Why its not Landau Fluctuation Su Dong Apr222016

Track ionization is surely well understood ? 1 mm 150 mm Landau 10 mm

Atomic Collisions in Silicon Collision cross section s(E)*E 2 45 Gev pion Moments M

Bichsel Model • Based the Bichsel RPM paper and Bichsel’s private COV software (FORTRAN!)

d. E/dx is a misleading concept Ionization loss spectrum is far from being scalable

Examples of COV based plots Investigation of thin silicon detectors for signal spectrum and

Parameterizing Bichsel Function • Single atomic collision spectrum COV code • Use the COV

Remarks on Simulation • Implementation based on Bichsel single collision spectrum by Qi Zeng

Slides: 8

Download presentation

Why it’s not “Landau” Fluctuation Su Dong Apr/22/2016 1

Track ionization is surely well understood ? 1 mm 150 mm Landau 10 mm 300 mm May be understood by a few, but not picked up by most simulations. Various kluges (e. g. Landau convoluted with Gaussian) help at limited thickness range and breakdown for very thin layers. Not just a problem for silicon. Drift-chamber d. E/dx for a single wire layer typically suffer similarly. The only known method works for very thin silicon: H. Bichsel, Rev. Mod. Phys. Vol 60, 663 (1988) Apr/22/2016 2

Atomic Collisions in Silicon Collision cross section s(E)*E 2 45 Gev pion Moments M 0 = S s(E) M 1 = S E s(E) • ~4 collisions per mm but mostly very low loss (~17 e. V) • Total E loss and fluctuation driven by the sporadic large E loss collisions. Apr/22/2016 3

Bichsel Model • Based the Bichsel RPM paper and Bichsel’s private COV software (FORTRAN!) now can be obtained from: http: //www. slac. stanford. edu/~sudong/silicon/ (see README file there) • Single atomic collision cross section for different particles at different momenta. • Using very thin layers (one collision) to ‘fold’ up to any particular thickness by doubling thickness at each stage. Apr/22/2016 4

d. E/dx is a misleading concept Ionization loss spectrum is far from being scalable to different thicknesses as the naïve term d. E/dx might suggest. Apr/22/2016 5

Examples of COV based plots Investigation of thin silicon detectors for signal spectrum and threshold effects. Apr/22/2016 6

Parameterizing Bichsel Function • Single atomic collision spectrum COV code • Use the COV output as table a few bg to interpolate to all momenta. • Simulation can generating individual collisions but needs to watch CPU time with conscious strategy. Apr/22/2016 Single p The key difficulty is handling different bg factors. 7

Remarks on Simulation • Implementation based on Bichsel single collision spectrum by Qi Zeng (SLAC) going into ATLAS pixel full simulation. • Bichsel model is particularly important for very thin sensors. This includes detailed full simulation of irradiated detectors with great variation of behavior at different depths to necessitate fine slicing the silicon depth for simulation. • Spectrum cut off needs careful coordination with G 4 d-ray threshold. Apr/22/2016 8