Coupling value in Madgraph5

Dear Liron,

> I try to add a vertex to 2HDM model (WZH+).
> I got a working model in MG4 but the model syntax was pretty different.
> In MG4, I have:
> gwwh3=ee*TH3*wm/sw*dcmplx(+1,Zero)
> gzwhc=ee*wm/(sw)*dcmplx(+1,Zero)
> I assume that in MG5, if I take the GC_171 (particles = [ P.w__minus__, P.w__plus__, P.h3 ],), I can use the same value just dropping the TH3.
> Looking in couplings.py, I see that the value = (ee**2*complex(0,1)*TH1x3*v)/(2.*sw**2), does it mean for my vertex I should have: value = ‘(ee**2*complex(0,1)*v)/(2.*sw**2)', ?

I’m not an expert of the 2hdm model, so my only comment is that this would be the equivalent of what you did in MG4 model.
Note that MG5 supports MG4 model (but not for all mode of MG5), and therefore you should be able to compare both model.

> One more stupid thing, can I set the BR in param_card.dat which is in the Cards directory or do I have to do it in the model directory like was done in MG4?

Actually MG never use the information about the branching fraction.
The current way to handle this is to use madwith: http://arxiv.org/abs/arXiv:1402.1178
to compute the width in a fully automatic way. This tools also write the partial-width such that they are available for pythia or any other code which needs them.

Cheers,

Olivier

On Mar 26, 2014, at 2:36 PM, Liron Barak <email address hidden> wrote:

> New question #246085 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/246085
>
> Dear All,
>
> I try to add a vertex to 2HDM model (WZH+).
> I got a working model in MG4 but the model syntax was pretty different.
> In MG4, I have:
> gwwh3=ee*TH3*wm/sw*dcmplx(+1,Zero)
> gzwhc=ee*wm/(sw)*dcmplx(+1,Zero)
> I assume that in MG5, if I take the GC_171 (particles = [ P.w__minus__, P.w__plus__, P.h3 ],), I can use the same value just dropping the TH3.
> Looking in couplings.py, I see that the value = (ee**2*complex(0,1)*TH1x3*v)/(2.*sw**2), does it mean for my vertex I should have: value = '(ee**2*complex(0,1)*v)/(2.*sw**2)', ?
>
> One more stupid thing, can I set the BR in param_card.dat which is in the Cards directory or do I have to do it in the model directory like was done in MG4?
>
> Many thanks in advance
> Liron
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Dear Olivier,

Thanks for the prompt answer!
I've tried to use the the model (from MG4), by doing: import model_v4 new.
It worked and I got;
Running...
done
But when I tried to create the events (./generate_events), I got compilation problems:
MadGraph5Error : A compilation Error occurs when trying to compile /mnt/Lustre/agrp/liron/MadGraph/MadGraph5_v1_5_12/HplusSWZ/Source.
.
.
.
            collect2: ld returned 1 exit status
            make: *** [../bin/internal/combine_events] Error 1

Is there anything that need to be done to use the model from version4 except from import model_v4 instead of import model?

Many thanks
Liron

Hi,

Usually you have to change the compiler by hand since most of the v4 model
use g77 while MG5 usually use gfortran (but if not installed on the machine)

I also found a bug for some model that I have fixed in 2.1.1 (not yet officially release).
You can try with that version if this is still not working.
to download that version you need to do:
bzr branch lp:~maddevelopers/mg5amcnlo/2.1.1
in a shell

Cheers,

Olivier

On Mar 26, 2014, at 4:01 PM, Liron Barak <email address hidden> wrote:

> Question #246085 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/246085
>
> Liron Barak posted a new comment:
> Dear Olivier,
>
> Thanks for the prompt answer!
> I've tried to use the the model (from MG4), by doing: import model_v4 new.
> It worked and I got;
> Running...
> done
> But when I tried to create the events (./generate_events), I got compilation problems:
> MadGraph5Error : A compilation Error occurs when trying to compile /mnt/Lustre/agrp/liron/MadGraph/MadGraph5_v1_5_12/HplusSWZ/Source.
> .
> .
> .
> collect2: ld returned 1 exit status
> make: *** [../bin/internal/combine_events] Error 1
>
> Is there anything that need to be done to use the model from version4
> except from import model_v4 instead of import model?
>
> Many thanks
> Liron
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Hi Olivier,

I've actually managed to switch to MG5 and introduced the changes to the model, thanks!
I do have now another question.
I've added the vertex (and its value)to the 2HDM model.
I then tried few options for the process (VBF (WZ) to produce charged Higgs that decays to WZ):
generate p p > h- j j, (h- > W- Z) - took me 7 hours for 10 events
generate p p > h- j j $$ w+ w- z / a QCD=99 QED=4 HIG=0, h- > W- Z - took me 4 hours for 10 events

Unfortunately, in neither of them the charged Higgs wasn't produce in VBF but from quarks.... the LHE file has all the particles (H+->WZ + jj) BUT in the diagrams you see it is not the right production mechanism.

What am I doing wrong? What should I do?

Thanks a lot
Liron

Dear Liron,

Did you define both
w- h+ z vertex?
and
w+ h- z vertex?

Cheers,

Olivier

On Mar 27, 2014, at 11:46 AM, Liron Barak <email address hidden> wrote:

> Question #246085 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/246085
>
> Liron Barak posted a new comment:
> Hi Olivier,
>
> I've actually managed to switch to MG5 and introduced the changes to the model, thanks!
> I do have now another question.
> I've added the vertex (and its value)to the 2HDM model.
> I then tried few options for the process (VBF (WZ) to produce charged Higgs that decays to WZ):
> generate p p > h- j j, (h- > W- Z) - took me 7 hours for 10 events
> generate p p > h- j j $$ w+ w- z / a QCD=99 QED=4 HIG=0, h- > W- Z - took me 4 hours for 10 events
>
>
> Unfortunately, in neither of them the charged Higgs wasn't produce in VBF but from quarks.... the LHE file has all the particles (H+->WZ + jj) BUT in the diagrams you see it is not the right production mechanism.
>
> What am I doing wrong? What should I do?
>
> Thanks a lot
> Liron
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Dear Olivier,

Yes, I added:
V_227 = Vertex(name = 'V_227',
               particles = [ P.w__plus__, P.z, P.h__minus__ ],
               color = [ '1' ],
               lorentz = [ L.VVS1 ],
               couplings = {(0,0):C.GC_318})

V_228 = Vertex(name = 'V_228',
               particles = [ P.w__minus__, P.z, P.h__plus__ ],
               color = [ '1' ],
               lorentz = [ L.VVS1 ],
               couplings = {(0,0):C.GC_318})

And defined:
GC_318 = Coupling(name = 'GC_318',
                  value = '(ee**2*complex(0,1)*v)/(2.*sw**2)',
                  order = {'QED':1})

Is that enough?

Thanks
Liron

Dear Liron,

yes this sounds good.
If you look at the diagram drawn did you see the associate diagram?
If yes, this probably means that the contribution is subdominant.
To have an idea of the importance of that diagram, you can look at the following topic to see to which channel of integration it is linked. In the HTML output, if you click on the cross-section you have the contribution for each diagram which should help:
https://answers.launchpad.net/mg5amcnlo/+question/187387

Also I realize that your claim are based on the LHE file, but boson on T channel are never written inside that file. (since they are not onshell, the shower doesn't need them and it is actually better to not include them).

Cheers,

Olivier

Dear Olivier,

No, I only see production with quarks, I don't see diagrams with production of WZH+ vertex.
Regarding the LHE, I wasn't clear, the LHE is just fine.... I have there:
        1 -1 0 0 501 0 0.00000000000E+00 0.00000000000E+00 0.30784075595E+03 0.30784075595E+03 0.00000000000E+00 0. -1.
        3 -1 0 0 502 0 0.00000000000E+00 0.00000000000E+00 -0.19225302491E+03 0.19225305358E+03 0.10499999672E+00 0. -1.
      -37 2 1 2 0 0 -0.17217614472E+01 0.47055757847E+02 0.10275516090E+03 0.32027128791E+03 0.29966292056E+03 0. 0.
      -24 1 3 3 0 0 -0.13021529122E+02 -0.98662361381E+02 0.19933534574E+02 0.12912721907E+03 0.79829011825E+02 0. 0.
       23 1 3 3 0 0 0.11299767675E+02 0.14571811923E+03 0.82821626331E+02 0.19114406884E+03 0.91187599182E+02 0. 0.
        2 1 1 2 501 0 0.29015348380E+02 -0.16830036217E+02 0.88085754130E+02 0.94256249879E+02 0.00000000000E+00 0. -1.
        3 1 1 2 502 0 -0.27293586933E+02 -0.30225721630E+02 -0.75253183990E+02 0.85566271751E+02 0.10499999672E+00 0. -1.

That exactly what I need in VBF too (I know I wouldn't see the WZ of the production) but when I looked on the diagrams, I saw there are all without WZH+ production.

Thanks a lot for all the help
Liron

Hi Liron,

So for
generate p p > h- j j, (h- > W- Z)
This is normal since by default madgraph takes only the leading QCD contribution.
if you want only the EW one, you have to specify QCD=0

generate p p > h- j j $$ w+ w- z / a QCD=99 QED=4 HIG=0, h- > W- Z - took me 4 hours for 10 events

For this one, I would have expect the diagram to be present since you specify the various order.
Are you sure that they are not present in that case?

Cheers,

Olivier

Hi Olivier,

The one in he main page is ud>h+ and when I checked many of the diagrams in the page (hundreds) and in all the H+ couples to quarks.
Can I force the WZ>H+ production?

Thanks
Liron

Could you send me the model to my professional email (<email address hidden>)?

Such that I can try to see what happens.

Cheers,

Olivier