How to check values of Higgs partial widths in MG5

Dear Giuseppe,

It looks to me that you can generate your process as follows

import heft
define p = p b b~
define j = j b b~
generate p p > t h j, h > a a

and output the madevent code to generate events.
You need to set WH = the total width of the Higgs to the appropriate value.
To compute this value, you can type

import heft
generate h > all all
output h_width
launch -i
calculate_decay_widths

Cheers,
Pierre

Dear Pierre,

 thanks for the reply. I have followed your recipe, but I am a bit puzzled.
This is what I get:

# PDG Width
DECAY 25 4.596193e-03
# BR NDA ID1 ID2 ...
8.984828e-01 2 5 -5 # 0.0041296
5.403821e-02 2 15 -15 # 0.00024837
4.569434e-02 2 21 21 # 0.00021002
1.784651e-03 2 22 22 # 8.2026e-06

It would seem to suggest that the BR(h->bb) is ~89%.
This is quite off from the SM expectation for mH=125 GeV, e.g. 5.77E-01 (see https://twiki.cern.ch/twiki/bin/view/LHCPhysics/CERNYellowReportPageBR2#Higgs_2_fermions )

Also: this doesn't output the values for the bosons, particularly di-photon in which I am interested.

Thanks for any clarification/further help

  G

see previous comment

Hi Giuseppe, Pierre,

Sorry for this late answer but I was in Holliday. Thanks Pierre for your very good answer, but looks like I need to clarify some additional points.

Let me first remind that the following interactions: h a a , h w+ w-, h z z, h z a are NOT Leading order interactions but loop induced
interactions. Since MadEvent is a Leading Order Event generator, those interactions are not included in the standard model.

As Pierre pointed in his answer, the correct model to use in your case is the "heft" one, which correspond to the standard model,
plus the Dimension 6 operator which approximates the loop interactions.
All information on this model are given on the following link:
https://cp3.irmp.ucl.ac.be/projects/madgraph/wiki/Models/HiggsEffective

If I interpreted correctly your last answer, you were still using sm model.

>The question is: how do I get the value of the partial width for H->gam gam implemented in MG5?

In MG5, the TOTAL width is a free parameter that the user can edit freely in order to set it to the NLO width if required.
The partial width are not free parameter and are in fact NEVER used by MadGraph. The rate of each production mode
is dependent of the coupling value and of the total width.

As Pierre pointed in his previous answer, the leading order total and partial width can be computed by MG5.

Cheers,

Olivier

On Apr 17, 2013, at 2:21 PM, Giuseppe Salamanna <email address hidden> wrote:

> Question #226823 on MadGraph5 changed:
> https://answers.launchpad.net/madgraph5/+question/226823
>
> Giuseppe Salamanna posted a new comment:
> Dear Pierre,
>
> thanks for the reply. I have followed your recipe, but I am a bit puzzled.
> This is what I get:
>
> # PDG Width
> DECAY 25 4.596193e-03
> # BR NDA ID1 ID2 ...
> 8.984828e-01 2 5 -5 # 0.0041296
> 5.403821e-02 2 15 -15 # 0.00024837
> 4.569434e-02 2 21 21 # 0.00021002
> 1.784651e-03 2 22 22 # 8.2026e-06
>
> It would seem to suggest that the BR(h->bb) is ~89%.
> This is quite off from the SM expectation for mH=125 GeV, e.g. 5.77E-01 (see https://twiki.cern.ch/twiki/bin/view/LHCPhysics/CERNYellowReportPageBR2#Higgs_2_fermions )
>
> Also: this doesn't output the values for the bosons, particularly di-
> photon in which I am interested.
>
> Thanks for any clarification/further help
>
> G
>
> --
> You received this question notification because you are a member of
> MadTeam, which is an answer contact for MadGraph5.

Dear Olivier
 thanks a lot for the clarifications. However, I am still in doubt - forgive me - that we may not have gotten to the point that is helpful for me.

I AM using heft.
I also think I fully understand the rest of your answer. But my question, in other words, is: if I set the total HIggs width to a certain value (e.g. the one expected by SM, 0.00407 GeV, then how do I know that I obtain the right BR (H->gam gam)? Here by "right" I mean SM value.
I want to be sure I generate events according to SM expectations, so I have to set the value of total width and coupling coherently, as you say. But I do not know how to extract the value of both from MG5! That is all of my question.
Again, I am using heft
  thanks, G

Hi,

> # PDG Width
> DECAY 25 4.596193e-03
> # BR NDA ID1 ID2 ...
> 8.984828e-01 2 5 -5 # 0.0041296
> 5.403821e-02 2 15 -15 # 0.00024837
> 4.569434e-02 2 21 21 # 0.00021002
> 1.784651e-03 2 22 22 # 8.2026e-06
>
> It would seem to suggest that the BR(h->bb) is ~89%.
> This is quite off from the SM expectation for mH=125 GeV, e.g. 5.77E-01 (seehttps://twiki.cern.ch/twiki/bin/view/LHCPhysics/CERNYellowReportPageBR2#Higgs_2_fermions )
>
> Also: this doesn't output the values for the bosons, particularly di-
> photon in which I am interested.

This is due to your last comment that I thought that you were using the sm.
In fact the di-photon is present in the above output: this is the last line.
The missing one are the WW and ZZ, the reason is that those are not two body decay since at least one of the two need to be off-shell.
In order to include those you need to add the associated three body decays.

> But my question, in other words, is: if I set the total HIggs width to a certain value (e.g. the one expected by SM, 0.00407 GeV, then how do I know that I obtain the right BR (H->gam gam)? Here by "right" I mean SM value.

If you use the method propose by Pierre, then of course the BR is not physical since this one is provided by the following formula:
Partial_width/\sum Partial_width
Since some decay channel (like z a) are not available on the heft model, you can't have the correct BR.
As pointed above this method is correct only if you include all relevant decay channel (including relevant three body decay)

On the other hand if you compute something like this:
generate p p > h , h > a a
add process p p > h, h > b b~
Then in fact, you don't care of the value of the BR (MadGraph didn't use those). The only important value is the ratio of
BR(h-> a a) / BR(h -> b b~) = Partial_width(h -> a a) / Partial_width(h-> b b~)
which control the relative abundance of the photon and b in the final states. This ration doesn't depend (in narrow width approximation)
of the the total width value that you put as parameter.

The only impact of the total width is on the total cross-section (especially since the Higgs width is smaller than LHC resolution).

> I want to be sure I generate events according to SM expectations, so I have to set the value of total width and coupling coherently, as you say.

If you use a NLO computation to rescale your total cross-section, then clearly you don't care of the actual highs width.
If you don't, then I would go to use the use a total width which is NOT the sum of the partial width of the model in order to have a realistic (but still LO) cross-section.

Cheers,

Olivier

On Apr 21, 2013, at 10:26 PM, Giuseppe Salamanna <email address hidden> wrote:

> Question #226823 on MadGraph5 changed:
> https://answers.launchpad.net/madgraph5/+question/226823
>
> Status: Answered => Open
>
> Giuseppe Salamanna is still having a problem:
> Dear Olivier
> thanks a lot for the clarifications. However, I am still in doubt - forgive me - that we may not have gotten to the point that is helpful for me.
>
> I AM using heft.
> I also think I fully understand the rest of your answer.
> I want to be sure I generate events according to SM expectations, so I have to set the value of total width and coupling coherently, as you say. But I do not know how to extract the value of both from MG5! That is all of my question.
> Again, I am using heft
> thanks, G
>
> --
> You received this question notification because you are a member of
> MadTeam, which is an answer contact for MadGraph5.

hi.

I not only want the right bb/gamgam ratio, but also that each of them separately produces a number of events that, given the LO cross-section, is the expected yield in the SM. So sure the independent BR (or else Partial_width/\sum Partial_width) will count here?
If I assume that bb and gamgam are in the same ratio, but both are twice as big as the SM, then I am screwed, no?

> If you don't, then I would go to use the use a total width which is NOT the sum of the partial width of the model in order to have a realistic (but still LO) cross-section.

And which would be the numeric value of it?

thanks
   Giuseppe

Hi Giuseppe,

In MadGraph you ask for a given number of event.
So if you ask for 10000 events and if the partial width for the bb and gam gam is the same then you will have (up to statistical fluctuation)
5000 bb~ and 5000 gam gam events. Completely independently of the value of the cross-section (and therefore of the total width that you choose as input).

Now the question is: How many events do you expect for a given Luminosity?
Then you need to have accurate cross-section. Which is the last point of my last email.

>> If you don't, then I would go to use the use a total width which is NOT the sum of the partial width of the model in order to have a realistic (but still LO) cross-section.
>
> And which would be the numeric value of it?

That's depend of the mass, coupling that you use for the sm parameter…. So this is not possible to give you a precise value.
I would personally use the sm calculator available online on the MG5 website
(tools -> Calculators -> Standard Model (MG5))
But you can use PDG value, or a reference paper value, ...

Cheers,

Olivier

On Apr 22, 2013, at 10:51 AM, Giuseppe Salamanna <email address hidden> wrote:

> Question #226823 on MadGraph5 changed:
> https://answers.launchpad.net/madgraph5/+question/226823
>
> Status: Answered => Open
>
> Giuseppe Salamanna is still having a problem:
> hi.
>
> I not only want the right bb/gamgam ratio, but also that each of them separately produces a number of events that, given the LO cross-section, is the expected yield in the SM. So sure the independent BR (or else Partial_width/\sum Partial_width) will count here?
> If I assume that bb and gamgam are in the same ratio, but both are twice as big as the SM, then I am screwed, no?
>
>> If you don't, then I would go to use the use a total width which is NOT the sum of the partial width of the model in order to have a realistic (but still LO) cross-section.
>
> And which would be the numeric value of it?
>
> thanks
> Giuseppe
>
> --
> You received this question notification because you are a member of
> MadTeam, which is an answer contact for MadGraph5.