One of the most well known particles we detect with ATLAS is the electron. We know that it will create a track within the Inner Detector (ID) and deposit all of its energy into the Electromagnetic Calorimeter (EMCal). How do we reconstruct this?
The ID is composed of multiple layers of silicon detectors plus the straw tracker layer. As charged particles travel through the silicon and excite the material to create an electrical signal. This shows up as “hits” as the charged particle passes through the ID. With this collection of hits, we can use combinatorics in order to find the most probable set of hits to create individual tracks.
Once the electron leaves the ID, it reaches the calorimeters. When the electron hits the calorimeter, it creates an “electromagnetic shower” of electrons and photons. Electrons create photons through Bremsstrahlung and photons create electrons through pair production (there are other interactions, but these are the two main ones). These processes continue until the full energy of the electron is absorbed by the calorimeter material.
These EM showers are a lot smaller than the equivalent “hadronic showers” created by hadrons, meaning they are contained in the first part of the calorimeter, which is optimized for reconstructing such showers and is hence called the EM calorimeter (EMCal). EM objects (electrons and photons) can be identified by having such shorter/narrower calorimeter deposits.
We then take the information from both subsystems in order to reconstruct the electron or photon. As shown in the Object Introduction, if we can spatially connect a track to a large deposit in the calorimeter then we have an electron! … Right?
Well … our electron reconstruction and identification aren’t exactly perfect. There are a number of ways we can reconstruct a “fake” electron, the software thinks an artifact is an electron but it is not actually one. The EGamma group helps define requirements for analyzers to set in order to have confidence that the electron object is truly an electron.
Several different “working points” are defined to allow analyzers to choose how strict of a requirement they want to use when identifying electrons. These working points typically range from ‘Very Loose’ to ‘Tight’. As the working points get tighter, the rate of “fake” electrons decreases, but so does the efficiency of selecting real electrons. This is always the tradeoff when doing any kind of classification.
Even if you choose the tightest available working point, it is very likely that you will still have some background from events with “fake” electrons. Choosing a tighter working point simply reduces the size of such backgrounds. In many analyses, it is necessary to use “data-driven” techniques to estimate the “fake” lepton background. This is often accomplished by comparing electrons that pass different working points in data.
In some analyses, the choice of triggers that you will use may require you to use a certain minimum working point.