN1987A in any way besides the time that we detected them? Supernovas and other cosmic events which cause neutrino bursts must happen every day. How often do we detect 24 neutrino bursts? If neutrinos travel at ANY speed but their standard predicted speed, it seems highly unlikely that the observed burst was related to SN1987A at all.

    It also seems that the simple fact that we observed any burst coincident with SN1987A does not prove that neutrinos must go a certain speed. Any burst even remotely close in time to the SN1987A observation that we “linked” to SN1987A would perforce be “measured” at a speed extremely close to the speed of light regardless of whether it came from half a universe away from SN1987A.

    If an uninformed observer watching an explosion take place from a distance of three hundred miles at the same moment that a mischievous elf set off a firecracker 5 feet away, that observer might be compelled to state that the speed of sound was something close to the speed of light because the two events arrived within moments of each other. And turning away from his telescope he might not note the arrival of the ACTUAL sound some time later because the two “events” are not apparently “related”.

    At the distances involved with SN1987A, it would seem that if we had chosen to “link” the State of the Union speech that was given that year to the explosion of SN1987A we could “measure” the time it took to reach us and conclude that it must have travelled that great distance at something very close to the speed of light (ok, I haven’t done the math, but is that even an exaggeration?)

    Anyways, have subsequent neutrino bursts been successfully used to predict an observed supernova (even post facto)?

  26. #26 Ethan Siegel
    September 23, 2011

    Thomas,

    As far as I am aware, SN1987a is the only supernova that occurred close enough to us that neutrinos from it have been able to be detected.

    A neutrino detector can give you both the energy and the initial direction (from kinematics of the reconstructed Cerenkov photons) of the neutrino, so we “know” that the 23 neutrinos associated with SN1987a were, in fact, coming from the direction of the LMC.

    (Incidentally, a small number of neutrinos were observed a few hours earlier, and those most definitely do not correlate with SN1987a.)

    No, we did not detect the neutrinos four years before we should have; the OPERA results are woefully inconsistent with the other observations. Reading the OPERA paper, and looking at their figures, I would in no way draw the conclusion that their 60 ns result is robust. I fully expect that this will fail to be reproduced.

  27. #27 Thomas Fisher
    September 23, 2011

    Thinking more about the issue, the critical evidence linking the particular 24 neutrino spike (as opposed to the 5 neutrino spike also observed that day) to SN1987A is how well the scatterplot of neutrino detections fits into our models of stellar collapse, and our “understanding” about how fast neutrinos travel.

    It seems apparent that the neutrino burst that we have decided is associated to SN1987A has done a lot to set boundaries on our models of stellar collapse. If the CERN data stands up (which it likely will not) then SN1987A is almost certainly bad neutrino science, and according to Wiki at least, it is responsible for setting “upper bounds on neutrino mass and charge, as well as the number of flavors of neutrinos and other properties.”

    So again, if we were back in 1987, and our thinking is unpolluted by those observations, how often do neutrino spikes happen that are consistent with what we expect from an event like SN1987A? And what sort of detection time could still be explained comfortably in those models?

    The 3 hour difference worked obviously. What if the detection had happened 5 hours before the visual observation, or 2 hours before? At what points do the models break down and we say any observation outside of a given range could not have been consistent with the visual data. That range of the possible “good” observations should give us the probability of getting a false positive if we know how often a neutrino burst happens that would be “good” enough to fit our models occurs.

    Bad math example alert – if a neutrino burst consistent with models randomly occurs once a day, and the models give us a 6 hour range when a burst can happen coincident with a visual observation, what are the odds that we will get a burst at a “good” time coincident with an actual visual observation. 25% I think, but its late and I never was good at math. Still I am pretty sure that its non-trivial.

    Would the 5 neutrino burst that we actually recorded that day have been excluded if not for the 24 neutrino burst that happened a few hours later? What if we had received a 50 neutrino burst instead of a 5 neutrino burst? Would that have been excluded, or would we have excluded the 24 neutrino burst?

    Since I personally don’t know the answer to these questions, it seems really problematic to me that we would point to SN1987A as somehow disproving Opera all by itself, if there is a non-trivial chance that we could have randomly detected a neutrino burst that day that we could “link” to the visual observations.

  28. #28 Thomas Fisher
    September 23, 2011

    @Ethan yay, I just wasted my time with that last post then :P

  29. #29 Thomas Fisher
    September 23, 2011

    Maybe you can suggest a change to the Wiki article which states that “Approximately three hours earlier, the Mont Blanc liquid scintillator detected a five-neutrino burst, but this is generally not believed to be associated with SN 1987A.”

    When I read “generally not believed to be associated” it leaves me with an impression that is quite far from “those most definitely do not correlate with SN1987a” :)

  30. #30 javad
    September 23, 2011

    Thank you Ethan. Please update us when you reach to a new point or conclusion regarding this subject.

  31. #31 Thomas Fisher
    September 23, 2011

    Looking at an abstract of one of the initial detections they do in fact determine the area of the sky that the neutrinos came from.

    “The signal consisted of 11 electron events of energy 7.5 to 36 MeV of which the first 2 point back to the Large Magellanic Cloud (LMC) with angles 18°±18° and 15°±27°.

    As taken from http://www.sciencedirect.com/science/article/pii/0375947488908445

    A 36°x54° range seems like a sizable chunk of sky. I’ll wait for them to figure out exactly how the Opera folks screwed up, but I’m still concerned that if potentially “valid” events happen often enough, there was a non-trivial chance of a false positive if neutrinos don’t travel as fast as we think they do.

    Anyways, thanks for the response. When someone puts out something like this, it puts me into a stream of consciousness mode and causes me to start posting uninformed plea’s for data like these ones :P

  32. #32 Thomas Fisher
    September 23, 2011

    Ok. I *THINK* that a chunk of sky 36°x54° represents only about 1.5% of the potential sky. Its really been so long since I’ve had to do real math that I have about the same amount of confidence in that answer as the value I came up with. If it IS 1.5% I guess that isn’t so bad from an odds standpoint when determining the chance of a random burst being mistakenly linked to SN1987A.

    There would have to be a broad range of acceptable bursts and they would have to happen quite often to have a significant chance of one also appearing in the tiny window in the sky that was available based on the observations. I think :P

  33. #33 Thomas Fisher
    September 23, 2011

    Not to mention that a foundational theory would have to be wrong. Very low odds indeed. Time for bed! :)

  34. #34 xhunterko
    September 23, 2011

    Ehem.
    “She’ll make .5 past light speed.”
    *flees*

  35. #35 bonbon dondon
    September 23, 2011

    i like tits!

  36. #36 a@b.com
    September 23, 2011

    Perhaps they just got the distance between the two points wrong ?
    Maybe Tidal effects move the earths crust a little.
    Moybe there was an earth quake that shifted the region just a little ?

  37. #37 Sinisa Lazarek
    September 23, 2011

    I believe that the general attitude should be of major congratulations to all of the people of OPERA instead of scepticism and saying they got it wrong. Got wrong wha texactly? No one is claiming anything here, except the media. The thing is that this is more likely than not a major scientific discovery even if it turns out that neutrinos don’t travel faster than the speed of light. I would recommend to all to read through the published paper, the link can be found on the original BBC article.

    The people involved in OPERA are not your armchair physicist wannabes, these are all hard core, top of the bunch, experimental physicists. And it’s not a group of 3-4 people in the garage. Like Ethan said, it’s a LARGE group.. around 50-60 physicist from the most prestigious of the world’s institutes. Some credit should be given, a lot of credit in fact.

    On the other hand, some wrote here that maybe this and that happened, so it might be fluke. You failed to read that this experiment has been going on for 5 or so years, and is still going. The experiment has been in fact repeated 16.000 times!!!! And you get the results which they got, over and over and over again, seems like the neutrinos are traveling faster than “c”. And like any other sane person, when something like this happens, you tend to start believing that something is seriously wrong here. And that what they are doing. They published the findings and after years of trying to find the error, and are saying “please help us find the mistake, if there is one”. They are not claiming that Einstein is wrong, only media is hyping that.

    Why I said that this might be a major discovery? Well, this sounds to me awfully like Penzias and Wilson experiment when they unintentionally discovered CMBR. OPERA wasn’t intended to test weather or not neutrinos travel faster than speed of light. Their measurement of this was only a side process which got them curios from the start. So maybe this is in fact a discovery of something totally new about subatomic particle behavior, or about space and time, or it may turn out that some of the math and formula are not good. Any of these will be great discoveries. And who knows maybe even that “c” is not the upper limit. Again, I think praises for these scientists should be in order.

  38. #38 john werneken
    September 23, 2011

    Appreciate the article (mal, at least here) plus the discussion (not normal on the internet)

  39. #39 mk
    September 23, 2011

    Also from the humor dept….

    http://xkcd.com/955/

  40. #40 Andrew Foland
    September 23, 2011

    Seems my earlier attempt to post this ended in moderation purgatory somehow.

    The now-public internal Opera note on geodesy.

    The location of the detector would have to be off by 60 feet to account for what they see. The GPS very plausibly has the low distance uncertainties described. The weak link is that they don’t have GPS in the experimental hall at the detector, but outside the lab. They carry the coordinates inside, from GPS markers outside the lab into the detector hall. It’s about 1.8 km from the “GPS monuments” outside to the detector. There’s really no plausible way (in my limited surveying experience) to mis-survey the absolute magnitude of this distance by 60 feet; 20 cm would seem very generous.

    The one thing that I can even remotely see is the accuracy of the angle of the displacement. It works out that you’d have to mis-align the survey coordinate system to the GPS coordinate system by about 2 degrees. Again, in my limited survey experience that’s a gigantic amount for a survey to miss by.

    But if I have to choose between a big survey goof and superluminal neutrinos, my Bayesian prior for the former is certainly higher.

  41. #41 Alex Besogonov
    September 23, 2011

    Also, if neutrinos DO travel faster than light – then we’re in the uncharted waters.

    It might be possible, that distance between neutrino detector and emitter can affect the speed of neutrinos. Hey, once you reject causality even stranger things can happen!

  42. #42 g724
    September 23, 2011

    I’m thinking that this was caused by a problem with the apparatus, that produced consistent but erroneous results. Not a “measurement error” in the procedural sense, but something subtle that’s built into the hardware or the data collection software.

    By analogy and from experience, there is an entire industry producing wind resource survey instruments for the wind energy industry, that are used to calculate the probable output of sites for utility-scale wind farms. These instruments all had the same inherent flaw in the way they integrated the raw data. That flaw produced the consistent result of over-estimating usable wind on wind farm sites, with the result that the actual wind turbines that were installed, appear to under-perform. Yet the output of the flawed instruments was “certified” and was taken as evidence in court in a couple of cases where investors sued claiming that the underperformance of the wind sites (compared to the survey results) was evidence of some kind of fraud or mismanagement. But in fact the problem wasn’t what management or the engineers did with the data, it was in the hardware and software of systems that were producing consistent but erroneous results.

    As above, so below, thus it wouldn’t surprise me if something similar was at work in the present case. But whatever comes of this is going to teach us something interesting.

  43. #43 Collin
    September 23, 2011

    Ironic that just minutes ago I posted a thought going the other way. Oh well, learning trumps guessing…

    If they’re traveling faster than c, and we know that they change flavor at a real rate, does that mean they’re Majorana particles?

    And if there is such a thing as a Majorana particle, how does it respond to gravity?

  44. #44 Wow
    September 23, 2011

    “Neutrinos pretty much fail to interact with anything, so (if massless) travel at the speed of light”

    It may be that if Higgs Bosons exist and cause “mass”, that maybe neutrinos interact so poorly that they have mass if you can get them to hang around long enough, but the faster they go, the less the Higgs is informed of the existence of that neutrino and therefore the less mass it has.

    Sort of the way that the naked electron charge in QM is higher than -e in the schoolbooks because virtual particles (which have positive and negative charges in their pairs) shade the naked charge of the electron until we get the observed electron charge at a distance.

    Given we don’t know WHY going faster than the speed of light makes things more massive, just that it works if we count it so, I don’t see any a priori need to say this is an extraordinary claim to make.

    Surprising, but no more extraordinary than the claim of wave-particle duality: the result of the model we use to explain the observations of the universe existing.

  45. #45 Wow
    September 23, 2011

    “So, as an uninformed observer, I’ll add my belief that its much more likely that there is a problem with the Opera observations than there is with a theory as well tested as Special Relativity.”

    We already know of several problems with SR..!

    I would agree that it’s more likely that there’s an incorrect assumption, but since we already know that GR has a problem at the very small scales, we already know that many models we have are stupendously successful (read: accurate), they do break down in odd places.

    Think of this as the scientists’ pot of gold: the sentence most treasured to be heard from any inquiring scientist is “Oh, that’s odd!”.

    It’s only when we spot something odd we have the opportunity to learn something new.

    Even if that something is just new ways of getting it wrong!

  46. #46 Wow
    September 23, 2011

    “As far as I am aware, SN1987a is the only supernova that occurred close enough to us that neutrinos from it have been able to be detected.”

    It’s the only one we spotted that had the require characteristics (as an astrophysicist :-) )

    1) We have detectors for neutrinos that can spot these
    2) We saw the supernova early enough to catch most of the explosion
    3) It’s close enough for us to make a deduction

    NOTE: #2 means that if there were a pulse of fast neutrinos at the beginning, then we weren’t looking for them at the same time.

    As to “four years earlier”, well we could have seen a pulse of neutrinos four years earlier, not seen visually anything that could have caused them, disregard them as equipment error and start again.

    The Sanduleak A supernova only has to produce subluminal neutrinos too to make it fit with superluninal ones.

    The way to find out independently if it’s wrong is to look at the conclusions. If they travel at FTL speeds, what happens when they DO interact with matter? Could this happen but be limited to 5cm at 5% faster than light, 10cm at 50% faster, … etc?

    In other words, if it were true, what would be the consequences? Can we see those consequences around us? Were we looking but didn’t understand (cf the photoelectric effect before the corpuscular nature of light was proposed).

    I expect that “No superluminal neutrinos” is going to be the answer. But it’ll be interesting to find out either way.

  47. #47 Wow
    September 23, 2011

    “38

    Also from the humor dept….”

    Of course, you can always insult someone and then, if they complain, just go “I was only kidding!”.

    Just pointing out that “from the humour department” isn’t really much of a come-back. “OK, sorry, that was meant to be humerous, please take it in that fashion” would work better than a smartarse “here’s XKCD!”.

  48. #48 bsci
    September 23, 2011

    While this is way out of my areas of understanding, what’s the possibility that this is actually showing we slightly underestimated the speed of light in a vacuum? Would minutely raising the speed of light cause fundamental problems with other physics theories or just required us to adjust all universe distances? As #6 John noted, photons do interact with more stuff than neutrons. What would it mean if it turned out the “The speed of light” whatever that means wasn’t defined as the speed of photons in a vacuum?

  49. #49 Rob Knop
    September 23, 2011

    Before I say anything, I have to say that I too very strongly suspect that we’re dealing with a subtle and hard-to-find systematic error rather than Lorentz violation.

    That being said….

    If I’m not mistaken, the neutrinos from 87A were electron neutrinos. That’s certainly mostly what you get when making a neutron star. It would be cool if it turned out that some, but not all, flavors of neutrinos turned out to be superluminal particles.

  50. #50 Wow
    September 23, 2011

    “we slightly underestimated the speed of light in a vacuum”

    We’re able to measure the speed of light (ignoring Tembla’s iconoclastic reading of “defined quantity”) to less than 1m/s accuracy.

    The difference here is smaller than that, so I don’t believe this is likely (to the three sigma limit). Not *impossible* (five sigmas), but very unlikely.

    It’s a little more likely we have the speed of light in interstellar or intergalactic space “wrong”, since we don’t have a lab big enough. By “wrong” here, I mean a velocity assumption that gives us an assertion of reality not in accord to the objective reality.

  51. #51 IW
    September 23, 2011

    Why should extraordinary scientific claims be held to a different standard than any other scientific claim? All scientific claims require sufficient evidence to establish their veracity, no more.

    Having said that, thanks for blogging on this. Phil Plait has some good observations, too:
    http://blogs.discovermagazine.com/badastronomy/

  52. #52 Paul
    September 23, 2011

    If the effect does turn out to be real, it seems that the easiest reconciliation with with 1987A data is that the transition between flavors of a neutrino involve a change in position. For example, we could imagine a neutrino as a four dimensional egg shape intersecting our 3 dimensional space. The egg moves at almost c, but if it rotates from its short plane intersecting our perception of space to its long plane, then the leading edge has traveled faster than c. From our perspective you’d have, say, a muon neutrino at t=0, then suddenly at t=epsilon you’ve got a electron neutrino 50 cm further along. If that were the case, you’d consistently measure certain neutrinos arriving 60ns early, regardless of the distance between emitter and detector.

    I agree that a measurement error does seem like the simplest explanation pending replication at other sites, but it’s interesting to consider how this could be possible.

  53. #53 Wow
    September 23, 2011

    It’s in the definition of extraordinary in science, IW.

    An extraordinary claim is one which contradicts directly another well-established case.

    This one could go either way, depending on how you rate the extremely low mass of the neutrino and whether you count the previous standard claim that neutrinos went at the speed of light.

    If your claims rebut another, better established (as in having proven itself accurate to predict results), your evidence had better be as good or better than all the evidence that the incumbent theory has.

    In other words: not only do you have to prove yourself right, but how and why the other is wrong.

    (Note, just like the problems the deniers have with their theories: they are unable to prove their theory right and how and why the IPCC proposition wrong)

  54. #54 Donovan
    September 23, 2011

    Thanks for an explanation I can understand. So isn’t the key issue here the precision of the distance measurement? (as pointed out by @dcastelvecchi Davide Castelvecchi ‘Would Opera results be more proper titled “Distance measurement using neutrinos, the most accurate available, shows other methods were off”?’ and RT’d by @SciAm)

  55. #55 Percy
    September 23, 2011

    Looking at this story, I do not doubt it is possible or even likely. Consider that if you take all the science in the world, and match it up to everything we don’t know about the universe, about the quantum physics and how we humans play a part in it all.

    I point to the Observer Effect of course. We assume what we see is correct, because we believe it to be correct based on what we have observed and what we believe true through what people like Elbert Einstein have said. However, I point to you that it is more likely we have changed and limited ourselves within a narrow band of thought as well and when a truth exist we refuse to see it for what it actually is unbiased by our “logical” minds.

    I would love to see these long standing theories about physics disproved.

  56. #56 Wow
    September 23, 2011

    “I point to the Observer Effect of course.”

    The what?

    There’s a bias of a human observer, but that’s what peer review and skepticism in science is there to counter.

    “I would love to see these long standing theories about physics disproved.”

    Why?

    *I* would love to see our understanding of the universe deepen. Note too that the replacement theory of General Relativity did not disprove Newtonian gravitational theory.

  57. #57 Scott
    September 23, 2011

    I clearly am randomly guessing based on my undergraduate physics knowledge and I googled negative neutrino mass. I saw this paper talking about neutrinos being superluminal fermions published in 2000.

    I’m sure it probably isn’t relevant but interesting. http://arxiv.org/abs/hep-ph/0009291

  58. #58 daedalus2u
    September 23, 2011

    Do they assume Euclidean space for the baseline length calculation? Since the path is somewhat deeper in Earth’s gravitational potential well, the path should be a little bit shorter than what you would calculate connecting the points via low potential paths.

  59. #59 daedalus2u
    September 23, 2011

    When they calibrated the response time of the photomultiplier tubes, they used a picosecond UV pulsed laser. The response time of a photomultiplier depends (in complex ways) on the magnitude of the optical pulse received. A large photon flux causes greater electron emission which can cause greater electron space charge in the photodynodes and slow the response time.

    A slower response time in the calibration with the high intensity pulsed UV laser then results in an artifact where the actual photomuliplier response to neutrino photons is faster because there are many fewer of them.

  60. #60 daedalus2u
    September 23, 2011

    What they could do to test the Euclidean space thing is get a coil of optical fiber and measure the transit time in it at one gravitational potential and then measure it in another gravitational potential.

  61. #61 cruf
    September 23, 2011

    A pretty straightforward generalization of Einstein’s rule is that nothing can accelerate to the speed of light, either from above or below. So subluminal things like snails and rocketships and electrons can’t speed up to c and superluminal things (mabe like these neutrinos they’re detecting) can’t deccelerate down to c. The problem with superluminal particle is it’s very hard to iinteract with them and detect them. Maye that’s why neutrinos are so hard to detect in the first place. Incidentally, the “rest state” of superluminal particles ought to have infinite velocity with respect to our reference frame. So they are basically everywhere at once. This sounds suspiciously like dark energy to me – we can’t detect it but it is apparently all over the place.

    If 73% of the total mass-energy in the universe is dark energy and another 23% is dark matter, that only leaves 4% for everything traveling slower than the speed of light. It would seem more natural for there to be equal mass-energy on either side of c. If the velocity of all the tachyons (> c) relative to us was infinite, they would appear to be everywhere all the time and their portion of the mass-energy budget would tend toward 100%. The fact that it is only 96% is a signature of their actual velocity distribution.

    So tachyons appear to have more energy than they actually do. And the excess grows as their speed increases, asymptoting to infinity as their velocity (relative to us) goes to infinity. This is probably a result of how we are interpreting the indications of dark energy’s presence – using inappropriate subluminal physics.

  62. #62 NotoriousRoscoe
    September 23, 2011

    So according to Einstein, the nutrino would be of infinite mass once accelerated to the speed of light, right? I wonder if the involved physicists have unwittingly discovered the wild nutrino, unique among subatomic entities, sheds its mass to zero as it accelerates toward light speed so as to exceed it, “flipping” into the “spiritual realm” for lack of a better way to say “non physical” — and reacquires its tiny mass upon deceleration, “flopping?”, below the physical universe’s speed limit? Perhaps that’s the one thing we need to know so we can understand how to “fold space” and communicate and travel using interdimensional transitions to bypass light speed. (Wild dreaming.)

  63. #63 RiverC
    September 23, 2011

    Is it possible that the neutrinos travel at true ‘c’ because they are unimpeded?

  64. #64 RiverC
    September 23, 2011

    What I mean is that, maybe it’s not mass that increases with speed, but the effect of that mass as dictated by how it interacts with other masses. Any mass which does not interact with other masses normally can go ‘c’ with impunity, provided there is some way to give it that velocity.

    Since we measure mass by an interaction, there doesn’t seem to be a distinction in the definition I learned in physics between effective mass and real mass. But if a neutrino has real mass but goes ‘c’, then maybe it’s not strictly mass that’s important but how that mass interacts with surrounding masses?

    Just my 2c.

  65. #65 NotoriousRoscoe
    September 23, 2011

    I shouldn’t have said the physicists “unwittingly” discovered new properties of the wild neutrino. A happy accident is more like it — took a whole boatload of wits to get here. But mightn’t they also have created something that happens naturally? Wouldn’t the neutrinos’ everyday, “flip,flop” access to a “spiritual” or “non-physical” infinite-speed shortcut go a long way toward explaining the instant communications between “pairs of electrons” or particles, or whatever, which weirdly exhibit identical behaviors and traits though they may be at opposite ends of the universe? (Wild dreaming, continued.)

  66. #66 Jake
    September 23, 2011

    Sinisa Lazarek points out something to keep in mind – the OPERA team repeated the experiment a large number of times over several years with consistent results.

    While a measurement error is certainly the simplest explanation, it’s also important to note that the observations of SN1987A were a single event – they have never been reproduced because, as Ethan pointed out, there haven’t been any other supernovas close enough for us to measure. Additionally, the relationship (distance, intervening materials, etc.) between the source and the detector in the OPERA experiments has been controlled as much as possible, while the relationship between SN197A and the detector was completely uncontrolled.

    The big question, as always, is whether the results can be duplicated by other scientists using separate equipment, and preferably in a separate location. But (once measurement error is eliminated as a possibility) I would trust a repeated controlled experiment over observations of a single, unrepeated, uncontrolled, unanticipated natural event.

  67. #67 Sinisa Lazarek
    September 23, 2011

    You can all watch the complete webcast of the seminar given earlier today at CERN by the science coordinator of OPERA.

    http://cdsweb.cern.ch/record/1384486

    It seems to be at 6 sigmas limit. But you should just watch and decide for yourselves. The OPERA team isn’t claiming anything at the moment, they urge the community to replicate the experiment and colaborate. I’m sure Fermi is gonna start soon.

  68. #68 Mark M
    September 23, 2011

    Hypothetically… if this ends up being a genuine effect, what would it mean?

    Photons have rest-mass? They’re actually heavier than neutrinos? (Negative mass?)

  69. #69 daedalus2u
    September 23, 2011

    I think they need to move the beam so they can aim it at other neutrino detectors. The one in Japan would be a good target. The distance is much greater, if there is an effect, it should be much larger.

  70. #70 Mark Baranoff
    September 23, 2011

    Fun discourse

    There’s some good added insight and analysis at
    http://cosmologyscience.com/cosblog/?p=1878#more-1878

  71. #71 Thomas Fisher
    September 23, 2011

    @daedalus2u I think any “check” of this data by replicating the experiment will take years, and a lot of money. As has been pointed out, the Opera results are unique because they have the equipment to measure this much more precisely than other facilities do. And even so the experiment had to run for many years to get to this point.

    The only “quick” resolution to this will be found in the guts of the Opera methodology. Until those checks have been made, it will be hard to convince people to spend millions on experiments to validate this.

  72. #72 Julia Telier
    September 23, 2011

    Could the value of the supernova’s distance from earth be incorrect if it’s based on a flawed understanding of the speed of light..?

  73. #73 Torbjörn Larsson, OM
    September 23, 2011

    So … has everybody caught where they goofed yet?

    It is an easy one. According to the paper the distance measurement procedure use the geodetic distance in the ETRF2000 (ITRF2000) system as given by some standard routine. The european GPS ITRF2000 system is used for geodesy, navigation, et cetera and is conveniently based on the geode.

    I get the difference between measuring distance along an Earth radius perfect sphere (roughly the geode) and measuring the distance of travel, for neutrinos the chord through the Earth, as 22 m over 730 km. A near light speed beam would appear to arrive ~ 60 ns early, give or take.

    Of course, they have had a whole team on this for 2 years, so it is unlikely they goofed. But it is at least possible. I read the paper, and I don’t see the explicit conversion between the geodesic distance and the travel distance anywhere.

    Unfortunately the technical details of the system and the routine used to give distance from position is too much to check this quickly. But the difference is a curious coincidence with the discrepancy against well established relativity.

  74. #74 Torbjörn Larsson, OM
    September 23, 2011

    “We don’t allow FTL neutrinos here”, said the barman. A neutrino walks into a bar.

    [HT Miscience]

  75. #75 daedalus2u
    September 23, 2011

    Thomas, the best “check” is to duplicate the result with completely independent equipment. Unless they find a mistake that obviously explains it, it will have to be replicated.

    There are now plenty of neutrino telescopes around. Most of the cost in generating a neutrino beam is in the proton accelerator which CERN already has.

    No one will believe this unless and until it is replicated and over a longer beam line.

    If this result is correct, this is the most important finding in physics in 50 years or longer. Or it could just be a simple (or a not so simple) mistake.

  76. #76 daedalus2u
    September 23, 2011

    Torbjörn, if that is the answer, people are going to be very embarassed.

  77. #77 Jake
    September 23, 2011

    “Of course, they have had a whole team on this for 2 years, so it is unlikely they goofed.”

    You’re probably right. On the other hand, I remember NASA crashing a Mars probe because someone forgot to convert between standard and metric. It is rather interesting that you calculated a difference equal to what OPERA is seeing.

    On the gripping hand, I would hope that one of the first things they would check is that they calculated the distance correctly.

    We shall see, I guess. This is why peer review is so important.

  78. #78 Joe Fogey
    September 23, 2011

    I expect that there is some error of calculation, or procedure, or something. But I hope it isn’t.

  79. #79 Neil Bates
    September 23, 2011

    Ethan, you wrote: “One experiment based out of Chicago, a few years ago, found marginal evidence that neutrinos might move just a tiny bit faster than the speed of light, at 1.000051 (+/- 0.000029) c.

    Of course, this result is consistent with neutrinos moving at or slower than the speed of light; the errors are not significantly smaller than the measured difference from c.”

    Uh, no. If that error margin is about right (is itself so more than say around+/- 30% of its stated value – the error on the error, we might say – then that previous experiment showed superluminal neutrinos because 29 < 51. Same point about the current claim: their margin of error is about 1/6 of the margin over c. Well, maybe they're wrong about the margin of error, but you need to attack that directly and not pretend that the result is inconclusive "withing the margin of [stated] error."

    Second, I think the OPERA group knows about the supernova results. They are saying, that *they* produced neutrinos going a bit over c, not that all of them do (it certainly would depend on their energy, wouldn't it? - although it's hard to see how to calculate energy at above c. What is the gamma factor? Do such neutrinos have imaginary or complex mass? etc.

    And, maybe supernovas produce tardyonic (v < c) neutrinos, and certain other processes produce tachyonic (v > c) neutrinos. Nevertheless, I don’t blame you for being skeptical. But why would this group make such a claim unless it was warranted (I know, not the same as having to true), they don’t seem like cranks or gold-diggers, etc.

    PS: could we get a “remember personal info” button? (Or, is it there and my weird Linux SeaMonkey can’t see it?)

  80. #80 Neil Bates
    September 23, 2011

    [Sorry to repeat, but some of my previous comment got clipped, oddly in the middle.]

    Ethan, you wrote:
    “One experiment based out of Chicago, a few years ago, found marginal evidence that neutrinos might move just a tiny bit faster than the speed of light, at 1.000051 (+/- 0.000029) c.

    Of course, this result is consistent with neutrinos moving at or slower than the speed of light; the errors are not significantly smaller than the measured difference from c.”

    Uh, no. If that error margin is about right (is itself so more than say around+/- 30% of its stated value – the error on the error – then that previous experiment showed superluminal neutrinos because 29 < 51. Same point about the current claim: their margin of error is about 1/6 of the margin over c. Well, maybe they're wrong about the margin of error, but you need to attack that directly and not pretend that the result is inconclusive "withing the margin of [stated] error."

    Second, I think the OPERA group knows about the supernova results. They are saying, that *they* produced neutrinos going a bit over c, not that all of them do (it certainly would depend on their energy, wouldn't it? - although it's hard to see how to calculate energy at above c. What is the gamma factor? Do such neutrinos have imaginary or complex mass? etc.

    And, maybe supernovas produce tardyonic (v < c) neutrinos, and certain other processes produce tachyonic (v > c) neutrinos. Nevertheless, I don’t blame you for being skeptical. But why would this group make such a claim unless it was warranted (I know, not the same as having to true), they don’t seem like cranks or gold-diggers, etc.

  81. #81 Neil B.
    September 23, 2011

    Is there some weird thing in here about the less-than symbol? HTML? Then how can we write that kind of math statement? I meant to write, because 29 [is less than] 51, (also to add:
    Same point about the current claim: their margin of error is about 1/6 of the margin over c. Well, maybe they’re wrong about the margin of error, but you need to attack that directly and not pretend that the result is inconclusive “withing the margin of [stated] error.”

  82. #82 Neil B
    September 23, 2011

    Wow, what do you mean “problems with SR”?

  83. #83 Rick Whitten
    September 23, 2011

    Perhaps the problem here is the uncertainty principle itself. If the experimental set-up inadvertently measured the momentum of a neutrino accurately, the position would have been smeared. The investigators should check to see if they have somehow made a precise measurement of neutrino mass without realizing it? Seems more likely that superluminal neutrino velocity.

  84. #84 Neil Bates
    September 23, 2011

    Rick, they don’t measure the position directly anyway, they use the center of probable “send” and “hit” signals to measure elapsed time directly.

  85. #85 Ethan Siegel
    September 23, 2011

    Neil, the +/- errors that they give indicate one-sigma statistical significance. So the MINOS experiment reported seeing a superluminal effect at about 1.75 sigma statistical significance, which is absolutely, by any reasonable standard, insignificant.

    I have read the OPERA paper thoroughly by this point, and I have a very strong hunch as to where one very likely source of error lies. However, I want to give the story some time to settle and see what others say, but if no one comes out and says what I’m thinking, you’re very likely to see another post from me on this topic next week.

  86. #86 Daniel de França MTd2
    September 24, 2011

    Ethan, aren’t the neutrinos from the supernova electron neutrinos while these are muon neutrinos? Muons are heavier than eletrons and in the case of tachyons, the bigger the mass, the faster they are. Won’t they test tau neutrinos?

  87. #87 Daniel de França MTd2
    September 24, 2011

    Note: I made a mistake talking about speed vs. mass. I am too tired here…

  88. #88 forrest noble
    September 24, 2011

    In my opinion the most likely culprit of inaccuracy is the distance. The distance was calculated based upon the GPS system. Each GPS location point is supposed to be accurate within 3 meters. For two points the combined error could be 6 meters. It only takes a 20 meter shorter distance to account for what they reported. The GPS system may also be prone to other possible errors that I won’t go into because they’re speculative. But I would bet a 12 pack that inaccurately determined distance is the problem.

  89. #89 bob goodwin
    September 24, 2011

    So I am an amateurs amateur. But it always struck me that if light can be converted into energy, it must have something akin to mass, and that it must travel slightly slower than the ideal limit. If it is harder to detect neutrinos than light, might that mean its mass is lower?

    Sorry to state the question so stupidly, but I never got comfortable with this in college physics decades ago…

  90. #90 Sinisa Lazarek
    September 24, 2011

    @forrest noble:

    if you read through the paper you find that in fact they use an “augmented” GPS signal processing (P3) which is in laiman’s terms basically a millitary grade GPS system.

    You can find more on wiki: http://en.wikipedia.org/wiki/Global_Positioning_System#Applications

    but basically..
    Carrier phase tracking (surveying)

    Another method that is used in surveying applications is carrier phase tracking. The period of the carrier frequency times the speed of light gives the wavelength, which is about 0.19 meters for the L1 carrier. Accuracy within 1% of wavelength in detecting the leading edge, reduces this component of pseudorange error to as little as 2 millimeters. This compares to 3 meters for the C/A code and 0.3 meters for the P code.
    However, 2 millimeter accuracy requires measuring the total phase—the number of waves times the wavelength plus the fractional wavelength, which requires specially equipped receivers. This method has many surveying applications.”

    And that’s what OPERA team also says, that they got the distance correct to 20cm. These are not your car GPS devices ;)

    But the issue of distance is an important one. It really stands to question how well can we measure anything? I think that OPERA really did go an extra mile as far as distance and time measurement, using the very best our science and technology has to offer in this decade. I don’t believe there are general scientific devices in use that could measure distance and time more presicelly than the one they used. And all of their measurements of those were checked and double checked by other institutes. So I don’t think that is off. But it might very well be, that our technology is not good enough, for now.. :)

  91. #91 Sinisa Lazarek
    September 24, 2011

    @Torbjörn Larsson:

    The way I understood yesterday’s presentation at CERN is that they used GPS tags at CERN and Gran Sasso labs to get accurate times (~1ns) from knowing precise time at both stations, they can calculate the distance taking into account relativistic effects. Of course this is under the assumption that the measurement of position and our knowledge of earth’s dimensions is correct. OPERA says that they took into account earth’s rotation, the different positions on earth crust of both labs, the fact that one is rotating a bit faster than the other, moon rotation, moon’s gravity effect on earth and distorsions of there of, crust shifts, time of day, seasons, temperature… etc. etc.. But yes, they might have missed one, perhaps not of the most obvious ones

  92. #92 mark
    September 24, 2011

    Tesla saw it coming around 80 years ago.

  93. #93 Hannes
    September 24, 2011

    Fermilab’s MINOS experiment has also measured the speed of neutrinos and claimed a faster than light speed of neutrinos.

    The distance over which they measured was almost the same as at CERN, 734 km.

    They used other type of GPS receivers – TrueTime model XL-AK.

    Both MINOS and CERN rely on the same method of measuring distance, potentially showing the source of the problem.

    I personally think that the we are getting a hint that shows a possible problem with GR, certainly not SR.

    The first thing I would do is to check the distance again in a different way, for instance with a portable radio. This to exclude errors with the GPS system.

    If there is a to be found a difference in distance between the two methods than they will have to look at the GPS systems naturally. But if there are no technical errors detected in the GPS?

    Einstein’s GR will certainly be under fire in that case.

  94. #94 Stewart Hinsley
    September 24, 2011

    The positions are measured by GPS. But, if I understand correctly, both source and detector are underground, and presumably can’t receive a GPS signal. How do they get from the position of the GPS receivers to the position of source and detector?

    (If they can detect a GPS signal at source and detector then presumably they have to correct for the refractive index of rock.)

  95. #95 Nalliah Thayabharan
    September 24, 2011

    All of my investigations seem to point to the conclusion that they are small particles, each carrying so small a charge that we are justified in calling them neutrons. They move with great velocity, exceeding that of light – Nikola Tesla 1932

    Neutrinos have a small non-zero rest mass. Einstein’s theories of relativity say that nothing with non-zero rest mass can go faster than the speed of the light. But zero rest-mass particles can go faster than the light. We should understand the relation between local and nonlocal events like the dynamics of universal structure. In any physical theory, it is assumed that there is some kind of nonlocal structure and this nonlocal structure itself violates causality. Experimental tests of Bell’s inequality have shown that microscopic causality must be violated, so there must be faster than light travel – hence faster than light interactions are a necessity and they provide the non local structure of the universe. If Neutrinos are traveling faster than light, then Neutrinos must be on the other side of the light barrier going backwards in time – an inverted universe, where the future can interact with the past.

  96. #96 Jack Aidley
    September 24, 2011

    I’m really not a physicist, so I’ll make no great claims to understanding the detail here but I’ve looked over the paper and I can’t see any discussion of the effect of the Earth’s movement on the distance measurement.

    Surely the distance they should be using is not the distance between CERN and the Gran Sasso detector but the distance between CERN at the point the neutrino was generated and the point where the Gran Sasso detector will be when the beam arrives?

    Am I right? Should this be what they are doing? Are they not? And, if so, is that a large enough effect to explain the difference they detect?

  97. #97 Sinisa Lazarek
    September 24, 2011

    @Jack Aidley

    The distance in question is the distance between beam emiter and detector (straight line), not the air or land distance between two labs. The question for all geodesics people out there would be how accurate can you measure straight line distance between two underground objects? Hoping there will be more on that in coming time.

    In the webcast, there is a questions part where they go into more detail about measurments.. it starts around 1h05min into the video. Much more detail than in the paper.
    http://cdsweb.cern.ch/record/1384486

  98. #98 daedalus2u
    September 24, 2011

    Jack, the movement of the source and detector shouldn’t matter. As long as both are in the same “inertial frame” (that is they are not moving relative to each other and are not experiencing acceleration relative to each other), then the light travel time should be determined by the distance.

    The problem is that while the two locations are not moving relative to each other, they are at different gravitational accelerations due to the Earth’s field, the Sun’s field and the moon’s field. The neutrino beam passes through regions where the acceleration of gravity and the gravitational potential are different too.

    Light that does pass into and out of a gravitational potential well is delayed relative to light that does not. This has been observed in what is called the Shapiro delay.

    http://en.wikipedia.org/wiki/Shapiro_delay

    This should delay light and neutrinos and by the same amount. They didn’t mention taking this into account in the OPERA paper, so I assume they didn’t and it is small.

    They did look at the data comparing different times of day, different times of year, and different phases of the tides and didn’t see any systematic effects.

    But I think this effect would show up as the neutrinos appearing to move slower than light, not faster.

  99. #99 giorgio.parisi
    September 24, 2011

    @daedalus2u I guess that you are right. The general relativity effect should produce a delay that is small, or the order of one ns.(I hope that I have got it correct). In any case it goes in the wrong direcetion

    This is due to the fact the the gravitational field is higher inside the earth.

  100. #100 Nalliah Thayabharan
    September 24, 2011

    All of my investigations seem to point to the conclusion that they are small particles, each carrying so small a charge that we are justified in calling them neutrons. They move with great velocity, exceeding that of light – Nikola Tesla 1932

    Experimental tests of Bell inequality have shown that microscopic causality must be violated, so there must be faster than light travel. According to Albert Einstein’s theory of relativity, nothing with nonzero rest mass can go faster than light. But zero rest mass particles can go faster than the light. Neutrinos have a small nonzero rest mass. Faster than light interactions are a necessity and they provide the non local structure of the universe. We should understand the relation between local and nonlocal events like the dynamics of universal structure. In any physical theory, it is assumed that there is some kind of nonlocal structure violates causality. If neutrinos are traveling faster than light, then neutrinos must be on the otherside of the light barrier going backwards in time, where the future can interact with the past.

    – Nalliah Thayabharan

  101. #101 daedalus2u
    September 24, 2011

    giorgio, it is the gravitational potential that increases inside the Earth. The gravitational field decreases (and goes to zero at the center). The gravitational potential increases all the way to the center.

  102. #102 Christina
    September 24, 2011

    Someone said that we would’ve detected the neutrinos from SN1987A 4 years in advance if these findings were correct, but did we even *have* neutrino detectors in 1983? So, *would* we have detected superluminal neutrinos from 1987A if they’d existed?

    Of course, the results from 1987A indicate that at least some neutrinos must be subluminal, so the question would still remain of why we apparently only found superluminal ones from this experiment. So, either way, if the results hold up, it’ll create some very interesting questions. Do certain processes only create superluminal neutrinos and others only subluminal neutrinos, for example?

  103. #103 giorgio parisi
    September 24, 2011

    @daedalus2u You are write, but it was typo. I meant the gravitational potential. I want to recheck the computation and ask Opera.

  104. #104 giorgio parisi
    September 24, 2011

    @daedalus2u I am sorry. I checked the computation. I missed a square. The effect is much smaller that the picosecond! of the order of a femtosecond. Not relevant!!!

  105. #105 Chris Schmidt-Harms
    September 24, 2011

    The long base line interferometry people in the radio astronomy community know how to accurately measure distances between locations separated by hundreds or thousands of km.
    A list of accurately mapped stations can be found at:

    http://ivscc.gsfc.nasa.gov/stations/ns-map.html#maps

    I guess the OPERA teams has checked their GPS procedure against a known baseline from this list. The International VLBI Service for Geodesy and Astrometry can provide technical help.

  106. #106 michel deket
    September 24, 2011

    consciousness still quicker!

  107. #107 Daniel
    September 24, 2011

    There is a problem with SN1987A’s neutrinos, anyway. If neutrinos have a mass, even a tiny one, why did they arrive at the same time (almost) than light? They should be slower than light and, therefore, would arrive after a while, wouldn’t they?

  108. #108 Kafpaŭzo
    September 24, 2011

    The explanation about the neutrino speed inferred from supernova 1987A is strange and surprising.

    According to the explanation, neutrinos raced away from the nova immediately, while the shock wave that produced photons reached the surface of the nova three hours later. No problem there. But then it gets weird.

    After leaving the nova, the neutrinos, closely followed by the photons, spent 168,000 years racing through the 168,000 light-year distance to the Earth. And after this very long time, the neutrinos were still three hours ahead!

    This would mean that the neutrinos travelled just as fast as the photons, i.e. at the speed of light. Since there’s strong evidence that neutrinos have mass, this should be impossible.

    Unfortunately I don’t know the math that is involved here, so I can’t say anything about how much slower than light the neutrinos should be. But is it reasonable for them to remain three hours ahead, even after 168,000 years?

  109. #109 Kafpaŭzo
    September 24, 2011

    Wow, I should have checked the very latest comments before clicking “Post”. I now see that by some surprising coincidence, Daniel posted the same question while I was typing mine.

  110. #110 Bob Brand
    September 24, 2011

    Interesting discussion, many good points.

    I wonder whether it is wise to concentrate on the distance measurement that much: it seems pretty unlikely that the geodesy people would be 18 meters off, after all that effort – 18 to 22 meters would be needed to explain the effect. Even the GPS in my car does (a bit) better than that.

    Wouldn’t the clocks and the timing procedures be a more likely suspect?

  111. #111 Bob Brand
    September 24, 2011

    Two differences between SN1987A and OPERA which may be relevant (or maybe not):

    1 – Opera measures the flight time of *muon* neutrinos, while SN1987A concerns only *electron* neutrinos?

    2 – the beam from CERN to CNGS did travel thru a solid medium instead of a vacuum, like with SN1987A.

    The first diff. might be more relevant than the second, though.

  112. #112 Bob Brand
    September 24, 2011

    Oops… third difference:

    3 – mass, with Opera we’re talking 3 GeV, while the SN1987A electron neutrinos are at ~ 6 MeV.

  113. #113 daedalus2u
    September 24, 2011

    I figured it out.

    The scintillation material that they are using is polystyrene doped with primary (p-Terphenyl) and secondary (POPOP) fluors. The plastic strip is co-extruded with a TiO2 containing reflector. TiO2 is the white powder that is used in everything to make it white. It is white because it has a very high dielectric constant.

    Unfortunately, while TiO2 is highly reflective, it is also photo-active. It does absorb photons and generate electron-hole pairs. These can be stable for variable periods of time femtoseconds to microseconds and then decompose releasing the energy which is then emitted as a photon. If there are multiple reflections (which there are), then the time delay for each reflection gets added and there is a cumulative delay.

    http://pubs.acs.org/doi/abs/10.1021/jp982210s

    http://pubs.acs.org/doi/abs/10.1021/jp9944381

    The Fermi Lab result that showed FTL neutrinos used the same type of scintillation.

    The recombination time depends on the “details” of the electronic states of the TiO2 and what ever is adsorbed on the surface. They used standard TiO2 polystyrene blend, so there could be traces of just about anything in there. This is non-linear, so the time delay could depend on the number of photons.

    When they “calibrated” the system, they used a picosecond UV laser system. That may have produced a much larger signal, and a longer (shorter) delay. In the two different energy tests, the 42.9 GeV neutrinos showed a slightly longer delay than the 13.9 neutrinos. The number of photons produced is proportional to total energy, so if more energy gives a longer delay, then the very high energy pulsed UV laser will give an even greater delay so the “calibration” is what screwed up the measurements.

    Unfortunately to fix this they have to change out their scintillation plastic and replace the TiO2 with reflective metal. They should be able to quickly attempt a calibration with variable light intensity which will (my hypothesis) show variable delays.

  114. #114 Christina
    September 24, 2011

    @Kafpaŭzo #108: As I understand it, the delay would’ve been *more* than 3 hours at the initial supernova, and over 168,000 light-years, the photons would’ve closed the gap by whatever the difference between the original delay and 3 hours, which would require a velocity just a tiny fraction under the speed of light (probably on the order of 1 m/s slower or so!)

  115. #115 Glenn Chapman
    September 25, 2011

    Unfortunately there several important holes in the argument that the supernova 1987a neutrino detections are a more precise measurement that contradicts the Opera faster than light measurements. It is true that supernova observations were made by 3 detectors: Kamiokande (11 detections), Irvine–Michigan–Brookhaven (8 detections) and the Baksan Neutrino Observatory (5 detections) which has been taken as strong evidence because they saw these events at about the same time of 3 hours before the SN 1987a light detection. However this only says they all saw the same process creating neutrinos, it does not mean they are from SN 1987a – we only assume that because the supernova was a observed event in close approximation. As noted by Ben Still
    http://neutrinoscience.blogspot.com/2011/09/arriving-fashionable-late-for-party.html
    the Opera results would suggest a 4.14 ± 0.97 years earlier detection. The point is the dates covered are then Jan 6, 1983 (4.13), Jan. 18, 1982 (+0.97) to Dec. 26, 1983 (-0.97). However a quick check shows the Kamiokande was completed in April 1983 and upgraded to detect the solar neutrinos in 1985. Hence it would have missed almost all the expected detection period (all if the upgrade was needed to detect them). The IMB detector got its first results in 1982 (month not known in the data I can find) so it is again not certain to be detecting an earlier neutrino pulse. The point is that both of these dates are within the error of the Opera results. Only the Baksan detect (first operating 1977) was in operation. In a report in 2002 the Baksan researchers say that only the SN1987a was significant detection in 20 years of observations. However they noted that there was considerable time taken to understanding the detector operation and that continuous observations have only taken place since the mid 1980’s
    [Twenty Years of Galactic Observations in Searching for Bursts of Collapse Neutrinos with the Baksan Underground Scintillation Telescope, Alekseev, E. N., Alekseeva, L. N., J. Exp. Theor. Phys. 95 (2002) 5-10]
    http://arxiv.org/PS_cache/astro-ph/pdf/0212/0212499v1.pdf
    Thus even the Baksan detector may not have been fully running at the critical period.

    My point is that there is a big difference between saying that observations in the earlier period also saw nothing with the same detectors which would certainly strength the argument against the Opera results. However we must recognize that we really do not have those earlier observations for the whole period, if at all. This really weakens the SN1987a data as a counter to the Opera measurements. You cannot say that “look if this was true we would have seen the results 4 year before” when we really were not able to properly observe during all that period.

  116. #116 Raymond
    September 25, 2011

    Here are the issues: Nothing with mass can travel at the speed of light because the amount of energy needed to move the mass becomes infinite… A neutrino has mass and therefore cannot possibly travel faster than light.
    But let us suppose for some reason it can.
    One’s perception of time (relative to say, a man waiting for a bus) slows down the faster you move through space until you achieve ‘C’ and there time (as we know it) stops…
    So theoretically, anything that travels faster than light is actually traveling backwards in time! Therefore for a Neutrino to travel faster than light, wouldn’t the effect HAVE to precede the cause?
    I’m pretty sure this breaks a few commandments in the physics world.

    See, that is the problem. All evidence is indirect evidence in the sub-atomic world. All that we can see when we look at an atom is the photon that bounces off of the electron. We smash atoms into each other to see what they are made of, and all we ever see are varying frequencies of photons…. But I digress.

    We have been stuck in a quantum rut for the past 80 years, looking for all these different invented particles that we cannot find that have to exist to fill in the unknown variables to equations built on what is most likely huge misconceptions in the sub-atomic world! Wrong theories built upon wrong theories… because though the mathematics may work out, the physicists don’t actually understand WHAT is happening at the sub-atomic level of reality. Fallacies as big as believing the Earth is the center of the universe and inventing reasons why the planets move in predictable but eccentric motions. The equations worked out just fine …but ultimately they were wrong.
    The true answer in science is always revealed when you can shift your perspective and the Universe always seems to reveal itself to be much simpler than we previously thought.

    The Sun is the center, not the Earth. We need to shift our perspective on the quantum level.

  117. #117 Daniel
    September 25, 2011

    To Kafpaŭzo#108 (and myself#107!) and Christina#114:

    I indeed see that, if we consider neutrinos going 1m/s less than c, 168000 years times 365 days times 86400 seconds by day times 1 meter by second divided by c = 299792458 meters by second, divided by 3600 seconds by hour gives:
    168000 * 365 * 86400 / 299792458 / 3600 = 4.90899607621

    Only 4.9 hours delay between light and neutrinos when arriving on earth. I thought it should be more, for only 1m/s behind the speed of light, but no.

  118. #118 giorgio parisi
    September 25, 2011

    @daedalus2u

    Brilliant idea! However they say.

    The delays in producing the Target Tracker signal including the scintillator response, the propagation of the signals in the WLS fibres, the transit time of the photomultiplier [8], and the time response of the OPERA analogue frontend readout chip (ROC) [36] were overall calibrated by exciting the scintillator strips at known positions by a UV picosecond laser [37]. The arrival time distribution of the photons to the photocathode and the time walk due to the discriminator threshold in the analogue frontend chip as a function of the signal pulse height were accurately parameterized in laboratory measurements and included in the detector simulation. The total time elapsed from the moment photons reach the photocathode, a trigger is issued by the ROC analogue frontend chip, and the trigger arrives at the FPGA, where it is time-stamped, was determined to be (50.2 ± 2.3) ns.

    The conclusion is that the effect you are discussing may account only for a fraction of 50 ns. Part of the 50 ns will come from the cables and the electronics so the total effect cannot explain the bulk of the observation, however it may contribute to reduce the discrepancy.

    By the way, in the two different energy tests, the 42.9 GeV neutrinos showed a slightly longer ADVANCE (not DELAY) than the 13.9 neutrinos.

  119. #119 Steven Boyd
    September 25, 2011

    @Raymond : Er no – we see the tracks of particles that have been produced i our detector. We can measure their mass, their momentum, their charge. We understand how they decay and interact. We can predict such things to 1 part in a trillion. It is certainly NOT the case that “all we ever see are photons”. Of course the evidence is indirect – but then so is every probe. I don’t see how you make such a sweeping generalisation that everything is wrong, and you’ve given us no experimental evidence to back up your claim.

    The OPERA result might be correct. It might not be. It certainly requires a lot more investigation and verification by other experiments. If true we will do some serious rethinking of the theory. I applaud these guys courage for coming out with a result that seems to disprove one of the fundamental tenets of modern physics. I especially applaud them for not claiming a discovery and for saying that they want the rest of the physics world to disprove them, to find the error. That’s how science is done.

  120. #120 David Kahana
    September 25, 2011

    As to the (anti)-neutrinos detected from SN1987A: there are a couple of points that should be made. First, about the origin of the huge pulse of prompt neutrinos and anti-neutrinos that come from a core collapse supernova: there was some confusion shown on the part of some of the posters above. These prompt neutrinos and anti-neutrinos carry away almost the entire gravitational binding energy which is released in the collapse of the stellar core, amounting to about 100 foe or 100×10^51 ergs, while a very healthy supernova explosion ends up producing only about 1 foe in kinetic energy of outward going stellar matter and supernova light. In fact the light from the supernova is actually only about 1% of that final 1 foe. So the supernova process can be seen to be in some sense very inefficient. Only a tiny percentage of the available energy goes into matter and light – almost all of the energy goes into neutrinos which then for the most part just free stream through the universe.

    The prompt (anti)-neutrinos and neutrinos are produced by the mechanism of e+ e- annihilation during the early stages of the core collapse, not for the most part by the neutronisation of the core. Neutronisation does in fact begin during the collapse and it likely proceeds to a Z/A of about 1/3 during the initial core collapse, at which time the core reaches some 2-4 times nuclear matter density. This degree of neutronisation, the very high density, and the high temperature all mean that one can have no direct experimental knowledge of the nuclear matter equation of state in the collapsing core, and that lack of knowledge is actually the single largest uncertainty in the theory of these supernovae. However, temperature in the collapsing core can be estimated to reach about 10-20 MeV.

    So there is plenty of thermal energy available to create electron-positron pairs and these are indeed created very copiously, and they then also annihilate copiously into neutrino+antineutrino pairs, whose numbers build up. At several times nuclear matter density, neutrinos deep in the core have very short mean free paths, and at first they are effectively trapped as the core falls in and the density increases. But as the core begins to bounce and an outward going shock wave forms, it is only the neutrinos that can transport at all.

    Second: all three flavours of neutrinos and anti-neutrinos are certainly made in a core collapse supernova. The electron neutrinos that come from neutronization are less copious and come gradually, later on, as the collapsed core settles down and finally forms a neutron star, possibly with a transition to strange matter in its innermost core on a timescale of a few seconds, where densities may finally reach 8-10 times nuclear matter density. Such a transition to strange matter could be associated with a fresh burst of neutrinos – there was actually some evidence for a second burst in the data for SN1987A, though the lack of an accurately synchronised clock at one of the three detectors that definitively saw the (anti)-neutrinos made the interpretation somewhat difficult.

    Finally, with this theoretical knowledge: that the prompt neutrinos are expected to have energies on the order of 10 MeV one can easily estimate their speeds, guessing that all of the neutrino masses are on the order of 1-2 eV. This implies a gamma of 5,000,000 or so, thus a v/c of about 1-2*10^-14. That is, the SN1987A neutrino speeds are expected to be light speed to within 2 parts in 10^14. So over 168,000 years of travel time, the light pulse would catch up with the neutrinos by about 3 one billionths of a year in time … or by about 0.01 seconds.

  121. #121 RJM.Corbet
    September 25, 2011

    First you say that it’s unrealistic to expect a result to 7 decimal places to be accurate, in the case of the opera finding, and then you say that the speed of the supernova neutrinos was measured accurately to 10 decimal places.

    You can’t have it both ways …?

  122. #122 Sinisa Lazarek
    September 25, 2011

    Was thinking a bit about basic Special Relativity, and one thought is bugging me. Maybe someone could explain.

    Seems to me that something is illogical. We know the basic mass calculation from SR where we have 1- v squared over c squared, all under square root etc.. And we were always thought that v over c is 1 because nothing can travel faster… because you get infinity.. or vice verca.. like chicken and the egg. But consider what happens “hypoteticaly” if v>c. Well then we get a negative number. And we were again, always taught that yo can’t have a negative number number under square root, because it’s not possible. But we do have “i” in math. So again, in theory, you could have a square root of any negative number expressed as a certain value of “i”. Yes they are imaginary numbers, but at least you can apply some math to it. You can’t with infinity.

    So what I don’t understand is why would this violate the SR formula.. it just opens new doors, doesn’t break the formula. Imaginary numbers are used in math all over the place.

  123. #123 Lost in Space
    September 25, 2011

    If the results do turn out to be true, perhaps supersymmetry can explain it. The energy of the CERN neutrinos is higher than that detected from sources such as SN 1987a. If the energy is great enough to exceed c or what has been considered as c so far, the mass of the particle will be greater too. Perhaps the energy input has created a superparticle which is tachyonic because its energy has become negative on passing c. It could also be the reason why until now no other superparticles have been observed experimentally.
    It’s hard to conceive that the current criticisms which abound related to this experiment have not been previously considered and anticipated by the scientists involved. It will be interesting to see if these results can be independently verified.

  124. #124 Neil Bates
    September 25, 2011

    Re SN 1987A and the travel time: Neutrino mass is so small that yes, even over that extreme distance, the delay is a matter of hours (and REM again the delay in collapse reaching the star’s surface.) They calculated correctly but note that this only showed upper bounds on neutrino mass. From that Wikipedia article http://en.wikipedia.org/wiki/SN_1987A#Neutrino_emissions:
    “The neutrino measurements allowed upper bounds on neutrino mass and charge, as well as the number of flavors of neutrinos and other properties.[6] For example, the data show that within 5% confidence, the rest mass of the electron neutrino is at most 16 eV.”

    Interesting, “upper bounds” … And surely there is enough uncertainty in something like the exact moment a core collapses compared to when we see the flash from the surface, that maybe those neutrinos were also superluminal, even more than Ethan imagined? Note also the possible distinction between electron-neutrinos and muon-neutrinos.

    In discussions at a picnic yesterday loaded with capable types like a Jefferson Lab (VA) nuclear physicist, he noted that maybe photons themselves don’t even travel at the physical limit c, because of a tiny mass (or, as I thought, interaction with background “syrupyness” of space like caused by dark energy and quantum effects, that might actually slow down photons more than neutrinos.) No, it doesn’t violate the semantics of “speed of light” because we would treat “c” as a physical parameter of limit based on the absence of other factors, as what would determine actual time dilation, causal factors, etc. even if light is inhibited from going that fast.

  125. #125 Hannes
    September 25, 2011

    Well, there might be a more mundane explanation to it all.

    Fermilab’s MINOS experiment used the Symmetricon TrueTime model XL-AK GPS system. With known problems:

    Some XLi Time and Frequency Systems with GPS Option module generate an
    incorrect date (year, month, day-of-month, day-of-year) on and after June 23, 2009 when
    operating in GPS mode. The error manifests in a negative 19.7 year offset to the XLi clock.

    The Symmetricon Xli 1PPS used by OPERA from CERN is another model but from the same producer. Its output represents the reference point of the timelink to OPERA. This point is also the source of the “General Machine Timing” chain (GMT) serving the CERN accelerator complex.

    http://www.symmetricom.com/media/files/support/ttm/fsb/098-41620-002k.pdf
    http://www.symmetricom.com/media/files/support/ttm/fsb/098-50620-020_%20REV%20B.pdf

  126. #126 Neil Bates
    September 25, 2011

    Here’s another paper, from back in 1988 but of interest regarding issues of neutrino emission from supernovae:
    http://www.sciencedirect.com/science/article/pii/0370269388906892

  127. #127 Neil Bates
    September 25, 2011

    Sinisa: Not only is it hard to imagine what physical properties a mass of “i” grams means (what gravity, what inertia, what energy equivalent, etc), but speed > c leads to causality paradoxes. However that only happens if we try to make simultaneity consistent with there being no preferred frame of reference. Not so bad if there is a PF (like, the cosmic clock background of equal time from the big bang, also equivalent to isotropic averaged CMB.)

  128. #128 Neil Bates
    September 25, 2011

    [And I break after this third.]
    BTW, if some neutrinos are FTL and others not (question: what energy range and what type, and be superluminal?), we can call them newtrinos. Or maybe, tachtinos. (PS: if they are superluminal, we have a problem defining their energy of course. I wonder who is thinking about that.)

  129. #129 Sinisa Lazarek
    September 25, 2011

    @Neil Bates

    Thank you for the reply. I agree completely with you that it’s hard to imagine. But i.e. in Quantum Mechanics imaginary, or rather complex numbers are used all the time. To describe wavefunctions, magnitutes, probability distributions… etc. So why not have a complex number describing mass or velocity once v>c. It would point to something trully great. Surpass c and you enter the realm of quantum laws again, where again you can’t measure anything with precision, only state probability factor of something. Of course, this is pure theoretical idea :)

  130. #130 Neil Bates
    September 25, 2011
  131. #131 daedalus2u
    September 25, 2011

    Giorgio, The 50.2 ns delay is from when photons hit the photocathode of the PMTs. The delay I am talking about is between when energy is deposited in the scintillator and when photons hit the photocathode. The TiO2 delayed response makes it worse (i.e. moves the neutrino detection earlier), not better.

    However, in looking at the details more carefully, I can’t find where they have calibrated a delay between scintillation events and PMT response, it looks like they just assume it to be less than 9.4 ns (figure 6).

    This assumption is wrong because the maximum delay could be considerably longer. The wavelength shifting fibers are polystyrene which has a refractive index of 1.6. The speed of light in such a fiber is c/1.6 or 0.625c.

    The length between the illumination point and the PMT is variable, but is up to ~9 meters. 9 meters of 1.6 refractive index fiber has a light transit time of 48 ns vs 30 ns for 9 meters of vacuum. But again, this makes the problem worse not better. I don’t know how they could get 9.4 ns, unless they divided by 1.6 instead of multiply (9/2 * 3.3 / 1.6) = 9.3

    There is another potential effect. When a neutrino is detected, it generates light in the scintillator which travels to the PMT and triggers it. Multiple PMTs get triggered, and the trigger length of an “event” is about 60 ns (the output pulse length of the PMT following the fast wave shaper). There is also another wave shaper, the slow wave shaper. It has a time constant of 160 ns. When multiple PMTs are triggered in an “event”, the first one is assumed to be the arrival time of the neutrino. But then they say they simulate all neutrino events and then calculate what the estimated arrival time is, so that should eliminate spurious triggers. But if they used the wrong sign,

    These PMTs do have significant dark counts, mean ~2.45 Hz per channel over 63,488 channels. There were 16,000 neutrinos detected in 10 microsecond proton bursts, for a total time of 0.16 seconds. That means there were 4,160 dark counts during that time. If a dark count came just before a “real” neutrino, that could bias the “arrival time” of the neutrino to be sooner than actually occurred.

    There is considerable variability in the number of dark counts per PMT. Some were as high as 300. They have records as to which ones have a high dark count, so they could check to see if the FTL neutrinos were first “detected” by those high dark count PMTs.

  132. #132 bilal saad
    September 25, 2011

    hi guys i know that fact that rule is wrong before 2 years ago because i’m Muslim in our Islam our profit Muhammad travel form MEKE (KSA) to Palestine and then to the seventh sky and return back to his bed and it is stile warm actually that faster then speed of light, that first prof and the second prof in Quran the chair of Suleiman travel form Palestine to Yemen and return back all that happened like a glance of your eye so that faster then light. thank god i’m Muslim.

  133. #133 Sascha Vongehr
    September 25, 2011

    The supernova data tell us that if the new data hold up (no stat/syst error), the early arrival time must be due to superluminal processes at the origin inside CERN:
    http://www.science20.com/alpha_meme/neutrinos_can_go_faster_light_without_violating_relativity-82950
    What do you think?

  134. #134 forrest noble
    September 26, 2011

    @Sinisa Lazarek,

    Yes they claimed a 20cm accuracy, but such distance measurements should now be otherwise confirmed by laser shots by a surveying team without using the GPS system for any part of it. I’d bet a six pack that the GPS system is the culprit. Presently my best guess is that the error relates to the differences in altitude.

    From what I’ve read the GPS system has some cool algorithms for correction but I don’t expect this kind of accuracy requirement concerning 8,300 ft. difference in altitude has ever needed this kind of accuracy or been tested before concerning big differences in altitude between two points and then confirmed by surveying the distance.

    My expectation is that future experiments in Japan and the U.S. on a level plane will show no such discrepancy. I think these muon neutrinos were traveling at a speed a little less than the speed of light and the calculated distance was about 20 meters less than what the GPS system indicated. My prediction is that it is a fault in the GPS programming related to presently unknown aspects of gravity as it relates to the speed of light.

  135. #135 Wow
    September 26, 2011

    “in our Islam our profit Muhammad travel form MEKE (KSA) to Palestine and then to the seventh sky and return back to his bed and it is stile warm actually that faster then speed of light”

    So he wasn’t a prophet. He wasn’t even a man. He was, in fact, a neutrino.

    Is that what you’re telling us?

  136. #136 Eric Lund
    September 26, 2011

    Torbjörn @72: I’d like to think that’s the answer, but when I do the calculation myself I don’t get the same result.

    The two stations are separated by rφ (where r is the radius of the Earth and φ the angular separation) along the spherical surface of the Earth. If we approximate the difference in altitude as being small compared with the radius of the Earth, then the straight line separation between the two stations is 2r sin(φ/2). The difference Δ between the two will be given roughly by the third order term in the expansion of the sine, or Δ = 2r * (φ/2)^3/6, which reduces to rφ^3/24. Plugging in 732 km for rφ and 6370 km for r, I get Δ = 403 m.

    I haven’t checked the second order iteration of this problem, which would be to correct for the fact that the stations are at different altitudes (there is a claim upthread of an altitude difference of 8300 ft, which is about 2700 m). If this introduces an error proportional to φ^2, then we are in the right ballpark (φ = 0.115 radians) if they took the first order correction into account but not this second order correction.

  137. #137 daedalus2u
    September 26, 2011

    What they should do is bore a hole from the surface down to the experimental hall. Then they can have line-of-sight surveying from the surface. There is ~1400 meters of rock above the lab. That wouldn’t be very expensive to drill through. A 6 inch diameter bore hole would be plenty. Put a modest laboratory at the surface, and the location could be determined very precisely. They could even do laser ranging to the moon and interferometry via radio telescopes.

    They could run the power and stuff up through the hole, so they wouldn’t need to run power lines. Make the shaft big enough and they could put an elevator so the staff could get their via the elevator and not have to go around.

  138. #138 Mu
    September 26, 2011

    I’m just surprised that it’s such a big deal. The deviation from the accepted speed of light is a fraction of a percent, so it’s not going to invalidate everything measured so far. It’s not stating that there isn’t a speed limit in the universe, only that it’s not the so far stated photon in vacuum speed. Maybe photons get slowed by dark matter, so our measured speed isn’t truly the unimpeded speed? I’d thought the new experiments would excite physicists, not send everyone in a frenzy of “that can’t be right”.

  139. #139 Sinisa Lazarek
    September 26, 2011

    @daedalus2u no.137

    One of the guys at CERN conference asked the exact same question. The answer was that the deviation in drilling such a deep hole is around 5%. Which is in turn a much greater offset in the end. 5% of 1500m = 75m

    Am starting to think that this might be first tachyon detection ;)

  140. #140 Bob Brand
    September 26, 2011

    Mu @137

    You said: “I’d thought the new experiments would excite physicists, not send everyone in a frenzy of “that can’t be right”

    They ARE excited. At the same time it is vitally important to stay skeptical and triple-check everything to make sure there are no mistakes.

    A number of good ideas have been mentioned above (I particularly like the suggestions by Daedalus about the detector system), and there may be much subtler ones (like what happens at the CERN end to produce the neutrinos… and exactly *when* and *where* that happens).

    The search for mistakes may even deliver clues for the new physics if the effect turns out to be real. Maybe about any energy dependence, or about the oscillations?

  141. #141 Bob Brand
    September 26, 2011

    Oops, that’s Mu @138.

    Must have been a neutrino post.

  142. #142 Norman Birkett
    September 26, 2011

    Many thanks to the blog author and the many commenters for this discussion. It has been immensely interesting–one might even say illuminating–for this layman.

    But I’m curious about one statement that as far as I can tell hasn’t been discussed. cruf @61: “A pretty straightforward generalization of Einstein’s rule is that nothing can accelerate to the speed of light, either from above or below.”

    I’m sure I’ve heard similar claims before from at least one physics authority (sorry–can’t remember who or where; lost most of my papers and notes in a disaster not long ago). I don’t know enough physics to be able to evaluate the claim.

    Is the claim correct?

    If it is, then why would it not be possible for some particles to begin their existence with v>c, provided they never decelerated to sublight velocities?

  143. #143 Wow
    September 26, 2011

    “Is the claim correct?”

    Well, not the “either from above” bit.

    Above light speed, your mass is imaginary according to the maths. What this does to you being decelerated is a little odd. Accelerating faster would reduce your imaginary mass.

    “why would it not be possible for some particles to begin their existence with v>c, provided they never decelerated to sublight velocities?”

    Aye, they could do that. But his claim in 61 wouldn’t preclude that anyway.

  144. #144 Radkrishna
    September 26, 2011

    Excellent article.

  145. #145 Ethan Siegel
    September 26, 2011

    For anyone who wants further detail on what’s in the paper, what the data looks like and what some of the mistakes they could be making are, you’ll definitely want to check this out:

    http://scienceblogs.com/startswithabang/2011/09/are_we_fooling_ourselves_with.php

  146. #146 Thomas Moore
    September 26, 2011

    Stock exchange companies are going to acquire proton accelerators and giant photomultiplier arrays, and someone’s going to accidentally irradiate NYSE.

  147. #147 OKThen
    September 26, 2011

    Ethan
    Excellent summary. Thanks. Yes the supernova data most be reckoned with. Thanks for that evidence.

    @124 Neil Bates
    Thanks for your picnic thoughts. Nice. And I agree that the supernova conclusions may not be iron clad.

    @133 Sascha Vongehr
    Thanks, yes extra dimensions makes sense as one of several possible alternative; if results stand.

    All of the obvious errors have been ruled out; so I’ll keep my arm chair analysis to myself. (e.g. Shut up self, they couldn’t be that stupid! Shut up before I strangle myself!!)

    CERNs results have been scrutinized far beyond almost any previous physics experiment. CERN published reluctantly and in disbelief of their own results.

    I await an explanation from the experimental and theoretical experts.

    If a duplicate experimental setup is available; great!

    If this result is the only game in a year; well the game of New Physics is already afoot.

    I’m betting on New Physics; I don’t expect any significant error to be found in a week. Maybe 6 months. But some genius might just explain why the kettle is black. And we’ll all say, thank you very much.

    CERN scientists are experimentally and theoretically of the highest calaber. To write a paper with such restraint demonstrates the highest scientific integrity; e.g. they modestly conclude, “Despite the large significance of the measurement reported here and the stability of the analysis, the potentially great impact of the result motivates the continuation of our studies in order to investigate possible still unknown systematic effects that could explain the observed anomaly. We deliberately do not attempt any theoretical or phenomenological interpretation of the results.” Amazing.

    I await; new interpretation, new physics, something very interesting.

  148. #148 crd2
    September 27, 2011

    All this talk of bad measurements and faulty GPS systems. Do you guys have any idea how long it must take to get off that little raft and to a bathroom? (see 5th image)

    $100 bucks says one of the service tech’s pee’ed in the neutrino detector.

  149. #149 forrest noble
    September 27, 2011

    “…$100 bucks says one of the service tech’s pee’ed in the neutrino detector.’

    Oh no, not that! Shroedingers cat explains the whole problem. Once we observe the neutrino particles they loose location, NOT :)

  150. #150 JBJ
    September 27, 2011

    From a hobby astronomer/physicist: If the results prove correct, I’m starting to think Dark Energy for several reasons. First explanation could be that negative or imaginary mass could or would produce repulsive gravitational force? Second as a “frame dragging” effect due to the “speed limit violation” in that the neutrinos “forces” the “fabric of space” to expand? An additional thought: With the extreme density of neutrino radiation carrying away all that energy; Where does it “end up”? Will the universe be drained for this energy in that it will never more be accessible for matter etc., or might it be that it is used to “stretch” the universe?

  151. #151 ex0du5
    September 29, 2011

    I’m surprised no one has mentioned what it actually points to in the paper (available on arXiv): that neutrinos may potentially traverse higher dimensional paths on world-branes that allow for the traversal faster than the path on the 3D projection. Or that this has a developed theoretical background, going back several years in papers. Or really, any of the actual models being proposed. The discussion in the blog and comments seems to have nothing to do with an actual evaluation of the paper except for very basic attacks on it’s error bounds (pretty well explained in the paper) or posturing without models.

  152. #152 John F Moore,
    September 29, 2011

    While GPS can be made to work that accurately, even if the two sites are at different altitudes, there may be an error because the the radio signal from the satellite to the Gran Sasso detector must pass through an average of 4600 feet of rock. That means that the radio path can be refracted, just as a light beam is refracted when passing from air to water. Since the neutrino is claimed to be about 60 nanoseconds early, that corresponds to only about 59 feet. A refraction angle from the West of only three-quarters of a degree would be enough to make the distance between sites appear longer than it really is by that much, explaining the apparent FTL speed.

  153. #153 johno
    September 29, 2011

    Let me get this straight, these impossible results are impossible, since its impossible for anything to exceed the impossible speed of light? – I see. The next few months are going to be highly interesting and as a layman, i’am not nervously clutching at straws.

  154. #154 OKThen
    September 30, 2011

    Relative to:
    @55 “I point to the Observer Effect of course.”
    and
    @56 “The What?”

    I’d never heard the term “Observer effect”; so the recent readable summary caught my eye.

    The observer effect by Massimiliano Sassoli de Bianchi
    Sept 30 2011, http://arxiv.org/PS_cache/arxiv/pdf/1109/1109.3536v2.pdf

    A few quotes, “observation is not interpretation… Observation is typically associated to the experimental process of data collection, whereas interpretation is associated to the creation of explanatory models and theories. However, a radical distinction between observation and interpretation is not possible. There is indeed a lot of objectivity in our theoretical interpretations of the experimental data… But there is also a lot of subjectivity and conventionality (i.e., of interpretation) in our alleged objective observations. This is so because we necessarily see reality through the lenses of our theories, which determine what, how, where and when to observe… but also our comprehension of reality is strongly biased by our understanding of what an observation is, or is meant to be… What Heisenberg thus realized is that when we pretend from reality a very accurate reply to our interrogatives, we are not anymore guaranteed that these interrogatives will be without consequences on what is being observed… A common prejudice is to believe that the problem highlighted by Heisenberg with his analysis of the gamma microscope is only pertinent to the micro-world. In other terms, it is usually believed that the inevitable and irreducible disturbance of the observer on the observed (the so-called observer eect ) is only one of the many strangeness of the micro-world, but that nothing of the sort can truly happen in our everyday macroscopic reality… the real strangeness is the fact that we had to wait for Heisenberg to realize that our observations, whether concerning microscopic or macroscopic entities, cannot be reduced to mere acts of discovery. Observations are much more than that. Indeed, one thing is to observe real

  155. #155 Bharath kumar jinka
    October 23, 2011


    physics is great because of our scientists

  156. #156 Hartwig Thim
    March 15, 2012

    Richtig: CERN hat die spezielle Relativitätstheorie wider-legt, denn nach dieser Theorie kann such kein Teilchen schneller als Licht bewegen. Die Neutrinos waren aber schneller, also ist die spezielle Relativitätstheorie vom Tische gefegt.