Query regarding H->WW->lvjj processes for off-Shell Higgs decay width analysis

Hi,

Obviously, I'm not an expert on what is the relevant diagram for each process, so just take this as suggestion, draw the diagram that are generated and check if this correspond to what you want.
But this would be my suggestion (based on my understanding of your question):

1) gg->WW continuum background:
generate g g > w+ w- /h [noborn=QCD]

2) SBI
generate g g > w+ w- [noborn=QCD]

Note that for both syntax such loop have an integrable singularity at pt=0 so you are are force to set a cut on the pt of the W to be able to generate events.

Cheers,

Olivier

Thanks Olivier!

That answers my question.

Best Regards,
Sadhana

Dear Olivier!

In the above processes If I want to decay :

1) one w into lv (lepton and neutrino) and other w to 2 jets ( qq'). ( any of the w can decay to lv and qq')
2) one w into lv (lepton and neutrino ) and other w to lv (lepton and neutrino). ( same here any of the w (either w+ or w-) can decay to lv and qq')

What would be the syntax for that in standalone madgraph for the processes provided below:

1) p p > h > w+ w- [noborn=QCD]

2)
generate g g > w+ w- /h [noborn=QCD]

2) SBI
generate g g > w+ w- [noborn=QCD]

Please note we are using the Madgraph v2.6.5.

Thank you!

Best Regards,
Sadhana

The best is likely to let the parton-shower to do such decay

Cheers,

Olivier

Thanks Oliver !

I will give it a try.

Best,
Sadhana

Dear Oliver !

I have a doubt if I use:

p p instead of g g in all above process then we get few processes having feynaman diagrams which consists quarks too.
So, which one is exactly I should use?

The reason is I am getting same diagram for S [1] and SBI ( interference ) [2]. Background process ( B) [3] is fully invalid , I tried to generate in the standalone mode.

1) g g > h > w+ w- , w+ > l+ vl, w- > j j

2)
generate g g > w+ w- /h, w+ > l+ vl, w- > j j

3) SBI
generate g g > w+ w- , w+ > l+ vl, w- > j j

The model I am using *heft* model in which the “loop”, which is dominantly a top loop, is treated as an effective coupling, i.e. loop replaced by one point.

Please let me know is it feasible to use p p instead of gg for gluon gluon fusion process.

Also I have a question in mind:

as any of the w in dileptonic can decay to j j then I have to use both process.

generate g g > h > w- w+ , w- > j j, w+ > l+ vl
generate g g > h > w+ w-, w+ > j j, w- > l- vl~

My doubt is :
Can we do like this :

generate g g > h > w- w+ , w- > j j, w+ > l+ vl @0
add process g g > h > w+ w-, w+ > j j, w- > l- vl~ @1

Does this syntax makes sense?
I also have doubt why do we use @0 and @1. I actually tried this syntax from an example.

Thanks and Regards,
Sadhana

Hi,

The "@X" syntax does not have any meaning in this context and is mainly a remnant of a sytntax of madgraph4 which was allowing to force Madgraph4 to not overwritte process. This does not happen anymore since MG5, and therefore those @X can be dropped.
So the example that you use is likely inspired by madgraph4 compatibility and/or user.

 For the other question, I'm really confused now about what you want to do.
You can try to use "p p", then check the diagram and see if this is what you want or not (personally I have no clue)

Cheers,

Olivier

Hi Oliver !

Thanks, so you mean that if I remove @X then process is fine ?

Now, on the other question.

I would try to get in touch with MC contact and if any MG related query , I will get back to you.

Thanks a lot for your time and doubts clarification.

Thanks and Regards,
Sadhana

Yes they are fully fine without the @X.

Actually, to be complete, I forgot an impact on the output format, in the lhef, a process number is associated to each event and this number can be controlled by the @X. Pythia normally report separate different cross-section for different process number/...
This is likely irrelevant in most case.

Olivier

Hi @Olivier!

I am again opening this chat. Thanks for helping me throughout for gluon fusion process.

Now I would like to get the syntax for the same process but this time I would like you to provide me the syntax for Vector Boson Fusion Production mode for all three cases. I would be grateful for you. Please let me know if I need to provide further information.

Thanks and Regards,
Sadhana

Again, do not use those syntax without checking that the diagram correspond to what you are looking for (and without checking the limitation/meaning of such syntax). That being said, here is my proposal here if you restrict yourself to Leading Order simulation (both signal alone and background should be easy to do at NLO (in 4 flavor)), the SBI will be more challenging but should be doable as well)

1) So for the WW continuum background without any Higgs interference term:
The question is which background are you interested in, since you said VBF, I guess that you are not insterested to the QCD background and therefore the syntax is likely:
generate p p > j j w+ w- /h QCD=0, w+ > l+ vl, w- > l- vl~

If you want to include the QCD background on top, then the syntax would be:
generate p p > j j w+ w- /h QCD<=4, w+ > l+ vl, w- > l- vl~

If you want only the QCD background part:
generate p p > j j w+ w- /h , w+ > l+ vl, w- > l- vl~
or equivalently (but more clear syntax)
generate p p > j j w+ w- /h QED=0, w+ > l+ vl, w- > l- vl~

2) For the signal alone
generate p p > j j h, h > w+> l+ l- vl vl~

3) for the Signal + Background + Interference:
generate p p > w+ w- > j j l+ l- vl vl~ QCD=0
or
generate p p > w+ w- > j j l+ l- vl vl~ QCD<=4
(asking only for QCD background will not make sense here)

NLO syntax would be (in that case I would avoid the QCD background for simplicity):
1) generate p p > j j w+ w- /h QCD=0 [QCD]

2) generate p p > j j h [QCD]

3) generate p p > w+ w- > j j l+ l- vl vl~ QCD=0 [QCD]

For 1 and 2, you need to do the decay in Madspin
for 1: decay w+ > l+ vl
          decay w- > l- vl~
For 2:
set spinmode none
decay h > w+> l+ l- vl vl~

Cheers,

Olivier

Thanks, Olivier!

Though I have just looked at the message and I would like to understand as its Off-Shell Higgs. ( where two Ws will be on-shell)

you have suggested me this syntax ( would it be okay as I remember for off-shell Higgs sometimes we use $h , but just consider it as my doubt for clarity)
-------------------------
2) For the signal alone
generate p p > j j h, h > w+> l+ l- vl vl~
---------------------

also I would like to just clear one more thing, I need lvjj final state that means one w decays to lv and one decays to jj. I think it won't change any of the provided syntax here ( just I need to change one w to j j , is that right?)

Thanks and Regards,
Sadhana

Regarding Background, I am interested in continuum background, specially that means its interfering background (qqWWqq) having similar initial and final states, the only thing is just it will not include Higgs.

Thanks and Regards,
Sadhana

Dear Olivier !

Two things for now:
Rest, I am having a look on the diagrams and syntax.
1) Regarding Background, I will confirm you after a while. I need to confirm things again with relevant persons.

2) Another thing is:
For the signal alone, let's say if the syntax (for NLO and h is off shell) is the provided one is true, this is also the syntax for this process in on shell mode. so difference would be nothing but maybe I could be wrong.

generate p p > j j h, h > w+> l+ l- vl vl~

Thanks and Regards,
Sadhana

Ah,

I guess I was not understing correctly what you mean by signal here.
If you neither want the onshell contribution, nor the interference with background.

Then the syntax might be
> generate p p > h w+ > j j l+ l- vl vl~ $h

Cheers,

Olivier

> On 19 Mar 2024, at 12:35, Sadhana Verma <email address hidden> wrote:
>
> Question #708757 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/708757
>
> Sadhana Verma posted a new comment:
> Dear Olivier !
>
> Two things for now:
> Rest, I am having a look on the diagrams and syntax.
> 1) Regarding Background, I will confirm you after a while. I need to confirm things again with relevant persons.
>
> 2) Another thing is:
> For the signal alone, let's say if the syntax (for NLO and h is off shell) is the provided one is true, this is also the syntax for this process in on shell mode. so difference would be nothing but maybe I could be wrong.
>
> generate p p > j j h, h > w+> l+ l- vl vl~
>
> Thanks and Regards,
> Sadhana
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Hi Oliver!

Let me just focus and explain for now on signal process as for Background, I am not so sure. So I would get back to you tomorrow or later today.

-- 1) So for signal :

I have two quarks decaying to H via WW and then further one W decays to lv and one decays to qq ( or can say 2 jets ) . But since its VBF process so we also get two more jets in final state.
So for particularly for my case, higgs is now considered as off-shell Higgs ( where two w go on shell) . So I would say this syntax is not right in some way.

But may be I am not clear in conveying, so I am sharing here the syntax of one process available for the background ( which is the background qqWWqq with noTop ) though its from another analysis (in which both w decays to lv) using similar kind of background.

generate p p > w+ w- j j $ t t~ QED=2 QCD=99, w+ > l+ vl, w- > l- vl~
output WWjj_OS_noTop_qcd

Thanks and Regards,
Sadhana

Ah you want the W to be always onshell? This will indeed simplifiy things:
Then you can do:
generate p p > h > j j w+ w- $h, w+ > l+ vl, w- > j j
add process p p > h > j j w+ w- $h, w+ > j j, w- > l- vl~

and for the signal+background I would likely go for this syntax then (that is still forbidding the Higgs to be onshell)
generate p p > j j w+ w- $h QCD=0, w+ > l+ vl, w- > j j
add process p p > j j w+ w- $h QCD=0, w+ > j j, w- > l- vl~

Cheers,

Olivier

Hi @Olivier !

Thanks for your inputs.
I have a question the syntax for the signal that you have provided to me seems giving fine results. But I would say that its cross-section is 0.00535. It seems me quite low.

Is it okay to go with such a low cross-section.

Please suggest me.

Thanks and Regards,
Sadhana

Dear Olivier !

Thanks a lot !
I am getting very nice distribution for signal.
I clarified that we do not need QCD background, we want simple non-QCD interferring background.
May I know as you have provided me the separate syntax for LO and NLO.
---------
Lets say in last message this syntax was provided, Is it now for NLO or LO ? ( as when you decayed with madspain that syntax you had mentioned is NLO) , how its being determined just only by syntax?
------------------------------------------

Then you can do:
generate p p > h > j j w+ w- $h, w+ > l+ vl, w- > j j
add process p p > h > j j w+ w- $h, w+ > j j, w- > l- vl~

and for the signal+background I would likely go for this syntax then (that is still forbidding the Higgs to be onshell)
generate p p > j j w+ w- $h QCD=0, w+ > l+ vl, w- > j j
add process p p > j j w+ w- $h QCD=0, w+ > j j, w- > l- vl~

----------------------------------------------
also I would like to go for NLO samples, what would be the syntax of the background in that case along with all other samples ( S, B,SBI) ?
PS: Process is VBF H->WW->lnu jj

Thanks and Regards,
Sadhana

Concerning, the syntax question:

In madgraph, you can handle three type of computation
1) LO
2) NLO
3) loop induced (which can either be seen as a piece of the NNLO computation but is also sometimes seen as LO)

To simulate (2) or (3), the syntax should contains [QCD].
If the process before the [QCD] fails to be associated to any tree level amplitude then the code will generate a process of type (3) --which will be handle like a LO process in term of interface/...-- otherwise you will do a NLO generation (type 2).

If you do not have any [QCD] then this is a LO computation.

Concerning your request of syntax for NLO:
In that case the decay of the Ws need to be handle by MadSpin.
If I remember correctly your need, I would likely go for the following syntax:

S+B: generate p p > j j w+ w- QCD=0 [QCD]
S: generate p p > h > j j w+ w- QCD=0 [QCD]
B: generate p p > j j w+ w- / h QCD=0 [QCD]

None of those syntax will allow a higgs onshell since they require w onshell.
Now a problem can occur in madspin since the decay of the W can be done offshell.
If you allow the BW of the W to go too far offshell, madspin will be super slow since NWA is not valid anymore and that madspin will (try) to to migrate (many) events to have an onshell higgs. So in that case you need to be sure that BW cut is not too large (i.e. maximum at (MW-MH/2)/width_W) so less than 9). You can set that cut either in the run_card or in the madspin_card.

Hi Olivier !

Thanks for your response. I was waiting for these from a long time.

I have one question for VBF process that we are discussing here that in case we also want to keep spin correlations. Usually in on shell samples in HWW ( in VH process ) my group people used spin none mode ( for madspin).

Would default settings here of madspin would be able to keep Spin Correlation or I should also go for `spin none` mode.

Thanks and Regards,
Sadhana

default mode of madspin should work here.
Another mode that could work is the "onshell" mode where the W are kept exacty onshell and where you do not have the issue with the breaking of NWA that I mention above (but better to have the full mode with a "small" window allow for the offshell effect).

Cheers,

Olivier

Though I should have opened two separate topics for ggF Off Shell and VBF Off Shell in two separate emails. But as its now I hope will converge sooner. Let's discuss it here.

--------------
Now, I have a little doubt as the loop-induced process is not compatible with madspin default and onshell mode.
I had tried spin `none mode` for ggF fusion process ( You had advised me to decay using pythia last time for it ) but I also tried to do with madspin. As stated in last mail for on shell samples in HWW ( in VH process with similar final states ) my group people used spin none mode so I also tried that.

1)) In that case should I use none mode here as well for VBF off shell?
2) or I can't use `none mode `in ggF too ( that means I can't use any mode - because that was the only one that was allowed for loop induced processes.)

----
3) Here from Madgraph website link.

https://cp3.irmp.ucl.ac.be/projects/madgraph/wiki/MadSpin

 I am a bit confused as it says ``NO OFF Shell effect`` in `none` mode. What does it mean by NO OFF Shell effect (maybe that could clear few things).

4) Also as none mode also states that Higgs is a scalar so spin correlations won't matter much. so in that case what would be the best mode for both processes here in my case.

Your insights regarding this would be useful.

Thanks and Regards,
Sadhana

For 1/2:
 I do not see why you should use a worse prediction for VBF because you can not use a better one for the gluon fusion case

For 3:
It means that the mass of the decaying particle (W in your case) will always be exactly onshell (so 80.4 something) never above, never below.

For 4:
You do not use MadSPin for the decay of the Higgs but for the decay of the W.
So this is a spin one where spin correlation occur between production and decay.

Cheers,

Olivier