$\tau$ALP

This repo contains the work for the following publication:

Citation

@article{Ema:2025bww,
    author = "Ema, Yohei and Fox, Patrick J. and Hostert, Matheus and Menzo, Tony and Pospelov, Maxim and Ray, Anupam and Zupan, Jure",
    title = "{Long-lived Axion-Like Particles from Tau Decays}",
    eprint = "2507.15271",
    archivePrefix = "arXiv",
    primaryClass = "hep-ph",
    reportNumber = "CERN-TH-2025-123, FERMILAB-PUB-25-0408-T, N3AS-25-011",
    month = "7",
    year = "2025"
}

Description

• 0_generate_tau_events.ipynb: generates dataframes of tau events from existing Pythia8 simulation files. Run the first few cells to generate .parquet files of tau events for faster event rate evaluation. To include less Pythia events, simply replace:

  NUMI_files = ['pythia8_events/soft_120_GeV',
              'pythia8_events/soft_120_GeV_3e3',
              'pythia8_events/soft_120_GeV_2e4']
  

  by, for example,

  NUMI_files = ['pythia8_events/soft_120_GeV']
  

  or

  NUMI_files = 'pythia8_events/soft_120_GeV_0'
  

• generate_rate_tables.py: this is where the event rate sensitivities are calculated with a parameter scan. Note that this is quite a slow and memory-intensive task. You can reduce the number of events used for lighter and faster evaluation (see above how to regenerate new .parquet files or simply pass the pythia files to Experiment class).

• 1_plot_kinematics.ipynb: self-explanatory.

• 2_plot_alp_properties.ipynb: self-explanatory.

• 3_lfv_main_plots.ipynb: LFV event-rate sensitivities using the results of the parameter scan.

• 4_lfc_main_plots.ipynb: LFC event-rate sensitivities using the results of the parameter scan.

• 5_lfv_uncertainty_study.ipynb: some tests and studies with the simplified approach to generate taus.

The Pythia event generation is performed in .cpp and generate_taus.py files.

Requirements

• numpy
• pandas
• scipy
• vector