In this illustration, one photon (purple) carries a million times the energy of another (yellow). Some theorists predict travel delays for higher-energy photons, which interact more strongly with the proposed frothy nature of space-time. Yet Fermi data on two photons from a gamma-ray burst fail to show this effect. The animation below shows the delay scientists had expected to observe. Credit: NASA/Sonoma State University/Aurore Simonnet

Einstein Still Rules, Says Fermi Telescope Team

Article Updated: 24 Dec , 2015


While the Fermi Space Telescope has mapped the gamma ray sky with unprecedented resolution and sensitivity, it now has been able to take a measurement that has provided rare experimental evidence about the very structure of space and time, unified as space-time. Einstein’s theory of relativity states that all electromagnetic radiation travels through a vacuum at the same speed. Fermi detected two gamma ray photons which varied widely in energy; yet even after traveling 7 billion years, the two different photons arrived almost simultaneously.

On May 10, 2009, Fermi and other satellites detected a so-called short gamma ray burst, designated GRB 090510. Astronomers think this type of explosion happens when neutron stars collide. Ground-based studies show the event took place in a galaxy 7.3 billion light-years away. Of the many gamma ray photons Fermi’s LAT detected from the 2.1-second burst, two possessed energies differing by a million times. Yet after traveling some seven billion years, the pair arrived just nine-tenths of a second apart.

“This measurement eliminates any approach to a new theory of gravity that predicts a strong energy dependent change in the speed of light,” Michelson said. “To one part in 100 million billion, these two photons traveled at the same speed. Einstein still rules.”

“Physicists would like to replace Einstein’s vision of gravity — as expressed in his relativity theories — with something that handles all fundamental forces,” said Peter Michelson, principal investigator of Fermi’s Large Area Telescope, or LAT, at Stanford University in Palo Alto, Calif. “There are many ideas, but few ways to test them.”

Artist concept of Fermi in space. Credit: NASA

Artist concept of Fermi in space. Credit: NASA

Many approaches to new theories of gravity picture space-time as having a shifting, frothy structure at physical scales trillions of times smaller than an electron. Some models predict that the foamy aspect of space-time will cause higher-energy gamma rays to move slightly more slowly than photons at lower energy.

GRB 090510 displayed the fastest observed motions, with ejected matter moving at 99.99995 percent of light speed. The highest energy gamma ray yet seen from a burst — 33.4 billion electron volts or about 13 billion times the energy of visible light — came from September’s GRB 090902B. Last year’s GRB 080916C produced the greatest total energy, equivalent to 9,000 typical supernovae.

More images and videos about the Fermi Space Telescope.

Lead image caption: In this illustration, one photon (purple) carries a million times the energy of another (yellow). Some theorists predict travel delays for higher-energy photons, which interact more strongly with the proposed frothy nature of space-time. Yet Fermi data on two photons from a gamma-ray burst fail to show this effect. The animation below shows the delay scientists had expected to observe. Credit: NASA/Sonoma State University/Aurore Simonnet

Source: NASA


34 Responses

  1. Hon. Salacious B. Crumb says:

    “Physicists would like to replace Einstein’s vision of gravity β€” as expressed in his relativity theories β€” with something that handles all fundamental forces,” said Peter Michelson.

    Take that on the chin AnaCONda!!

    Considering you don’t believe in most of the four forces nor of the importance of their interactions, this must be another fatal nail in that coffin of EU.
    Of course, you will come up with an excuse or another nitwit argument, but it shows much of everyday physics is close to reality.

    Science and Einstein and left us a wonderful legacy. Pity you just can’t see it! πŸ™

  2. Jon Hanford says:

    Greatt news, Nancy, on results from the first year-long study of the VHE Gamma-ray all sky flux from the Fermi satellite. As longer exposures are undertaken in The VHE/UHE portion of the EM spectrum, surely the it presents an opportunity to study and speculate as to the astrophysical origins, evolution and formation these highly energetic objects.

  3. Navneeth says:

    Nancy, did you mean to say [i]simultaneously[/i] at the end of the first paragraph, instead of [i]at instantaneously[/i]. Or was it, perhaps, [i]at the same time[/i]?

  4. Navneeth says:

    I’m still in BB mode and forgot all about HTML tags. 😳

  5. DrFlimmer says:

    NewScientist also included in this story that this measurement contradicts an earlier experiment by the MAGIC collaboration. They had a time delay of 4 minutes! But their object was much closer than this one here, so Fermi is probably a little more right than MAGIC.

    At least, I hope so – because a difference of 4 minutes between two light rays to arrive that should have been sent instantanously is quite a time, even on cosmological scales!

  6. Navneeth – yes, thank you for figuring out what I meant!

  7. Lawrence B. Crowell says:

    Great! This will send Smolin and Amelino-Camelia back to their offices and white boards. String theory makes a beautiful prediction that this chaotic foam is valenced from measurable physics, or in a sense renormalized out. This is really good news.


  8. Lawrence B. Crowell says:

    The research paper is at:

    I have been having long hard arm wrestling sessions over this. This makes me absolutely giddy! Bye bye loop quantum gravity breaking of the Lorentz group at small scales — YIPEEE!!! There are litterally thousands of theory papers now which are worth less than toilet paper. God it feels good. πŸ™‚

    And to think I delayed reading this blog post and spent so much time on the global warming kerfuffle. Hon. Salacious B. Crumb should sharpen his rapier against the deniers over there. Anaconda is just a minor voice of nonsense. Anti-AGW people are more serious.


  9. Astrofiend says:

    Lawrence B. Crowell Says:
    October 28th, 2009 at 4:08 pm

    Don’t be too humble in victory or anything Lawrence!

    …but I was about to say the same thing. There will be a lot of crest-fallen physicists out there at the moment, realising that they’ve just spent a fat wad of their career on a whole bunch of wrong.

  10. Lawrence B. Crowell says:

    Oh for sure. I think I heard a number of big thuds of bodies falling. I put Smolin on that list, but that guy is smart enough and did not put all his eggs in this basket. He dropped a couple of eggs though. The person who I hope really hurts is … , well that would get personal.

    Loop Quantum Gravity is not dead though. In fact I have been working, rather unsuccessfully to be honest, on how LQG could be a system of constraints which exist in a map from the string/M-theory AdS_5 spacetime to the physical (the real universe) de Sitter spacetime. String theory works on the anti-de Sitter spacetime, but that is not the universe. It maps to the de Sitter spacetime under a Wick rotation. A BTZ black hole on AdS spacetime makes this rotation-map difficult, but I think the BPS charge of that black hole maps to the LQG constraints on physical spacetime. This problem has proven to be abominably difficult though. But I think something like this has to happen. String/M-theory stuck forever in the AdS domain does not do as much good for understanding this universe.

    Which ever is the case this does adjust a lot of things. Various models of quantum gravity, except string/M theory πŸ™‚ , have spacetime broken into struts and links near the Planck scale. This results in the breaking of Lorentz symmetry (spacetime symmetry of relativity) at very short scales. These models include LQG and Dynamical Traingulated Causal Nets and so forth, have these features. The one alternative to string theory which fly through this problem is twistor theory, which in what I am working on with string/M-theory does emerge. These various models are going to have to be either seriously revised or scrapped.

    I keep up with this and some other sites so I can keep abreast of these developments without having to golpher through the observational papers, which I find more difficult to read (to be honest) than theory/math papers.

    Lawrence B. Crowell

  11. Astrofiend says:

    “I keep up with this and some other sites so I can keep abreast of these developments without having to golpher through the observational papers, which I find more difficult to read (to be honest) than theory/math papers.”

    Yeah – I wish my math were that good. Unfortunately for me, I have very limited natural ability or intuition in maths. All of my math ability has been hard won, and even then I think I’m pretty much reaching the limits of what I can understand mathematically without an order of magnitude increase in the amount of time I spend trying to get my head around it.!

    So I guess I’ll let you guys take care of the ultimate basis of reality, and I’ll ply my mind to other problems in astronomy!

  12. rudeyd says:

    I must ask this question and hopefully won’t be crucified for my ignorance. Be gentle people, for I am a very curious soul.

    The short version if you can,
    what is theTwistor (twister) theory? M-Theory?

  13. Lawrence B. Crowell says:

    A twistor is a pair of spinors (well what’s a spinor?, you might ask) where if one represents a flagged coordinate, the other is a momentum value. As such the position spinor x is related to the momentum spinor p by a “motion,” similar to x = x(0) + vt. Position and momentum are conjugate variables in Hamiltonian classical mechanics and quantum mechanics. The pair of spinors or the twistor constructs a manifold according to its projective structure, which defines light cones and so forth.

    As spinor is a variable which has algebraic properties identical to that of angular momentum or spin.

    M-theory is something Ed Witten advanced back in 1996. The M stands for a range of things, from membrane to matrix and some have said it is an upside down W for “Witten.” Ed Witten is bloody smart, and does garner this sort of respect. The membrane is a nonlinear sigma model (well what’s that? you might ask) for an extended object. A one dimensional membrane, called a D1-brane, is a string, a two dimensional membrane (D2-brane) is a two dimensional dynamical evolving object, and this goes up to D10 branes of 10 dimensions existing in 11 dimensions. These D-branes interact with each other, and in particular open strings with end points attach to D-branes and the dynamics of the D-brane can convert one type of string into another. There exists a range of string types, and the M-theory tells how these strings are transformed into each other according to S or T dualities.

    String theory and M-theory get a bit deep in many ways. The theory is vast as well. I have ideas that the D-branes have spinor structure similar to electrons which constrain some of that structure in ways. I also spectulate that LQG might constrain some of this as well.

    I also have said that if we understood the acoustical physics of a Stradivari violin we would make progress on M-theory. Nonlinear sigma models are wave mechanics which are nonlinear and involve dynamics which are strange.

    To really understand some of this getting a PhD in physics is probably advised. Popular books on this by Susskind and Randall are on the book shelves. I think that checking these out would help for the nonspecialist.

    Cheers LC

  14. Astrofiend says:

    rudeyd Says:
    October 29th, 2009 at 2:38 am

    “hopefully won’t be crucified for my ignorance”

    Jeez – it’s not that scary in here, is it?! No one gets crucified for ignorance here, or for just asking questions…

    And speaking of which, It’s a pretty advanced question – certainly not something I’m too knowledgeable about, though I’m sure there are a couple of people in here that could help in far more detail than I (and probably more accurately).

    I’m not quite sure what the question was – were you asking if Twistor theory IS M-theory, or were you asking what both Twistor theory and M-theory are?

    Anyway, from what little I think I know about this, twistor theory is not M-theory. It is pretty hard to explain the gist of both in a small comment box on this site – you may be better off googling them, but here is an extremely brief spiel:

    Twistor theory was originally proposed a long time ago now, I think in the early 60s. I believe it’s original formulation was proposed in order to attempt a theory of quantum gravity. It is set in a four dimensional complex space, as opposed to the 3+1 dimensional Minkowski space of normal, flat spacetime. It provides a mapping between the real coordinates of spacetime to the complex valued coordinates of twistor space called, funnily enough, twistors. Beyond this, I know not much, like how exactly this leads to a better understanding of quantum gravity. I think the theory is particularly adept at describing massless fields (which the hypothetical graviton is supposed to be), but I’m really not too sure from there.

    M-theory is a super(symmetric)-string theory first laid out in the mid 90’s that is set in 11 dimensional space. In it, all of the fundamental particles of nature are viewed as harmonics of little ‘strings’ vibrating around in this higher dimensional space. Prior to M-theory, string theory had been progressing on a number of different fronts, but there were a number of different string theories, and nobody really knew why that was or whose ultimately would prove to be correct. Mr Ed Witten came along though, and (not entirely by himself) came up with M-theory. It showed in one fell swoop that all of these other string theories were actually special cases of the more general m-theory case – essentially different manifestations of a more complete theory.

    At first there doesn’t really seem to be too much of a connection here, but relatively recently, Witten (again) found some way of expressing string theory (not M theory, I think) within the mathematical framework of Twistor space. Hence there is potentially some connection there, but I don’t really know if anyone has come to an understanding of exactly what that is yet, how deep it is or whether it will end up being a fruitful thing to pursue…

  15. Astrofiend says:

    Ahh – I see Mr. LBC has beat me to the punch.

  16. Lawrence B. Crowell says:

    Astrofried wrote a reasonable description of things. Twistors were proposed by Roger Penrose back in 1967, the same date that Weinberg and Salaam propsed the electroweak theory. It is on C^4 or in real dimensions M^8 and it has a structure which gives rise to projective systems that correspond to light cones.

    There are types I and II strings of subtype A and B and the closed heterotic string. An open string is attached to a D-brane, think of the loops of thread in shag-carpet attached to the rug. The dynamics of the string is such that it can pinch off and release a closed string. So the membrane is a sort of anchor for the open string which permits this.

    Witten did write a rather long paper on connections between twistors and M-theory a few years ago. That paper makes for some challenging reading. It hinges upon Plucker coordinates on D-branes which also connect with twistor theory.

    Lawrence B. Crowell

  17. Torbjorn Larsson OM says:

    But this is old (and good!) news, why the late or absent reaction from the LQG people? [Scratches head]

    The first piece of good news is, as I understood the paper at the time, that it is a Planck scale test! And people said such bests didn’t exist.

    The second piece of good news is that Lorentz invariance goes well into the Planck scale. Rather modestly with the significant observations, but hinting at possibly orders of magnitude; love to get a repeat of that. So spacetime is an emergent property, at least in GR, but despite numerous dimensions relative positions are behaving nicely. No froth or “spacetime quantization”, as far as I understand.

    The third piece of good news is that ideas of gravitation that mistakes this emergent property for a real object, such as LQG, are dead in the water. (I believe that is what LCB is trying to tell me as well.) It just doesn’t make sense to me to try to build physics with axiomatic principles when physics seems to be algorithmic.

    These loopy ideas don’t pass elementary tests AFAIU, such as showing how one can quantize axiomatically or how to extract dynamics. (LQG famously can’t produce even a harmonic oscillator or lower energy bounds. This is one of the very few things that are relatively easy to check for an outsider, so I have dug in a few times – nothing, nada, zip.)

    Apparently they aren’t falsifiable as regards the various variants, is that how I should interpret “Loop Quantum Gravity is not dead”? Because I don’t see how you can do “loops”, putting quantum structure on space-time as if it was a simple Newtonian field, without breaking Lorentz symmetry. But then again, how is it then physics?

    Speaking of questions, here is one more raised by reading the comment thread:

    Feynman path integrals are IIRC said to work by Wick rotation mapping. Apparently they work on the physical universe, because it is said to be a completely valid though at times awkward QM theory IIRC. So why would one claim that string/M doesn’t “work” on the universe? Naively it should work as well as path integrals.

    [Or, I believe, as well as the holographic principle is supposed to work. Isn’t the conformal theory equivalent to the string theory just by Wick rotation? I.e. AFAIU the conformal theory isn’t “physical” but it is physics. Path integrals aren’t “physical” but it is physics. And string theory isn’t “physical” in dS but it is physics.

    But so are, say GR. An “emergent classical theory”, not “physical” in all the senses one could wish, but still physics.]

  18. Manu says:

    @ Torbjorn:

    “it is a Planck scale test”

    I’d like to understand this!

  19. Lawrence B. Crowell says:


    The biggest problem LQG has is gauge ambiguity and the Barbero-Immirzi parameter problem. The breaking of Lorentz symmetry is another problem, which I think has just been killed.

    In the end LQG is basically a system of constraints. The Wheeler DeWitt equation gives us that

    NH = 0, N^iH_i = 0

    which are not dynamics! These are just constraint equations or Lagrange multipliers. These can be canonically quantized. Most of what LQG does is to impose a constraint on sources which obtain from Gauss law on Sen connections, and a conformal constraint.

    LQG has a problem with bounding quantum states. The reason is there is no fundamental concept of time. The reason there is no time operator conjugate to the Hamiltonian operator is the eigen-spectra of the Hamiltonian is bounded below. The means any putative time operator can only be at best a sort of periodicity operator. The Hamiltonian constraint NH = 0 defines a diffeomorphism structure on spatial manifolds by the lapse function N, but this is not identical to time. As such the energy of any quanitzed system on the manifold will have an improperly specified periodicity, which means improperly bounded spectra.

    These constraint systems might be of some value in transforming from a BTZ-AdS black hole spacetime to a dS spacetime with its gravitational sources. The problem is troubling though. This idea might be stillborn in the end.

    String theory of course has its problems, but what might be called the “theory space” is huge — far larger than that of LQG. The anti-de Sitter spacetime with group SU(4),Wwick rotated from SU(2,2) ~ SO(4,2) in 6 dimensions has a boundary which is equivalent to conformal field theory on S^5. This is the celebrated result of Maldacena 11 years ago. The Wick rotation from SO(4,2) to SO(5,1) is a de Sitter spacetime, with two additional dimensions. The physical universe has the local Lorentz group SO(3,1) which embeds in the dS spacetime.

    One can of course have string in ordinary spacetime under an appropriate target map. It is just that the full structure of M-theory and string is with the AdS spacetime. It also has to be pointed out that gravitation in string theory is a sort of large length (longer than the string length) where the curvature emerges on a back ground. This of course annoys people, in particular the LQG people who boast of no background. Yet now in a sense there is a natural spacetime for string-gravity, which is the AdS.

    Cheers LC

  20. Jon Hanford says:

    Another informative, illuminating, thought provoking discussion on a topic. Hopefully this rational discourse catches on here at UT for the sake of interested readers and long-time UT followers. Thanks to all who took the time to elucidate for others as to the real significance of this finding.

    Special thanks to LBC for pointing me to the less technical , more accessible books on this subject from Susskind and Randall.

  21. Lawrence B. Crowell says:

    I have read Susskind, which for me is more interesting for the historical and stroy telling aspects than the physics. The physics I already understand. Lenny Susskind does lay out a decent layman’s case for conservation of information in black holes. Randall’s book I have yet to read.

    A good technical book on string theory and black holes is also by Lenny Suskind “An introduction to black holes, information and the string theory revolution,” which you can find at:

    String theory has its own difficulties of course. In particular the curvature Lagrangian is the standard sqrt{-g} R plus an expansion on this classical background of the form alpha’R^2 + higher orders. The LQG people point to this as a main complaint with string theory. It is an expansion around a background. I do think I have found a way around this problem πŸ™‚ however. The most general Lorentzian spacetime is the exceptional J^3(O) matrix in 26 dimensions, reduced from 27 with a light cone condition or the “infinite momentum” frame. There are three E_8’s there which under a triality define as well an 8 + 3 = 11 dimensional space, which under this condition is compactified to 10. This has as its automorphism groups F_4 (whose root space is a minimal tessellation in 4 dimensions) and G_2. The G_2 acts as a holonomy of M^7 and its transformations are tied to those of F_4 in a way which reduces gauge-like theory for the gravitational multiplet to an abelian Skyrme theory. The basis elements of the automorphisms are e_a = ?^dagger gamma^a? (gamma^a a Dirac matrix in the Clifford Cl_{7,1}) which defines connection terms according to internal fermion degrees of freedom.

    It think these internal fermion degrees of freedom behave as Landau electrons around a quantum critical point. This renormalizes the cosmological constant to a small value.

    Cheer LC

  22. Jon Hanford says:

    What a small world, I already own the 2005 book by Susskind (“An introduction to black holes…”) and could only follow the maths in a few sections. I’ll seek out the other Susskind and Randall books to help get my feet wet in this field. Then maybe I can get through the entire 2005 Susskind book! Thanks again, LC.

  23. Lawrence B. Crowell says:

    It is too bad this thread has fallen off the first page. If you are interested in an overview Susskind’s “Black Hole Wars” is entertaining.


  24. Why the two different photons did not arrive simultaneously?.
    Why the photons of less energy arrived first?.

  25. IVAN3MAN says:

    @ Alfonso,

    The same reason why the sky is blue and not pretty pink! πŸ˜‰

  26. Did photons travel to speed c? The difference in energy of a million times is significant? What the outcome would between the higher energy photons of gamma rays and photons of visible light?

  27. pellis says:

    “Fermi detected two gamma ray photons which varied widely in energy; yet even after traveling 7 billion years, the two different photons arrived almost simultaneously.”

    If the article means, literally, *two* photons, then how can they tell that the two photons started from the same (or extremely close) points in spacetime?

    Why can they dismiss the possibility that the two photons came from very different points, and merely arrived at the detector at nearly the same time?

  28. DrFlimmer says:

    @ Alfonso:

    Photons ALWAYS travel with the speed of light in vaccum. That does not depend on their energy. The energy only says something about the wavelength and the frequency of the photon.

    @ pellis

    Detectors in these days are able to say, from which direction the photons came. So this reduces the possibility that those photons arrive at the same time by chance. Still you cannot rule out such a possibility, because there could be two sources close to each other on the plain of the sky. But Fermi looks for gamma ray sources, and those are seldom and only rather strong sources are detectable at all. So you can assume that it is very likely that the source is the same.

  29. According to Quantum Physics standard and Relativity standard in vacuum the photons and the gravitons travel to the speed c, because they do not interact with vacuum. But, in some alternatives theories no, because they interact with vacuum. For example, in 2001, Dimitri Nanopoulos, Nikolaos Mavromatos and John Ellis found a new expression to speed of the photon, depending of their frequency; this is, of their energy. In other alternative theory, the LQG, it says too.

    Since 2006, we was expected that GLAST, today Fermy, would show that alternatives theories are false. But, sadly the great day arrived and this does not occurred. However, some readers of the news say that if.

    I wrote to distinguished professor Dimitri Nanopoulos and in E-mail, 1 of November 14:38, he said. β€œNo it is not!”.

    In my opinion, Fermy detected that two gamma ray photons which varied widely in energy, differing by a million times; yet even after traveling 7 billion years, the two different photons arrived almost simultaneously; and the team of Fermy showed that β€œto one part in 100 million billion, these two photons traveled at the same speed. Einstein still rules”. Not more!.

  30. pellis says:

    To Dr Flimmer

    Thanks for the clarification – I agree that knowing the direction and high energy of the photons reduces the chance of line of sight coincidence of a second source, but for alternative theories to be ruled out by observation I would hope for more convincing proof, or at least an estimate of the likelihood of the result not being a coincidence, such that the probability might be reduced by successive observations.

    I would not throw out LQG just yet…

  31. DrFlimmer says:

    My knowledge about String Theory, LQG and all that things is far too rudimentary. So I don’t want to judge them, and neither do I want to say something specific.

    I still hold to: c is the same. If some other theories are right, we will see that in the future.

  32. Lawrence B. Crowell says:

    Nanopoulos and co. have tendencies to make some outlandish claims. The photon does not couple to the vacuum state in the same way a photon does in a medium so as to display dispersion. I think this is fairly clear.


  33. To test alternatives theories, in moment only there are two registers on radiation gamma; they are based on very different facts:
    – Register of Fermi is based in an phenomenon that Astronomers think is cause of explosion when neutron stars collide , to 7.3 billion light-years away, and with photons possessed energies only differing by a million times.
    – The register of Magic is based in a blazar to notable minor distance, that showed an indication of a 4Β±1 min time delay between the peaks of F(1.2 TeV).

    Magic, 3 years after, it uses to probing quantum gravity; result is: β€œThe probability of the zero-delay assumption relative to the one obtained with the ECF estimator is P = 0.026. The observed energy-dependent delay thus is a likely observation, but does not constitute a statistically firm discovery.”.

    I think that register Fermi is affected strongly by Universe expansion, changes in density of energy of vacuum; and probably changes in permittivity and permeability. This is much minor in Magic. This will give to speculate. For example, in register Fermi, if photons interact with vacuum in two ways, that is, photon-vacuum and vacuum-photon then can that higher photon travel faster that low photon inside of an unknown lapse?. But, result of carrier is that pair arrived just nine-tenths of a second apart.

    Of other hand, I have a great respect to both teams. Team of Magic is compound by a great number of scientists and in the test on quantum gravity also are included the emeritus professors D. Nanopoulos, John Ellis, N. Mavromatos, A. Sakharov, E. Sarkisyan.

  34. warnings: – By F(>1.2 TeV) read F(1.2 TeV).
    – By carrier read race.

Comments are closed.