In (perhaps even in older versions) version FeynCalc 9.0.0 sometimes, on the amplitude level, one gets expressions of the form

$\epsilon^{\alpha q\varepsilon^*(p)\varepsilon^*(q)}$ where $\alpha$ is a Lorentz index, $p,q$ are four-momenta and the $\varepsilon^*(p)$ are polarization vectors of out/incoming vector particles (some massive some massless).

I'm sure there's some logic behind this notation but my question is more why does the LeviCivita remain in some terms even after I issue Calc[] and/or Contract[]? That really does not make sense to me:

In the squared amplitude (after summing over polarizations with DoPolarizationSums[] and after FermionSpinSum[] has been issued) I can get terms like

$$\tag{1}a + b\epsilon^{kpql}$$ where $k,p$ are incoming and $q,l$ are outgoing four-momenta.

How can I make sense of eq $(1)$ or get rid of the LeviCivita?

Also I should perhaps add that there's a bar over all fourmomenta in FeynCalc 9.0.0.


1 Answer 1


OK I think I got it. The final LeviCivita[] that remains has (of course) only Momentum[] as indices. In my case (as in most other cases) four-momentum conservation implies $$\tag{2} k+p = l+q $$ and I've got something like $$\tag{3}\epsilon^{klpq}.$$ Now substituting e.g. $k = l+q-p$ and EpsEvaluate[] the LeviCivita I get zero which is what I was looking for.

In other words, unless I define Eq. $(2)$ somewhere and (perhaps?) FeynCalc recognizes this and automatically sets Eq. $(3)$ to zero, one has to set it to zero by hand.

  • 1
    $\begingroup$ Thanks for the answer, in my case, the expression involves some auxiliary vectors, therefore it is not zero. How can I tell Mathematica to make the expression explicit? $\endgroup$
    – zyy
    Dec 9, 2021 at 8:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.