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

I Energy-momentum tensor

  1. Apr 23, 2017 #1
    As you may know from some other thread, I was interested through the week in finding a general way of express the energy-momentum tensor that appears in one side of the Einstein's equation.
    After much trials, I found that

    $$T^{\sigma \nu} = g^{\sigma \nu} \frac{\partial \mathcal{L}}{\partial \partial_{\mu} \phi} \partial_{\nu} \phi - g^{\sigma \nu} \delta^{\mu}{}_{\nu} \mathcal{L}$$ where ##\mathcal{L}## is the Lagrangian density.

    As I haven't found an expression like this one on web, I'm unsure about its validity in Relativity.

    I would appreciate if someone could tell me whether this expression is valid or not.

    I can post here links to the pdf's I have read and how I arrived at the above result.
    Last edited: Apr 23, 2017
  2. jcsd
  3. Apr 23, 2017 #2


    Staff: Mentor

    What you are doing is not general enough; you're assuming that the matter is a scalar field (that's what ##\phi## is).

    The most general way of expressing the stress-energy tensor that appears on the RHS of the Einstein Field Equation is the Hilbert stress-energy tensor, given here:


    Note that this expression just involves ##\mathcal{L}_{\text{matter}}##, i.e., the Lagrangian density for matter, without making any assumptions about the form that Lagrangian density takes.

    Even leaving out the above, your expression is not valid on its face because the indexes don't match up; you have ##\sigma## and ##\nu## as free indexes on the LHS but ##\sigma## and ##\mu## as free indexes on the RHS.

    However, as above, the more general problem with your expression is that it assumes that the matter is a scalar field, which you shouldn't do if you want the most general expression possible.

    Even for matter as a scalar field, though, your expression is not correct. You can use the general Hilbert form I linked to, plus the correct form of ##\mathcal{L}_{\text{matter}}## for a scalar field, which is not just ##\phi##, to derive what the correct stress-energy tensor is for a scalar field.
  4. Apr 23, 2017 #3
    Thanks Peter.
    I'm sorry, it was a typo. The correct would be inserting ##\mu## instead of ##\nu## on the left hand side.

    I thought we could always represent matter as a scalar quantity, represented by a scalar field. Why it is not the case?

    Unfortunately the Wiki article don't provide a derivation for it. But it's interesting to see its form. The Lagrangian density is dependent of what quantities?
  5. Apr 23, 2017 #4


    Staff: Mentor

    I have no idea why you would believe that. In any case, it's false. All ordinary matter is fermions, which are spinor fields (spin 1/2), not scalar fields (spin 0). All gauge bosons in the Standard Model are spin-1 fields. The only spin-0 field in the Standard Model is the Higgs.

    More generally, at the macroscopic level, the simplest model of matter is a perfect fluid, which is not the same as a scalar field, as is evident from looking at their respective stress-energy tensors.

    It depends on what kind of matter you are talking about. There is no general rule. That's why the Hilbert stress-energy tensor makes no assumptions about its form or what it depends on.
  6. Apr 23, 2017 #5


    Staff: Mentor

    Click on the "Einstein-Hilbert action" link.
  7. Apr 23, 2017 #6
    That's because I've seen many times phrases like "the field has a mass m"; "a massive field"; etc... (as when one talks about the Klein-Gordon Equation). So I related these things in my mind.
  8. Apr 23, 2017 #7


    Staff: Mentor

    Sure, the scalar field can have a mass. But that doesn't mean the scalar field is the only field that can have a mass, which is what your belief that all matter can be described by scalar fields implies. You will see similar statements made about fermion fields and the Dirac equation in many texts. You will also see many texts talking about massive spin-1 fields in the context of showing what kinds of experiments might tell you whether photons have mass or not.
  9. Apr 24, 2017 #8


    User Avatar
    Science Advisor

    By the way, if you want to understand the coupling of spin 1/2 to gravity, you need the vielbein-(or tetrad)formalism. See e.g. Carroll on GR or Samtleben on SUGRA.
  10. Apr 24, 2017 #9


    Staff: Mentor

    Actually, even this isn't broad enough. Massless fields can also contribute to the stress-energy tensor. Typically the term "radiation" is used for massless fields (such as photons in the early universe), but as far as GR is concerned, stress-energy is stress-energy, and all types of stress-energy appear in the Einstein Field Equation. In other words, "matter" as the term is usually used is not general enough to describe all of the possible sources in the EFE.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted