Physicist suggests speed of light might be slower than thought

[Steve001]

(Phys.org) —Physicist James Franson of the University of Maryland has captured the attention of the physics community by posting an article to the peer-reviewed New Journal of Physics in which he claims to have found evidence that suggests the speed of light as described by the theory of general relativity, is actually slower than has been thought.
The theory of general relativity suggests that light travels at a constant speed of 299,792,458 meters per second in a vacuum. It's the c in Einstein's famous equation after all, and virtually everything measured in the cosmos is based on it—in short, it's pretty important. But, what if it's wrong?


Read more at: http://phys.org/news/2014-06-physicist-slower-thought.html#jCp

Read more at: http://phys.org/news/2014-06-physicist-slower-thought.html#jCp
Read more at: http://phys.org/news/2014-06-physicist-slower-thought.html#jCp

 

[ElMondoHummus]

Folks, I'm not physicist. I only took 2 undergrad courses in college so I'm as casual a layman as you can get. But that said, the paragraph that stands out is this one:

... Franson wonders, what if light slows down as it travels due to a property of photons known as vacuum polarization—where a photon splits into a positron and an electron, for a very short time before recombining back into a photon. That should create a gravitational differential, he notes, between the pair of particles, which, he theorizes, would have a tiny energy impact when they recombine—enough to cause a slight bit of a slowdown during travel. If such splitting and rejoining occurred many times with many photons on a journey of 168,000 light years, the distance between us and SN 1987A, it could easily add up to the 4.7 hour delay, he suggests.

Is he really saying c is incorrect? Or that in terms of practical travel, a photon occasionally stops to split itself before recombining and resuming travel at c? I know that sounds like a nitpick, but honestly, that's what I'm taking from it: It's not that c is wrong as an ultimate limit, it's just that it won't measure 100% because of these "interruptions". Is that the proper interpretation of the theory?

Any criticisms of my conclusion are sincerely welcome.

 

[ehcks]

Folks, I'm not physicist. I only took 2 undergrad courses in college so I'm as casual a layman as you can get. But that said, the paragraph that stands out is this one:


Is he really saying c is incorrect? Or that in terms of practical travel, a photon occasionally stops to split itself before recombining and resuming travel at c? I know that sounds like a nitpick, but honestly, that's what I'm taking from it: It's not that c is wrong as an ultimate limit, it's just that it won't measure 100% because of these "interruptions". Is that the proper interpretation of the theory?

Any criticisms of my conclusion are sincerely welcome.

If that's how it is, I'm not sure it's even that big of a change. There are already experiments showing you can slow the transfer of light by putting a dense fluid in the way. The photons interact with electrons in their path, cease to exist as photons, and are then re-emitted.

The photons themselves are always traveling at exactly c.

 

[Ziggurat]

I don't get it. I thought the neutrino/light delay was already understood: neutrinos are emitted from the core of a star going supernova, and they pass through the outer layer of the star without slowing down. Photons from the core of the star, however, cannot reach the surface of the star, because stars are opaque. No change in the light emitted from the star will occur until a shockwave (which may travel much slower than c) can propagate from the core to the surface.

 

[BowlOfRed]

I don't get it. I thought the neutrino/light delay was already understood: neutrinos are emitted from the core of a star going supernova, and they pass through the outer layer of the star without slowing down. Photons from the core of the star, however, cannot reach the surface of the star, because stars are opaque. No change in the light emitted from the star will occur until a shockwave (which may travel much slower than c) can propagate from the core to the surface.

Exactly what my understanding was.

The phys.org article itself seems to be hyperbole. If this is a real effect and there is an error, it would seem to be on the order of 4.7hours/168000years or 3.2x10^-9. The breathless "everything is wrong!!!" really annoyed me.

Still a few billionths seems to be within the precision that we could measure in a lab. Or is the idea that this effect is zero at "short" distances? Would it affect different energy level photons differently (something for which we appear to have contradictory evidence)?

 

[Reality Check]

J D Franson does point out that his idea is nonphysical:
Apparent correction to the speed of light in a gravitational potential

The effects of physical interactions are usually incorporated into the quantum theory by including the corresponding terms in the Hamiltonian. Here we consider the effects of including the gravitational potential energy of massive particles in the Hamiltonian of quantum electrodynamics. This results in a predicted correction to the speed of light that is proportional to the fine structure constant. The correction to the speed of light obtained in this way depends on the gravitational potential and not the gravitational field, which is not gauge invariant and presumably nonphysical. Nevertheless, the predicted results are in reasonable agreement with experimental observations from Supernova 1987a.

(my emphasis added)

The Phys.org is not quite right: That the speed of light is constant is a postulate of special relativity.

 

[ben m]

The article appears to be wrong from the get-go.

The author looks at a Feynman diagram of a photon generating a virtual e+e- loop. The author treats these as real electrons of mass m and simply tosses a gravitational potential energy term (2 m Phi) into the Hamiltonian. This is flatly incorrect; GR doesn't act on the basis of rest-mass, it acts on energy and momentum. The energy-momentum of the e+e- loop is exactly the same as the photon energy-momentum and in GR it is incorrect to put the electron rest mass in there anywhere.

The author seems to think that his equation is using GR correctly, so this is a problem. If the author thinks he's proposing a modified version of GR, his modified version appears subject to experimental tests other than "let's make some assumptions about the source timing of SN1987A".

It is sort of astounding that someone could have the idea "maybe standard GR needs this new term" and not immediately follow it up with " ... let me see what the new term does to high-precision tests of GR". The author's discussion of the equivalence principle seems to miss the point, treating equivalence principle as a thought-experiment-on-a-train way and not bothering to consider (nor even showing awareness of) the actually-existing high-sensitivity experiments that compare the gravitational and inertial masses of bodies of different compositions. (Without doing the math, his theory would appear to predict that a lead brick falls faster than a carbon brick---extremely stringently ruled out by experiment.) I mean: just because "this term affects the speed of light" is the first thing you can think of doesn't mean it's the only thing that happens or the only way to test the theory.

The statement that something about the dynamics "depends on the potential but not the field" is bizarre. If you put a potential in the Hamiltonian, then, um, yes the dynamics are going to depend on the potential gradients!

More generally, there are some oddities that suggest to me that the author has never really dealt with these equations before, either as a reader or an author. When has a theory paper felt the need to write out the (standard) Dirac spinors explicitly?

 

[angrysoba]

Physicist suggests speed of light might be slower than thought

No, thought is much slower than the speed of light. Especially in the morning.

 

[McHrozni]

4.7 hours in 168,000 years?

That's an error at the 10th significant digit. He's arguing whether the speed of a photon isn't really 299,792,458 m/s, but rather 299,792,457.8 m/s, assuming our definition of meter remains as it is. Meter is defined by the speed of light, so this complicates things a bit. But the point stands, the difference is so small it's hardly worth loosing sleep over.

McHrozni

 

[uke2se]

4.7 hours in 168,000 years?

That's an error at the 10th significant digit. He's arguing whether the speed of a photon isn't really 299,792,458 m/s, but rather 299,792,457.8 m/s, assuming our definition of meter remains as it is. Meter is defined by the speed of light, so this complicates things a bit. But the point stands, the difference is so small it's hardly worth loosing sleep over.

McHrozni

On the contrary! This obviously means Plasma Cosmology is correct.

Not directed at OP, simply a general reflection.

 

[epepke]

The author looks at a Feynman diagram of a photon generating a virtual e+e- loop. The author treats these as real electrons of mass m and simply tosses a gravitational potential energy term (2 m Phi) into the Hamiltonian. This is flatly incorrect; GR doesn't act on the basis of rest-mass, it acts on energy and momentum. The energy-momentum of the e+e- loop is exactly the same as the photon energy-momentum and in GR it is incorrect to put the electron rest mass in there anywhere.

Winner.

/thread

 

[Ziggurat]

4.7 hours in 168,000 years?

That's an error at the 10th significant digit. He's arguing whether the speed of a photon isn't really 299,792,458 m/s, but rather 299,792,457.8 m/s, assuming our definition of meter remains as it is. Meter is defined by the speed of light, so this complicates things a bit. But the point stands, the difference is so small it's hardly worth loosing sleep over.

McHrozni

If there was any evidence that there was an error in the 10th significant digit, that would indeed be major. If there was any evidence of an error in the 20th significant digit, that too would be major. Any nonzero difference between c (the limiting local velocity of the universe) and the speed at which light travels in a vacuum would be gigantic news for physics.

But there is no evidence. He screwed up the calculations, and his interpretation of observations is wrong. He has neither theoretical nor experimental grounds for his claim. That is why this is a nothingburger, not because the difference is so small.

 

[McHrozni]

If there was any evidence that there was an error in the 10th significant digit, that would indeed be major. If there was any evidence of an error in the 20th significant digit, that too would be major. Any nonzero difference between c (the limiting local velocity of the universe) and the speed at which light travels in a vacuum would be gigantic news for physics.

I red his argument as photons exist as particles with mass for brief periods of time, and travel at speeds lower than c during that time. In light of that I don't see this as an issue. I may be wrong though.

McHrozni

 

[The Greater Fool]

*IF* this is true... It would seem not to impact our scale (distances in our solar system), but most certainly would affect calculations of how far away other stars and galaxies actually are. Certainly, 4.7 hours would impact the amount of red shift, so minimally the universe would be expanding at a slower pace than currently believed. (?)

 

[ben m]

4.7 hours in 168,000 years?

That's an error at the 10th significant digit. He's arguing whether the speed of a photon isn't really 299,792,458 m/s, but rather 299,792,457.8 m/s, assuming our definition of meter remains as it is. Meter is defined by the speed of light, so this complicates things a bit. But the point stands, the difference is so small it's hardly worth loosing sleep over.

McHrozni

Yeah, no, the point wouldn't be "wow actually Alpha Centauri is only 3.99999999999 light years away". The point would be "spacetime and gravity behave fundamentally differently than we thought it did", including the loss of the equivalence principle among other fundamental things.

It's like if a biologist discovered that 0.00000001% of your DNA consists of, not nucleotide pairs, but tiny malevolent leprechauns holding hands. Nobody would say "no big deal, 99.9999999% is organic chemistry like we already knew".

 

[Ziggurat]

I red his argument as photons exist as particles with mass for brief periods of time, and travel at speeds lower than c during that time. In light of that I don't see this as an issue. I may be wrong though.

McHrozni

From a practical point of view, it may not matter if the difference is sufficiently small. But from a theoretical point of view, it would be a very big deal, if it were true (but there's no reason to think it is). Among other problems, it screws up our current scheme of defining the meter in terms of c, since we use light to determine c, and I don't see how you can experimentally separate the instantaneous top speed of a photon from its average speed. It would also mean that light should be slightly dispersive even in a vacuum (that is, different wavelengths should travel at different speeds).

 

[DuvalHMFIC]

I just want to know, as a current engineering student, if this is going to have effects on my homework or not. 3 x 10^9 m/s is pretty easy to remember, I don't need another Avogadro dancing in my head.

 

[Ziggurat]

I just want to know, as a current engineering student, if this is going to have effects on my homework or not. 3 x 10^9 m/s is pretty easy to remember, I don't need another Avogadro dancing in my head.

3 x 10^9 m/s is indeed easy to remember. Unfortunately for you, however, c is approximately 3 x 10^8 m/s, not 3 x 10^9 m/s.

 

[Jrrarglblarg]

Didn't Lord Kelvin flail on about setting the final decimals or something?

 

[ben m]

I don't get it. I thought the neutrino/light delay was already understood: neutrinos are emitted from the core of a star going supernova, and they pass through the outer layer of the star without slowing down. Photons from the core of the star, however, cannot reach the surface of the star, because stars are opaque. No change in the light emitted from the star will occur until a shockwave (which may travel much slower than c) can propagate from the core to the surface.

That's what I thought he was objecting to, but it's not quite.

Kamiokande II and IMB both saw the effect you're describing, and it's explained as you say. The (small) LSD experiment at Mont Blanc claims to have seen five neutrinos, very close to their threshhold, four hours earlier. This is normally attributed to non-neutrino noise, basically---if there was really a huge 4-hour-early neutrino pulse Kamioka would have seen it. LSD wasn't a great detector, the events weren't great (near the noise threshhold, not too well documented, etc.) so this is not hard to believe.

The author is saying that 1987A "went off" with a neutrino pulse; four hours later it produced light and a second neutrino pulse. (He has a citation to one paper presenting a double-bang model like this). He claims that this double-bang model, plus normal light propagation, would have let LSD see the first pulse but the second pulse (Kamiokande+IMB) should have arrived at the same time as the light. (I haven't read the citation to see if this is true.) The entire point of his new model, then, is to insert an additional light-travel delay after the 2nd neutrino pulse.

So, it's crazy and poorly motivated, but it's not as "did you not even read the first paper you cited" poorly motivated as I first thought.