MadGraph tutorial and WZ VBS guide

Note: The Cards folder is outdated as of now. The commit was made before I knew the madgraph configuration.

Included in this repo is an example of:

1. script to generate events
2. run card
3. param card
4. analysis script

The cards have changes from default values. In run card, the changes are in cuts. The param card needs to be changed corresponding to which variable one would like to assess.

Full process directories can be found on the follow google drive link (https://drive.google.com/drive/u/2/folders/1dfcebeW1XrWNl6G8tJm_6dAHqft2VYmC)

For reference, here are the slides: https://docs.google.com/presentation/d/1k3s4BwQIYrz0HTTbY98qibA8hlAi6nlqUY4xTcI-Kb4/edit#slide=id.g88e0da113c_0_31

Instructions

I highly reccomend people using Windows to set up a linux install or linux virtual machine.

1. To begin, make sure you have a fortran compiler like gcc-fortran, a c++ compiler like gcc, and python version above 2.6 and below 3 installed. The installation process is dependent on the OS you are using. On linux, you can use your distributions package manager to install the compilers. Now install madgraph v2.6.x (https://launchpad.net/mg5amcnlo/+download). Extract the files from the tarball to your desired location, and cd into the new directory. To start madgraph use

python2 bin/mg5_aMC

You should be in the madgraph program, the terminal should display an input line that looks something like this

MG5_aMC>

2. For analysis, you will require root (https://root.cern.ch/) or madanalysis5 (https://madanalysis.irmp.ucl.ac.be/wiki/tutorials). You must install root using either a provided binary or from source before it can be used for analysis. If you have to build root from source, here is a list of prereqs (https://root.cern.ch/build-prerequisites), which can all be installed through a package manager. MadAnalysis requires matplotlib to be able to make plots. Install either analysis package using the following in madgraph.

install MadAnalysis5

install ExRootAnalysis

Introductory tutorial

madgraph has a good built in tutorial for p p > t t~ (top anti-top quark production). To begin this tutorial, simply type tutorial into madgraph.

tutorial

It should print steps to help guide you to determine the cross section of this process.

If you would like, you can edit the beam energies in run_card.dat and see how that effects the cross section. The top quark a very massive quark, so how would you expect the cross section to depend on the beam energy? The most recent LHC run was at a center-of-mass energy of 13 TeV, but older runs were lower energy. The earliest run was about 900 GeV beam energy (or 1.8 TeV center-of-mass energy), what would the cross section be in this case? How about higher energies?

To analyze the results, you can install MadAnalysis5, or ExRootAnalysis. Most groups will use root in their analysis, but you can stick with what you prefer for this.

For root, you will need to know the particle id's (http://pdg.lbl.gov/2007/reviews/montecarlorpp.pdf) where the anti-particles take the negative value of the particle ID. The example script reads many madgraph runs and plots a histogram of differential cross section as a function of the transverse momentum of particles with id +11 and -11, which correspond to electrons and positrons.

If using mad analysis, making kinematic plots is straight forward. First, cd to the run you want to analyze in the output directory you specified (output/Events/run_xx). Start mad analysis.

python2 ../../../../HEPTools/madanalysis5/madanalysis5/bin/ma5

After this starts import the unweighted_events.lhe.gz file, define the top quark multiparticle to include both top and anti-top quarks. Then specify the observable to plot, then submit the job.

define to = t t~

import unweighted_events.lhe.gz

plot PT(to)

submit

A window should appear in you browser after that is completed.

WZ VBS guide

1. Open smeft-sim-cw-tot . This is only a text file that contains madgraph commands. The set commands are setting up correct madgraph configuration. The "import model" line in the event generation script will automatically download the SMEFT model if you do not already have it. The define lines will make sure multi-particle definitions are consistent between runs. The generate line will generate all Feynman diagrams for the defined process. In this case, it is the inclusive cross section for purely leptonic WZ scattering.

You will need to change the NP value in line 16 to correspond with which term you would like to generate. More info found here (https://arxiv.org/pdf/1709.06492.pdf). NP stands for "new physics". SMEFT treates the standard model as an effective field theory, or a low energy approximation of another theory. The NP value dictates how strong the effects of the "new physics" is.

SM: NP==0,
Interference: NP^2==1,
EFT (quadratic): NP==1,
Total: NP<=1

Finally, set the output to your choice.

3. Run madgraph with the script.

To automate the setup of multiparticles, importing of the model, and generating the processes, use the following. Point mg5 to the script in this repository.

python2 ./bin/mg5_aMC $PATH_TO_SCRIPT

4. Now you will want to edit the cards. You can copy the param_card.dat and run_card.dat from here, or edit your own to match. The changes in run_card.dat are in the cuts section. param_card.dat should be edited based on your run preferences. To replicate results, set all coefficients except for the one of interest to 0.

5. cd to your output directory set earlier. run madevent and generate the events.

python2 ./bin/madevent

generate_events

You can set analysis packages here to the one of your choice. Reweight and Spin are not used and should be OFF. Next, you can also edit the cards here, but if you have done that before, you can skip this. After that, the events should be generating, which takes about 10-20 minutes for me on a mid-tier laptop.

6. The root file is stored in output/Runs/run_xx. The python script above will read electron and positron pT from the file, and make a plot using a SM, SMEFT total with cw=+-1. Make sure to adjust the directories to your own.

WZ scattering with EWdim6 model

The EWdim6 model will generate diagrams with dimension 6 operators, which allow for triple boson coupling, as well as dimension 4 operators. Generate the same process using this new model (no need to specify NP). There are 5 dim6 operators, check how each one effects the cross section. If using root, compare the differential cross section w.r.t pt for each. For everyone, try plotting electron/positron pt and muon pt distributions.