1. Overview of Ion Physics and Requirements for Space Applications Pete Truscott, Fan Lei, Ana Keating, Laurent Desorgher, Bart Quaghebeur QinetiQ Ltd, Farnborough; LIP, Lisbon; SpaceIT, Bern; BIRA, Brussels Petteri Nieminen ESA/ESTEC, Noordwijk Geant4 Space Users Workshop, Pasadena, CA 6 th -11 th November 2006 QinetiQ developments and research funded by ESA under contracts 17191/03/NL/LvH and 19770/06/NL/JD, and by the UK MOD under contract C/MAT/N03517 Aero 

2. 2 Background: Prediction of Single Event Effect Rates for >60,000 feet Below ~60,000 feet, SEE rates are largely driven by the secondary neutron and proton fields in the atmosphere undergoing nuclear interactions in Si At higher altitudes within the atmosphere, direction ionisation by cosmic ray ions and residual ion fragments become important Therefore need to predict ion interaction rates in upper atmosphere

3. 3 Background: Space & Terrestrial Space and aviation environments Nuclear-nuclear models are also applicable to single event effects prediction Evidence that these interactions in tungsten vias and (potentially) tungsten silicide layers give rise to high-LET nuclear recoils from low- LET incident particles (work of Reed, Weller et al at Vanderbilt University) (Data and picture from Weller & Reed, Vanderbilt University)

4. 4 (Pictures from Chipworks Inc.) Beware of the tungsten! In this examples, polycide = tungsten silicide The problems will likely get worse, as manufacturers replace Al metallization (tracks connecting transistors) with Cu  higher-Z recoils into the silicon

5. 5 Background: Interplanetary Assessment of influence of shielding on radiological effects of cosmic rays / solar particles –Long-duration manned spaceflight (interplanetary) –Investigating probability/possibility of extra-terrestrial life –Degradation of Martian “pregnancy test” kits - assessment of radiation on biological agents used to indicate presence of extra-terrestrial life NO LIFE I THINK NOT! (Picture courtesy of ESA)

6. 6 ESA MarsREM Project Collaboration of LIP, SpaceIT, BIRA and QinetiQ Primary objectives are: –Design, develop, implement and validate engineering tools, based on Geant4, to predict the Martian radiation environment for orbital spacecraft, and Mars planetary and moon landers or habitats –The tools shall be easy-to-use by mission designers and planners (rather than developed just for radiation experts), web-based and interfaced with existing radiation shielding and effects simulation tools at the SPENVIS web-site

7. 7 ESA MarsREM Project - dMEREM Developments such as: PLANETOCOSMICS (University of Bern) and MarsREC (LIP), detailed simulation models intended to provide precise predictions for scientific applications 5º Latitude 5º longitude

8. 8 ESA MarsREM Project - eMEREM The QinetiQ Atmospheric Radiation Model (QARM) is an engineering tool developed to allow rapid calculation of the GCR- and SEPE- induced atmospheric radiation environment on Earth for ground level and aircraft altitudes

9. 9 ESA MarsREM Project In support of the above, assess existing physics models in Geant4 covering energetic nuclear-nuclear and ion-electromagnetic interactions, and then develop, implement and verify additional or improved models Implement models to assess and compare the performance of passive and active radiation shielding, and apply to example cases to assess the performance of some active shielding cases

10. 10 Background and Requirements Data from W Schimmerling, J W Wilson, F Cucinota, and M-H Y Kim, 1998. Contribution of GCR and SPE ions very important to radiobiological dose in the interplanetary environment

11. 11 Requirements - Nuclear-Nuclear Species and energy range of source particles for interplanetary env. GCR: –Very wide range in species, with noticeable dips after He and Fe –Typical energy range of concern: 10’s MeV/nuc - ~100 GeV/nuc, although mean energy is several hundred MeV/nuc. Solar particle events 10’s MeV/nuc to ~1 GeV/nuc: –Impulsive, short-term events associated with solar flares have greater fraction of heavy particles –CMEs produce gradual events that are proton-rich and last longer Picture courtesy of Janet Barth, NASA

12. 12 Requirements - Nuclear-Nuclear Energies: –~100 MeV/nuc  10s GeV/nucessential –10s MeV/nuc  ~100 MeV/nucdesirable –100s GeV/nucdesirable? Projectile species: He  Fe essential, > Fe desirable Target species –Interplanetary missions A  60: metal alloys, plastics & composites, oxidants & fuels, consumables, Martial soil / regolith … but possibly U-238 desirable –Semiconductor radiation effects: standard semiconductor materials + tungsten, indium, silver, copper... Double-differential cross-sections, with treatment of kinematics of recoiling nuclear fragment & secondaries (though effects of prompt secondary nucleons not as important for semiconductor effects) Treatment of fragment de-excitation processes

13. 13 Background and Requirements Applicable target materials: –Man-made / transported materials such as: metal alloys of Al, Ti, Fe, Mg, Be; plastics and composites; fuels/oxidizers; deliberate shielding materials (polyethene, water); crew consumables/life- support –Mars atmosphere (CO 2, N 2, Ar, O 2 …..) –Martian or Lunar soil / regolith (O, Si, Al, Fe, Mg, Ca), including composites with man-transported materials to form solid radiation shields

14. 14 Existing Geant4 Nuclear-Nuclear Final State Models Binary Light Ions Photon Evap Multifragment Fermi breakup Evaporation Pre- compound Rad. Decay Ions Wilson Abrasion Electromagnetic Dissociation Thermal 1 MeV 10 MeV 100 MeV 1 GeV 10 GeV 100 GeV 1 TeV (/n) Ablation (from T Koi, 28/07/06) JAM interface Intrinsic G4 Interfaced JQMD interface EPAX model also compared with above but not interfaced (REAT-MS) INCL4/5-ABLA

15. 15 Existing Geant4 Nuclear-Nuclear Final State Models Abrasion model: –better at predicting nuclear fragment yield, especially if used with ablation model (75% of time within factor-of-two) –poor at predicting secondary nucleon spectra and kinematics of ion fragments questionable - not a microscopic model Binary Cascade modeL –does worst at predicting nuclear fragment yields –better for prediction of secondary nucleon spectra JQMD accurate and wide energy-range, but slow simulation Other than electromagnetic dissociation, no intrinsic models >5GeV/nucl EMD plays an important role for relativistic medium and high-Z nuclei (e.g. 25% of events for 28 Si on Ag), but currently does not include emission of particles heavier than nucleons No intrinsic models <100 MeV/nucl

16. 16 Comparison of the DDCS for neutron production in 12 C + C (600MeV/A)

17. 17 Comparison of the percentage of times the predicted cross-section for fragment production is within a factor of E of the experimental value (for various projectile nuclei on carbon target)

18. 18 Options - Nuclear-Nuclear HIGH ENERGY Treatment of relativistic regime essential (>5GeV/nuc): –Completion and testing of G4 QGSM - how difficult is this? –Interfaces with other high-energy nuclear-nuclear models DPMJET-III (also used by FLUKA) CORSIKA atmospheric shower code underpinned by VENUS (10 7 GeV max), QGSJET, DPMJET-II (10 11 GeV) - note implied lower-energy limit (lab-frame) of 80GeV/nuc for these models LAQGSM from LANL - results for Ne on Cu shown, and intended to treat 100’s GeV/nuc. LAQGSM to be included in next MCNPX release

19. 19 Comparison of GSI data and LAQGSM+GEM2 prediction for 80 Kr + 9 Be (1GeV/A). Predicted nuclide yields generally show good agreement with GSI data except for neutron-rich Rb (from SG Mashnik et al, arXiv:nucl-th/0404018 v1 7 Apr 2004) 238 U + 64 Cu (950MeV/A) However, limited access to detailed information about the LAQGSM model

20. 20 Comparisons from Ballarini et al www.fluka.org DPMJET-III >5GeV/nuc used in FLUKA is not directly available, but DPMJET II.5 is Comparison measured and FLUKA- predicted yields from Pb on C, Al, Cu, Sn, Au and Pb (158 GeV/A) DPMJET-III Expt

21. 21 Comparison of neutron spectra from 12 C + C (100MeV/A) and Fe + Pb (400MeV/A) The PHITS code is an integration of JQMD and JAM, and gives an indication of the performance of the very latest versions of these models Results show excellent agreement for secondary nucleon spectra and nuclide yields in <GeV/nuc regime (from Niita et al, Radiation Measurements - in press) 40 Ar + Cu (230MeV/A)

22. 22 Options - Nuclear-Nuclear HIGH ENERGY (CONT.) –Abrasion-ablation model in ABRABLA, based on Gaimard & Schmitt - results given up to 2GeV/nuc by G&S, but expected to treat higher-energies? Other abrasion-ablation physics … but limited information about upper-energy limits: –Townsend (optical fragment) abrasion model - not clear about energy regime (published tests over 0.77 - 1.65 GeV/nuc) –Cucinotta microscopic abrasion model - avoids reliance on parametric fits and reports more accurate secondary neutron production (published tests cover 0.74 - 1.6 GeV/nuc)

23. 23 Options - Nuclear-Nuclear Require a better description of the kinematics of nuclear recoil - especially for microdosimetry - abrasion models will have difficulties here Extension of electromagnetic dissociation so that fragments heavier than protons & neutrons treated - Not high priority for low-Z LOW ENERGY Boltzmann Transport Equation –Describes thermalisation of nucleus from N-N collisions –Appears to be tested over 11-30 MeV/nuc, i.e. very much solution for low- energies - Not a high priority for interplanetary radiation environments Extensions of pre-equilibrium physics

24. 24 Comparison of measured neutron spectra and Boltzmann Master Equation (BME) prediction for 20 Ne on 165 Ho (11, 14.6, 21, and 30 MeV/nuc) Predictions are in excellent agreement with experimental results (from M Cavinato et al, Phys Lett B, 1996 & 1997) DD cross-sections for neutrons from 36 Ar + 197 Au (35MeV/A) calculated using BME FLUKA has recently released with BME model

25. 25 G4BinaryLightIonWrapper - uses G4BinaryLightIonReaction and biases the nuclide yields by the cross-section calculated by the Wilson Abrasion Model Maintains secondary nucleon spectra, but improves nuclear fragment prediction Comparison of the DDCS for neutron production in 12 C + C (600MeV/A) G4BinaryLightIonWrapper G4HadFinalState *ApplyYourse lf G4BLI yields G4BinaryLightIonReaction G4WilsonAbrasionCrossSection A simple solution currently being implemented (REAT-MS) Appropriate for fluence calculations May be unsuitable for microdosimetry Use with EPAX? Return

26. 26 G4BinaryLightIonReaction Assessment of the suitability of using G4BinaryLightIonReaction to determine total inelastic interaction cross-section G4BinaryLightIonReaction arranges nucleons within the projectile and target and samples impact parameter space G4BinaryCascade first samples if a collision between nucleons has occurred before sampling final state R t +R p -R t -R p R t +R p Target Projectile

27. 27 Comparison of total nuclear cross-sections from experiment (symbols) and determined from G4BinaryLightIonReaction, Tripathi and Shen models - interactions of 12 C with 56 Fe, 40 Ca, 27 Al and 12 C 56 F e 40 Ca 27 Al 12 C

28. 28 Conclusions For Mars environment analysis and Earth atmospheric radiation propagation urgent need to address rel. and ultra-rel. regimes Several microscopic models already cover this regime: –Access to the detailed theory behind the models is problematic –In some cases (DPMJET-II.5) easier to develop interface rather than re- engineer the physics –Would this provide a significant improvement over G4  PHITS/JAM interface? Use of pragmatic solutions as well may be valuable to addressing Mars radiation analysis –Parametric models based on detailed modelling results for limited material set Use other abrasion models? How accurate in rel. and ultra-rel? Addressing high-energy does not help the micro-/nano- dosimetrists amongst us Discuss...

29. 29 Backup Slides

30. 30 Other models The “THERMINATOR” event generator from WUT: simulations demonstrated for Au+Au @ 200GeV/nuc N S Amelin Universal Kinetic Model (UKM & UKM-R) - sub-rel regime –Attempting to get this model