Generate Higgs > Z gamma process with MG5

Hi,

They are two ways of doing it:
1) computing the associated loop:
import model loop_qcd_qed_sm
generate h > z a [noborn=QED]
as explained in arXiv:1507.00020

2) You can also used contracted loop as it was implemented in the following model:
http://cp3.irmp.ucl.ac.be/~alaureys/h_za/h_za.html
Note that the latest version of the code, should support such type of model and you should not need the specific version of MG5aMC

Cheers,

Olivier

Hi Olivier,

Thanks for your reply and help!

Here is a following question. I can get the cross-section and width value with the syntax "generate h > z a [noborn=QED]". But I will get the error message when I want to generate events:

"Command "generate_events run_01" interrupted with error:
ValueError : invalid literal for float(): 0.6953319-309
Please report this bug on https://bugs.launchpad.net/mg5amcnlo
More information is found in '/afs/cern.ch/work/z/ziwang/public/MadGraph/MG5_aMC_v2_5_5/test_0828/run_01_tag_1_debug.log'.
Please attach this file to your report."

I noticed that it will always happen when I simply use "h > anything " syntax. But it works well when I use something like " g g > h > a a ".

However, it seems that the loop-induced process doesn't support the decay chain syntax. So, may I know the reason of this error, and the possible solutions?

Best,
Zirui

Hi Olivier,

(Sorry for my maybe-naive questions. You have already helped me a lot!)
I also tried the option 2 you gave. I think it is exactly the model I need.

However, I just follow the syntax on the webpage:
    ./bin/mg5
    >import model h_Zgamma
    >generate p p > h > z a
    >output pp_h_za

it always return to the error:
 Survey return zero cross section.
   Typical reasons are the following:
   1) A massive s-channel particle has a width set to zero.
   2) The pdf are zero for at least one of the initial state particles
      or you are using maxjetflavor=4 for initial state b:s.
   3) The cuts are too strong.
   Please check/correct your param_card and/or your run_card.

I try to set no_parton_cut in the run card. But the error stays the same.
May I know if I forget to do something in this procedure?

Hi Zirui,

Did you put some cut for the
generate h > z a [noborn=QED]

I've run it with 2.6.0, and the partial width computed in my case is much bigger:
3.993e-06 +- 1.087e-09 GeV

Your problem is a typical formatting issue when width/cross-section are much too small (lower than 1e-100).
But this should not occur in this case (But in presence of cut).

For the other model, looks like that they are a problem with the model.
When running the debugger on the process, I have the following:
Process 93862 stopped
* thread #1, queue = 'com.apple.main-thread', stop reason = EXC_ARITHMETIC (code=EXC_I386_SSEEXTERR, subcode=0x1021)
    frame #0: 0x0000000100038da2 madevent`vvs5_0_ at VVS5_0.f:242
   239 FCT3 = 1D0/(MDL_MT)
   240 FCT4 = 1D0/(1D0/4D0 * FCT3*FCT3*(FCT2-TMP4))
   241 FCT5 = 1D0/(TMP4)
-> 242 FCT6 = SQRT(DBLE((-1D0)*(+4D0*(MDL_MT*MDL_MT*FCT5)+ -1D0)))

Now I have to see if this is a bug of the model (in which case, I can not do anything) or a problem of ALOHA, in which case I would be able to provide a patch.

Cheers,

Olivier

Hi Zirui,

For the effective model, the bug was indeed in aloha.
You can find the associated patch here:
http://bazaar.launchpad.net/~mg5core2/mg5amcnlo/2.6.1/diff/276?context=3

I have check that if you set the same input parameter (b mass/top mass/...) in both model, you get a comparable partial width
with the effective theory and with the loop-induced module.

Cheers and thanks,

Olivier

Hi Oliver,

I come back to this topic since I'm happy that the model "h_Zgamma": http://cp3.irmp.ucl.ac.be/~alaureys/h_za/h_za.html is a suitable model for my requests.

However, there is an issue when I use it in the latest version of MG5. I follow the syntax on the webpage:
    ./bin/mg5_aMC
    >import model h_Zgamma
    >generate p p > h > z a
    >output pp_h_za

it always return to the error:
 Survey return zero cross section.
   Typical reasons are the following:
   1) A massive s-channel particle has a width set to zero.
   2) The pdf are zero for at least one of the initial state particles
      or you are using maxjetflavor=4 for initial state b:s.
   3) The cuts are too strong.
   Please check/correct your param_card and/or your run_card.

It says on the webpage that " a specific version of MadGraph5 is dedicated to this process to support formfactors". And I found the version of madgraph on the webpage works.
But since I need to use this model to generate samples using Madgraph distributed on grid, to produce some "official" samples. May I know if there is a way to edit the model to make it work under the common version madgraph?

Thanks!
Zirui