Talk:Virtual particle

From Wikipedia, the free encyclopedia

WikiProject Physics This article is within the scope of WikiProject Physics, which collaborates on articles related to physics.
Portal:Physics
 Physics Portal
??? This article has not yet received a rating on the assessment scale. [FAQ]
??? This article has not yet received an importance rating within physics.

Please rate this article, and then leave comments here to explain the ratings and/or to identify the strengths and weaknesses of the article.

Contents

[edit] Real vs. virtual

I've got a question for the physicists here. I've always wondered about so-called "virtual photons." Is there REALLY any difference between real photons and virtual photons? What I mean is this: when we detect a photon with our eye or some instrument, the photon coming in excites some electrons into some higher energy states, or causes induced relaxation of some excited state, or whatever...but basically all the photon did before it "died" was move a charged particle (the electron). This is the same as what a virtual photon does! So, the only difference that I can see is that with a virtual photon, we consider that we observed a charged particle change its trajectory, whereas with a "real" photon, we claim that we observed a photon hitting our detector. But, really, they're the same thing? Comments? Ed Sanville 04:32, 23 Dec 2004 (UTC)

As described in the article, virtual particles occur strictly as intermediate steps in interactions. If your photodetector was triggered, something other than the virtual particle provided the initial energy for the triggering. The collision of two electrons is a good example for this: the force that the two electrons experience as they approach each other is carried by disturbances in a sea of virtual photons, but the change in trajectory of one electron is caused by exchange of energy and momentum with the other electron. The virtual photons acted solely as intermediaries. (ObCaveat: I'm still hazy on where the energy recovered from the work done by the pressure produced by the Casimir effect between plates whose separation distance is being changed comes from.)--Christopher Thomas 06:53, 12 Jun 2005 (UTC)
"virtual particles occur strictly as intermediate steps in interactions" But doesn't that depend on what you characterize as "the interaction?" For example, in my above example of a photon hitting a detector, you could consider the interaction to be between the particle that emitted the photon, and the particle in the detector that absorbed it. If you considered that to be the interaction, then it would seem that the photon has now become a virtual photon... is there some other requirement for a photon to be "virtual?" Ed Sanville 21:39, 17 April 2006 (UTC)
There are formal definitions of real vs. virtual photons. If you're using Feynman diagrams, virtual particles are ones that don't enter or leave the diagram (or alternatively, particles that do enter or leave the diagram are the real ones). If I recall correctly it was User:Linas who gave another description that treated virtual particles as the mathematical artifacts you get when you quantize behavior of forces in a certain way (via perturbation theory), with alternate ways of treating the problem removing the virtual particles, but leaving the real ones. My best advice for telling which is which is to treat any interaction over a distance as being mediated by real particles, as the range of interaction for virtual particles is very, very short (the simplified explanation which I was given was that they couldn't exist long enough for it to be _possible_ to measure them, due to uncertainty principle effects providing the loophole in conservation laws that the simplified explanation uses to handwave their existence in the first place). This is mostly-true down to scales approaching the size scale at which the Casimir effect becomes important, for photons, or scales comparable to the size of a nucleon, for things like mesons and gluons. One of the lurking physicists can give you a much clearer answer than I can. --Christopher Thomas 23:39, 17 April 2006 (UTC)
Ed's reasoning is correct: in a certain philosophical sense, there is no distinction between real and virtual photons. Chris Thomas' answer is correct as well: in practical matters, there's a clear distinction, and one has to cook up intriguing and confusing thought experiments to blur the distinction. linas 04:39, 20 April 2006 (UTC)
OK, thank you Chris Thomas and linas, this is one of those things I've always wondered about, and I wanted to get the consensus from the people who actually study these things. Ed Sanville 12:27, 20 April 2006 (UTC)

I would just like to thank the person who wrote this: "Furthermore, in the photon's frame of reference, no time elapses between emission and absorption. This statement illustrates the difficulty of trying to distinguish between "real" and "virtual" particles as mathematically they are the same objects and it is only our definition of "reality" which is weak here." Excellent quote! Itistoday 06:52, 19 June 2006 (UTC)

[edit] More info

Where can I get more info on this (like where are the references for this article?) ChadThomson 1 July 2005 07:15 (UTC)

Google for "physics FAQ", or take a look at a first-year physics text of the same type I suggested at Talk:Big Bang. An introductory quantum mechanics text should have a more detailed description (virtual particles show up a lot when dealing with Feynman diagrams), but be warned that QM texts tend to have daunting amounts of math, so be prepared to brush up on relevant areas before such a text will make much sense.--Christopher Thomas 1 July 2005 15:18 (UTC)

[edit] Casimir effect dispute

See Talk:Casimir_effect -- Intangir 08:13, 26 December 2005 (UTC)

This page denies the validity of the common VP explanation of the Casimir Effect without citing a source for this. This is the same dispute as the one at that page- please discuss there instead of here.

Text under dispute:

Note that it is often believed that the Casimir effect has to do something with virtual particles. This is not directly true since the Casimir-effect results from an external perturbation of the zero point oscillations of the electromagnetic quantum field through two plates placed in the vacuum. -- Intangir 21:40, 26 December 2005 (UTC)

Yeah, this was based on a misunderstanding of the Casimir effect. Now fixed. I re-wrote the article entirely. linas 02:15, 28 December 2005 (UTC)

[edit] Ontological status

It doesn't seem to me to be terribly difficult to fix any perceived ontological problems with virtual particles. They are, of course, off shell because they can violate special relativity's speed limit and more. However they are still considered compatible with relativity because they violate this kind of stuff too briefly to be directly observed(naturally as they themselves exist too briefly to be observed). As they last longer(I mean word line interval) they become more probable to obey the rules. In essence, this is just saying that these rules are only approximate- they apply on all but the most incredibly short scales. Moreover, there need not be a binary classification between virtual particles and real particles where the particle is either one or the other. There is a continuum, as the worldlines get longer virtual particles smoothly become more 'real'. We don't need to postulate some new kind of unobservable thing. All we need do is postulate that some of the rules which we see to be followed on larger scales might not be followed on scales smaller than we have verified these rules.
As for the perturbative stuff, it may be the pertubation is only an approximation. However, it seems perfectly possible that there could be a maximum perturbation. Of course, the general failure of perturbative QCD casts doubt on this. --Intangir 00:21, 30 December 2005 (UTC)

I'm not sure what to say. Most practicing physicists understand (fairly clearly) that virtual particles are "fictitious", they're the side-effect of perturbation theory. The article tries to report this common, accepted understanding. One can try to "fix the ontological problem", but there's lots of dangerous reefs on that sea: one must steer towards philosphy, the interpretation of quantum mechanics, and "original research". Trying to somehow "clarify" this concept won't bring you new insight, nor will it bring you new experimetal predictions, and so the concept kind of gets put aside; one moves on to more promising ideas. linas 01:11, 30 December 2005 (UTC)

Interesting. Clearly "fictitious" is virtually in their very name. Hehe. However, it occurs to me that they are the insight. They seem to me to be an attempt to restore the various phenomena that we call 'particles' to the status of the fundamental building blocks of reality(atoms if you will). Through this "fixing of the ontological problem" behind particles, Feynman visualized the formulation of QED, ultimately enabling him to make new predictions. However, you can't really push this kind of issue aside. Whether you take waves, particles, or fields as fundamental, the issue is always which to choose and is always there. The insight is that it really does matter which.
I know that the philosophers of physics still coherently debate whether or not particles are fundamental. I would find it tragic if what you suggest is true, that physicists themselves have almost invariably given up on interpreting particles as potentially fundamental.
Well, actually, yes. For starters, there are many different interpretation of quantum mechanics, many of which are fraught with technical and philosophical difficulties. There is a small group of active workers who try to design experiments that distinguish between the different interpretations, but the work is hard, slogging and confusing. Most practicing physicists are "instrumentalists" (instrumentalist interpretation Gack!. Well, there is such a thing, but probably not what this article describes. I'm exhausted trying to keep WP honest about quantum.) in that they believe only in what the formulas predict, and don't ask "but what does the formula mean?". Its a very safe approach, keeps you on solid ground, keeps you from making errors.
As to "particles" ... good lord, left that ballpark long ago. For starters, particle-wave duality discarded the notion of a "particle" 80 years ago, and recognized the demolition of the particle with a nobel prize for the demolisher (de Broglie). You may argue that QED and QFT and Feynman resurrected the particle ... but good lord, the world has moved on since then. Take a look at Kaluza-Klein theory, supergravity, loop quantum gravity, string theory, 't Hooft loops, Wilson loops, BRST ... one measures "particles" in particle colliders, but that's not what the theories talk about.
You were thinking of instrumentalism, which is very close to my personal philosophy of science. From my point of view a 'particle' (these days) is a metaphor for a set of phenomena where events are discretely quantized but their probability distributions exhibits interference even 'self-interference'. The 'particle' in this case is a self-interacting one where the lack of independance between possibilities forces the abandonment of simple probability. Without localization, normal probability can't be used. What I mean is just this 'event' kind of phenomena. It makes good sense to directly reason from it, as Feynman essentially did. Especially since it can build up such fabulously predictive models without making any crazy assumptions. I know of the first four of those theories. None of them seem to have actually 'succeeded', only inspired. But, come on, extra dimensions, entire families of missing particles, and stringy super difficult math be damned- are we really that desperate? Has any of these paths achieved some new level of precision? Surely there are still plenty of more plausible paths to consider? Did U(4) and SU(4) just fail spectacularly or something? Are we fresh out of 'simple ideas', or have physicists forgot to induct over theories and not just experimental evidence? The fact that the world has moved on is what is precisely troubling. With so many caught up in a desire to create, who is trying to fix the old stuff? Oh right, thats below, hehe. --Intangir 05:35, 31 December 2005 (UTC)
Certainly virtual particles have bothersome, unresolved technical issues. But at this time you really can't argue that virtual particles are mere artifacts of a perturbative approach to the more fundamental field theory. As I see it, quantum field theory seems to have inherited the very same issues. Haag's theorem shows that the standard approach to QFT is just as "dippy" as renormalizing perturbative QED in the sense that QED couldn't rigorously handle the interaction picture, and so neither can QFT. QFT is just better at sweeping this same old inconsistancy deeper under the carpet(This is the same kind of success that Feynman said he earned the Nobel Prize for, hehe).
A lot of excitement and effort is going into non-perturbative models these days, which don't have "feynman diagrams" in them, because they are exactly solvable. Basically, over the last five decades, there were a few outstanding non-perturbative solutions (see Category:exactly solvable models) that have shed light on how to go about solving QFT "correctly". These include the Potts model, the KdV equation, the non-linear Shroedinger equation, Toda field theory, and to a lesser degree, the Skyrmion, the WZW model: all of these models have many aspects that are a lot like QFT ... much of the math in statistical mechanics resembles QFT, and so there is now mounting hope that these methods can be extended and pushed much farther. Modern work thus focuses on Quantum groups, Kac-Moody algebras, Virasoro algebras, Loop algebras. A lot of the excitement is coming from the fact that rapid strides are being made, with many disparate branches of mathematics yielding surprising, astounding, fore-head slapping, who-ever-would-have-thought-that-this-is-related-to-that insights.
Here's a real simple example: did you know that SL(2,C) is the Lorentz group, is the Mobius group? Did you know that the subroup SL(2,Z) aka the modular group is central to number theory and modular forms? Why are the symmetries of space-time (the lorentz group) so closely related to number theory? Its astounding and forehad-slapping! Some of this wildness, truth-is-stranger-than-fiction stuff got named monstrous moonshine because of its wildness. linas 03:47, 31 December 2005 (UTC)
Yeah, I don't know much of anything about these guys, but doing it 'correctly' does seem like the likely way to make real progress. Thanks for mentioniong this stuff- now I know some good nooks to look into. Also, understanding the symmetry stuff seems to me to be a great endeavor. --Intangir 05:35, 31 December 2005 (UTC)
To bring up a point of contention between us, this kind of issue is why I find Stochastic Electrodynamics fascinating. By taking both fields and particles as fundamental, it bypasses the mathematical inconsistancy which confounds QED/QFT. Particles are used to describe what they describe best- the old 'grains of sand' notion. While a field is added to model the quantum behavior. The advantage derived is an even more coherent handling of the old self-energy problem, one of the most important advantages that QED had over the old wave-based approaches. Of course, I find it unlikely that both particles and fields are truely seperate and fundamental since each seems to be able to mimic the other. However, SED shows promise in someday making the kind of predictive progress that QED derived theory cannot seem to do till the old "dippy" problem is resolved. --Intangir 12:55, 30 December 2005 (UTC)
My allergy to SED is multi-fold. First, it makes such bold claims that it don't sound believable. When I look into it, I seem to find only some fairly simple-minded math, which seems to be getting misused, and doesn't seem to be terribly powerful or insightful. I don't walk away slapping my forehead "but of course" with some grand new realization or insight. It seems to lack any sort of deep connections to anything that is well-known or well-understood; it doesn't touch base. What I've read just comes off as being shallow, unconvincing, not particularly bright, and smells "not even wrong" at times. The fact that some of it is associated with infamous cranks certainly doesn't help. My advice: before you get hooked on SED, crawl through some of the links I've posted above. linas 04:09, 31 December 2005 (UTC)
SED may very well be another predictive dead-end. However, it doesn't seem to postulate any a priori outrageousness. Its rather simplistic formulation makes it another easy study for non-physicists like myself. Part of the reason why I like virtual particles so much is that they are the only formulation I have been able to 'get a handle on' so far. It is icing on the cake that they end up working so well. The crazy SED claims about gravity and mass are just that- crazy. Yet another symptom of GUT-lust. The formulation itself seems to make good, mathematical model-style sense; what the researchers do with it is another story. --Intangir 05:35, 31 December 2005 (UTC)
BTW, I liked your antigravity of anti-matter article. I'd show more support of it, except that recently I've rubbed User:CH the wrong way, and am trying to keep his stress levels down by mostly avoiding things he gets tangled in. But I did like your article. linas 04:09, 31 December 2005 (UTC)
Thanks! I also admire your recent work on this article and for the Casimir Effect. You've gone above and beyond merely supporting my original dispute and have done much to improve these pages. --Intangir 05:35, 31 December 2005 (UTC)

[edit] Fermion virtual particles

I don't think this statement:

However, in order to preserve quantum numbers, most simple diagrams involving fermion exchange are prohibited.

is really appropriate. You can, of course, have a simple diagram involving fermions: electron-photon scattering (Compton scattering). In a sense, this is a good diagram to mention virtual particles with - you can think of it as an electron absorbing a photon, and going off-mass shell (becoming virtual). That virtual particle is then, in a sense, unstable (with the timescale determined by Heisenberg) and reradiates a photon, pushing the electron back on mass shell. The confusion with "what is real, and what is virtual" is obvious here, because as the energy transfered becomes very, very small, the timescale becomes infinitely long, which means any electron could be virtual, so long as it's sufficiently close to mass-shell for the purposes of the discussion.

I think the key word here may be exchange. e\gamma \to e^* \to e\gamma isn't the exchange of a fermion, although it does proceed by way of a virtual fermion. Does that make sense? -- SCZenz 19:47, 25 January 2006 (UTC)
True, true. But exchange diagrams aren't the best way to explain virtual particles, in my mind, because virtual particles are just off-shell particles. Exchange diagrams have the "creation" of a virtual particle, which somewhat implies the particles could be "different". They're not, and I think that's made explicitly clear by a Compton scattering diagram, in which a normal particle "becomes" virtual, and then real again.
I'm not sure I understand your point. In electron-positron scattering, the diagram with e^+e^- \to \gamma \to e^+e^- is fundamentally different from the one where they exchange a photon. In the former, the virtual photon is timelike. In the latter, the photon is spacelike; thus where it started and ended depends on your reference frame, so we just draw the thing horizontally (if time goes up). But the point is, I think it's reasonable to call the latter an exchange diagram, and I believe that's the terminology. Does that help, or can you clarify the issue? -- SCZenz 21:04, 30 January 2006 (UTC)

[edit] Oneloop diagram

Off-topic, but we really need a one-loop diagram with two legs, not three. Know where any of these are lying round? linas 01:11, 31 January 2006 (UTC)

I could make one. What exactly do you want? -- SCZenz 01:18, 31 January 2006 (UTC)

[edit] not so good

I've reverted recent additions by Enormousdude. The additions do not seem to be correct. -lethe talk + 02:38, 21 March 2006 (UTC)

[edit] todo

The article should be expanded to cover the topics that Enormusdude is trying to get at, with his recent edits:

  • Spontaneous emission is an interaction between the "virtual particles" of the electromagnetic vacuum and the excited state of an atom. (I just fixed that article, but a discussion should be made here as well).
  • The static electric field is an infinite ocean of "virtual photons" in the super-duper-deep infrared. (I cut the discussion from that article, it should be more precisely defined here).

These two, together with Casimir energy and pair-production (spontaneous decay of the vacuum) are the complete list of QFT/QED effects which require integration over the full phase space of the fields. Am I forgetting any other effects?

These four effects should be summarized by one or two sentences in the introduction. linas 15:27, 28 April 2006 (UTC)

Wait, what? Complete list? If you're looking for a complete list of field-theoretic physical phenomena modelled using virtual particles, you're missing hundreds of stuffs. Every interaction in nature comes exchange of virtual particles, so every interaction has virtual particles at the heart of its description. On the other hand, it's not clear to me what the Casimir effect or what a static field has to do with virtual particles. But I think maybe the Lamb shift should be mentioned along side the vacuum polarization. -lethe talk + 19:51, 28 April 2006 (UTC)
Thanks, yes. I dunno. In the back of my mind, I was thinking of "the usual field-theory type calculations", of which there are hundreds and I didn't want to try to describe, and the "unusual" effects, which would get reviewed. But maybe there's a big grey area in the middle, as you point out with Lamb shift. As to Casimir effect, read the article. Its essentially a one-loop effect, its just that its "unusual" in that instead of being an integral over momentum with propagators in the integrand, its a sum. As to "const electric field", you will occasionally find QFT textbooks that state that a const electric field can be written as a coherent state of zero energy photons, or some such (take a Glauber state for a laser, and run the laser frequency down to zero). A synonym is infrared catastrophe, referring to the infinite number of photons that arise. Clearly Enormusdude has heard of the idea; he tried to add it to the article on the electric field; I removed it just today mostly because it was confusingly presented, and is more appropriately detailed in this article instead of that. I see what he was trying to get at, and was hoping to just say it more elegantly and accurately. linas 03:10, 29 April 2006 (UTC)
Hmm, I think I see your point. Ah, we could understand it this way: every interaction in field theory requires virtual particles for its full nonperturbative description, but most can be understood already at the tree level, which is usually similar to the classical description (like how the tree level e-μ scattering coincides with classical Rutherford), and therefore is not a purely virtual particle description. Is that what you have in mind? I could get on board the list, in that case. There could just be a single item as the last entry of the list which said something like "additionally, every physical interaction requires the use of virtual particles for its full description, though simplified pictures without virtual particles are possible" or something like that.
Edit':uh, I'm not sure what I was thinking above. Even tree level diagrams have virtual particles. Maybe the tree level virtual particles are "less weird" in some sense, since there's only a single particle exchange, rather than a superposition of virtual states at each momentum? -lethe talk + 03:34, 29 April 2006 (UTC)
As for static fields and Casimir. Well, I'm familiar with coherent states. They're eigenvalues of the annihilation operator, and usually look like ea†|0>. Because of the exponential, you can view this as a superposition of 0, 1, 2, 3 ... particle states. It's not exactly right to say it has an infinite number of particles, better to say it's just not an eigenstate of the number operator N=a†a. Anyway, those states are purely kinematic, they have nothing to do with interaction, and you can construct them even for free fields or the harmonic oscillator, so there isn't (or at least need not be) any perturbation involved. In short: I fail to see what virtual particles have to do with it. I am dubious about virtual particles being at all related. As for the Casimir effect, I was under the impression that this was a vacuum energy effect. But maybe vacuum energy is understood as a 1-loop effect? I admit I'm not crystal clear on that, so maybe you're right, it is a purely virtual particle effect. And the Lamb shift, I guess that's just a manifestation of vacuum polarization, but it's an important one, so it ought to get mentioned. -lethe talk + 03:29, 29 April 2006 (UTC)

Yes, Lamb shift should be added. You're right about the coherent states, they're not really "virtual", they're rather unusual real "particles", and illustrate the treachery of the concept of a "particle" in QFT. The vacuum contains all loop diagrams of arbitrary order. The vacuum, in perturbation theory, is a strange beast. I like to think of the vacuum as being sort-of like the partition function in dynamical systems: dull, flat, empty, boring, until you realize that it holds everything and everything else can be teased out of it with the appropriate derivative, coupling, limit, etc.

Re your earlier comment: my take on what this article should contain is this: if some pop lit book or technical article ascribes some effect to "virtual particles", using possibly loose, non-standard or layman's language, we should acknowledge that kind of a description here, in some way. So I was fishing for a "complete list" of the types things a Popular Science /Scientific American /New Scientist etc article might ascribe to "virtual particles". linas 05:14, 30 April 2006 (UTC)

I've removed static fields from the list. Looking at the list, I see spontaneous emmission, Casimir, pair production, van der Waals, and Hawking radiation. It seems to me that all of these phenomena are manifestations of vacuum polarization. I wonder therefore if the distinction you want to draw, between normal interactions which involve virtual particles and SciAm virtual particles, may just be the distinction between disconnected diagrams (virtual particles in a vacuum) and connected diagrams (virtual particles involved in an interaction). -lethe talk + 17:16, 30 April 2006 (UTC)
Wow, reading this article more closely, it seems to me that its intro is not very good. Seems to me that since the mass shell condition is a perfectly precise concept for a particle to satisfy, the definition of virtual particles is fairly precise. I guess the author was trying to get at the following fact: no interacting particle can be exactly on its mass shell (since it is just an internal leg between interactions). Therefore, no particle is exactly a real particle, and the distinction between real particles is arbitrary and not precisely defined. This does not change the fact the definition is precise, however, and I wonder how many times the intro has to say "loose" or "vague". -lethe talk + 17:23, 30 April 2006 (UTC)

[edit] new intro

I've written a new intro for the article. Comments welcome. I think some of the subsequent paragraphs need considerable help as well. -lethe talk + 18:21, 30 April 2006 (UTC)

I've reverted Enormousdude's revisions to the new intro. Some of my complaints are:
  1. "for which product of mathematically entangled quantities is less than required by the uncertainty principle". The phrase "mathematically entangled" is bad, the uncertainty principle says nothing about products of quantities, and violating the HUP is not a good way to think.
  2. "strictly speaking can not be required to comply with known laws - say, conservation laws". Since Feynman, no one claims that virtual particles violate conservation laws. This is an old-fashioned view that is still found in pop science books, but pop science books are no place to learn particle physics.
  3. this bit about "origin of all known forces" leaves a bad taste in my mouth. Virtual particles don't explain the origin of anything. They are nothing more than a calculational tool for understanding terms in perturbation theory. -lethe talk + 00:37, 2 May 2006 (UTC)

[edit] Proposed addition

I would really like to add something like:

"Real" particles represent so called energy eigenstates of the quantum fields, which means time-independent, unchanging situations. Anything that cannot be described as a superposition of such unchanging energy eigenstates must be described in terms of virtual particles (if one insists on imposing a particle description at all). Therefore, in a sense, whenever "anything interesting" happens – anything interacts with anything else – it by necessity involves virtual particles.

but I would like to hear comments from other physicists first. - Mglg 03:21, 16 June 2006 (UTC)

I think this is rather misleading. A virtual particle isn't a state vector, it's a matrix element. Virtual particles are made up of particle states (which are eigenstates of the number operator/energy operator of the free field) just as much as real particles are. The difference is the kinematics of the matrix element (real particles have momenta fixed by observation and the Euler-Lagrange equation, virtual particles are integrated over all momenta, even off shell). -lethe talk + 03:34, 16 June 2006 (UTC)

[edit] Virtual Particles Gravity Theory

A man wrote me this: Gravity is not a pulling force, it is a pushing force, it is the virtual particles surrounding a planet, pushing down the objects. Matter is mostly space but some objects have farther between the smallest particles, and therefore they are lighter - there is less resistance, less particles to be hit by the virtual particles. Say a piece of lead for example, is very heavy simply because it is dense!. So what you think of that? /Minoya 12:17, 14 December 2006 (UTC)

I think your friend was pulling that out of his ass. Ed Sanville 22:46, 14 December 2006 (UTC)

Retrieved from "http://en.wikipedia.org../../../v/i/r/Talk%7EVirtual_particle_8a35.html"