The CRME Tools on the Internet Marcus H. Mendenhall, Brian - - PowerPoint PPT Presentation

the cr me tools on the internet
SMART_READER_LITE
LIVE PREVIEW

The CRME Tools on the Internet Marcus H. Mendenhall, Brian - - PowerPoint PPT Presentation

The CRME Tools on the Internet Marcus H. Mendenhall, Brian Sierawski, Robert A. Weller, and Robert A. Reed Department of Electrical Engineering & Computer Science and Institute for Space and Defense Electronics Vanderbilt University


slide-1
SLIDE 1

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010 1

The CRÈME Tools on the Internet

Marcus H. Mendenhall, Brian Sierawski, Robert A. Weller, and Robert A. Reed

Department of Electrical Engineering & Computer Science and Institute for Space and Defense Electronics Vanderbilt University

Acknowledgements: – NASA MSFC: Advanced Avionics and Processor Systems (AAPS), formerly RHESE – The Geant4 collaboration, especially Makoto Asai, Dennis Wright and Vladimir Ivantchenko – DTRA Basic Research and Radiation Hardened Microelectronics Programs – NASA GSFC: NASA Electronic Parts and Packaging (NEPP) Program – LANL and TRIUMF, for collaborative work

slide-2
SLIDE 2

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

What is the Crème site?

✦Omnibus web site for modeling of energy deposition

in materials due to heavy particle radiation

✦Provides access to:

✴Legacy Creme-86 and Creme-96 RPP models ✴New, Geant4-based Crème-MC Monte-Carlo radiation

transport in multilayer stacks

✴Secure and simple management of files on site ✴Plotting and download of data from all simulations

2

slide-3
SLIDE 3

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Site Feature Status

  • Naval Research Laboratory was shut down CREME96 on July 19, 2010
  • Users should register with CRÈME: https://creme.isde.vanderbilt.edu
  • Public release scheduled for November 2010
  • Normal Registration does not provide access to MC tools... email a request

to be a beta user

3

Feature Status

CREME96 Modules Public Updated GCR Model Public Multiplanar Stack Beta MC Sensitive Volumes Beta CREME86 Alpha Lunar Neutron Albedo Model Alpha Probabilistic Solar Proton Models Development HZETRN Radiation Transport Development Radiation Transport Through Spacecraft Development Revised Geomagnetic Cutoff Model Development

slide-4
SLIDE 4

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Why is Crème-MC important or useful?

✦For simple geometries, provides highly detailed

particle transport and energy deposition

✦Allows Creme-96 GCR, trapped proton and

geomagnetic transmission models to generate particle fluxes to transport with Geant4

✦Has multiple-device coincidence capability built in ✦Has weighted-sensitive-volumes built in to account for

partial charge collection from regions distant from device center

✦Includes CEM03 and LAQGSM nuclear models for

high-fidelity breakup of heavy nuclei at cosmic-ray energies

4

slide-5
SLIDE 5

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Radiation Environments

✦ Legacy models ✴ CREME86 and CREME96 ✦ Updated Galactic Cosmic Ray model (ISO 15390:2004) ✦ Lunar Neutron Albedo model (Adams, 2007) ✦ Probabilistic Solar Proton Models ✴ Worst-Case Peak Flux Spectra ❖ ESP model for protons (Xapsos et al., 1998) ❖ Extend to heavy ions ✦ Worst-Case Event Fluence Spectra ✴ ESP model for protons (Xapsos et al. 1999) ✴ Extend to heavy ions ✦ Worst-Case Cumulative Mission Spectra ✴ Psychic model extended to include heavy ions (Xapsos et al., 2007) 5

slide-6
SLIDE 6

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Crème-MC technology

✦Web site is controlled via Plone to provide secure

environment and file management

✦Using much custom Python glue code, site produces

a set of files which are digested by Vanderbilt MRED code

✦MRED is a python-wrapped Geant4 application

  • ptimized for small (semiconductor-scale)

geometries

6

slide-7
SLIDE 7

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

MRED Architecture

7

MRED C++ Geant4 C++ JQMD/ PHITS

C++ C++ C++ SWIG Fortran

G4Core C++ Wrapper MRED C++Lib Geant4 C++Lib

JQMD/PHITS FortLib Interface C++Libs

Python distutils G4Core.py _G4Core.so

mredPy.py

G4Support.py run_mred.py

HDF5

C++

HDF5 C++Lib

OpenDX Grace AIDA

mred

LAQGSM/ CEM03

Fortran

LAQGSM/CEM

FortLib PENELOPE FortLib Fortran PENELOPE

Fortran Interfaces

slide-8
SLIDE 8

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Extended nuclear physics codes

8

✦Heavy-ion driven nuclear fragmentation drives many

important types of microelectronic events

✦National labs, etc., have very detailed nuclear

physics models, mostly written in FORTRAN, which can provide high fidelity reactions.

✦Need to be able to use these codes in our

framework

✴These codes are not designed to be executed inside the

framework of other codes. They are stand-alone.

✴Common heritage of codes means many COMMON

blocks and variables have same names

slide-9
SLIDE 9

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Tool for importing FORTRAN codes

9

✦ Automated Python script collects and rearranges code and

renames structures to avoid conflicts

✴ All separate files collected into a single big file ✴ Main code body moved into FORTRAN ʻmoduleʼ ✴ BLOCK DATA statements collected and moved to end, renamed

with unique names e.g. BLOCK DATA constants -> BLOCK DATA constants_cem

✴ COMMON blocks renamed with unique names e.g.

COMMON reaction -> COMMON reaction_cem

✦ c++ & python interface generated ✦ Automated procedure guarantees 2 things:

✴ low probability of bugs introduced via typos ✴ updates in master code easily reincorporated

✦ This is a unique tool -- no one else has this machinery

slide-10
SLIDE 10

module cem03 private contains subroutine cem03event common / adbf_cem / * amf, r0m, ijsp, nhump common / ajsbar_cem / * ainit, zinit, eb(30,70,100), egs(30,70,100) ... end subroutine end module cem03 block data bd1_cem common / coefa_cem / * ankj(4,4,29) common / coefbc_cem / * bnkj(4,4,8), ckj(3,8) c j = 17; pi- + p --> pi0 n or pi+ + n --> pi0 + p Charge exchange c scattering; Tlab <= 0.08 GeV: data ((ankj(n,k,17),n=1,4),k=1,4) / & 1.4988d-1 , 2.8753d+0 , -5.3078d+0 , 6.2233d+0 , &-5.9558d+0 , -1.6203d+2 , 4.3079d+2 , -6.2548d+2 , & 1.2875d+2 , 3.1402d+3 , -7.9189d+3 , 1.0983d+4 , &-8.5161d+2 , -1.8780d+4 , 4.4607d+4 , -5.8790d+4 / end block data

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Sample rearranged code

10

subroutines and functions collected inside ʻmoduleʼ BLOCK DATA not allowed in module, automatically moved to end

slide-11
SLIDE 11
  • 1. Si3N4 [400 nm]
  • 2. SiO2 [1000 nm]
  • 3. aluminum [840 nm]
  • 4. SiO2 [600 nm]
  • 5. aluminum [450 nm]
  • 6. tungsten [400 nm]
  • 7. aluminum [450 nm]
  • 8. SiO2 [600 nm]
  • 9. silicon [250 nm]
  • 10. silicon [10 µm]

Schematic (x-z) view Schematic (y-z) view Scale (x-y) view Scale (x-z) view

10 10

1

10

2

10

3

10

4

10

5

Kinetic Energy (MeV/nucleon) 10

  • 9

10

  • 7

10

  • 5

10

  • 3

10

  • 1

10

1

10

3

Flux (m

2-s-sr-MeV/nuc)

  • 1

1-H 4-He 12-C 14-N 16-O 56-Fe

10

  • 4

10

  • 3

10

  • 2

10

  • 1

10 10

1

10

2

Energy (MeV) 10

  • 16

10

  • 15

10

  • 14

10

  • 13

10

  • 12

10

  • 11

10

  • 10

10

  • 9

10

  • 8

10

  • 7

10

  • 6

Cross section (cm

2) All devices plug Bare

Integral Cross Section of helium plug high stat

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Conceptual framework of Crème-MC

11

+ +

slide-12
SLIDE 12

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Capabilities

12

✦ Create stacks with arbitrary number of layers of

materials commonly found in electronics

✦ Create either monochromatic beam or ʻspace

environmentʼ as radiation source

✦ Define multiple ʻdevicesʼ, each consisting of a set of

RPPs or ellipsoids with specified collection weight

✦ Manage range cuts for either

✴high detail of delta ray tracking (LET mode) ✴lower detail, to allow very large numbers of incident ions

(nuclear reaction mode)

✦ Produce histograms of energy deposition and of

integral cross section

✦ Compute coincidence rates between devices

slide-13
SLIDE 13

✦ Weighted sensitive volumes relate spatial ionizing energy

deposition with charge collected at a circuit node

✦ Volumes may be rectangular parallelepipeds or ellipsoids

✴ Each have a location within the multilayer stack, size, and

efficiency

✴ Volumes may overlap or be disjoint

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Sensitive Volumes

13 100% 25%

slide-14
SLIDE 14

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Multiple Device Models

✦ Represent class of failures requiring multiple circuit nodes to

collect charge

✦ Multiple cell upsets, DICE latches, etc ✦ Sensitive volume models are specified for each device and

given upset threshold

✦ Cross sections and SEU rates are provided based on

frequency of events meeting coincidence requirement

14 Device 1 Device 2

slide-15
SLIDE 15

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010

Extended Services From ISDE

✦ Website is backed by the electronics expertise of ISDE

✴Largest university-based microelectronics radiation group

worldwide (?)

✴Can provide expertise in setting up model systems, running

simulations, and interpreting results

✴Problems which are beyond capabilities of the web

interface can be migrated to full MRED sims, with almost unlimited flexibility via Python interface.

✴Maintain close working relation with G4 collaboration,

especially SLAC group and EM physics.

✦ ISDE can coordinate full accelerator-based tests to

validate calculations

15

slide-16
SLIDE 16

The Institute for Space and Defense Electronics (ISDE) The Radiation Effects and Reliability (RER) Group

Vanderbilt University, Nashville, TN

ANALYSIS & SIMULATION DESIGN SUPPORT RADIATION TESTING

  • Test and characterization capabilities
  • Total Ionizing Dose
  • Single Event Effects
  • Displacement Damage
  • Extensive set of characterization

hardware

  • ISDE has access to a suite of radiation

sources and a fully equipped parts analysis laboratory

Modeled Response Space Radiation Transport Models
  • Device Effects
  • Circuit Effects
System Effects

Ground Based Experiments

Design Decisions Space Radiation Environment Models Orbit

ISDE & RER Capabilities

http://www.icknowledge.com/ trends/p4(2).jpg National Electrostatics Corporation

Radiation Environments Devices Supercomputer

/* int patmat(pattern, string) char *pattern, *string; patmat() examines the strings 'pattern' and 'string' for equality and returns YES (=1) if they are equal and NO (=0) if not. The string 'pattern' may contain '*' and '?' characters which will match any substring and any single character
  • respectively. Either symbol may appear at any location in the pattern. The '*'
will also match a NULL string -- that is, the absence of any characters in the designated position. The '?' character always requires that there be an explicit (but arbitrary) character present in the test string. */ #define YES 1 #define NO int patmat(pattern,string) char *pattern, *string; { register char *p, *s; register int match = NO; p = pattern; s = string; while(*p != '*') { if(*p != *s && *p != '?') goto retNO; if(*p++ == '\0') /* Then *s must also. This is a match. */ goto retYES; if(*s++ == '\0') /* No more string, but still some pattern */ goto retNO; /* which is not a '*'. No match. */ } if(*++p != '\0') /* '*' at the end matches anything, even a */ while(!patmat(p,s)) /* NULL string. */ if(!*s++) goto retNO; retYES: ++match; retNO: return(match); } /* Robert A. Weller */ /* December, 1990 */ Croc image: http://crocodilian.com/crocfaq/

Algorithms Do it once; do it right. Leverage supercomputer scaling.

Supercomputer-Based Radiation-Effects Analysis

Laser Testing Custom cold temperature dewar for in-situ heavy ion testing In addition to in-house test capabilities, ISDE has extensive experience working with outside radiation test facilities High Speed Probe Station ARACOR

  • Rad-aware compact modeling: ICs & discretes
  • Process modeling with radiation effects
  • Test chip design
  • RHBD systems design & training
  • Software tool development & automation

Radiation Tolerant

Systems

Analog/Mixed Signal Digital RF Optical Nano

Environments Technologies

Single-event Effects SEU SET Burnout Transient Ionizing Dose Total Dose CMOS Bulk SOI Bipolar Optical Displacement Damage Nanotechnologies BiCMOS B E C S D G Radiation Enabled Behavioral/Degraded Models Design, Simulation, and Topology Hardening Layout Hardening Techniques Final Radiation Hardened Design

RH - PDK

Radiation Test Data Model Development and Calibration Custom Test Chip Design Technology Characterization: TCAD Simulations

VANDERBILT ISDE & RER

Radiation Effects Research (RER) Group Institute for Space and Defense Electronics (ISDE)

World’s largest university-based radiation effects program

  • Faculty fellows with extensive expertise in radiation-effects
  • Vanderbilt Beowulf supercomputing cluster and ISDE mini cluster
  • Custom software codes
  • EDA tools from multiple commercial vendors
  • Multi-million $ aggregate annual funding
  • Test and characterization capabilities and partnerships
  • 15 full time engineers
  • 2 support staff
  • Limited access, ITAR compliant, IP
protection
  • Document control, milestone tracking,
structured management
  • Task driven support of specific radiation
effects engineering needs in government and industry, both large and small
  • 9 faculty with decades of radfx experience
  • 30 graduate students
  • A few undergraduate students
  • Open access
  • Hundreds of technical publications
  • Basic research and support of ISDE
engineering tasks
  • Training ground for rad-effects engineers

marcus.h.mendenhall@vanderbilt.edu 19/Aug/2010 Geant4 Space Users Workshop, 2010