Cannot reproduce heft model

Asked by Duarte Azevedo on 2017-10-24

I'm having difficulties in the following. I'm trying to create my own model using FeynRules. I'm studying a process that considers gluon fusion for Higgs production, as such I'm adding up the heft model available in http://feynrules.irmp.ucl.ac.be/wiki/HiggsEffectiveTheory to the SM .fr file (with LoadModel("SM.fr","HiggsEffective.fr","mymodel.fr").

To check if things are ok, im trying to generate 10000 events of p p > h , h > b b~, but just considering the LoadModel("SM.fr","HiggsEffective.fr") , so my model is not included, and I'm getting these results:

p p > h -----------|||---------- xs = ~17pb (13TeV)
p p > h , h > b b~ -----------|||--------- xs = 22.9 pb (13TeV)

Whereas if I use the heft model already implemented in MG5 I get:

p p > h -----------|||---------- xs = ~17pb (13TeV)
p p > h , h > b b~ -----------|||--------- xs = ~11 pb (13TeV)

The UFO file has the same couplings and lorentz structure in both. I'm really not getting the problem.

This is what I do to write the UFO files:

<mathematica>
$FeynRulesPath = SetDirectory[ "/home/bla"];
<< FeynRules`
SetDirectory[$FeynRulesPath <> "/Models/mymodel"]
LoadModel["SM.fr", "HiggsEffective.fr"]

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Olivier Mattelaer
Solved:
2017-10-30
Last query:
2017-10-30
Last reply:
2017-10-30

What is the total width of the Higgs?

Cheers,

Olivier
> On 24 Oct 2017, at 16:08, Duarte Azevedo <email address hidden> wrote:
>
> New question #659880 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/659880
>
> I'm having difficulties in the following. I'm trying to create my own model using FeynRules. I'm studying a process that considers gluon fusion for Higgs production, as such I'm adding up the heft model available in http://feynrules.irmp.ucl.ac.be/wiki/HiggsEffectiveTheory to the SM .fr file (with LoadModel("SM.fr","HiggsEffective.fr","mymodel.fr").
>
>
>
> To check if things are ok, im trying to generate 10000 events of p p > h , h > b b~, but just considering the LoadModel("SM.fr","HiggsEffective.fr") , so my model is not included, and I'm getting these results:
>
> p p > h -----------|||---------- xs = ~17pb (13TeV)
> p p > h , h > b b~ -----------|||--------- xs = 22.9 pb (13TeV)
>
> Whereas if I use the heft model already implemented in MG5 I get:
>
> p p > h -----------|||---------- xs = ~17pb (13TeV)
> p p > h , h > b b~ -----------|||--------- xs = ~11 pb (13TeV)
>
> The UFO file has the same couplings and lorentz structure in both. I'm really not getting the problem.
>
>
>
> This is what I do to write the UFO files:
>
> <mathematica>
> $FeynRulesPath = SetDirectory[ "/home/bla"];
> << FeynRules`
> SetDirectory[$FeynRulesPath <> "/Models/mymodel"]
> LoadModel["SM.fr", "HiggsEffective.fr"]

Duarte Azevedo (yuhuanknot) said : #2

After using

compute_width h

I get the following in the param_card.dat of my model:
# PDG Width
DECAY 25 6.758431e-03

Cheers,
Duarte

Hi,

My point was, is the Higss width used in both computation identical?
The fact that you have production+decay cross-section bigger than the production only one is typically a sign that your width is not set correctly.

Cheers,

Olivier

Duarte Azevedo (yuhuanknot) said : #4

Hello,

Regarding the widths:

heft: 6.758 e-3
my implementation of heft: 6.7328e-3 (diagonal CKM and light quarks are massless)

Weird thing, first time I generate p p > h , h > b b~ it gives me 22.9pb, second time it gives me ~15pb. I'm using MG5_aMC 2.2.3 . was there a bug of some sort with this version, that explains this behaviour?

Best,
Duarte

Duarte Azevedo (yuhuanknot) said : #5

Correction:

It actually just gives me a fair value of the cross section when I execute: compute_width h and then generate my events. I though the whole thing was automatic. I must admit that i found nothing in the manual about the procedure after importing a model in MG5. Maybe I'm missing something?

Best,
Duarte

Hi Duarte,

2.2.3 is so old that I do not remember of the fixes that have been done associated to that module.
If you want to have a look, you can check the UpdatesNotes to see how many stuff have changed and if anything might be related to that. I would not do that myself since I do not see the point for providing any support for such old version.
(If you are stuck with such old version, I guess that you have a valid reason for that and therefore you have to live with the associated bug)

Could you be more specific on what you mean by "Weird thing, first time I generate p p > h , h > b b~ it gives me 22.9pb, second time it gives me ~15pb."
You mean that the same computation run twice returns different result? If that's the case, this is likely to be resolved in a more recent version of the code.

Cheers,

Olivier

Duarte Azevedo (yuhuanknot) said : #7

Hello Olivier,

Yeah I'm stuck to that version due to many analysis codes being written for Pythia6. Anyway, I've corrected my previous comment by stating that only after I run: "compute_width h" that i get a cross section that makes more sense. I was wondering if it is necessary to do so every time I implement a new model and if there is something else that I should do that I'm not mentioning.

Best,
Duarte

Hi,

Your cross-section (with the decay) is proportional to the inverse of the width.
If your model that you generate does not have the correct value for that width then the result will be wrong.
One method to fix that is indeed to run that command which returns the value for the wdith.

Note that in general that command returns a wrong value for the Higgs due to the loop-induced decay in that case.
However in the case of heft those decay are included and therefore that automatic computation of the width for the Higgs make sense for that model.

Cheers,

Olivier

Duarte Azevedo (yuhuanknot) said : #9

Thanks Olivier Mattelaer, that solved my question.