Matching and merging for a dim 8 operator

Hi,

My advise is to increase the gap between qcut and xqcut.
By experience for dimension 6/8 the automatic gap that works for the SM is sometimes too small in presence of higher operator.

Cheers,

Olivier

> On Jan 4, 2016, at 11:12, Sebastian Bruggisser <email address hidden> wrote:
>
> Question #280434 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/280434
>
> Description changed to:
> Dear All,
>
> In the framework of a recast of a search for dark matter in monojet
> (-like) events I am trying to simulate a dimension 8 operator describing
> the interaction between a real scalar Dark Matter particle \phi (SM
> singlet) and quarks q:
>
> \partial_\mu \phi \partial_\nu \phi \bar{q} gamma^\mu partial^\nu q
>
> I did the implementation of the model in FeynRules and I exported it as
> a UFO file. I am running into troubles when I am trying to match and
> merge the samples (ME level and Pythia). I do something along the lines
> of:
>
> generate p p > phi phi @0
> add process p p > phi phi j @1
> add process p p > phi phi j j @2
>
> I run an pass this to pythia where I am using a mlm matching
> prescription with kt showering. The problem is that I cannot find a
> value for XQCUT for which my djr plots seem reasonable (tested from
> XQCUT=10 to XQCUT=500 for a mass of the particle \phi of 100). The djr
> plots show bumps where respectively the 1- and 2- jet samples dominate.
>
> As a cross check, I also tried the simulation with a UV-model including
> a heavy spin 2 particle which carries the interaction. The djr plots
> show the same features as in the EFT case described above. However I am
> not able to reproduce the same cross sections as in the EFT case.
>
> So my question is: Is there any reason why matching and merging for dim
> 8 operators should work differently than in the lower dimensional cases?
> How can I meaningfully match and merge in the case of this dim 8
> operator?
>
> Many thanks in advance.
> Cheers,
> Sebastian
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Hi Olivier,

Thanks for your reply. I tried to increase the gap between qcut and xqcut in different ways and found that the configuration which gives the smoothest DJR plots is XQCUT=60 and QCUT=200. This seems to be a huge gap. Is this still reasonable or should I try to get them somewhat closer?

Cheers,
Seb

Hi,

This sounds a very large gap and a quite high energy for QCUT given that your scale is ~200GeV.
It sounds unlikely that the parton-shower is trustable at so high energy. I would really expect to have QCUT <=50

Now one potential problem with dimension 8 operator is that they should in principle be added to the pardon-shower as well.
In order to have a coherent treatment of the matching/margin procedure. But including those additional kernel in the pardon-shower is typically not done.
You might be in a case where such inclusion is actually required. This is one of the many potential problem of matching/merging with BSM physics.

Cheers,

Olivier

> On Jan 6, 2016, at 17:17, Sebastian Bruggisser <email address hidden> wrote:
>
> Question #280434 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/280434
>
> Sebastian Bruggisser posted a new comment:
> Hi Olivier,
>
> Thanks for your reply. I tried to increase the gap between qcut and
> xqcut in different ways and found that the configuration which gives the
> smoothest DJR plots is XQCUT=60 and QCUT=200. This seems to be a huge
> gap. Is this still reasonable or should I try to get them somewhat
> closer?
>
> Cheers,
> Seb
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Dear Olivier,

Thank you allot for this answer. I guess that including this operator in the parton-shower is not a straight forward task?

Cheers,
Seb

Thanks Olivier Mattelaer, that solved my question.