Leading Particle Biasing Overview

1. March 2007 Leading Particle Biasing Overview Jane Tinslay, SLAC

2. Overview of Techique • Classic electromagnetic leading particle biasing • Applications where high energy particles initiate electromagnetic showers may spend a significant amount of time in analogue shower simulation • Most important processes contributing to EM shower at high energies are bremsstrahlung and pair production - ie, two secondaries produced in each interaction • Reduce computing time by preferentially tracking the highest energy secondary - highest contribution to energy deposit • Hadronic leading particle biasing • Hadronic interaction can produce many secondaries of same type and with similar characteristics • Reduce computing time by discarding a predetermined fraction of them which don’t significantly contribute to shower • Can also enhance production of interesting secondaries Jane Tinslay, SLAC

3. Side Effects • Lateral shower profile not reliably reproduced • Shower fluctuations not fully modeled • Possible to end up with large weight given to a few low energy particles • Energy deposit fluctuations • Codes recommend use of weight windows to control weight fluctuations Jane Tinslay, SLAC

4. Applications • Radiobiological doses • Heating effects • Radiation damage • Estimating shower punch through • Reduce time spend simulating hadronic cascades • Reduce time spent simulating high energy EM showers Jane Tinslay, SLAC

5. Leading Particle Biasing Summary Jane Tinslay, SLAC

6. EGS4/EGS5/EGSnrc • EM Leading particle biasing for e-/e+/ initiated showers • When bremsstrahlung/pair production event occurs, continue to track only one of the two remaining particles • Given: • R = random number between 0 and 1 • F = fraction of kinetic energy assigned to the lower energy particle: • If (R < F) keep lower energy particle • If (R> F) keep higher energy particle • I.e, preferentially keep higher energy particle, but keep lower lower energy particle some some of the time, to keep the game fair Jane Tinslay, SLAC

7. Assign surviving particle a weight • Manual states that speed of shower calculations improved by factor of 300 at 33GeV • Have problems with large weights reducing efficiency • Generally get factors of 20+ Jane Tinslay, SLAC

8. Fluka EM Leading Particle Biasing • EM Leading particle biasing for e-/e+/ initiated showers • Derived from the EGS4 implementation • Modified to account for annihilation photons produced from e+e- annihilation • Secondary particle selection probability proportional to useful energy rather than kinetic energy • Useful energy e-/ = KE • Useful energy e+ = KE + 2*me • Selected particle assigned weight which is inverse of selection probability • Same as EGS4, with useful energy taken into consideration Jane Tinslay, SLAC

9. Supports multiple configurations • Process combinations: • Bremsstrahlung and pair production • Bremsstrahlung • Pair production • Positron annihilation at rest • Compton scattering • Bhabha & Moller scattering • Photoelectric effect • Positron annihilation in flight • Energy thresholds for e-/e+/ • Region dependent • Recommend using weight windows to deal with large weight fluctuations Jane Tinslay, SLAC

10. Fluka Multiplicity Tuning • Leading particle biasing for hadrons/muon/photon photonuclear interactions • Define a factor by which average # secondaries should be scaled • Always retain leading particle • If factor < 1, play Russian Roulette to reduce # secondaries • If factor > 1, split secondaries (duplicate particles, split weight) • No Russian Roulette played if # secondaries < 3 • Adjust weight as appropriate • Configuration: • Mixed in with importance sampling configuration • Region by region basis • Possible to apply tuning to primary particles only • Recommend use weight window to control weight fluctuations (region defined) Jane Tinslay, SLAC

11. Geant4 Hadronic Leading Particle Biasing(Current) • Built in utility for hadronic processes • Keep only the most important part of the event along with representative tracks of given particle types • Always keep leading particle • Of remaining tracks, if a particle type exists, select one from each of Baryons, 0’s, mesons, leptons • Adjust weight as appropriate • Question: Which frame leading particle determined in ? Jane Tinslay, SLAC

12. MCNPX Secondary Particle Biasing • Similar to Fluka multiplicity tuning • Applies to any particle • Effectively combined EM/Hadronic leading particle biasing • Define a factor Sn equivalent to Fluka scale factor • Store appropriate weight • Didn’t see any mention about keeping the leading particle • Possibly implied ? • Supports multiple configurations • Secondary particle type • Secondary particle energy • Creator particle Jane Tinslay, SLAC

13. References • BEAMnrc Users Manual, D.W.O. Rogers et al. NRCC Report PIRS-0509(A)revK (2007) • The EGS4 Code System, W. R. Nelson and H. Hirayama and D.W.O. Rogers, SLAC-265, Stanford Linear Accelerator Center (1985) • History, overview and recent improvements of EGS4, A.F. Bielajew et al., SLAC-PUB-6499 (1994) • THE EGS5 CODE SYSTEM, Hirayama, Namito, Bielajew, Wilderman, Nelson SLAC-R-730 (2006) • The EGSnrc Code System, I. Kawrakow et al., NRCC Report PIRS-701 (2000) • Variance Reduction Techniques, D.W.O. Rogers and A.F. Bielajew (Monte Carlo Transport of Electrons and Photons. Editors Nelso, Jankins, Rindi, Nahum, Rogers. 1988) • NRC User Codes for EGSnrc, D.W.O. Rogers, I. Kawrakow, J.P. Seuntjens, B.R.B. Walters and E. Mainegra-Hing, PIRS-702(revB) (2005) • http://www.fluka.org/course/WebCourse/biasing/P001.html • http://www.fluka.org/manual/Online.shtml • http://geant4.web.cern.ch/geant4/UserDocumentation/UsersGuides/ForApplicationDeveloper/html/Fundamentals/biasing.html • MCNPX 2.3.0 Users Guide, 2002 (version 2.5.0 is restricted) • PENELOPE-2006: A Code System for Monte Carlo Simulation of Electron and Photon Transport, Workshop Proceedings Barcelona, Spain 4-7 July 2006, Francesc Salvat, Jose M. Fernadez-Varea, Josep Sempau, Facultat de Fisica (ECM) , Universitat de Barcelona Jane Tinslay, SLAC