MadGraph5_aMC@NLO

generating interference term for gg>ZZ process

Asked by Zhuoni

Hi Olivier and MadGraph experts,

I would like to generate at LO (one-loop) level for the gg>ZZ process : 1) higgs-mediated with triangle diagrams (signal), 2) the continuum background with the box diagrams, and 3) the interference between them. I am able to do 1) and 2), or 1)+2)+3) the total contribution with the model "loop_sm" with no problem.

Now since I want to study and generate 3), the interference contribution alone. I modify the "loop_sm" model by adding a "HYUK" coupling_order in "coupling_orders.py" file, and change the hbb, htt coupling from "QED" to "HYUK" in "couplings.py" file. Then I think I am able to generate the interference term alone by defining the process as:
gg > ZZ [QCD] QED^2=3 HYUK^2=1
Indeed it gives me the correct 28 diagrams.
But the distribution came out is wrong. This should always be a destructive interference with c.o.m above the Higgs mass (for example see JHEP07(2016)087) . The generated events indeed have a negative total cross section. However the events have both negative and positive weights. Especially if I check its mzz distribution, it is positive and seem to peak at the ZZ threshold and then drop quickly to pass zero around mzz=280 GeV, and then it stays negative and peak negatively around mzz=420 GeV.

I don't know how it could went wrong with my treatment, that making this LO one-loop calculation not compatible with standard results.
Do you know if such simple modification of the "coupling_orders.py" and "couplings.py" in the UFO file could cause any unwanted problem? Also I don't think Goldstone coupling could cause any issue here either.
Or, could you give advice on how to generate the interference term correctly?

P.S. I also tried with gg>e-e+mu-mu+ [QCD] QED^2=7 HYUK^2=1 which gives me the same wrong distribution.
and I have tried with MG_v3_5_7, MG_v2_9_18, MG_v2_4_3

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Olivier Mattelaer
Solved:
Last query:
Last reply:
Revision history for this message
Best Olivier Mattelaer (olivier-mattelaer) said : 		#1

Hi,

The most likely answer is that you did not change the model correctly,
one tricky part to change is to correctly change all the R2 counter-term that enter the loop.

What you need is to manually track all your yukawa term and correctly assign the new coupling order according such coupling in all the R2 counter-term. It is quite tedious to do. Maybe the easiest is actually to regenerate the SM model from Feynrules/NLOCT with such type of coupling definition such that all R2 counter-term are correctly flagged.

Cheers,

Olivier

Revision history for this message
Zhuoni (sdalyqian) said : 		#3

Hi Olivier,
Thanks for your answer. That solved my problem.

I accordingly changed all the coupling_order to HYUK for the yQ proportional terms in the "CT_couplings.py" file. This includes all UV and R2 terms for QQH, Goldstone couplings G0G0Q and massive quark corrections to ggH, ggHH etc. The last ones probably contribute the main misses in my treatment before.

So I change the orders of all such yQ proportional terms from QED to HYUK, now the generated event distribution matches expectation.

Thank you again for your help.