NASA - 2: MRED. Robert A. Weller

Transcription

1 NASA - 2: MRED Robert A. Weller Vanderbilt: R. A. Reed M. H. Mendenhall B. D. Sierawski K. M. Warren R. D. Schrimpf L. W. Massengill A. F. Witulski D. R. Ball J. A. Pellish C. L. Howe A. D. Tipton M. L. Alles A. L. Sternberg B. E. Templeton SLAC: M. Asai T. Koi D. H. Wright NASA GSFC: K. A. LaBel M. A. Xapsos J.-M. Lauenstein NASA MSFC: J. H. Adams J. W. Watts Sponsoring Agencies: NASA, DTRA, AFOSR Geant4 Japan 13/Feb/

2 Goal To develop a system for predicting radiation effects in arbitrary semiconductor devices and systems from first-principles. Founded on basic physics. Verified experimentally. Implemented on supercomputers. Useful now and improving continuously. 2 2

3 VU-NASA Programs With Goddard Space Flight Center Undertake research leading to improved methods and models for SEU prediction in selected technologies. With Marshall Space Flight Center Implement a CREME96 update with Monte Carlo capability using VU technology and leveraging NASAsponsored research. 3 3

4 Progression of SEU Analysis CREME96 VU Then VU Research Integral over path length Distribution + critical charge Multi-volume Calorimetry + Charge-collection models + Critical charge Calorimetry TCAD or SPICE in the loop coincidence analysis 4 4

5 MRED Analysis Modalities 5 5

6 Structure of MRED System language: Python. MRED: A Python module mredpy. Target machine: VU Linux cluster. Python: Writes scripts. Controls job execution. Merges results. SWIG writes: lines of c lines of Python. Python SWIG MRED C++ Geant4 6 6

7 Calorimetry Si-Nitride 0.4 µm SiO2 1.0 µm TiN 0.1 µm Al 0.84 µm TiN 0.1 µm SiO µm TiN 0.1 µm Al 0.45 µm SiO2 or W 0.6 µm TiN 0.1 µm Al 0.45 µm TiN 0.1 µm SiO2 0.6 µm Si 0.25 µm In spite of advances, calorimetry remains the key thread linking radiation and response. 50 µm 2x2x2 µm 3 Sensitive Volume LET Contribution 7 7

8 Multiple Bit Upsets MRED tracks energy deposited in all layers Energy at each sensitive node is calculated Sensitive Nodes α n+si Cell Array C+3n+2p + +3α 8 8

9 Nested Sensitive Volumes Weighted sensitive volumes General charge collection model Calibrated with measured data Q = N a i E i + b ij E i E j +... i=1 i, j =1 E i = Sensitive Volume Energies N Collector SV1=90% SV2=13% SV3=4% Base Emitter Multiple Sensitive Volumes 9 9

10 Nested SVs in Complex Systems NMOS PMOS NMOS 12-T DICE Latch Kevin Warren, VU ISDE 10 10

11 CREME-MC Home Directory 11 11

12 CREME-MC Tools Environment Modules Content Area Simulation Modules 12 12

13 Multilayer Planar Structure User-Specified Back-End-of-Line 1 Sensitive Volume 1. Warren et al. TNS

14 Deposited Energy Histogram de/dx Peak 14 14

15 Event Rate SER vs Qcrit 15 15

16 MRED Priorities Comprehensive geometry system and improved visualization. Improved generic analysis interface. (TCAD, SPICE, other ) (Adequately) Correct nuclear-reaction physics. Improved description of initial e-h distributions. Physics and computational strategies for nmscale volumes. MRED program completeness and code stability

17 Use of MRED to Predict SEGR Sensitivity in Power MOSFETs - J.-M. Lauenstein, NASA GSFC Single-event gate rupture modeling is needed for: Reduction of ground testing costs Determination of failure rate Analysis of high-energy response Guidance of on-orbit bias limits Radiation Environment Response Model under Development Virtual Irradiator Modified from Allenspach, et al., IEEE TNS,

18 GDML Input MSFC Spectrometer (John Watts, MSFC) 18 18

19 TCAD GDML 100k Events fet_msh.grd 1k Events fet_msh.gdml 1k Events fet_msh.gdml 100k Events Random No errors! 19 19

20 Validation: Nuclear Physics Literature However, available data are sparse! 20 20

21 Summary MRED code development is converging. Improved physics and high-level integration are now priorities. Research to predict circuit and system behavior is on-going in multiple technologies. CREME-MC web site is on target for general beta testing in mid