FeynCalc TID PaVeLimitTo4 option leads to divergence

Question

I'm trying to compute the Higgs to gluon pair decay at 1-loop using Feyncalc and FeynHelper (to connect with Package-X).

I first define variables and loop part as:

FCClearScalarProducts[];
SP[k1, k1] = 0;
SP[k2, k2] = 0;
SP[q, q] = mH^2;
SP[k1, q] = mH^2/2;
SP[k2, q] = m\[Chi]^2/2;
loop = DiracTrace[(GAD[5] . (GSD[k - q] + m) . 
     GAD[\[Nu]] . (GSD[k - k1] + m) . GAD[\[Mu]] . (GSD[k] + m)), 
   DiracTraceEvaluate -> True] FAD[{k - q, m}, {k - k1, m}, {k, m}]

Then I run

TID[loop, k, ToPaVe -> True, PaVeLimitTo4 -> True] // 
  DiracSimplify // Simplify

It gives me "Infinite expression 1/0 encountered". I have to run

loopPaVe = TID[loop, k, ToPaVe -> True] // DiracSimplify // Simplify
PaXEvaluate[ChangeDimension[loopPaVe, 4] // DiracSimplify, 
 PaXImplicitPrefactor -> 1/(2 \[Pi])^D]

so that I can get the result.

The result seems to be correct, but is my code correct (I'm assuming that PaVeLimitTo4 should be the correct way)? Could that be just a coincidence that I get the correct result in this way? I feel dangerous since I'm not sure whether ChangeDimension is the right way to take the dimension back to 4 in dimensional regularization.

@vsht Thank you. It works! Just wondering is this the most common way that the designer want us to do to calculate loop integral? I mean the way that 'define loop in D-dim' -> 'TID to PaVe and take PaVeLim to 4' -> 'PaXEvaluate with implicitfactor 1/(2pi)^D'. — Isaac Wang, Mar 07 '24 at 23:15
PaVeLimitTo4 can be dangerous because it doesn't check for IR divergences. Since anything containing photons, gluons or massless fermions in the loops can have IR poles, I wouldn't use it unless you are 100% sure what you are doing. TID and PaXEvaluate after should be good enough. — vsht, Mar 09 '24 at 12:05

FeynCalc TID PaVeLimitTo4 option leads to divergence

0 Answers0