Interference for 2HDM gg->A/H->ttbar

Hi,

If I understand correctly, you should generate your process like
> generate g g > t t~ QED=99, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j)

Cheers,

Olivier
On Jun 20, 2014, at 9:31 PM, Madalina Stanescu-Bellu <email address hidden> wrote:

> New question #250527 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/250527
>
> Dear MG authors,
>
> I am studying interference effects for gg->H->ttbar, and I have troubles at generation. Would you be so kind to take a look at the description below? Thanks !
>
> So this is how I generate signal+background+interference:
>
> 1)
>
> import model Higgs_Effective_Couplings_FormFact
> define l+ = e+ mu+ ta+
> define l- = e- mu- ta-
>
> generate g g > t t~, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j)
> add process g g > t t~, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > l- vl~)
> add process g g > t t~, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > j j)
> add process g g > t t~, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > l- vl~)
>
> output testAll
>
> launch
> set cut_decays F
> set MT 172.5
> set MH 750
> set MP 750
> set WH 22.5
> set WH1 22.5
> set nevents 200000
> set ebeam1 4000
> set ebeam2 4000
>
> I should theoretically see in the diagrams: gg->tt, gg->H->tt, gg->A->tt, but I only see gg->tt. If I check the unweighted.events.lhe text file , I also don't see any particles with PDG_ID 25 (the code for H in this model), or 9000006 (A). MG gives xsec=92.55pb and HIG=0, QCD=2, QED=4.
>
> If I replace the processes with :
>
> 2)
>
> generate g g > h1, ( h1 > t t~, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j))
> add process g g > h1, ( h1 > t t~, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > l- vl~))
> add process g g > h1, ( h1 > t t~, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > j j))
> add process g g > h1, ( h1 > t t~, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > l- vl~))
>
> to get only gg->A->tt signal, MG gives xsec=0.08009 pb, HIG=1, QCD=0, QED=5.
>
> So considering the xsec ratios, for 200k events at run 1) there must be ~170 signal events already.
>
> If I run :
>
> 3)
>
> generate g g > t t~, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j)
> add process g g > t t~, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > l- vl~)
> add process g g > t t~, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > j j)
> add process g g > t t~, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > l- vl~)
>
> add process g g > h1, ( h1 > t t~, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j))
> add process g g > h1, ( h1 > t t~, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > l- vl~))
> add process g g > h1, ( h1 > t t~, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > j j))
> add process g g > h1, ( h1 > t t~, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > l- vl~))
>
>
> this definitely gives me A events, but there won't be any interference simulated, as I read here: https://answers.launchpad.net/mg5amcnlo/+question/228557/ #6.
>
> I reproduced p p > e+ e- from http://indico.cern.ch/event/234296/material/slides/2?contribId=4 slide 15, upper left plot, import model sm, and there I see signal diagrams and events (for pp->Z->e+e-), so I thought I should get signal events at run 1) .
>
> Would you have any clue how to add both S+I+B? maybe play with HIG/QCD/QED ?
>
> Thank you for any hint or help !!!
> My regards,
> Madalina
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Thanks Olivier, it works !

Now I have all A and H included, so I must split them. Do you think the following would be correct ? (where B=background gg->tt, SA= signal gg->A->tt, SH=signal gg->H->tt, IA=interference for A, IH=interference for H, bwcutoff=15 everywhere)

B+SA+IA : generate g g > t t~ QED=99, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j) with MH=30000 (to disable H), MA=750 (will I really get rid of IH as well ?)

B+SH+IH : generate g g > t t~ QED=99, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j) with MA=30000 , MH=750

B: generate g g > t t~, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j) with MA=20000, MH=20000

B+IA: generate g g > t t~ $h $h1 QED=99, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j) with MA=750 , MH=20000 (not sure here)

B+IH: generate g g > t t~ $h $h1 QED=99, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j) with MA=750 , MH=20000 (not sure here)

And would there be a possibility to set m(ttbar) for B around 750GeV (like for SA/SH), to save computational time? Maybe playing with the mmjj ?

My regards,
Madalina

But B+IA and B+IH I don't really need. I have SA, SH, and once I have B+SA+IA, B+SH+IH and B, I can calculate the rest by subtracting scaled with xsec.

Dear Olivier,

Your previous suggestion was very helpful, thanks again, and I made lots of progress at my intereference studies ! Now I would need again some advice from you please (you can ignore my last 2 questions).

When I want to generate Signal+Background+Interference, for gg -> tt, I want to do it separately for scalar and pseudoscalar, mainly because they have different couplings and MadGraph takes as input only 1 set of couplings. So let's say for pseudoscalar 600GeV (in this model MH , WH are mass and width for scalar H; MP, WH1 are mass and width for pseudoscalar A ). I basically want to disable the signal gg->H->tt and its interference as well, and leave only gg->A->tt, background gg->tt and their interference. Couplings I take from a calculator http://www.feynhiggs.de/ . I tried 2 ways:

1. I set the H at 10 TeV. The output diagrams still have gg->h->tt, but no events in the output LHE files, so I wonder if interference for H is also removed.

import model Higgs_Effective_Couplings_FormFact
define l+ = e+ mu+ ta+
define l- = e- mu- ta-

generate g g > t t~ QED=99 QCD=99, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j)
add process g g > t t~ QED=99 QCD=99, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > l- vl~)
add process g g > t t~ QED=99 QCD=99, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > j j)
add process g g > t t~ QED=99 QCD=99, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > l- vl~)

output ggtt
launch -i
multi_run 100
set cut_decays F
set MT 172.5
set MH 10000
set MP 600
set WH1 6.19838
set YMTAU 3.6674
set YMT 80.7267
set YMB 8.66613
set nevents 100000
set ebeam1 4000
set ebeam2 4000

2. I remove H with / h . I let MH at default level (120GeV). gg->h>tt diagrams disapear now. But I read in a tutorial from you that / h is to be avoided, it's not gauge invariant, so is my approach correct? I also wonder if letting the default 120GeV is a mistake.

import model Higgs_Effective_Couplings_FormFact
define l+ = e+ mu+ ta+
define l- = e- mu- ta-

generate g g > t t~ / h QED=99 QCD=99, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > j j)
add process g g > t t~ / h QED=99 QCD=99, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > l- vl~)
add process g g > t t~ / h QED=99 QCD=99, ( t > b w+, w+ > j j), ( t~ > b~ w-, w- > j j)
add process g g > t t~ / h QED=99 QCD=99, ( t > b w+, w+ > l+ vl), ( t~ > b~ w-, w- > l- vl~)

output ggAlltt600-w1-A-final
launch -i
multi_run 100
set cut_decays F
set MT 172.5
set MP 600
set WH1 6.19838
set YMTAU 3.6674
set YMT 80.7267
set YMB 8.66613
set nevents 100000
set ebeam1 4000
set ebeam2 4000

My regards,
Madalina

At 1 I meant "but no gg->H->tt events in the output LHE file, only gg->A->tt and background, with a 1:10000 proportion"

Dear Madalina,

1. I set the H at 10 TeV. The output diagrams still have gg->h->tt, but no events in the output LHE files, so I wonder if interference for H is also removed.

Yes the interference is taken into account. With a such heavy H, that particle will never be on-shell and therefore will never be written in the output file.
More details on when a particle is written in the lhe file:
https://answers.launchpad.net/mg5amcnlo/+faq/2173

2. I remove H with / h . I let MH at default level (120GeV). gg->h>tt diagrams disapear now. But I read in a tutorial from you that / h is to be avoided, it's not gauge invariant, so is my approach correct? I also wonder if letting the default 120GeV is a mistake.

if you forbid it to be in the diagram, it does not matter which mass you use. In this case, the / h should be equivalent to put him at an infinite mass which mass that the interference is going to be suppressed and that you should not have problem with gauge invariance.

Cheers,

Olivier

so if I understand correctly, case 2 is the correct way ? I could do case 2 with MH=10TeV, but I guess it's redundant.

Thanks Olivier Mattelaer, that solved my question.