A Detailed Simulation of the CMS Pixel Sensor

This note describes a detailed simulation of pixel sensors called PIXELAV. It is not fast and is not intended as a replacement for the CMSIM/OSCAR pixel simulation but rather as a partial replacement for the test beam. It incorporates much of the currently known physics of charge deposition and transport in silicon. Some additional test beam data are still necessary to validate the simulation and to verify that the electronics works as designed, but hopefully, the simulation will reduce our reliance upon (expensive) test beam running. The simulation is intended to aid in: the fine-tuning of the pixel system design; the development of more realistic reconstruction algorithms; the tuning of the fast simulation to more accurately model the physical pixel system; and perhaps most importantly, the continuing calibration of the charge-sharing functions (needed for simulation and reconstruc tion) as detector is radiation damaged during operation. The simulation is already contributing to many of these goals. It gives a good description of charge sharing measurements made by the Atlas Collaboration and helps to resolve a mild controversy with the Atlas Collaboration regarding charge sharing after irradiation. The simulation indicates that the readout chip saturation does not limit but actually enhances the detector resolution by suppressing large fluctuations. The simulation demonstrates the importance of reversing the fan blade angles in the two endcaps (it gains a factor of 2 in resolution). It provides the charge sharing functions needed to reconstruct short cluster hits. It helps select the best long pixel reconstruction algorithm and provides bias corrections after radiation-induced trapping degrades the determination of one cluster end. It can be used to provide an ind ependent calibration of the forward pixel charge sharing functions after irradiation and will probably provide similar information for the barrel reconstruction.