Simulation

Current Involvements:

• Tracker upgrade layout studies. The present Inner Detector needs to be replaced for the Super LHC (sLHC) where a peak luminosity of 10^35 is expected. The new tracker will likely be an all silicon detector: pixels near the interaction point to handle the high rates and strips further out. There are many considerations in optimizing the layout, e.g. high tracking efficiency, low fake track rates, good impact parameter resolution, etc. We are making studies using the LCSIM package whose ease of geometry definition is well suited to this task. There are ~2 people from SLAC in this study.
• Cavern background. The intense beams and large cross sections at the LHC give rise to high flux of background throughout the ATLAS detector cavern. This background, composed mainly of low energy neutrons and photons, can cause radiation damage to detector elements and front-end electronics. The induced hits increase detector occupancy. It is important to properly simulate this background so we can understand the sources and attempt to shield them. Cavern background is also crucial for sLHC upgrade. The scope of muon spectrometer upgrade depends critically on the level of background, ranging from replacing the small-radius part of endcap chambers to replacing the entier muon detector system. We have a group of ~5 people at SLAC working on cavern background.
• Pile-up validation. When the LHC beams cross in ATLAS, each crossing has on average over 20 hadronic interactions. They are mostly low-pT and "uninteresting" events. However, they will necessarily be superimposed on the high-pT and "interesting" events which we normally simulate. Pile-up refers to the mechanism of adding these additional events to the high-pT one in order to simulate what we actually observe. We have ~2 people from SLAC who are developing tools to validate that it is done correctly.
• Optimization of Geant4 simulation. Full simulation using Geant4 gives very good agreement with data; however, the level of details required to achive this makes the code rather slow. The optimization can be divided into three broad categories. The first one is strictly technical where the output is not changed at all, e.g. recoding an algorithm to be more efficient. The second one is where we expect the output to change but at a level where physics is not affected, e.g. relaxing the calculational precision so results are still dominated by electronics noise. The third category is where the precision is dominated by the calculation, but this deterioration is acceptable for specific physics studies that choose to use this approach. At present, there are several people from SLAC working on various aspects.

Past Projects:

• Parameterized shower. The Geant4 based full simulation agrees well with data, but consumes a lot of CPU time. The largest fraction is from following each electromagnetic (EM) shower particles in the Liquid Argon calorimeter (LAr). Parameterizing the EM shower response instead of tracking each shower particle in detail can save significant CPU time with minimal impact on physics precision. This project was initiated by SLAC in 2006. It has been completed and is now part of the standard ATLAS code as an option to be used for appropriate physics use cases.
• Muon detector description. The detector description in the ATLAS simulation has clashes, i.e. two or more entities occupy overlapping regions in space. The problem is particularly severe in the muon detector system because of its complicated geometry. There was an extensive campaign in 2008 to eradicate this problem. In addition to providing expert advice to others, SLAC's Geant4 experts worked directly on the active detector elements to eliminate clashes.