Why gravitons?

[Hari Seldon]

Can someone enlighten me why gravitons are needed and what they do? The essence of general relativity is that matter curves space and so-called gravitational attraction is just objects following geodesics in space-time (unless, of course, some object like the earth's surface prevents it). So where do gravitons come in?

[JWT Kottekoe]

Gravitons are a product of the attempts to unify quantum mechanics and general relativity. All other known forces are described by quantum field theories, where the force is transmitted by the exchange of virtual particles. When you try to quantize general relativity, you are led to the notion of gravitons, which are the virtual particles that give rise to the gravitational force. Unfortunately, no one has succeeded in unifying gravity and quantum theory. Even if such a theory is found, gravity is so weak that we will likely never be able to detect gravitons. In fact, some have argued that gravitons are essentially undetectable, even with ridiculously impractical detection schemes.

[Chronos]

Even if such a theory is found, gravity is so weak that we will likely never be able to detect gravitons. In fact, some have argued that gravitons are essentially undetectable, even with ridiculously impractical detection schemes.

To clarify a bit: In the same way that an electromagnetic wave can be regarded as a beam of photons, a gravitational wave could likewise be regarded as a beam of gravitons, and we're quite close to detecting gravitational waves (expect the first detection sometime in the next few years). The tricky part is that it takes an astronomically large number of gravitons to be able to get a detection: What we probably won't ever detect is individual gravitons.

And yes, I really do mean "never", there, and I really do understand the full implications of that word. As in, over the entire span of time, the probability is extremely low that humans or any species descended from humans will ever even once manage to detect individual gravitons. It's just that hard a problem.

[Hari Seldon]

Okay, I see how gravitational waves are carried by gravitons. But I have seen them described in many places as the carriers of the "force" of gravity, which is a different matter.

I also do understand how the presumed emission of gravitons explains how pulsars slow down.

[MikeS]

Originally Posted by Hari Seldon

Okay, I see how gravitational waves are carried by gravitons. But I have seen them described in many places as the carriers of the "force" of gravity, which is a different matter.

This is the notion of "virtual particles". Basically, the forces in the Standard Model (electromagnetism and the two nuclear forces) can be explained if the objects exerting forces on each other exchange these virtual particles with each other. Two electrons that repel each other via the electromagnetic force, for example, can be viewed as sending "virtual photons" back & forth, and the recoil from these photons is what we view as the force they exert on each other. These virtual photons are a little different from "real" photons, in that the virtual ones have momenta and energies that wouldn't be allowed for "real" photons. But they're all waves of the same photon field.

In the case of gravity, then, "real" gravitons are the ones we would experience as gravitational waves, and "virtual" gravitons are the ones which are exchanged between two objects exerting a gravitational force on each other.

I also do understand how the presumed emission of gravitons explains how pulsars slow down.

The basic argument is thus: gravitons (or gravitational waves) carry energy with them. Energy is conserved, so where did that energy come from? It had to have come from the pulsar itself. Actually figuring out how the gravitational field "pushes back" on the pulsar to slow it down is another question entirely, and is notoriously difficult to figure out.

[Chronos]

Incidentally, the notion of virtual particles being responsible for forces is extremely well-tested for electromagnetism: QED, the theory which describes electromagnetism in this way, has been tested to greater precision than any other scientific theory in history, and has passed all of the tests swimmingly. So it's natural for physicists to want to be able to extend that success to other fields.

[Derleth]

Is there any way, even in principle, to replicate the double-slit experiment with gravitons?

[Chronos]

In principle, sure, but it'd be the mother of all engineering challenges. And it wouldn't be interesting anyway, since all it'd show would be gravitons behaving like waves, and we already understand gravity as waves just fine anyway.

[JWT Kottekoe]

See for example:

Can Gravitons be Detected

Freeman Dyson has questioned whether any conceivable experiment in the real universe can detect a single graviton. If not, is it meaningful to talk about gravitons as physical entities? We attempt to answer Dyson's question and find it is possible concoct an idealized thought experiment capable of detecting one graviton; however, when anything remotely resembling realistic physics is taken into account, detection becomes impossible, indicating that Dyson's conjecture is very likely true. We also point out several mistakes in the literature dealing with graviton detection and production.

[Lumpy]

Originally Posted by MikeS

The basic argument is thus: gravitons (or gravitational waves) carry energy with them. Energy is conserved, so where did that energy come from? It had to have come from the pulsar itself. Actually figuring out how the gravitational field "pushes back" on the pulsar to slow it down is another question entirely, and is notoriously difficult to figure out.

I thought it was explained by the same General Relativity equations that predict gravity waves in the first place: the emission of gravity waves alters the frame of reference that says that the pulsar is rotating, changing it to one where the pulsar is rotating more slowly.

[jayjay]

One thing I've never been able to understand...we're SITTING on what should be a massive emitter/receiver of gravitons, and damn regular ones, at that, given that we have one of the most relatively massive satellites in the solar system in orbit around us, exchanging ITS gravitons with us. So why is it so hard to detect these things? It's not like we actually have to wait for a graviton from some light-years distant sun to intersect our little blue dot.

[Exapno Mapcase]

Gravity is 42 orders of magnitude weaker than electromagnetism. So your detector has to be 42 orders of magnitude more sensitive. That's a million billion billion billion billion times more sensitive. That's (very loosely) like scaling up from the diameter of a proton to the diameter of the observable universe.

When your detectors have to be universe scale, humans decide it's not even theoretically possible.

[Chronos]

It's worse than that, even. Not only do you have to contend with the weakness of the interaction, but you also have to contend with the extremely low energies involved. Gravitons would have the same relationship between energy and frequency as photons do. Visible light, for instance, has a frequency of about 10¹⁵ Hz, and even microwaves, which are about the lowest-energy radiation where we can detect individual photons, are about 10¹⁰ or 10¹¹ Hz. Meanwhile, the highest-frequency gravitational waves we expect to exist in the Universe are only a few kilohertz.

And the Earth-Moon system would indeed be emitting gravitational waves, but far too weakly for us to hope to detect them or their effects. For a realistic chance of detection, you need something like a pair of neutron stars that are about to merge, or the like.

[jayjay]

That makes sense. Thanks, folks!

[Hari Seldon]

But gravity isn't a force. It is just a warp in space-time. If you tell me that gravitons make gravitational waves, so be it. But if you tell me that exchange of gravitons causes a force that doesn't exist, I have a problem.

[iamnotbatman]

Originally Posted by Hari Seldon

But gravity isn't a force. It is just a warp in space-time. If you tell me that gravitons make gravitational waves, so be it. But if you tell me that exchange of gravitons causes a force that doesn't exist, I have a problem.

The exchange of virtual gravitons. That is worth emphasizing because the perturbative treatment of "forces" in quantum field theory makes use of a convenient mathematical formalism involving mathematical terms we find convenient to associate with gravitons. This should not be surprising, since the gravitational "force" is a result of space-time curvature, and gravitons are vibrations in that curvature. Therefore gravitons provide a natural set of basis states for describing the interaction between particles and space-time curvature. Of course, what is "really happening" is up for debate. Ultimately particles affect space-time curvature and that curvature affects their trajectories; there is an interaction, and that interaction can be most conveniently described mathematically by the exchange of virtual gravitons.

"But a particle just travels along the shortest path through space-time", you object. Gravity is not actually unique in this respect. In QED and QCD, particles follow paths of extemal action. In all of these cases, when you try to actually calculate the path or the interactions, the "forces" that define those paths/interactions are described perturbatively using virtual particles.

[Hari Seldon]

Originally Posted by iamnotbatman

The exchange of virtual gravitons. That is worth emphasizing because the perturbative treatment of "forces" in quantum field theory makes use of a convenient mathematical formalism involving mathematical terms we find convenient to associate with gravitons. This should not be surprising, since the gravitational "force" is a result of space-time curvature, and gravitons are vibrations in that curvature. Therefore gravitons provide a natural set of basis states for describing the interaction between particles and space-time curvature. Of course, what is "really happening" is up for debate. Ultimately particles affect space-time curvature and that curvature affects their trajectories; there is an interaction, and that interaction can be most conveniently described mathematically by the exchange of virtual gravitons.

"But a particle just travels along the shortest path through space-time", you object. Gravity is not actually unique in this respect. In QED and QCD, particles follow paths of extemal action. In all of these cases, when you try to actually calculate the path or the interactions, the "forces" that define those paths/interactions are described perturbatively using virtual particles.

That helps. Incidentally, I am glad someone besides me has noticed that photons follow paths of extreme action rather than least action. For example, reflection in a sufficiently concave spoon will follow the longest path. This becomes very clear from the explanation that light follows all intermediate paths, but fails to self-interfere only at extreme points.

[Chronos]

It can't be following the longest path, since the longest path would be a fractal of infinite length. At most, it might be some sort of "locally longest path", but you'd have to be very careful about how you defined "locally", or you'll end up with one of those fractals in your neighborhood.

[Hari Seldon]

Originally Posted by Chronos

It can't be following the longest path, since the longest path would be a fractal of infinite length. At most, it might be some sort of "locally longest path", but you'd have to be very careful about how you defined "locally", or you'll end up with one of those fractals in your neighborhood.

Try "the longest path between the light source, the spoon, and me. Actually, I once heard Feynmann give a lecture on this very point and he would have included all possible paths, including those that went to alpha centauri and back. But the only ones that didn't self-interfere were those that followed a path of extremal path length, usually the shortest, but sometimes the longest.

[D18]

Originally Posted by Chronos

Meanwhile, the highest-frequency gravitational waves we expect to exist in the Universe are only a few kilohertz.

So if we amplified them somehow, we could hear them?

[Chronos]

So if we amplified them somehow, we could hear them?

Well, and converted them to sound waves. But yes, you can find sound files of, say, simulated neutron star mergers online. They start off low, then chirp high, and abruptly cut off.

[Lumpy]

Originally Posted by iamnotbatman

The exchange of virtual gravitons. That is worth emphasizing because the perturbative treatment of "forces" in quantum field theory makes use of a convenient mathematical formalism involving mathematical terms we find convenient to associate with gravitons. This should not be surprising, since the gravitational "force" is a result of space-time curvature, and gravitons are vibrations in that curvature. Therefore gravitons provide a natural set of basis states for describing the interaction between particles and space-time curvature. Of course, what is "really happening" is up for debate. Ultimately particles affect space-time curvature and that curvature affects their trajectories; there is an interaction, and that interaction can be most conveniently described mathematically by the exchange of virtual gravitons.

"But a particle just travels along the shortest path through space-time", you object. Gravity is not actually unique in this respect. In QED and QCD, particles follow paths of extemal action. In all of these cases, when you try to actually calculate the path or the interactions, the "forces" that define those paths/interactions are described perturbatively using virtual particles.

So gravitons (and photons, etc.) follow the shortest path and the longest path possible? Could this be the answer to the GR question of inertia- every object's virtual gravitons are interacting with the rest of the universe?

[iamnotbatman]

It's really more proper to say "stationary" action (my bad) rather than minimal/extremal/etc, although it's really not a point worth worrying about. Basically the action must either be a minimum or a saddle point, but the action is never actually maximized. When you calculate the trajectory of a particle using the principle of stationary action, you do it by finding a path for which the action is stationary. Whether that path happens to maximize or minimize (or is just a saddle point of) the action is not really relevant. It is not trivial to make a general statement about when a path of stationary action will be a minimum/maximum/saddle-point, but it can be proven that it is never a maximum. It is also difficult to find simple examples (I think -- maybe someone here knows one) where the action is not minimized. I can think of one. Put a particle on a sphere. Pick two points on the sphere. There are two paths of stationary action. One is the path of shortest length: an arc along a great circle connecting the two points. The second path is also along a great circle connecting the two points, but going around the "long way". Both paths minimize the action, and indeed, we know the particle could just as well take the shorter or the longer path: either way it would be travelling in a straight line. But you'll notice that the longer path is not "the path of maximum length," because we could have added wiggles to the path to make it as long as we wanted, but those paths would not have stationary action.

[Hari Seldon]

Originally Posted by iamnotbatman

It's really more proper to say "stationary" action (my bad) rather than minimal/extremal/etc, although it's really not a point worth worrying about. Basically the action must either be a minimum or a saddle point, but the action is never actually maximized.

I gave a simple example of a (local) maximum: reflection from a sufficiently convex spoon (the curvature has to be greater than that of a paraboloid). What is so hard about that?

[iamnotbatman]

I don't get it. How can it be a local maximum if you can always add wiggles that make it larger?

[wolfking]

Originally Posted by iamnotbatman

It is also difficult to find simple examples (I think -- maybe someone here knows one) where the action is not minimized.

Well here is an example that occurs all the time in daily life. Whenever you look at an object in a mirror and you also have direct line of sight to that object, you are presented with two different paths, one(direct line of sight) is minimal and the other(through the mirror) is not. However they are both locally minimal.

[Lumpy]

An aside about gravitons: since we don't have a good quantum theory of gravity, what we surmise about gravitons is more or less by default of what we guess has to be true. Since quantum theory holds that all energy has to be quanticized, gravity is presumed to be too. And simply to not contradict the known behavior of gravity, gravitons would have to have zero rest mass, be electrically neutral and be bosons with a quantum spin of 2. But all this is simply by way of all the alternatives being ruled out; we don't have a positive theory yet that predicts them.

[Chronos]

Well, maybe. The reason why the String Model first started getting any attention is that versions of it which accounted for all of the Standard Model particles also all turned out to predict the existence of a massless spin-2 particle, despite that not being one of the things they were designed to predict. Of course, there's very little real progress in the String Model.