Skip to content

ycwu1030/2HDM_FR

Repository files navigation

2HDM_FR

The FeynRule 2HDM Files

About

Actually, in FeynRule model databasis, there is a model file for 2HDM. However, I'm not used to the convention there. So I made this to use another convention.

In current convention, I only consider CP conserving case. The input parameters are

$$v, m_h, m_H, m_A, m_{\pm}, \cos(\beta-\alpha), \tan\beta, m_{12}^2$$

Citation

If you use the model file in this repository, please cite as

@software{wu_yongcheng_2023_8207058,
  author       = {Wu, Yongcheng},
  title        = {UFO Model File for CP conserving 2HDM},
  month        = aug,
  year         = 2023,
  publisher    = {Zenodo},
  version      = {0.5.1},
  doi          = {10.5281/zenodo.8207058},
  url          = {https://doi.org/10.5281/zenodo.8207058}
}

Download

Either finding the corresponding files in RELEASE folder or checking release page.

Usage

I already made several UFOs and FA model files in corresponding folders.

For the UFO files, the name conventions are:

  • With EFF, the LO model file + Effective gluon-gluon-Scalar, gluon-gluon-Scalar-Scalar and gamma-gamma-scalar vertices with correct form-factor. With these Effective vertices, the numerical value for ggF production of scalar or scalar pair should be the same as NLO model file. (No charged scalar yet)
  • With NLO, the model contains the CTs for QCD NLO calculation in MG5
  • Without any of above, it is a LO model file. Loop-induced vertices can not be calculated.

TODO

  • Adding the loop-induced vertex as effective vertex with form factors, such that $h\to\gamma\gamma$ can be handled.
  • Adding the loop-induced vertex as effective vertex for charged scalar.
  • Adding CP-violation case.

WARNING

For the model with effective vertices, we don't fully check through those effective vertices. As far as we checked, for process not involving interference effects, they should be fine (we've checked the scalar-pair productions with triangle and box diagrams). But be careful in any case one wants to check the interference effects (we've checked the $t\bar t$ case)! For the interference effects involving effective vertices, one might also want to using NLO model and check hacking the MG5.