Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Covariant quantum field theories

  1. Jun 28, 2007 #1
    I am not as well read as most of the people in here so I thought I would ask you guys first. What work has been done in the way of developing a covariant field theory? I'm going to ramble for just a little bit so try to follow. It seems to me that QFT is built on two principles Poincare invariance and the Structure Decomposition Principle. A covariant field theory, in my opinion, would be built on basically the same two things except this time it would be General covariance and the Structure Decomposition Principle. The structure decomposition principle implies, in my mind, the need for creation and annihilation operators, and the covariant part would require the generalization of the concept of a field. I am referring to this generalization as a covariant field. The homogeneous Lorentz group is O(3,1) and the corresponding generally covariant homogeneous group, at least locally, is GL(R,4). I guess my question is: is GL(R,4) unitarizable? Have people come up with the concept of a covariant field? If they have, why did it not work? Is it non-renormalizable, power counting sense? The reason I ask is because I think I have a covariant field theory. It's not completely done yet.
     
  2. jcsd
  3. Jun 28, 2007 #2

    olgranpappy

    User Avatar
    Homework Helper

    Perhaps you might find the textbook "Quantum Field Theory in Curved Spacetime and Blackhole Thermodynamics" by Wald useful.
     
  4. Jun 29, 2007 #3

    Demystifier

    User Avatar
    Science Advisor

    I think that Jim Kata has something different in mind.

    By the way, I was also involved in a problem of constructing a covariant field theory:
    http://xxx.lanl.gov/abs/hep-th/0407228
    http://xxx.lanl.gov/abs/hep-th/0601027
    but it does not seem to be what he has in mind either.
     
  5. Jul 1, 2007 #4

    samalkhaiat

    User Avatar
    Science Advisor

     
  6. Jul 1, 2007 #5

    Haelfix

    User Avatar
    Science Advisor

    Im a little confused by what you mean Jim. Write down the difference between a lorentz invariant field theory and a covariant field theory explicitly. I think we have different terminology in mind.

    My main objection to not using SL(2,C) as the main isometry group of spacetime, is the fact that the irreducible unitary representations of whatever else you use (if they even exist) would either not match the known particle spectrum and properties, or imply extra unobserved transformation properties.
     
    Last edited: Jul 1, 2007
  7. Jul 2, 2007 #6
    Thanks for the help

    I posted this question on a different forum, and didn't really get a good reply, but I typed up my argument in Latex on there. So instead of re-typing the whole thing, I'll just give you the url to the question:

    http://www.scienceforums.net/forum/showthread.php?t=27168

    I have way more than what I've written in this link, and I am working on typing it in post script right now, but I suck at Latex so it's going to take me a while to finish it. But when I do, I'll let you take a look, and tell me a.) what I did wrong or b.) who did it before me.

    Another question, what tags to do you use to write LaTeX in these forums?
     
  8. Jul 2, 2007 #7

    Hans de Vries

    User Avatar
    Science Advisor

    It seems you are reading Weinberg volume I, (which is not really advisable
    as an introduction...) Weinberg uses the term "Structure Decomposition
    Principle" to state that experiments done in CERN are not correlated to
    experiments done in Stanford at the same time.

    All Quantum Field Theories involving interaction are covariant theories, and
    only free particles are Lorentz (Poincaré) invariant without covariant terms.
    I do not see why the General Linear groups should be related to covariant
    theories.


    Regards, Hans
     
    Last edited: Jul 2, 2007
  9. Jul 2, 2007 #8

    Hans de Vries

    User Avatar
    Science Advisor

    Covariance results in the most elementary sense from locality. To explain
    this a bit look at the following: Let a photon with a linear polarization in
    the x direction [itex]\epsilon = (0, A_x, 0, 0)[/itex] arrive at the wave function of an electron
    at rest. How will this modify the phase of the electron's wave function in
    the x-direction?

    Answer: It doesn't, because of locality. The increase in spatial frequency
    is nullified by the covariant interaction term:

    [tex] \partial_x \psi - ieA_x \psi \ =\ 0[/tex]

    One expects the charge density to be accelerated in the x-direction
    as a result of the changing vector potential [itex]E = -\partial A_x /\partial t[/itex], it does indeed
    accelerate but the corresponding spatial deBroglie frequency is nullified.

    If this wasn't the case then the phase would be dependent on an
    instantaneous integration over all of x, meaning that arbitrary distant
    events would be of direct influence on the phase of the wave function.



    This is similar to the well known fact that a charge in a central potential
    has an energy which is independent of the local potential. A particle
    descending in the potential well acquires a kinetic energy equal but
    opposite to the lowered potential energy.

    [tex] \partial_t \psi - ieV \psi \ =\ 0[/tex]

    In the latter case there is a phase change in space for instance
    corresponding with the angular momentum. In the first case the
    energy of the electron does change as well as the phase in the
    z-direction.



    This all means that the wave-function of the electron is not longer
    Lorentz invariant, but it does become so again when combined
    together with the interaction term. Covariant stands for: together
    being Lorentz invariant.

    Technically the derivatives have to be replaced by the covariant
    derivatives to make the interacting theory Lorentz invariant again:

    [tex] \partial_\mu \psi \ \rightarrow\ \partial_\mu \psi - ieA_\mu \psi [/tex]


    Regards, Hans
     
    Last edited: Jul 2, 2007
  10. Jul 4, 2007 #9
    response to Hans

    This is in response to Hans. What me and you are talking about are two different things, but in the same jest. What you wrote down is called the covariant derivative, but that's not the meaning of covariant to which I am referring. Maybe, I used poor terminology, but I referring to covariant in the sense that a vector transforms like [tex]\bar p^\alpha = \frac{{\partial \bar x^\alpha }}{{\partial x^\beta }}p^\beta [/tex]. The reason I'm working with [tex]Gl(4,\mathbb{R})[/tex] and not [tex]\frac{{SL(2,\mathbb{C})}}{{Z_2 }}[/tex] is because I'm considering coordinate changes of the the form
    [tex]\frac{{\partial \bar x^\alpha }}{{\partial x^\beta }}[/tex] not Lorentz transformations. And to my knowledge the group that represents [tex]\frac{{\partial \bar x^\alpha }}{{\partial x^\beta }}[/tex] is [tex]Gl(4,\mathbb{R})[/tex]. To illustrate what I mean by a covariant field theory, consider the following. All massive fields obey the the Klein Gordon equation [tex][\partial ^\mu \partial _\mu - m^2 ]\psi _l[/tex], well a covariant field theory would obey the same equation except you'd have [tex]
    [g^{\mu \tau } \partial _\mu \partial _\tau - m^2 ]\psi _l[/tex] not
    [tex][\eta ^{\mu \tau } \partial _\mu \partial _\tau - m^2 ]\psi _l[/tex]. I think my real problem is that I am not very well read on the history of the approaches to quantum gravity. I am sure this has been tried before.
     
  11. Jul 5, 2007 #10

    Haelfix

    User Avatar
    Science Advisor

    Hi Jim, theres something a little fishy about this business, but im not going to dwell on it b/c I think I might be off. It seems to me Gl(4,R) wants a linear non symmetric connection to transform properly, but again I think I have something different in mind.

    Anyway, I think what you are trying to do is to extend the poincare group to the general affine group GA(4,R) ~ GL(4,R) + translations.

    I looked up a reference and it seems Y Ne'eman has done a lot of work on this subject. You can try this:

    Ne'eman et al, GL(4,R) (bar) group topology, covariance and curved space spinors Int J Mod physics. A2 (1987) 1655

    I gather the difficulty with the whole program is getting the double cover to work out for fermions, but theres probably a reference there to the simpler boson case. Its likely old stuff tho, so you'd have to do some journal hunting. Good luck
     
    Last edited: Jul 5, 2007
  12. Jul 9, 2007 #11

    samalkhaiat

    User Avatar
    Science Advisor

     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?



Similar Discussions: Covariant quantum field theories
Loading...