Transcript Document
Low Energy Electromagnetic Physics http://www.ge.infn.it/geant4/lowE Maria Grazia Pia INFN Genova on behalf of Geant4 Low Energy Electromagnetic Working Group Monte Carlo 2005 Chattanooga, 18-21 April 2005 Maria Grazia Pia, INFN Genova Dark matter searches XMM Boulby mine From deep underground to galaxies Courtesy of NASA/CXC/SAO Bepi Colombo From crystals to human beings Brachytherapy Radiobiology Maria Grazia Pia, INFN Genova Radiotherapy Low Energy Electromagnetic Physics A set of processes extending the coverage of electromagnetic interactions in Geant4 down to “low” energy – 250/100 eV (in principle even below this limit) for electrons and photons – down to approximately the ionisation potential of the interacting material for hadrons and ions Processes based on detailed models – shell structure of the atom – precise angular distributions Specialised models depending on particle type – – – – data-driven models based on the Livermore Libraries for e- and photons analytical models for e+, e- and photons (reengineering Penelope into Geant4) parameterised models for hadrons and ions (Ziegler 1977/1985/2000, ICRU49) original model for negative hadrons Maria Grazia Pia, INFN Genova The process in a nutshell Rigorous software process – – – – Iterative and incremental model Based on the Unified Process: bidimensional, static + dynamic dimension Use case driven, architecture centric Continuous software improvement process User Requirements Document – Updated with regular contacts with users Analysis and design – Design validated against use cases Unit, package integration, system tests + physics validation – We do a lot… but we would like to do more – Limited by availability of resources for core testing – Rigorous quantitative tests, applying statistical methods Peer design and code reviews – We would like to do more… main problem: geographical spread + overwork Close collaboration with users Maria Grazia Pia, INFN Genova User requirements Various methodologies adopted to capture URs GEANT4 LOW ENERGY ELECTROMAGNETIC PHYSICS Elicitation through interviews and surveys useful to ensure that UR are complete and there is wide agreement User Requirements GEANT4 LOW ENERGY ELECTROMAGNETIC PHYSICS Joint workshops with user groups Use cases Analysis of existing Monte Carlo codes User Requirements Document Study of past and current experiments Status: in CVS repository Direct requests from users to WG coordinators Maria Grazia Pia, INFN Genova Version: 2.4 Project: Geant4-LowE Reference: LowE-URD-V2.4 Created: 22 June 1999 Last modified: 26 March 2001 Prepared by: Petteri Nieminen (ESA) and Maria Grazia Pia (INFN) OOAD Technology as a support to physics Rigorous adoption of OO methods openness to extension and evolution Maria Grazia Pia, INFN Genova Version 2 Testing 27 May 2001 The Role of Testing in the Software Process of the Geant4 Low-Energy Electromagnetic Physics Working Group P. Nieminen and M.G. Pia Introduction 1 Testing forms a vital part of the software process in developments as advanced and complex as those currently in progress in the Geant4 Low-Energy e-m physics Working Group. The purpose of this document is to outline the procedures to be followed regarding testing both during development of new software, and during updates and corrections to existing code. Integrated with development (not “something to do at the end”) Suite of unit tests (at least 1 per class) Testing objectives and goals 2 The objective of testing is to ensure the new, or updated, code performs as intended. Testing should reveal any potential deviancies from expected behaviour of the code both from physics and performance point of view. The goal is high-quality code ready for public release, ultimately leading to easier maintenance and substantial timesaving for developers in the course of the software lifecycle. 3 Cluster testing 3 integration/system tests Test designs and testing schedules 3.1 Test requirements Suite of physics tests (in progress with publications) 1. Testing should be performed according to agreed and documented procedures. 2. Traceability through requirements-design-implementation-tests should be implemented. 3. The design should be tested for satisfying the user requirements. 4. The code implementation should be tested for compliance with the design. 5. The code should be tested for correct functionality. Regression testing Testing process 6. The code should be tested for compliance with Geant4 coding guidelines. 7. The code should be tested for satisfactory quality, clarity and readability. 8. Every class of the lowenergy category shall be exercised in an appropriate system test (directly or indirectly). 9. The code should be tested on all Geant4 supported platforms. 10. The code shall be submitted to the entire set of tests above to be considered for release. - Testing requirements Testing procedures etc. 11. Tests and test tools should be documented. 12. The test code should be kept under configuration management (in Geant4 CVS repository). 13. Reference outputs, data sets for validation tests etc. should be kept in appropriate agreed locations, accessible to the whole WG. Physics validation 14. Test tools should be maintained. 15. Modifications of the tests (including test tools, reference outputs, data sets etc.) should be performed according to agreed and documented procedures. 16. The most recent test results should be made available to WG coordinators for code to be included in a monthly global tag or in a Geant4 public release, according to the guidelines described in the "Testing process" section. Maria Grazia Pia, INFN Genova XP practice “write a test before writing the code” recommended to WG developers! Photons and electrons: processes based on the Livermore library Based on evaluated data libraries from LLNL: – EADL (Evaluated Atomic Data Library) – EEDL (Evaluated Electrons Data Library) – EPDL97 (Evaluated Photons Data Library) especially formatted for Geant4 distribution (courtesy of D. Cullen, LLNL) Validity range: 250 eV - 100 GeV – The processes can be used down to 100 eV, with degraded accuracy – In principle the validity range of the data libraries extends down to ~10 eV Elements Z=1 to Z=100 – Atomic relaxation: Z > 5 (transition data available in EADL) Maria Grazia Pia, INFN Genova Calculation of cross sections Interpolation from the data libraries: log 1 logE2 / E log 2 logE / E1 log E logE2 / E1 E1 and E2 are the lower and higher energy for which data (1 and 2) are available Mean free path for a process, at energy E: 1 i E ni i ni = atomic density of the ith element contributing to the material composition Maria Grazia Pia, INFN Genova Photons Maria Grazia Pia, INFN Genova Compton scattering Klein-Nishina cross section: d 1 2 h2 h0 h 2 r0 2 2 4 cos d 4 h0 h h0 Energy distribution of the scattered photon according to the Klein-Nishina formula, multiplied by scattering function F(q) from EPDL97 The effect of scattering function becomes significant at low energies – suppresses forward scattering Angular distribution also based on EPDL97 Rayleigh scattering Angular distribution: F(E,q)=[1+cos2(q)]F2(q) – where F(q) is the energy-dependent form factor obtained from EPDL97 Maria Grazia Pia, INFN Genova Photoelectric effect Cross section – Integrated cross section (over the shells) from EPDL + interpolation – Shell from which the electron is emitted selected according to EPDL Final state generation – Direction of emitted electron = direction of incident photon – Improved angular distribution in preparation Deexcitation via the atomic relaxation sub-process – Initial vacancy + following chain of vacancies created g conversion Pair and triplet production cross sections The secondary e- and e+ energies are sampled using Bethe-Heitler cross sections with Coulomb correction e- and e+ assumed to have symmetric angular distribution Energy and polar angle sampled w.r.t. the incoming photon using Tsai differential cross section Maria Grazia Pia, INFN Genova Polarisation d 1 2 h2 h0 h r0 2 2 sin 2 q cos2 f Cross section: d 2 h0 h h0 x cos x sin q cos f sin x 1 sin 2 q cos 2 f N Scattered Photon Polarization h0 O h q a f A z C y 100 keV small large Maria Grazia Pia, INFN Genova 1 cos q ˆj sin q sin f kˆ sin N 1 1 ||' N ˆi sin2 q sin f cos f ˆj sin q cos q cos f kˆ cos N N 250 eV -100 GeV x ' q Polar angle f Azimuthal angle Polarization vector 1 MeV small More details: talk on large Low Energy Polarised Compton 10 MeV small Geant4 Low Energy Electromagnetic Physics large Other polarised processes under development Electron Bremsstrahlung Parameterisation of EEDL data – 16 parameters for each atom – At high energy the parameterisation reproduces the Bethe-Heitler formula – Precision is ~ 1.5 % Plans – Systematic verification over Z and energy Maria Grazia Pia, INFN Genova Bremsstrahlung Angular Distributions Three LowE generators available in GEANT4: G4ModifiedTsai, G4Generator2BS and G4Generator2BN G4Generator2BN allows a correct treatment at low energies (< 500 keV) Maria Grazia Pia, INFN Genova Electron ionisation Parameterisation based on 5 parameters for each shell Precision of parametrisation is better then 5% for 50 % of shells, less accurate for the remaining shells Work in progress to improve the parameterisation and the performance Maria Grazia Pia, INFN Genova Processes à la Penelope The whole physics content of the Penelope Monte Carlo code has been re-engineered into Geant4 (except for multiple scattering) – processes for photons: release 5.2, for electrons: release 6.0 Physics models by F. Salvat et al. Power of the OO technology: – extending the software system is easy – all processes obey to the same abstract interfaces – using new implementations in application code is simple Profit of Geant4 advanced geometry modeling, interactive facilities etc. – same physics as original Penelope Maria Grazia Pia, INFN Genova Hadrons and ions Variety of models, depending on – energy range – particle type – charge Composition of models across the energy range, with different approaches – analytical – based on data reviews + parameterisations Specialised models for fluctuations Open to extension and evolution Maria Grazia Pia, INFN Genova Hadrons and ions Physics models handled through abstract classes Algorithms encapsulated in objects Transparency of physics, clearly exposed to users Maria Grazia Pia, and INFN transparent Genova Interchangeable access to data sets Hadron and ion processes Variety of models, depending on energy range, particle type and charge Positive charged hadrons Bethe-Bloch model of energy loss, E > 2 MeV 5 parameterisation models, E < 2 MeV - based on Ziegler and ICRU reviews 3 models of energy loss fluctuations - Density correction for high energy - Shell correction term for intermediate energy - Spin dependent term - Barkas and Bloch terms - Chemical effect for compound materials - Nuclear stopping power Positive charged ions Scaling: 2 Sion (T ) Z ion S p (Tp ), Tp T mp mion 0.01 < < 0.05 parameterisations, Bragg peak - based on Ziegler and ICRU reviews < 0.01: Free Electron Gas Model - Effective charge model - Nuclear stopping power Negative charged hadrons Parameterisation of available experimental data - Model original to Geant4 Quantum Harmonic Oscillator Model - Negative charged ions: required, foreseen Maria Grazia Pia, INFN Genova Some results: protons Stopping power Z dependence for various energies Ziegler and ICRU models Ziegler and ICRU, Fe Ziegler and ICRU, Si Straggling Nuclear stopping power Bragg peak (with hadronic interactions) Maria Grazia Pia, INFN Genova Positive charged ions Scaling: 2 Sion (T ) Z ion S p (Tp ), Tp T mp mion 0.01 < < 0.05 parameterisations, Bragg peak - based on Ziegler and ICRU reviews < 0.01: Free Electron Gas Model - Effective charge model - Nuclear stopping power Deuterons Maria Grazia Pia, INFN Genova Models for antiprotons > 0.5 0.01 < < 0.5 < 0.01 Bethe-Bloch formula Quantum harmonic oscillator model Free electron gas model Proton G4 Antiproton Antiproton exp. data Antiproton from Arista et. al Maria Grazia Pia, INFN Genova Proton G4 Antiproton Antiproton exp. data Antiproton from Arista et. al Atomic relaxation See next talk Maria Grazia Pia, INFN Genova Geant4 validation vs. NIST database All Geant4 physics models of electrons, photons, protons and a compared to NIST database – Photoelectric, Compton, Rayleigh, Pair Production cross-sections – Photon attenuation coefficients – Electron, proton, a stopping power and range Comparison of Geant4 Standard and Low Energy Electromagnetic packages against NIST reference data – document the respective strengths of Geant4 electromagnetic models Quantitative comparison – Statistical goodness-of-fit tests See talk by B. Mascialino on Wednesday Maria Grazia Pia, INFN Genova Electrons: dE/dx Ionisation energy loss in various materials Compared to Sandia database More systematic verification planned Also Fe, Ur Maria Grazia Pia, INFN Genova The problem of validation: finding reliable data Backscattering low energies - Au Note: Geant4 validation at low energy is not always easy experimental data often exhibit large differences! Maria Grazia Pia, INFN Genova Applications A small sample in the next slides – various talks at this conference concerning Geant4 Low Energy Electromagnetic applications Many valuable contributions to the validation of LowE physics from users all over the world – excellent relationship with our user community Maria Grazia Pia, INFN Genova LINAC for IMRT Kolmogorov-Smirnov Test: p-value=1 Kolmogorov-Smirnov Test: p-value=0.1-0.9 Maria Grazia Pia, INFN Genova M.Piergentili, INFN Genova Dosimetry Superficial brachytherapy Dosimetry Interstitial brachytherapy Leipzig applicator Dosimetry Endocavitary brachytherapy Bebig Isoseed I-125 source Maria Grazia Pia, INFN Genova MicroSelectron-HDR source Hadrontherapy beam line at INFN-LNS, Catania G.A.P. Cirrone, G. Cuttone, INFN LNS Maria Grazia Pia, INFN Genova Bepi Colombo Mission to Mercury Study of the elemental composition of Mercury by means of X-ray fluorescence and PIXE Insight into the formation of the Solar System (discrimination among various models) Maria Grazia Pia, INFN Genova Shielding in Interplanetary Space Missions Aurora Programme GCR (all ion components) p O - 16 ESA REMSIM Project Maria Grazia Pia, INFN Genova Fe - 52 C - 12 Si - 28 Dose in astronaut resulting from Galactic Cosmic Rays a Conclusions New physics domain in HEP simulation Wide interest in the user community A wealth of physics models A rigorous approach to software engineering Significant results from an extensive validation programme A variety of applications in diverse domains Maria Grazia Pia, INFN Genova