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

1PI vertex conventions

  1. Nov 19, 2014 #1
    I assume that if your theory has an interaction term written as LI= ±λ*fields, where λ is the coupling, that the 1PI vertex is defined as:

    $$i(\pm \Gamma)=i(\pm \lambda)+loops$$

    That way to tree level, Γ is λ, and not -λ. The exception seems to be the self-energy. A mass term has -m2*field2, so this suggests that you should write -i Π whenever you insert a self energy. However, it seems textbooks write +iΠ for self-energy insertions. This leads to propagators that look like:


    where Π has the opposite sign of m2.

    Is it conventional to define the self-energy with +i instead of -i, so that in the propagator it has the opposite sign of the mass? Why is this so?

    Also, is the self-energy even a 1PI vertex? Don't you have to include the tree-level term? So really:

    $$i\Gamma^{(2)}=-i(p^2-m^2+\Pi(p^2) )$$

    Or should it be defined:

    $$-i\Gamma^{(2)}=-i(p^2-m^2+\Pi(p^2) )$$

    It's not clear how to logically define the sign of Γ(2). Should it be the same sign as m2, or Π(p2)?
    Last edited: Nov 20, 2014
  2. jcsd
  3. Nov 20, 2014 #2


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    No, usually you identify [itex]-\mathrm{i} \Pi[/itex] for a truncated 1PI self-energy diagram, then you get
    [itex]\mathrm{i} G(p)=\frac{\mathrm{i}}{p^2-m^2-\Pi}.[/itex]
    The real part of the self-energy is then a correction to [itex]m^2[/itex] as it should be. Note that I use the west-coast convenction of the metric, [itex]\eta_{\mu \nu}=\mathrm{diag}(1,-1,-1,-1)[/itex].

    It's different for massive vector bosons. There you have
    [tex]\mathcal{L}_{\text{free}}=-\frac{1}{4} F_{\mu \nu} F^{\mu \nu} +\frac{1}{2} m^2 A_{\mu} A^{\mu}.[/tex]
    This change in sign of the mass term is due to the west-coast convention of the metric, which is "mostly negative". Then the self-energy (or polarization) tensor as represented by a corresponding truncated 1PI Feynman diagram is [itex]+\mathrm{i} \Pi^{\mu \nu}[/itex], so that you get again a contribution to the mass with the correct sign.

    Of course, there are as many conventions as textbooks, and it's a pain in the a... to sort these signs out when comparing a calculation in one convention with one done in another :-(.
  4. Nov 20, 2014 #3
    In mark sredinicki's textbook, he has the self energy opposite the sign of the mass in the propagator. He uses the east coast metric however. But I don't see why using the east coast metric (-+++) should lead you to want to define the self energy as having opposite sign to the mass.

    I checked cheng and li's book, and they use West coast metric, and they have self energy same sign as mass, but they don't define interaction 1pI's with a prefactor of 'i' at all, so they have $$ \Gamma^4=-i lambda+loops $$.
  5. Nov 20, 2014 #4


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    As I said, there are as many conventions as books.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook