ucrl-lr-132042 enhanced surveillance program/67531/metadc784933/m2/1/high_res... · today. the...

Enhanced Surveillance Program

Enhanced Surveillance ProgramFY1998 Accomplishments

UCRL-LR-132042October 1998

AlliedSignal/Federal Manufacturing & Technologies

Lawrence Livermore National Laboratory

Los Alamos National Laboratory

Mason & Hanger Corporation

Oak Ridge y-12 Plant

Sandia National Laboratories

Savannah River Technology Center

This document was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor theUniversity of California nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy,completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights.Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute orimply its endorsement, recommendation, or favoring by the United States Government or the University of California. The views and opinions of authors expressedherein do not necessarily state or reflect those of the United States Government or the University of California, and shall not be used for advertising or productendorsement purposes.

This report has been reproduced directly from the best available copy.UCRL-LR-132042 • October 1998

Available to DOE and DOE contractors from the Office of Scientific and Technical Information, P.O. Box 62, Oak Ridge, TN 37831. Prices available from (423) 576-8401

Available to the public from the National Technical Information Service, U.S. Department of Commerce, 5285 Port Royal Rd., Springfield, VA 22161

Work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract W-7405-ENG-48.

Publication Editor: Stephanie P. Shang • Editors: Patricia M. Boyd, Sue Stull, Katie Walter, Gloria Wilt Designer: Kitty Tinsley • Illustrators: George Kitrinos, Ray Marazzi, Kitty Tinsley • Proofreader: Linda Null

Enhanced Surveillance ProgramFY1998 Accomplishments

October 1998

AlliedSignal/Federal Manufacturing & Technologies

Lawrence Livermore National Laboratory

Los Alamos National Laboratory

Mason & Hanger Corporation

Oak Ridge Y-12 Plant

Sandia National Laboratories

Savannah River Technology Center

UCRL-LR-132042

ii

AbstractThis report highlights the accomplishments of the Enhanced SurveillanceProgram (ESP), the highest-priority research and development effort instockpile management today. This is volume one of eleven, the unclassifiedsummary of selected program highlights. These highlights fall into thefollowing focus areas: pits, high explosives, organics, dynamics,diagnostics, systems, secondaries, materials-aging models, non-nuclearcomponents, and routine surveillance testing system upgrades. Principalinvestigators from around the DOE complex contributed to this report.

Enhanced Surveillance Program • FY1998 Accomplishments • Volume I

Table of Contents

iii


Acronyms and Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vForeword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viIntroduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2Program Tactics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Pits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5Focus Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

LL01/LL30/LL33/LL34/LA11 • Accelerated Aging of Plutonium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8LL03 • Experimental Studies of Ejecta and Debris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10LA09 • The Miniflyer (Pu and U): High-Strain-Rate Measurements of Pit Materials . . . . . . . . . . . . . . . 11LA12 • Dynamic Behavior of Plutonium and Uranium (Intermediate Strain Rates) . . . . . . . . . . . . . . . 12LA29 • Local-Structure Analysis of Gallium-Stabilized δ-Phase Plutonium . . . . . . . . . . . . . . . . . . . . . . 13LA32 • Measurements by X-Ray Synchrotron Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14LA34 • Resonant Ultrasound Study of Pu Aging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15PX01 • Pit Surface Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

High Explosives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Focus Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

LL06 • Material Degradation Model Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20LA35 • High Explosive Initiability and Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21LA37 • PBX 9501 Lifetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22LL07/LL40 • Detecting and Evaluating Age Effects on Explosives. . . . . . . . . . . . . . . . . . . . . . . . . . . . 23LL10 • Safety and Reliability of Stockpile-Aged Materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24LL11 • Enhanced Characterization of Stockpile-Aged Initiation Components . . . . . . . . . . . . . . . . . . . 25LL12 • Mechanical Properties of Aged PBXs and Organics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Organic Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Focus Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

LA06 • Polymeric Materials Aging Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29LA24 • Theory and Simulation of Polymer Aging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30KC07 • Cellular Silicone Cushion Aging Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31LL35 • Compatibility and Predictive Modeling of Weapon Organics and Polymers . . . . . . . . . . . . . . 32

Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33Focus Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

KC02 • Fiber-Optic Velocity Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35LL16 • Aging Effects on Explosive Performance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37Focus Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

LL26 • High-Resolution X-Ray Tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40LA19 • Advanced Ultrasonic Tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41LL19 • Enhanced Ultrasonic Characterization of Assemblies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42SR05 • Enhanced Inspection of Reservoir Girth Welds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43LA16 • High-Energy Computed Tomography Detector Array . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44LA21 • Development of SQUID Microscope for Surveillance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45LL13 • Solid-Phase Microextraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46KC10 • Sensor Technology Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47LL32 • Microsensors for Evaluation of Materials Degradation and Corrosion in Weapon Systems . . . . 48SN13 • Systems Engineering of an Arming, Firing, and Fuzing Testbed. . . . . . . . . . . . . . . . . . . . . . . 49

SN14 • Weapon State-of-Health Monitoring. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50LA26 • Identifying Damage with High-Resolution Vibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51LA20 • Low-Level Vibration and Advanced Signal Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52OR18 • Laser Penetration and Rewelding for Gas Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53OR04 • Infrared Surface and Gas Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54OR10 • Material Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55OR19 • Radiographic Parameter Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56OR09 • Evaluation of Hydrogen Balance Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57LL24 • Development of High-Energy Neutron Radiography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58LA18 • Bayesian Inference Tomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61Focus Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

OR01 • Historical Data Analysis at the Y-12 Plant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63LA03 • Advanced Diagnostics Development for Flight Test Scoring . . . . . . . . . . . . . . . . . . . . . . . . . . 64LL28 • Nuclear Explosive Package Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66SN20 • Advanced Telemetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67LL27 • Stockpile Information Group—Database Management Systems . . . . . . . . . . . . . . . . . . . . . . . 68LA14 • Online Material Property Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69LA17 • Enhanced Surveillance Information Management Program . . . . . . . . . . . . . . . . . . . . . . . . . . 70LA23 • Enhanced Reliability Methodology Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71SN15 • The Stockpile Surveillance/Data Management System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

Secondaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73Focus Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

LL20/LL22/LL23 • Effects of Solid–Gas Interactions on Component Aging . . . . . . . . . . . . . . . . . . . . . 75OR02 • Interrogation of Disassembled Units. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76LA15 • Uranium Corrosion and Aging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77OR14 • Material Properties Evaluation of Uranium Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

Materials-Aging Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79Focus Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

SN01 • Atmospheric Corrosion of Weapon Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81SN02 • Welds in Radioisotope Thermoelectric Generators. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82SN04 • Energetic Materials Predictive Capability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83SN05 • Predicting and Verifying Elastomer Lifetimes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

Non-Nuclear Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85Focus Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

SN23 • Age-Induced Degradation of Nuclear Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88SN24 • Component Prioritization and Stockpile Life Extension Program Integration . . . . . . . . . . . . . 89SN25 • Electromechanical Components—Stronglinks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90SN26 • Electronic Components and Subsystems—Capacitive Discharge Unit Firing Set . . . . . . . . . . . 92SN27 • Energetic Components and Subsystems— Slim-Loop Ferroelectric Firing Set . . . . . . . . . . . . . 94

Routine Surveillance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97Focus Area. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

SN17 • Enhanced Gas Analysis for Diagnostics and Surveillance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99SN16 • Enhanced Test Methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100


iv


Acronyms and Terms

v

ARIES Advanced Recovery and Integrated Extraction System

AS/FM&T AlliedSignal/Federal Manufacturing & Technologies

Cap capacitanceCCD charge-coupled-deviceCDU capacitive discharge unitCS compression setCSR compression stress-relaxationCT compact tensionCT computed tomographyCVA Canonical variate analysis∞D Bayesian inference engineDADS dynamic analysis and design systemDARHT Dual-Axis Radiographic Hydro-TestDF dissipation factorDLO diffusion-limited oxidationdpa displacements per atomDR dual revalidationDRIFT diffuse reflectance infrared fourier

transformDTM distributed telemetryEFI enhanced-fidelity instrumentEOS equation of stateEPDM ethylene-propylene rubberESD environmental sensing deviceESP Enhanced Surveillance ProgramEXAFS extended x-ray adsorption spectroscopyfcc face-centered cubicGa galliumGC/MS gas chromatograph/mass spectrometerGPC gel permeation chromatographyGWIS Global Weapons Information SystemHEs high explosive(s)HEAF High Explosives Application FacilityHERT high explosive radio telemetryIC integrated circuitINuMM integrated nuclear materials monitoringJTAs joint test assembliesKCP Kansas City PlantLANL Los Alamos National LaboratoryLDA Local density approximationLDMF laser-driven miniflyerLINAC linear acceleratorLLNL Lawrence Livermore National LaboratoryLTTD low-temperature thermal decompositionMC Military characteristics

MTS mechanical threshold stressMW molecular weightNDE nondestructive evaluationNDIGA nondestructive infrared gas analysisNEP nuclear explosive packageNWC Nuclear Weapons ComplexOUAL Ohio University Accelerator LaboratoryPBXs plastic bonded high explosivesPD partial dischargePDA partial discharge analysisPDF pair-distribution functionPDF portable document formatPDM product data managementPETN pentaerythritol tetranitratePLR percent load retentionPOC points of contactPSCL Pit Surface Characterization LaboratoryPTW Preston-Tonks-WallacePu plutoniumQAM quadrature amplitude modulationQET quality evaluation trackingR&D Research and developmentrf radio frequency RGA residual gas analyzerRTGs radioisotope thermoelectric generatorsRUS Resonant ultrasound spectroscopySCE subcritical experimentSCG subcritical crack growthSFE slim-loop ferroelectricSIG Surveillance Information GroupSNL Sandia National LaboratorySNM special nuclear materialsSNMS sputtered neutrals mass spectroscopySPAM Signal Processing And ModelingSPME solid-phase microextractionSQUID superconducting quantum interference

deviceSRTC Savannah River Technology CenterSS/DMS Stockpile Surveillance/Data Management

SystemSTM scanning tunneling microscopySTS Stockpile-to-target sequenceVOD velocity-of-detonationWR war reserveXPS photoelectron spectroscopyXPS x-ray photoelectron spectroscopyXSR x-ray synchrotron radiation


Foreword

vi

One Mission. One Program. One Stockpile.It is our distinct pleasure to present the accomplishments of the Enhanced Surveillance Program. The Enhanced

Surveillance Program is the highest-priority research and development (R&D) effort in stockpile managementtoday. The program provides an understanding of future weapon degradation, essential to maintaining thestockpile indefinitely in the absence of underground nuclear tests. We want to share with you some of its mostimportant characteristics.

1. There is one Enhanced Surveillance Program that spans the entire nuclear weapons complex. Teams of scientistsand engineers from all Defense Programs laboratories and plants conduct research in this program. The progress todate and the program’s future success relies on this collaboration, requiring the combination of skills and facilitiesspanning the weapons complex.

2. This is a goal-oriented program. As you review the progress over the last year, you will see that we aresystematically addressing specific stockpile concerns through achievement of the program’s predefined objectives.The program’s documented milestones and deliverables drive and gauge our progress. The program balances thedelivery of many low-risk, near-term tools with a few higher-risk tasks potentially offering a big payoff for thestockpile in the long term.

3. It is our intention that this be a genuine expansion of surveillance R&D. All organizations in the weaponscomplex have conducted surveillance R&D throughout their history. With this program, we are placing additionalemphasis on improving our surveillance capability. The Enhanced Surveillance Program Plan acknowledges theongoing R&D efforts and seeks to demonstrate how those actions support the achievement of enhancedsurveillance R&D goals.

We are very pleased with this program’s progress. We commend the many investigators for their outstandingcontributions, as well as the members of the Enhanced Surveillance Steering Committee for building quality teamsof performers. We are confident this program will continue to be executed with vigor and success.

Robin StaffinDeputy Assistant Secretaryfor Research and DevelopmentDefense Programs

Gene IvesDeputy Assistant Secretaryfor Military Application and Stockpile ManagementDefense Programs


Introduction

1

This Annual ReportYou are reading one volume of the Enhanced Surveillance Program (ESP) FY1998 Accomplishments. The complete report

of accomplishments consists of eleven volumes. Volume one includes an ESP overview and a summary of selected unclassifiedFY1998 program highlights. Volume one specifically targets a general audience, reflecting about half of the tasks conducted inFY1998 and emphasizing key program accomplishments and contributions.

The remaining volumes of the accomplishments report are classified, organized by program focus area and present intechnical detail the progress achieved in each of the 96 program tasks in FY1998. Focus areas are as follows:• Pits• High Explosives• Organics• Dynamics• Diagnostics• Systems• Secondaries• Materials-Aging Models• Non-nuclear Components• Surveillance Testing Program Upgrades

ESP in ContextThe only thing remaining constant in the Nuclear Weapons Program is its mission: to maintain a safe and reliable nuclear

weapons stockpile.In the past, that mission was accomplished on a large scale with growth. Stockpile systems were periodically replaced with

newer and better versions, a robust design and production capacity supported both stockpile modernization and the rapidimplementation of stockpile repairs, and confidence was assured with the certainty of an underground nuclear test.

Today, with the challenging constraints of producing no new nuclear weapons and conducting no underground nucleartests, the mission is founded in science-based stockpile stewardship and management. Current plans require systems toremain in the stockpile indefinitely, and therefore, confidence in the readiness of the stockpile now includes an uncertaintydriven principally by aging.

The Enhanced Surveillance Program (ESP) is central to the new strategy, focused on understanding and predicting theeffects of aging and ultimately reducing uncertainty.

As you read this report, keep in mind that ESP contributions transcend stockpile management. The new materials andaging information gained from the ESP support much more than just predicting component failures and scheduling repairs.This information extends the knowledge base supporting component design, fabrication, certification, safety analysis,surveillance, and dismantlement. With this new knowledge, our confidence in all aspects of stockpile support can be firmlyanchored to a broader and deeper scientific base, making the Enhanced Surveillance Program a key component of science-based stockpile stewardship and management.

Also, remember that the ESP is an important element of stockpile life extension. The Stockpile Life Extension Program,DOE’s planning framework for proactive management of system maintenance activities, relies on ESP to provide the technicalbasis for component reuse decisions during refurbishment and to provide the diagnostics required to detect degradation ofcomponents before they fail.

The ESP has implemented a strategic vision, based on requirements and focused on the stockpile. The program exploitsresources and capabilities across the nuclear weapons complex. We measure our progress against defined performanceexpectation as documented in the ESP Program Plan. This report is just one example of how ESP documents the delivery of results.

David V. FeatherEnhanced Surveillance Program ManagerOffice of Defense ProgramsU.S. Department of Energy


Overview

2

The objective of the ESP is to develop tools, techniques, and procedures to advance the capabilities of theDepartment of Energy (DOE) to measure, analyze, calculate, and predict the effect of aging on weaponsmaterials, components, and systems and to determine if and/or when these effects will impact weaponreliability, safety, or performance. These tools, techniques, and procedures are elements of two major ESPdeliverables: augmentation of the department’s surveillance program with age-focused diagnostics andprediction of service lives of safety and reliability components.

The ESP is an essential element of the United States’ strategy for efficiently maintaining a safe and reliablenuclear deterrent. Over the last decade, the nuclear weapons program has changed in two significant ways: apolicy of extending the lifetime of existing stockpile weapons (hence, no new weapon production) has beenadopted, and the nuclear weapons complex has been streamlined, drastically reducing both capacity andcapability. The weapons complex now relies on careful planning and scheduling to conduct the requiredcomponent production, surveillance, and maintenance with diminished resources.

The tools provided by the ESP alleviate the uncertainty of the future. The conclusions of the program serve tovalidate or dismiss concerns about weapons component failure and to identify previously unknown agingimpacts. With early identification of age-related failure mechanisms (through predictive models and age-focuseddiagnostics) the program enables efficient strategic planning, minimizing the demands for infrastructure andother resources and ultimately reducing the threat to national security. Finally, investments in ESP are recoveredwhen conclusions indicate components or materials do not require replacement.

The activities under this program are carried out at the DOE weapon production plants and designlaboratories: the Kansas City Plant (KC), Y-12 Plant (termed “OR” for Oak Ridge), Savannah River Site (SR),Pantex Plant (PX), Los Alamos National Laboratory (LA), Lawrence Livermore National Laboratory (LL), andSandia National Laboratories (SN). There is strong coordination and teaming among the laboratories andproduction sites for the planning, selection, and conduct of research projects for this program.

Enhanced Surveillance Program Funding Profile

FY96 FY97 FY98 FY99 FY00$28.8M $75M $60M $68M $84.5M

The ESP represents a bridge between the Defense Programs Office’s current core R&D and stockpilesurveillance programs and provides the necessary predictive models and age-focused diagnostics required toanticipate weapon refurbishment. The program is a necessary precursor of Stockpile Life Extension Programplanning. The materials and component service-life predictive capabilities resulting from ESP are focused onanticipating stockpile defects, allowing timely remedial action and refurbishment planning.

The ESP sponsors applied research and development related to the effects of material and component aging.The program leverages work conducted through core R&D and other initiatives such as Advanced Design andProduction Technologies and the Advanced Strategic Computing Initiative program. Examples of capabilities ESPis enabling include new age-focused system, component and materials performance tests, new flight testconfigurations, material characterization and modeling, nondestructive evaluation, and modeling of the weaponaging process to the extent practicable.


3

Program TacticsDuring the conception phase of the ESP, a committee composed of nuclear weapon design and production

subject matter experts reviewed each major subsystem within a generic weapon system (i.e. primaries,secondaries, and non-nuclear components) to determine where surveillance capabilities and technologiesrequired enhancement. This assessment is validated annually. Within the three focus areas (primaries,secondaries, and non-nuclear components), the results of the assessment fell into six categories.

• Material characterization and surveillance• Materials-aging model development• Component performance models• System performance models• Component surveillance and diagnostics• Enhanced system testingThe figure depicts the DOE’s surveillance program. Enhanced surveillance provides age-focused diagnostics

and age-aware models as an augmentation to traditional surveillance testing and performance prediction.Those areas covered by the traditional surveillance program are shown on the right side of the diagram.

The ESP is shown on the left of the dashed line. The ESP, as integrated with the traditional surveillance program,forms a coherent approach toward the goal of maintaining a reliable and safe stockpile.

The ESP was constructed to be a short-term (5–7 year), rapidly evolving R&D effort directed by an ESPsteering committee. The ESP Program Manager reviews and authorizes all projects. Less fruitful projects areterminated, and new advanced research concepts are added to the program through the steering committeereview and authorization process. A program plan is published annually. Full-scale implementation of ESPtechnologies into the traditional surveillance program occurs when those technologies are demonstrated anddetermined essential.

There are many examples of the products the ESP delivers. These products range from performance forecastson critical stockpile components and assemblies, the demonstrated early identification of life-limitingdevelopments, to the identification of the key signatures of aging and the deployment of new surveillance andassessment tools. In summary, the program provides the technical basis for risk-management decision making.

Materials aging

models

Materials characterization and surveillance

Component performance

models

Diagnostics for component

surveillance

Enhanced systems testing

Defect detection

Repair, refurbish

Meets requirements

?

MCs and STS*

Performance prediction

System performance

models

Unsatisfactory reports, quality assurance, and reliabity testing

or

No

Yes

Traditional

Enhanced

Requalify

Retire

Age-aware models

Age-focused diagnostics

*Military characteristics (MCs) and stockpile-to-target sequence (STS)

6


PITS

Focus AreaThe Pit Focus Area assesses the effects of aging on

the safety and reliability of pits. Tasks in this areainclude material studies, theoretical models, andsystem-level performance assessments. Material studiesinvestigate the effects of aging on static and dynamicproperties of pit materials in both engineering andnuclear applications. Physical measurements of materialproperties combined with experiments are used toassess effects on aspects of physics performance. Newtheoretical models provide a basic understanding ofaging that will lead lifetime predictions for pits. System-level experiments are being conducted to assure thevalidity of predictions and the safety and reliability ofthe pits in the stockpile.

Analysis, prediction, and mitigation of aging effectsin pits, specifically in plutonium, are key to enduringsafety and reliability. Changes resulting from aging areexpected to occur first in fundamental properties suchas composition, crystal structure, and chemical

potential. These changes may affect engineering andphysics performance characteristics such as equation ofstate, spall and ejecta formation, strength, density,geometry, corrosion resistance, and nuclear reactivity.Aging mechanisms that cause these potential changesinclude the in-growth of decay products, uranium recoildamage and associated void formation, phase stabilityconcerns, changes in surface chemistry, and a variety ofenvironmental changes including thermal cycling.Development of advanced characterization tools tomeasure changes in these properties is essential. Theresulting data analysis will expand our nuclear materialsknowledge base and form the basis for computationalmodels necessary for predictive assessment. Newmethods for evaluating the chemistry of external andinternal pit surfaces are also an area of emphasis. Thesecapabilities permit the collection of data on thepotential reactions of other chemical species in theweapon environment.

Pertinent TasksLA09 The Miniflyer (Plutonium and Uranium): High-Strain-Rate Measurements of Pit MaterialsLA11 Investigation of Aging in Pu Using 238Pu-Enriched MetalLA12 Dynamic Behavior of Plutonium and Uranium at Intermediate Strain RatesLA29 Modeling Stability of Aged PuLA32 X-Ray Synchrotron Radiation Measurements of Residual StressLA34 Resonant Ultrasound Study of Pu AgingLA40 Advanced Recovery and Integrated Extraction System (ARIES) Sampling and AnalysisLA41 Pu Microstructural and Surface CharacterizationLA50 Pu Sample PreparationLL01 Modeling of Void Formation and Other Aging MechanismsLL03 Experimental Studies of Ejecta and DebrisLL30 Positron Annihilation Study of Age-Induced Defect StructuresLL33 Accelerated Aging of Plutonium: A Joint study with Los Alamos National LaboratoryLL34 Pu Transmission Electron MicroscopyLL37 Gas Gun Experiments: Pu Spallation and Surface DynamicsLL42 Sensitivity AnalysisPX01 Pit Characterization Laboratory

7


Deliverables1. Measure and characterize the products and nature ofcorrosion processes occurring during weapon lifetimes.

2. Define the chemistry and stability of the gas phaseand its relationship to corrosion kinetics.

3. Identify important aging phenomena (e.g., heliumin-growth) and develop procedures to monitor andevaluate those changes.

4. Obtain stress-strain mechanical property data atintermediate- and high-strain rates on baseline weaponsmaterials and stockpile-returned plutonium anduranium to understand the compressive deformationresponse of these materials as a function of aging.

5. Measure the x-ray absorption fine structure spectra,synchrotron radiation-based x-ray diffraction/scatteringpatterns, and neutron scattering patterns of baselineplutonium alloys, aged plutonium samples, and otherplutonium. Prepare a 238Pu-spiked plutonium alloy foraccelerated aging experiments. Calibrate and comparethe spiked alloy against known, aged samples.Accelerate age to extended lifetimes and measure keymaterial properties to elucidate aging effects.

6. Develop models of plutonium metal aging processwith the goal of predicting changes to key materialproperties. In conjunction with the AcceleratedStrategic Computing Initiative and other programs,assess the impact of these changes on pit safety andperformance with the final goal of determining usefulpit lifetimes.

7. Use high-flux x-ray synchrotron radiationmicrodiffraction to nondestructively measure residualstress with micrometer-sized spatial resolutions withinpit materials.

8. Measure elastic moduli in a broad temperaturerange for new and aged polycrystalline and single-crystal samples of plutonium, and correlate structuraland mechanical properties with age.

9. Predict and measure directly the dimensionalchanges occurring in aged plutonium as a result ofhelium accumulation, radiation damage, and resultantvoid formation.

10. Determine the nature of irradiation-aging-induceddefects (distribution of voids, vacancies, and othermicrostructural characteristics) in the microstructure of plutonium using positron annihilation spectroscopydefect analysis.

11. Experimentally assess effects of aging on ejecta,spall, and equation of state, and assess impacts onprimary performance for stockpile weapons. Gathermaterial property data from pits dismantled in theARIES process in order to expand the age-correlateddatabase of applied plutonium properties.

12. Instrumentation for direct examination of theexternal surface chemistry of the clad metal on pits.

13. Expanded monitoring program to provide real-time aging data on pits stored at Pantex.

Accelerated Aging of PlutoniumThe Accelerated Aging of Plutonium program formally

began in October 1997 by Livermore and Los Alamosnational laboratories to develop the enrichment techniqueand to characterize aging behavior as a function of time ina manner that makes best use of the facilities and resourcesavailable at each laboratory. We have made steady progressaddressing theoretical aging issues, devising plans forfabricating and testing the enriched material, andresolving plutonium facility issues.

Compared with 239Pu, 238Pu has a lifetime shorter bya factor of 280. Therefore, it is possible to increase the rateof self-irradiation damage and accelerate the agingprocesses. The premise is that by enriching 239Pu with238Pu, the much faster decay from the 238Pu will createdamage in the plutonium at a much faster rate thanwould normally occur (half lives are 86.4 and 24,360 yearsfor 238Pu and 239Pu, respectively). More importantly, wewill be able to measure the properties of the plutoniumboth in the new condition and periodically as it ages.Once we know how aging manifests itself, we can look forsimilar behavior in aged stockpile material. By identifyingand characterizing the changes (or lack of them), we will

provide the information necessary to makerecommendations for the future.

Specific tasks are to:• Fabricate and characterize enriched materialand specimens made from it.• Design and construct storage equipment foraccelerated aging of specimens.• Design, develop, and procure equipment forexperiments on aged specimens.• Develop a database of fundamental parametersfor analyzing experiments.• Develop computational models to translateaging phenomena and effects from the enrichedto the weapons-grade materials.

The measure of accelerations (shown inFigure 1) is defined as the number of yearsrequired to reach a radiation dose of 10 displacements per atom (dpa). Weapons-grade material will take 100 years to reach thisdose, or about 10 years if 5% of 238Pu is added.While it is possible to enrich the material evenmore, handling and storing the enriched materialbecome limiting factors. All these considerationsled to the selection of 5% enriched material.

We have selected experimental tests (seetable) that are capable of measuring changesinduced by aging, most of which are already inplace at the laboratories. Approvals, processes,and equipment must be in place before we makethe first batch of enriched material because agingof the enriched plutonium begins immediately.We expect to make the first batch in FY1999.

Radiation-induced void swelling occurs in aspecific temperature window determined bydose rate and by vacancy diffusion. As the doserate increases, the window shifts to highertemperatures. As a result, the specimensenriched with 238Pu must be aged at anappropriately higher temperature. In order toevaluate this temperature shift, theoreticalestimates had to be made on the diffusioncoefficient for vacancy migration in delta-Pubecause no measured value is available. Thetheoretical value for the activation energy forvacancy migration was found to be 0.69 eV.Using a well-established model for radiation-induced void growth, we calculated thetemperature shift as a function of the dose rateor the enrichment (Figure 2). For the 5%enriched material, the temperature shift is 15°C,and this value was adopted for the design of theheater incubators used for aging the enrichedspecimens.

8 Pits • LL01/LL30/LL33/LL34/LA11


Accelerated aging of plutonium test program.

Test Technique Information

Analytical chemistry Atomic absorption, Impurity content,emission spectroscopy, alloying elements,radiochemistry ingrowth of decay products

Density Archimedes immersion Swelling, phase composition

Dilatometry Laser triangulation, Swellinglinear variable displacement transducers

Helium effusion Gas pressure determination Helium diffusivity

Mechanical testing Tensile testing, Yield strength,compression testing ultimate tensile strength,

strain hardening

Kolsky bar, Equation of state,40-mm gas gun shock-loading behavior

Microscopy Light microscopy Grain size, phase content,morphology of inclusions

Transmission electron Bubbles, microscopy dislocation structure

Oriented image Texture microscopy

Positron annihilation Gamma radiation Types of atomic defects, spectroscopy Location of helium atoms

X-Ray diffraction Diffractometry Interatomic lattice spacing,texture, phase composition

Low temperature Self-annealing effects

A roadblock for computing the properties ofdefects created by self-irradiation in Pu has beenthe lack of an adequate electronic structure for thismaterial. Some of the valence electrons in Puchange from itinerant to localized when densitychanges because of temperature or pressure. Anew method has been developed and applied toPu and its dilute alloys that gives a correct physicaldescription of these valence electrons and their rolein the properties of weapons-grade Pu.

This method, referred to as LDA+U, is a modifiedmodel of the traditional LDA (local densityapproximation) method. It has been implementedand tested by computing the lattice parameters inPu–Ga alloys. The results and predictions obtainedfrom the LDA method are shown in Figure 3. Theagreement of the LDA+U predictions with thepublished experimental data is very good, whilethe old LDA method fails completely.

A similar improvement was noticed for the heatof mixing (Figure 4). Whereas the old theorypredicts that the two elements Pu and Ga areessentially immiscible, the new theory gives thecorrect values and shape for the heat of mixing asinferred from the Pu–Ga phase diagram.

The new electronic structure method will enableus to generate effective interatomic interactionpotentials needed for radiation damagesimulations. In the meantime, we are performingradiation damage simulations on a surrogatematerial, Pb, to study the defects generated as afunction of increasing recoil energy and oftemperature. We found that in high-Z materials, adramatic change occurs in the structure of thedamage at high recoil energies (>30 keV). Most ofthe self-interstitial defects are retained as smalldislocation loops, in contrast to the damageproduced by neutrons in fission reactors thatconsists predominantly of isolated vacancies andinterstitial clusters. We anticipate that this trend willcontinue and that the 85 keV recoil damage in Puwill be very different from the fast neutron damagein stainless steels.

The microscopic defects created as a result ofthe displacement damage can be observedexperimentally with high-resolution electronmicroscopy techniques (LL33). We expect thatradiation-enhanced transport can lead to aredistribution of alloy elements that require achemical analysis capability. To measure bothsituations, we are purchasing an analyticalmicroscope with 300-kV electron energy capableof atomic resolution. We expect to obtain thisdedicated instrument by the end of 1998 (LL34).

Pits • LL01/LL30/LL33/LL34/LA11 9


10864201

10

100

0.0

0.2

0.4

0.6

0.8

1.0

Yea

rs t

o r

each

10

dp

a

Self

-hea

tin

g p

ow

er, W

/cm

3

Enrichment, at % of additional 238Pu

Figure 4.Calculated heatof mixing forbinary Pu–Gaalloys.

10864200

5

10

15

20

Enrichment, at % of additional 238Pu

Tem

per

atur

e sh

ift,

°C

1.00.80.60.40.20.04.0

4.1

4.2

4.3

4.4

4.5

4.6

4.7LDA LDA+U Experimental values

Composition x

Latt

ice

par

amet

er

1.00.80.60.40.20.0–30

–20

–10

0

10LDA LDA + U

Composition x

Hea

t o

f m

ixin

g, a

.u.

Figure 1.Acceleration ofaging andassociated self-heating ofenriched Pu.

Figure 2. Thetemperature shiftdetermines theappropriatestoragetemperature forthe enrichedmaterial relativeto the pittemperature.

Figure 3. Latticeparameters forPu–Ga alloys,where x is the Gaatomic fraction.

Experimental Studies of Ejecta and DebrisIt is known that material can be ejected from the free

surface of some metals when the metals are shocked. Suchejecta from shocked Pu are being characterized in a seriesof experiments in the U1a complex at the Nevada TestSite. Over time, we will examine ejecta from Pu samples ofvarying age, surface finish, and manufacturing technique;we will also study the effects of various temporal shockstructures and intensities. The experiments are beingperformed in high-explosive-driven shock systems verysimilar to that illustrated in the figure, which is theconfiguration for a Livermore subcritical experiment (SCE)code-named Bagpipe.

The first Livermore SCE, called Holog, was executed inSeptember 1997. All diagnostics fielded on Holog weresuccessful, except for the particle sizing measurementperformed by laser holography. The holographicmeasurement failed because scattered light severelyoverexposed the film in the detection system. A solution tothe overexposure problem was developed and successfullytested in an experiment at Livermore Site 300 on June 26,1998. The utility of this solution on Pu samples is beingverified with the Bagpipe subcritical experiment performedon September 26, 1998. If the overexposure problem hasindeed been eliminated, the Bagpipe experiment willprovide the particle size data that was lost in the Hologejecta experiment.

The Bagpipe experiment also measures the Pu surfacevelocity and Asay foil ejecta data that were collected inHolog. In addition to measuring the ejecta particle sizedistribution, the Bagpipe experiment is providing

information on the reproducibility ofexperiments and the diagnostic techniques usedon Pu samples.

Following Bagpipe is Clarinet, an SCEdesigned to characterize ejecta differencesbetween new and aged wrought Pu. Two ejectaexperiments are included in Clarinet, each ofwhich will be functionally equivalent to the oneillustrated in the figure below. One of theClarinet experiments will use a piece of wroughtPu from an over-30-year-old sample. The secondexperiment will use a new piece of wrought Pumanufactured with the same specifications asthe aged sample. Both experiments will beexecuted simultaneously and diagnosed withFabry-Perot velocimetry to measure both thesurface velocity and the velocity history of theAsay foils from which ejecta mass/velocitydistributions are inferred. Each experiment willalso be diagnosed with holographic particlesizing to obtain ejecta particle size distributions.Finally, depending on the quality of the resultscollected on Bagpipe, piezoelectric probes maybe used on Clarinet and could corroborate theAsay foil measurements. Clarinet is scheduled tobe executed in the first half of FY99; the date isdependent on the availability and capabilities ofLLNL’s plutonium facility.

A future SCE will include a third experimentof the exact configuration as the Clarinet, but

will use a Pu sample made by the castpit-rebuild process. This experimentwill provide a direct comparisonbetween cast and wrought Pu.Additional future experiments areoutlined in the Experimental Studiesof Ejecta and Debris plan.

Laser

Mask

Pu target: 1.5 mm thick, 25-mm radius, 46 g

Detonator

High explosive main charge: 20-mm radius, 20 mm long, 50 grams

Steel housing

Film

Tamper

Coarse Pu finish

Gold coating (2.5 µm)

Asay foil

Piezo probes• Infer ejected mass and velocity distributions

Holographic particle sizing• Measure particle size distribution • Infer particle velocity

Asay foil velocimetry• Infer ejected mass and velocity distribution

Fabry-Perot velocimetry• Free surface velocity

Schematic of the ejecta experimentalconfiguration designed for Bagpipe. Manyfuture ejecta experiments, like Clarinet, willemploy qualitatively very similar designs. Asappropriate, the experiments will bediagnosed using velocimetry of the Pu freesurface to measure the shock arrival time vsthe free surface velocity history, holographicparticle sizing to measure the ejecta particlesize distribution, velocimetry of Asay foils toinfer the ejecta mass/velocity distribution,and piezoelectric probes to obtain a secondmeasure of the ejecta mass/velocitydistribution.

10 Pits • LL03


The Miniflyer (Pu and U): High-Strain-Rate Measurements of Pit Materials

The laser-driven miniflyer (LDMF) experimental methodprovides the unique capability of evaluating small samples(<500 mg) and obtaining performance data that simulatespecial nuclear material dynamics. In particular, the LDMFwill quantify the effects of self-irradiation on the dynamicspall strength. Delta-phase Pu surrogates were tested thisyear, and additional tests are planned. These data assist inthe basic understanding of radiation-induced defects andwill be used in determining the useful life of a weapon pit.The preparation for the first Pu dynamic spall strengthmeasurement is in progress.

During the first half of FY1998, equipment was installedin the ESP Miniflyer Laboratory, located at Los AlamosNational Laboratory. The Nd:YAG laser-beam profile wasmeasured, and related diagnostic equipment was installed.Optical and electrical signals and timing circuits weresynchronized. The streak camera with optical fiber inputand charge-coupled device readout to a microcomputerwas interfaced for velocity interferometry signals. The90/10 optical splitters fabricated by the Kansas City Plantwere incorporated for electrical and optical recording ofvelocity interferometry signals. The electrical recordingprovides long (>3-µs) recording at 2-ns resolution. Thestreak camera has shorter record lengths at <200-pstemporal resolution. The small flyers and targets used inminiflyer experiments require the improved temporalresolution of the streak camera.

Experimental Results with non-Pu materialsInitial high-strain-rate spall experiments have been

conducted with copper, gold, 316SS, other specialalloys, and metal–polymer interfaces. Theseexperiments used laser-launched flyer plates (miniflyers)25–100 µm thick by 3-mm diameter at velocities from150 to 700 m/s. Each target’s thickness was twice theflyer’s thickness. The experimental results werecorrelated with spall experiments by traditionalexperimental methods using much larger flyer platesand targets. The extrapolation of traditional data to ourminiflyer data yielded expected results.

Experimental Results with δ-Pu surrogateA surrogate δ-Pu (Sδ-Pu) was prepared for use in the

LDMF to evaluate the dynamic spall strength effects of

self-irradiation damage. Two important materialproperties affecting radiation damage of metalswere studied: crystal structure and mass density.Different populations and types of radiation-induced defects can occur in different crystalsystems. Because a material’s mass density affectsthe distance that an energetic particle travelsthrough a solid, the concentrations of radiationdefects differ. Delta-Pu has a mass density of15.8 g/cc and a face-centered cubic structure.Surrogate δ-Pu samples were spalled using theminiflyer technique. Surrogate δ-Pu flyer plates20 µm thick were impacted into Sδ-Pu targets40 µm thick. The spall strengths of nonirradiatedδ-Pu targets and Sδ-Pu targets irradiated to anequivalent 20-year He concentration weretested. Four shots were conducted: three withnonirradiated targets and one with an irradiatedSδ-Pu target. In preliminary tests, the spallstrength was lower by approximately 30% forthe irradiated sample. Additional tests ofsurrogate Pu will be performed over the next fewmonths. Below, a spall plane in a target and arecovered flyer plate are shown. The spall planeis approximately 20 µm deep in a 40-µm target,as expected with a 20-µm-thick flyer plate of thesame surrogate δ-Pu. The connected voids in theplane are typical of the spall process.

A cross section (top)of a 40-µm-thicksurrogate δ-Pu targetafter impact of a 20-µm-thick surrogateδ-Pu flyer plate(bottom). Theconnected voids forma spall plane (plane ofmaximum tensilestress). Irradiationlowers spall strength.

Pits • LA09 11


Dynamic Behavior of Plutonium and Uranium(Intermediate Strain Rates)

The Kolsky-Hopkinson pressure bar is designed to measurethe uniaxial compression stress-strain characteristics of specialnuclear materials (SNM) in a glovebox environment at highstrain rates and over a wide temperature range (cryogenic upto near-melting temperatures). To date, plutonium sampleshave been tested at strain rates of 800 to 10,000 per secondat room temperature.

Experimental data from SNM are used to construct andvalidate constitutive material models. These models aredesigned to provide predictive capabilities for simulating thehigh-rate deformation response of weapon materials to meetweapons performance and safeguard program milestones. Inaddition, experiments provide a quantitative characterizationtechnique for evaluating the influence of aging, processing,microstructure, and composition on the mechanical behaviorof plutonium-based and enriched-uranium alloys. This datacannot be accurately extrapolated from either quasi-static orequation-of-state (EOS) measurements and must bemeasured directly.

Los Alamos National Laboratory’s Kolsky-Hopkinson BarDynamic Test Facility became operational in December 1997,and the first series of plutonium samples were successfullytested immediately thereafter. The Enhanced SurveillanceProgram provided the critical support needed to initiate thisunique operation, the only one of its kind in the U.S.

Cylindrical SNM samples, typically 5 mm in diameter by5 mm long, are “sandwiched” between two 10-mm-diameterby 1.3-meter-long, high-strength steel pressure barsinstrumented with strain gauges, as illustrated in the figure.To perform a test, a low-pressure gas breech drives a short“striker” bar of variable length into one end of the “incident”pressure bar. The striker impact generates a well-defined, 50-to 150-microsecond-long elastic compression pulse thatpropagates through the incident bar, the sample, and finallythe “transmitter” bar before its momentum is trapped. Adynamic test is considered successful when the stress pulseuniformly deforms the entire sample at a constant strain rate.

This facility is unique because SNM samples,along with short sections of the two pressure bars,are contained within a glovebox to containradiation and any potential safety hazards. Flexibleglovebox seals isolate and separate the remainderof the bar-breech system and the roomenvironment from the test area within theglovebox.

The uniaxial deformation behavior of SNMsamples is quantified using stress-strain strain-ratecurves calculated from waveform signals acquiredby strain gauges fixed midway along the two longpressure bars. Experiments at a given temperatureare performed using different striker velocities(controlled by the breech pressure) and striker barlengths that yield a series of stress-strain curves atdifferent applied strain rates and total strains,respectively.

The compressive stress-strain mechanicalbehavior of a number of weapons-relevant SNM(both “baseline” and stockpile-aged plutoniumand enriched uranium) are being measured over awide range of intermediate strain rates andtemperatures to support the development ofpredictive constitutive models and to allowassessment of the mechanical response of SNM asa function of age. Tests can be performed attemperatures down to –196°C (77°K) by placingthe sample in a bath of liquid nitrogen or attemperatures up to near the melting point byusing a resistance-heated furnace placed aroundthe specimen.

Thermally activated dislocation kinetics formthe physical basis for material constitutive models,such as PTW (Preston-Tonks-Wallace) and MTS(mechanical threshold stress), which are used atLANL in large-scale, finite-element system

calculations of weaponperformance. Taken collectively,

low-strain-rate data, Kolsky-Hopkinson bar results, 40-mm plutonium EOS tests,and experiments at theNevada Test Site U1acomplex form an integratedapproach to characterizingthe mechanical response ofweapons materials needed tosupport our surveillance,remanufacturing, andrecertification missions.

N2

εε

ε I

R

T

Fast digitizing oscilloscopes capture strain gauge signals at

a sampling rate of 107s-1

Low-pressure gas launcher

50 psi

Momentum stop Strain

gauge Strain gauge

Transmitter barIncident

bar

Striker bar (velocity=10–30 m/s)

Pu sample

Zone 1 glovebox

A schematic of the operation of the Kolsky-Hopkinson bar and the three bar strainwaveforms used in the analysis of samplestress, strain, and strain rate.

12 Pits • LA12


Local-Structure Analysis of Gallium-Stabilizedδ-Phase Plutonium

We are developing and applying new methods foranalyzing extended x-ray absorption spectroscopy (EXAFS)and x-ray and neutron pair-distribution function (PDF) tomore accurately and reliably extract local structuralinformation about aging in δ-stabilized Pu. We havefocused on possible local structural defects and/oradditional phases in the total model structure. We areanalyzing gallium (Ga)-edge EXAFS and the plutonium (Pu)EXAFS edge. The latter is more difficult to analyze becauseof weaker and noisier data. At low Ga concentrations, thereappears to be an additional peak at about 3.8 Å, which wehave shown cannot be attributed to the face-centeredcubic (fcc) structure nor to the atomic distortions aroundthe Ga atoms. The most likely explanation for this peak issmall domains of a second Pu phase and additionaldisorder. We have introduced these domains into our dataanalysis and are looking at their influence on the EXAFS(and PDF) spectra.

We are in the process of developing a new method tomore reliably extract all of the information available inEXAFS experimental data and to reduce effects of noise on the interpretation. Pu-edge data show that the peak at 3.8 Å in low-Ga samples is a true structural feature.

The fit to the Ga edge data [(a) in figure] in both new and aged samples shows that its local structureis fcc, but with the first three shells of Pu atomscontracted around the Ga atom. Multiple-scatteringcalculations indicate that the distortions are radialcontractions of these shells towards the Ga atom.The fit gives a 0.14-Å contraction of the first shelland about 0.05 Å for the second and third shells. At1 wt% Ga and above the three shells of Pu, atomsaround different Ga atoms should start to overlapeach other.

The Pu edge data for both old and new samples[(b) and (c) in figure] show fcc peaks, but for low-Ga-concentration samples the degree of disorder inthe different shells cannot be explained by thermaldisorder alone. The data also show an additional

well-formed peak in the Fourier-transformedmoduli of the EXAFS χ function. We performedsimulations of the Pu edge including thedistortions around the Ga atoms. As expected,the results show increased disorder in the fccshells but cannot account for the additional peak.A body-centered cubic (bcc) or body-centeredtetragonal (bct) structure with a similar density asthe fcc one can produce a second-shell peak atthis distance with its first-shell bond lengthapproximately equal to that of the first-shell ofthe fcc structure. We are currently studying thepossibility of an fcc structure with a distributionof small bcc or bct domains included in a modelconsistent with the presence of the distortionsaround the Ga atoms. Because the α phase has3.6-Å bond lengths and is expected to bepartially present, we are also modeling thispossibility. The most difficult aspect of thesesimulations is studying the disorder associatedwith the domain boundaries from one phase toanother. The existence of anomalous bondlengths can also come from regions involvingslips, stacking faults, or internal twins.

1 2 3 4 5 6 70

10

20

30

ExperimentFit (fcc)

1 2 3 4 5 6 70

10

20

| FT

[ k3

χ(k)

] |,

a.u

.

| FT

[ k3

χ(k)

] |,

a.u.

| FT

[ k3

χ(k)

] |,

a.u

.

1 2 3 4 5 6 70.0

0.1

0.2

0.3

0.4

0.5Experiment

Theory, afcc = 4.613 [Å]Theory, abcc = 3.76 [Å]

Pu, L III edgeGa, K edge

R, Å R, Å

R, Å

1 wt% Ga

(a) (b)

(c)

Pu, L III edge,

2 wt% Ga

0.5 wt% Ga

ExperimentFit (fcc)

Pits • LA29 13


Experimental Ga K-edge XAFS data (a) and a fit to itwith a radially collapsed local fcc Pu structure aroundthe Ga atoms. The region that was fit is in between theblack vertical lines. Experimental Pu LIII-edge XAFS data(b) and a fit to it with an fcc structure. Experimental PuLIII-edge XAFS data (c) and fcc fits with the three-shellPu distortion around the Ga atoms. Fits for a bccstructure are also given for comparison.

Measurements by X-Ray Synchrotron RadiationWe are developing capabilities that will yield reliable

benchmarks for evaluation of aging weapon-componentmaterials in beryllium, steel, and uranium. The spatial andthe in-gauge volume resolutions of the x-ray synchrotronradiation (XSR) diagnostic technique in bulk material canbe micrometers and cubic micrometers to millimeters,respectively. The penetration depth of XSR in beryllium iscomparable to that of neutrons. Using measurable materialproperties (such as XSR-measured changes in strains andpolycrystallinity in beryllium) and an appropriate modelingtool, we may be able to predict the service lives ofcomponents.

The purpose of this project is to develop a high-flux, XSRmicrodiffraction capability, in the energy range of 11 to80 KeV, for measurement of strain and stress, elastic andplastic deformation, and polycrystallinity in the materials ofinterest with an extremely high degree of resolution andprecision. The technical feasibility of measuring residualstrain and stress in thick specimens of beryllium under loadand no load conditions has been demonstrated.

A specific objective of this work is to conduct studies ofmicromechanical deformation (elastic and plastic) in thepolycrystalline beryllium and steels at the grain level. The

measurements of residual strain and stress aswell as polycrystallinity at grain and subgrainlevels in polycrystalline bulk beryllium and 304Lspecimens under load and no load conditionswill be of high priority in our task plans forFY1999.

The results from such investigations will beused to develop polycrystal modeling tools forimproving or modifying a state-of-the-art finite-element analysis code for predicting the strainand stress profiles in materials and components.

Through mapping of strain and stress andmeasurement of polycrystallinity in criticalregions of beryllium compact tension (CT)specimens of differing textures, theunderstanding of fracture mechanisms inberyllium near room temperature appears to have been advanced as a result of ourFY1998 work.

We have measured the strain field at andaround the crack-tip of fatigue precracked anduncracked specimens of 6-mm-thick S-200 Dand P31664 beryllium CT-LT and/or CT-TLspecimens under load conditions (T and Lindicate the grain orientations from texture inpolycrystalline beryllium). Analysis of the data isunder way. As a result of this work, we haveshown that strain fields can be mapped incritical regions of beryllium CT specimens underload with spatial resolution of 0.175 mm and in-gauge volume resolution of ~0.001 mm3. Withan XSR microprobe beam, the spatial resolutioncan be extended to ~1 mm.

We are able to measure strains over a grain(or subgrains) in polycrystalline berylliumspecimen under load. This opens up thepossibilities for the micromechanicaldeformation studies to be conducted inberyllium at the grain level.

Finite-elementcalculations of the(from top) hoop,radial, and axialstrains in aspecially-welded21-6-9-steel part.

14 Pits • LA32


Resonant Ultrasound Study of Pu AgingResonant ultrasound spectroscopy (RUS) is a precision

probe used to measure elastic properties. It provides boththe highest absolute accuracy and the best relative accuracyof any routine elastic modulus measurement technique andcan be used on smaller samples in less time thancompeting techniques. The Los Alamos National Laboratorysystems operate from 4°K to 900°K to measure thecomplete anisotropic elastic tensor of a material over thepractical engineering temperature range, as well as theentire range of solid Pu. This technique is being applied toPu polycrystal samples now; later it will provide the firstcomplete measurement of the elastic tensor of single-crystalPu over time and temperature.

The elastic tensor of Pu, in both single- and polycrystalsamples, is a particularly sensitive test for the effects ofaging. It is an input parameter for the modeling of Puperformance in weapons and a crucial first test formicroscopic models of Pu.

For the study of aging, the measurement of elasticstiffness is extremely sensitive to such quantities as Heingrowth, defect density, composition, and impurity levels.For example, the elastic stiffness of a system whose specificvolume has changed by 0.1% because of various agingeffects will exhibit as much as a 3% shift in elastic stiffness,30 times bigger than the volume effect. Thus, elasticitymeasurements can provide the first warning of changes. Inaddition, an RUS measurement is basically a frequencymeasurement; thus very high precision is achieved, asshown in the figure. It means that in a small sample, whereradioactive heating temperature can be controlledaccurately, it will be possible to observe, in real time,changes in elastic moduli over the course of a few days asradioactive decay proceeds.

For engineering applications such as evaluating theperformance of Pu under forming and casting operations,Pu presents unique problems as it ages or comes underdynamic stress. Because the shear stiffness of single grainsvaries by a factor of seven with direction—the largest shearanisotropy of any face-centered cubic metal—it is important

to understand the variation of this anisotropywith Ga concentration and temperature in poly-and single-crystal samples. The elastic moduli alsocontrol the compressional and shear soundspeeds, of importance to an understanding of thedynamics of Pu. Results of a recent measurementof Ga-stabilized δ-Pu show values within therange predicted from earlier work.

This experimental work will also provide thecalibration and input needed for microscopicmodeling of Pu. Something not included incurrent modeling efforts is how entropycontributes to Pu stability. In contrast to mostother materials, the low specific volume of Pucauses a very low Debye temperature. Thus if thefree energy, which determines the stable phases,is F=U–TS, where U is the zero-temperatureinternal energy usually calculated in microscopicmodels, T is temperature, and S is entropy, thenwe find that S is very large and dominated by thecontribution from the very low elastic moduli ofPu. Coupled with the ease with which Pu can becompressed (the bulk modulus is small) this maybe why Pu can contract on warming, and the keyto understanding the two highest-temperaturephases of pure Pu.

0.444 0.445 0.446 0.447 0.448 0.4490.0

0.2

0.4

0.6

0.8

Single peak

Q=5700 RUS measurement of mechanical resonances of Ga-stabilized δ-Pu polycrystal chip

Am

plit

ude,

mV

Frequency, MHz

Pits • LA34 15


A typical resonance of a polycrystalline sample of Pu.The very sharp resonance and the low noise indicatethat part-per-million precision is possible.

Pit Surface CharacterizationNew diagnostic techniques in the Pit Surface

Characterization Laboratory (PSCL) at the Pantex Plantwere used in FY1998 to evaluate the chemistry of theexternal surface of the pit cladding material. During themanufacturing process, while in storage, and during theassembly process, pits are exposed to many environmentsthat can leave residues on the external surface that maynot be detected when a weapon is built. The residuescould, under certain conditions during the lifetime of theweapon, react with other materials in the weaponenvironment causing corrosion of the pit cladding.

One example of the application of the new diagnosticcapabilities in the PSCL is the determination of thechemical composition of a milky-white, cake-likecontaminant on the surface of aged pits from the enduring

stockpile. This contaminant resulted fromcontact of the metal cladding of the pit with aneoprene gasket of the pit fixturing. The analysisrevealed the milky substance to be a cake-likematerial, primarily a mixture of zinc oxide andstearic acid, which is a mold-release agent usedin fabricating the neoprene gasket. Thiscontaminant is not corrosive to the pit and hasno detrimental effects on the metal cladding ofpits for the enduring stockpile; it will not affectthe weapon’s reliability, safety, or performance.

In some of the fixturings, neoprene rubbergaskets are pressed against the surface of pits.Neoprene is a chloropolymer. It is made by theaddition of HCl to vinylacetylene (CH=CCH-CH2), producing chloroprene [CH2=C(Cl)-CH=CH2], which will subsequently polymerize toneoprene. Residual chlorides will be trapped inthe polymer; the amount depends on theconditions in which the polymer is produced(e.g., humidity). The concern is potentialcorrosion of the pit cladding caused by thepresence of the chloride in the neoprene.

This study addressed whether corrosion had,in fact, occurred on pits from the enduringstockpile as a result of aging in the presence ofneoprene rubber gaskets. We examined pits thathad been in contact with the gaskets for over sixyears with the surface analytical techniquesavailable in the PSCL. Data from areas on the pitsurface that were touched by the neoprene werecompared to those of other areas away from theneoprene. In addition, samples lifted from thepit surfaces were analyzed elsewhere by energydispersive spectroscopy and x-ray photoelectronspectroscopy (XPS). These techniques will beavailable in the PSCL in the near future.

Pits from three different programs wereanalyzed. Upon removal of a pit from thefixturing, an imprint was found on the pitsurface. The imprint is a ring locatedapproximately halfway between the pole andthe equator (or waist) of the pit. The digitaloptical image of the imprint (left) shows that theband in the impression left by the neoprene ringis approximately 6 mm wide. Analysis of themilky-white, cake-like material revealed thesubstance to consist of zinc, carbon, and oxygen(zinc stearate). A small amount of silicone wasnoted. But no corrosion promoters (e.g.,chlorides) were detected on the milky-whitesubstance in contact with the pit surface, and nocorrosion of the metal cladding was detectedfollowing removal of the milky-white substance.

14000

5

10

15

Inte

nsi

ty c

oun

ts (

×100

0)

20

25

30

35

1200 1000Binding energy, eV

Carbon auger

Zinc 2p

Mg 1s

Zinc auger

Carbon 1s

Si 2sZinc 3s

Zinc 3pMg 2p

Mg auger

Zinc 3d

800 600 400 200 0

Mold release from neoprene

Si 2p

Oxygen 1s

Oxygen auger

Optical micrograph of neoprene imprint (top). XPS spectra from themilky-white material, a zinc stearate mold-release agent employed in themanufacture of neoprene gaskets (bottom).

16 Pits • PX01


Mold releasefrom neoprene

Celotex

0.5 mm

High Explosives

18


High Explosives

Focus AreaThis focus area is intended to assess the effects of

aging on the initiation, detonator/booster chain, andmain-charge high-explosive components used in nuclearweapons systems in the enduring stockpile. It includescharacterization of the physical changes that may occurwith aging (including physical, chemical, andmechanical, as well as safety and explosive performanceaspects), development of nondestructive examinationtechniques that allow early and accurate detection ofchanges that may occur, and the development ofmodels that allow projection of aging and performancebehavior into the future. These models will allowestimation of the lifetimes of existing components sothat reasonable planning for potential refurbishment orremanufacturing can be made within the Stockpile LifeExtension Program.

The work on initiation systems comprises thedevelopment of a new nondestructive examination

technique used to assess aging of initiation cables. Thedetonator/booster work consists of an evaluation of theperformance aspects of old detonators that remainrelevant to those in continuing service. A new test willevaluate divergence between the performance of“retired” detonators and those that will continue to bein the stockpile.

The effort surrounding main charges consists ofassessments of the initiability (how promptly and wherethe detonation reaction begins), energy (how the highexplosives move metal), and safety (the change ofsensitivity with time). The work also includes necessaryassessments of the high explosives as engineeringmaterials.

These tasks were chosen to enhance our ability topredict high-explosive aging phenomena and remainingservice lives. As appropriate, results of the work will betransferred to the routine surveillance testing program.

Pertinent TasksLA35 High Explosives Initiability and StabilityLA36 High Explosives Mechanical PropertiesLA37 High Explosives Lifetime PredictionLL06 Material Degradation Model DevelopmentLL07 Enhanced Detection of Environment-Driven Changes in Aged High ExplosivesLL08 Understanding the Effect of Damage and Aging on Initiation and Performance of High ExplosivesLL09 Detonator and Booster ReliabilityLL10 Safety and Reliability of Stockpile-Aged MaterialsLL11 Enhanced Characterization of Stockpile-Aged Initiation ComponentsLL12 Mechanical Properties of Aged PBXs and OrganicsLL13 Solid-Phase MicroextractionLL18 Changes in Shock Wave ProfilesLL38 Science-Based Modeling of the Gold-Indide Reaction and its Effect on the Performance of Aging DetonatorsLL40 Simulated Stockpile-Aged ExplosivesLL43 Detonator SafetyLL44 Advanced Analysis of Mechanical Safing and Arming Device Strong-Link Performance

19


Deliverables1. Partial discharge analysis test for initiation cablesdelivered to the routine surveillance program.

2. Analytical and test-fire comparisons of new and agedPETN, LX-16, LX-13, and ultrafine TATB explosives.

3. Models to predict future performance of PETN, LX-16, LX-13, and ultrafine TATB.

4. Development of a new divergence test for stockpileaged HE and the introduction of this new test into theroutine surveillance program.

5. Embedded magnetic impulse and velocitycomparison of new and aged main-charge highexplosives as well as transfer of the test to the routinesurveillance program.

6. Skid tests to evaluate aging effects on the safety ofmain-charge explosives.

7. Three-inch gun tests to evaluate effects of aging onthe safety of main-charge explosives; introduction ofthis compact and analyzable test into practice in theroutine surveillance program.

8. Characterization of aging effects on materials-relatedproperties of PBX 9501, PBX 9502, XTX 8004, LX-04,and LX-17. Tests include molecular weight andmechanical properties such as strength and ductility.

9. New nondestructive techniques to allowcharacterization of aging phenomena and theintroduction of these technologies into the routinesurveillance program.

10. Characterization of aging effects on the explosiveperformance of PBX 9501, PBX 9502, LX-04, and LX-17main-charge high explosives as well as development ofmodels to predict future performance behavior.

Material Degradation Model DevelopmentHigh explosives (HEs) with slow chemical reactions

require a long time to achieve chemical equilibrium, whichposes a challenge in developing accurate equations ofstate. The ideal explosive, LX-14, was studied as abenchmarking case, because it is known to have a shortreaction zone and therefore very little kinetic effect.Calculations were performed with no adjustment specific tothe bigplate experiment or the individual explosives used.These calculations are therefore predictions of the explosiveperformance based only on heat of formation and density.We found that we were able to reproduce the time ofarrival of the shock wave at the beam positions (0, 2, 4 and8 cm) very well. We have linked the Cheetah chemistrycode to Lawrence Livermore National Laboratory’s VHEMPhydrodynamic code. Cheetah linked to the hydrodynamiccode calculates the explosive equation of state as a functionof elemental composition, heat of formation, and density.This will give us the ability to predict how explosiveperformance in a warhead is affected by changes in densityor chemical composition upon physicochemical aging.

For LX-14 and also for stockpile explosives LX-04 andLX-17, we undertook an extensive comparison of theVHEMP-linked hydrodynamic code to results of the bigplateexperiment (LL16, which is the experimental analogue ofthis task). The bigplate experiment is designed to be asensitive test of the high explosive equation of state. Thetest measures explosive push at variable angles ofincidence. A 20-mm-thick disk of high explosive with a100-mm radius was initiated with a point detonator at thecenter. The other side of the explosive was covered with a0.5-mm-thick copper plate. Several Fabry-Perot laser

interferometer beams were aimed at varying radiion the plate.

We compared our calculations to bigplateexperimental results on LX-04, which is found inthe W62 warhead. In the figure below, wecompare the chemistry predictions with theexperimental values for the velocity of thecopper plate. LX-04 shows significant non-idealbehavior due to the relatively large amount ofbinder (15% by weight). We see reasonableagreement between the calculations and theexperiment. We have obtained similar results forLX-17, which is found in the B83, W84, and W87warheads. We are now implementing a kineticcapability into the linked code, which will takeinto account the effects of slow chemical agingreactions. We have performed trial calculationsusing kinetics in one-dimensional plate pushexperiments with very good results.

In conclusion, we have established andvalidated a capability to predict changes in HEperformance due to aging-related changes indensity or chemical composition. While resultsare generally promising, we find that theinclusion of chemical kinetics will be necessaryfor obtaining quantitative accuracy for LX-17 andLX-04. In the next year, we will develop modelsof observed subtle shifts in chemical compositionand/or density. Observed changes of chemistryand density have been very small to date.Density changes will be explored by developing

a model of small changes in part size anddensity due to mechanical loading overtime. Chemical changes will be exploredby developing models of binderdecomposition and high explosive agingchemistry. The aging models will beextrapolated forward in time, then fed intothe linked VHEMP/Cheetah code. Theresult will be an indication of whetherexplosive performance will changesignificantly in the foreseeable future.

0 5 10 15Time, microsecond

0

1

2

3

4

Vel

oci

ty, k

m/s

Experiment Thermochemistry predictions

0 cm

2 cm

4 cm 8 cm

The copper plate velocities for the LX-04bigplate experiment as measured by Fabry-Perot interferometry at four radii (0, 2, 4, and 8cm) are compared with the predictions of thelinked chemistry/hydro code. Good agreementbetween calculation and experiment wasfound, thereby validating the new linked code.


20 High Explosives • LL06

High Explosive Initiability and StabilityWe are characterizing the initiability, energy, and equation

of state (EOS) of stockpiled high explosives (HEs). First wegather state-of-the-art baseline data for reference andsubsequently characterize stockpile-returned samples foraging-related changes that might affect weaponperformance, reliability, and safety.

We completed a series of Steven tests (formerly known asspigot-gun experiments) that detect changes in HE sensitivityto low-amplitude mechanical environments. Ongoingactivities include: (1) skid tests to detect changes in sensitivityof bare HE charges, (2) development of sandwich test as apotential replacement for cylinder tests measuring HEperformance and EOS, (3) embedded electromagnetic gaugeexperiments to detect subtle changes in HE initiability, (4)hockey-puck experiments using proton radiography tocharacterize low-temperature (<–55°C) initiability anddetonation-wave spreading in PBX 9502 for hydrocode andJTF reactive burn model calibration, and (5) XTX 8004initiability/energy experiments at the Pantex plant.

The Steven test was originally developed atLawrence Livermore National Laboratory tocharacterize the threshold for HE response to a low-amplitude impact (few hundred feet per second)such as might be experienced during a transportationaccident. Los Alamos National Laboratory slightlymodified this test and added instrumentation tocharacterize the reaction timing profile and violence.We tested PBX 9501 samples in stainless-steel holdingfixtures, which were impacted by mild-steelhemispherical projectiles. Projectile velocity wassystematically varied to determine PBX 9501 reactionthresholds. Reaction violence was measured usingredundant blast-overpressure gauges and a ballisticpendulum developed for this test. Spigot projectilevelocity vs. the average energy release is normalizedto a full detonation of PBX 9501. The results indicateno significant difference in the sensitivity of baselinePBX 9501 and W76 samples after 142 and 182months in the stockpile. The slight differenceobserved for reaction thresholds has tentatively beenattributed to variations in sample densities.

We also developed embedded electromagneticgauge experiments as an ultra-sensitive measure ofthe initiability (i.e., ease of inducing a detonation fora given input shock) of HE materials. For this test, aparticle velocity/shock-tracker gauge package isinstalled in the HE sample to minimize shockperturbations. The test sample is located in amagnetic field and each of 11 gauge elements(located at 11 depths) records the local particlevelocity. The shock-tracker records the position ofthe shock front with time. A gas-gun projectile isused to introduce carefully controlled shockconditions (i.e., pressure/duration) into the HE

sample. We characterized the initiation behavior ofbaseline PBX 9501 at three different densities usingthree different input-shock pressures. As timepermits, we will also test PBX 9501 stockpilesamples; the bulk of stockpile testing will be inFY1999. Baseline data will be compared to dataobtained using PBX 9501 samples from stockpiledW76 and W78 warheads (which have very differentstockpile-to-target sequence environments) todetect age-induced changes in initiation behavior.Thus far, acquisition of baseline data is nearlycomplete and one W76 sample has been tested.Presently there are insufficient data to determinewhat effect, if any, age has on the initiation behaviorof PBX 9501. Baselining experiments will begin inFY1999 using PBX 9502 samples, and stockpiletesting will be completed in FY2000.

Ave

rag

e en

erg

y re

leas

e, p

erce

nt

Baseline 1.830 g/cc W76 1.840 g/cc (182 month) W76 1.840 g/cc (142 month)

Projectile velocity, ft/s

Part

icle

vel

ocit

y, k

m/s

1.75

50

40

30

20

10

0

–10

1.50

1.25

1.00

0.75

0.50

0.25

0

30 40 50 60 70 80 90 100

0.250 0.50 0.75Time, µs

1.00 1.20 1.50 1.75


Steven test results (top) indicate no significant difference in the sensitivityof baseline PBX 9501 and two W76 samples in stockpile. Baseline data andresults of W76 sample tests (bottom) show that there are insufficient data todetermine what effect, if any, age has on the initiation behavior of PBX9501.

High Explosives • LA35 21

PBX 9501 LifetimeThe work of this task provides information to continuously

update and improve the lifetime predictions for PBX 9501 inthe W76, W78, and W88 warheads. These activities werecompleted during FY1998:•Formulation of nominal PBX 9501 using flaked vs. pelletized Estane for comparisons and certification•Study of effects of formulation parameters on PBX 9501•Study of NP migration mechanism*•Characterization of NP migration into W78 shields•Thermal sensitivity study comparing freshly formulated and stockpile samples of PBX 9501*•Measurement of glass transition temperature of Estane, Estanewith NP and PBX 9501 historical samples•Measurement of affinity of PBX 9501 for H2O at differing relativehumidity•Measurement of radiation effects on PBX 9501 binder*•Estane aging study to determine degradation products•Study of the effects of “hot pressing” PBX 9501 on the molecularweight (MW) of Estane*•Study of effects of quasi-static compressive loading on Estane MW•Study of inhibitor lifetime in PBX 9501•Characterization of Estane MW in PBX 9501 in stockpile•Gathering of historical information regarding PBX 9501 manufacturing

Italicized activities were performed to support developmentof a PBX 9501 lifetime model. Details on selected activities arepresented below.

All 14 lots of stockpiled PBX 9501 wereformulated with flaked Estane from B.F. Goodrich.Goodrich now produces a pelletized Estane thatmeets all specifications for use in PBX 9501, but isslightly different. If it becomes necessary to replacestockpiled PBX 9501, a decision on whether touse reserved flaked Estane (acquired in 1989 andsomewhat deteriorated) or newly obtainedpelletized Estane will be required. Therefore,Pantex formulated PBX 9501 using both flakedEstane (500 lb) and pelletized Estane (1000 lb) fora direct comparison of properties and war reserve(WR) certification for the W76. Both lots of PBX9501 have been characterized analytically, andW76 main charges are being manufactured fortesting. LANL conducted a parametric study offormulation parameters for PBX 9501 to fullycharacterize manufacturing techniques.

The primary aging mechanism for pure Estaneand PBX 9501 outside (and perhaps inside)weapons is hydrolysis. The affinity of PBX 9501 forwater is critical for predicting lifetime. Therefore,pressed pieces and molding powder of PBX 9501were stored in air at five different humidities andanalyzed for water content (see the figure below).Pressed pellets retain more water than moldingpowder, a surprising result.

One first step in developing an aging model forEstane and PBX 9501 is identifying degradationproducts to gain insight into reaction mechanisms.A two-year study at Pantex examined Estanesamples aged at 50, 75 and 90°C under air orhelium, both dry and moist. Estane samples werefractionated and analyzed to qualitatively identifydegradation products.

All stockpiled PBX 9501 was characterized forEstane MW by Pantex and Los Alamos NationalLaboratory. Extensive Pantex investigationsrevealed correlations between historical reports ofEstane MW and specific instruments used at theplant. This correlation is seen by examining theEstane MW of one lot of PBX 9501 used forinstrument checkout at Pantex since the early1970s (lot 622-7) and historical stockpile data(see LA24). Lot 622-7, which is not stabilized andtherefore expected to age more quickly than WRPBX 9501, exhibits discontinuities in MW trendsassociated with instrument changes (see figure,left). These discontinuities lead us to concludesurveillance data alone cannot be used to quantifyany age-related increase in Estane MW. Instead,refined analytical techniques must be used todetermine if an MW increase occurs and theimportance of that aging mechanism.

H20

/PB

X 9

501,

µg

/g 1200

800

400

0

0% RH 20% RH 52% RH 76% RH 100% RH

PBX 9501 powder

PBX 9501 pellet

50000

90000

130000

170000

210000

Waters 200 Micromeretics Waters 150 Waters 150-11-51

Wei

gh

t av

erag

e M

W

0 5 10Age, years

15 20 25


PBX 9501 pellets retain more water than PBX 9501 power, indicatingthe former may age more quickly (top). Historical Estane molecularweight (MW) data from PBX 9501 lot 622-7 exhibits discontinuities inMW trends associated with instrument changes (bottom).

22 High Explosives • LA37

*ongoing

Two related tasks study the effects of age on highexplosive (HE): Task LL07, Enhanced Detection ofEnvironment-Driven Changes in Aged HE, and Task LL40,Simulated Stockpile Aging of Explosives.

Task LL07 deals with enhanced detection ofenvironmentally driven changes in aging HE; it continuesto evaluate and develop modern tools to measure thechemical and physical signatures of aging explosives. Theobserved aging phenomena are beginning to provide thebasis for simulating aging stockpile explosives—a focus forthe new Task LL40. The following techniques developed inTask LL07 represent promising metrics for the simulatedaging of LX-17:• The increase in the atomic pore volume previouslyobserved in the aging of LX-17 seems to occur when TATBis thermomechanically stressed. When the relativepopulations of the defect sizes are compared for all thesamples measured to date, we observe a power lawcorrelation. This may provide a method for determiningthe amount of thermomechanical aging that has occurredin a historic stockpile unit, as well as provide metrics forthe simulated aged explosives.• Thermal techniques allow the crystallinity of the binderin LX-17 to be determined; crystallinity influences thestiffness of the overall composite. These techniques canalso indicate how long the binder has been above its glasstransition temperature and provide a record of the LX-17’sthermal history since the last thermal excursion. Other PBXsystems with crystalline fluoropolymer binder systems(LX-04 and LX-16) can employ this technique as well.• Highly charged, ion-based secondary ion massspectrometry provides a means of making an imaged,solid analysis for trace contaminants in aged PBX. Thermaldecomposition products of TATB were quantified forpressed LX-17.

A new task for FY1999 is Task LL40, which isfocused on the production of explosives having thesame properties as those in the aging stockpile.These explosives will be used for tests in associatedtasks to determine age effects on materialproperties and explosives performance and safety.That data will then be incorporated into agingmodels, being developed in yet other tasks, toobtain a capability for predicting aging effects andlifetimes of explosives. The lack of adequate agedexplosives and the unavailability of aged samplesolder than 10–20 years have significantly deterredthe development of this predictive capability. Weplan to fill this void by producing explosives thatsimulate a variety of ages.

Simulated aged explosives will not be producedby simply accelerating thermal and otherprocesses. The explosives we produce will have thesame chemical and physical signature(s) as thosedetected in stockpile-aged explosives. Efforts arenow underway to study the detected signaturesand identify the relevant ones. We expect toinclude some or all of the signatures detected bytools and techniques developed in LL07.

The pathway to simulated aged explosives maybe entirely different from the actual agingpathway. For example, to study the effect ofobserved changes in binder crystallinity, we mayformulate explosives with binders having differentinitial degrees of crystallinity. The ultimate fidelityof the aging simulation will only be as good as thesignatures that we select. Within that constraint,however, this task offers the means to extend theunderstanding of explosives aging far beyondwhere we now stand.

Feature lifetime, ps1000

Sig

nal

inte

nsi

ty, %

1

10

100

Age/stress test seriesIrradiation seriesHeating, ratchet growth series

300

Stockpile life test

Ground test

Unaged/control materials

Power law correlation between TATB defectannihilation lifetime values and lifetime intensity.The apparent aging phenomena are unaffectedby irradiation and give the impression thatunder thermomechanical stress, the smalldefects coalesce into large ones.


Detecting and Evaluating Age Effects on Explosives

High Explosives • LL07/LL40 23

Impact ignition is a common measure of how we assessthe safety of explosive materials. Impact ignition generallyinvolves one or more of the mechanical processes offriction, shear, fracture, recompression of damaged orreacting explosive, etc. These processes are extremelydifficult to quantify in terms of the work done on theexplosive, the amount of explosive on which themechanical work was done, and the temperatures reachedwithin the heated regions called “hot spots.” Previousimpact tests, such as the drop hammer test, the skid test,and the Susan test, were developed to test specific hazardscenarios. But the results are not consistent between thevarious tests, most of which are not amenable toquantitative analysis and computer modeling. In contrast,the Steven impact test was developed to be consistent,instrumentable, and modelable. We completed a 55-experiment test series using this test and determinedthreshold impact velocities for observable exothermicreaction for aged HMX-based explosives to within a fewmeters per second. Previous work concentrated ondetermining threshold impact velocities for pristine anddamaged high explosive charges and inert modeling of themechanical aspects of the test. Concern has been raisedthat the handling, shipping, and environmental andchemical changes that main-charge high explosives (HEs)undergo during their stockpile lifetime may change their

impact sensitivity. Therefore, new and stockpile-aged charges of four HMX-based conventionalHEs and the two TATB-based insensitive HEs weretested in the Steven test configuration. Wedeveloped improved instrumentation to yieldquantitative data on the energy release ratesduring reaction. Reactive flow modeling iseffective for explaining these experimentalmeasurements and for developing predictivemodels for other scenarios, most importantlyhere, aging.

Unlike other impact tests, the Steven test andthe Los Alamos version of the Steven test, thespigot gun test, yield distinct critical velocities,below which no reactions are observed. Thecritical impact velocities increase in the usualorder of decreasing impact sensitivity: PBX 9404,LX-10, PBX 9501, and LX-04. Thus far, noimportant effects of aging on critical impactvelocity for LX-04 or the violence of thesubsequent reaction have been observed.Embedded manganin, carbon resistance, andcarbon foil gauges have yielded pressure historiesin the impacted explosive charges lasting forseveral hundred microseconds. The figure showstypical embedded gauge records in the impactregion and at the explosive charge boundaryobtained in a Steven test of a 95% HMXexplosive impacted near its threshold velocity.

We modeled the ignition of the explosive inthe Steven test in several ways. We added a shearignition mechanism to the ignition and growthreactive flow model in the ALE3D hydrodynamiccode, which yielded reactions under theprojectile at approximately the correct impactvelocities. We also performed quantitativeignition and growth reactive flow modeling forthe individual HMX-based explosives in theDYNA2D code using an ignition criterion similarto the 0.37 cal/cm2 frictional work criterion ofChidester and the growth of reaction parametersdetermined for low-pressure shock initiation inLX-10 and LX-04. The figure shows excellentagreement between the pressure histories froman ignition and growth calculation and theexperimental pressure gauge records in terms ofthe pressure history in the impact region and thetime to violent reaction. We will soon bepublishing details of the experimental andmodeling results. The ignition and growthreactive flow models for LX-04 and LX-10 arebeing used to predict hypothetical impactscenarios on aged material that can not betested experimentally.

3.0

2.5

2.0

1.5

1.0

0.5

0

Pres

sure

, kilo

bar

s

5004003002001000Time, microseconds

Ignition and growth model results (calculated reaction time = 335 µs)

Carbon foil gauge record in projectile impact region

Carbon resistance gauge record located at edge of explosive (measured reaction time = 360 µs)

Experimental and calculated pressure histories for the impact ignition of a 95%HMX explosive near its reaction threshold velocity.


Safety and Reliability of Stockpile-Aged Materials


In the past, the nuclear weapons complex has used apass/fail detonator cable high-pot test to determinedetonator cable electrical integrity. Unfortunately, this testprovides no qualitative information. In contrast, partialdischarge analysis (PDA) is a qualitative, nondestructiveelectrical technique for verifying cable integrity andforecasting cable life. This technique applies a voltage tothe cable under test and records micro-discharges (partialdischarges) in the cable. The charge distribution,frequency, and location of the partial discharge (PD)signatures convey information about the internalstructure of the cable. This technique can be used toanalyze the structure of a cable and forecast the life of thecable. It can also be used to screen new production andto determine if assembly/disassembly operations aredamaging the detonator cables. This project is developingthe tools and database necessary for technology transfer.

PDA can be conducted between the two conductorsof the cable or from one or both conductors to ground.Through FY1998, we demonstrated the viability ofconductor-to-conductor measurements, includingdielectric degradation of material through electrical agingand stockpile weapon aging.

The figure shows three snapshots of PD signatureswhile a cable has voltage applied for an extended period.This figure clearly shows that the signatures tend towardhigher charge and smaller peak counts as the componentages. Testing also identified the weakest points of existingdesigns and quantified degradation with time in cablesaged in storage and in weapon environments.

Conductor-to-ground measurements are inherentlymore difficult because metal parts are attached after the

assembly is manufactured, leaving voids andPD sites even when the cable assembly ispristine. In FY1999, we will eliminate sourcesof extraneous PD signatures by performingthe test in a vacuum that is below the corona region.

The requirement to compare signatures ofcables at various stages of life is pivotal to thePDA project. As a result, it is crucial to havesignatures of new and aged cables available.Based on FY1998 results, AlliedSignal/FederalManufacturing and Technology hasincorporated PDA into their manufacturingquality assurance process and is acquiring thenecessary database of new cable signatures.

Analysis of aged cables is also essential toan effective forecasting capability. In additionto surveillance returns, a large number ofcables made available from the Life ExtensionProgram will be tested in FY1999. Analyseswill include mechanical damage testingrepresentative of possible weapon assembly/disassembly mishandling, as well as materialdegradation testing.

Partial discharge analysis is a powerful toolthat has many potential applications in theweapons complex, including forecasting ofminimum cable life, identification of designweaknesses (points to improve future designs),improvement of quality assurance of newcables, and quantification of age- and damage-related changes to the cable insulation.

This shows theprogression of the partialdischarge signatures as acable is agedelectronically. Note thatthe signatures containcounts of higher chargebut lower peak counts asthe cable ages. There areno counts for a typicalpristine cable.


Enhanced Characterization of Stockpile-Aged Initiation Components

0

200

400

600

800

1000

1200

1400

1600

0 20 40 60 80Picocoulombs

Co

unts

Pristine Early partial discharge (PD) signature Mid-time PD signature PD signature near failure

High Explosives • LL11 25

Mechanical Properties of Aged PBXs and OrganicsWe are establishing the capability to conduct

mechanical testing required for characterizing advancedmaterial aging constitutive models for plastic bonded highexplosives (PBXs) and other organic materials used in thestockpile. Our objective is to compile a database of time-dependent changes in the mechanical properties of PBXs asa function of their respective stockpile aging conditions.The mechanical properties of interest include stress/strainto failure as a function of strain rate and temperature,short-term creep tests as a function of temperature andstress, long-term creep tests in weapon environments,stockpile-to-target sequence thermal-mechanical simulationand acceleration; biaxial shear/compression, and fracture.

Although present surveillance testing was not designedto generate such a database, we have made excellentprogress toward this goal.

In FY1997, we began building a mechanical testingfacility within Lawrence Livermore National Laboratory’sHigh Explosives Application Facility (HEAF). We alsodeveloped the PBX environmental test unit, which permitscontrolled environmental exposures to the mechanical testspecimen during loading.

In FY1998, we also began an extensive program togenerate the advanced materials database. Because only

small amounts of material are available, wecontinue to perform most development workusing the conventional high explosive LX-14

(95/5 HMX to urethane binder) and mock highexplosive (HE). Some mechanical data have beengenerated on LX-04 (85/15 HMX to Viton A),and a very limited amount of data have beengenerated on LX-17 (92.5/7.5 TATB to Kel-F 800)at Pantex. A thermal strain rate testing matrix hasbeen developed and is being systematicallygenerated for each PBX critical to LLNL programs.We presently have the capability to test in thestrain rate ranges of 10–5 ≤ ε/sec ≤ 100 andtemperatures from –50 ≤ T ≤ 120°C. Mechanicaldata for LX-04 are shown in the figure.

Compression testing of PBX has proven to beextremely difficult to do well. We are working toimprove specimen alignment because we believealignment and machining tolerance errorsassociated with this test have been principalsources of scatter in the compression data in thehistorical stockpile data. We are helping thePantex Plant, Los Alamos, and the U.K. AtomicWeapons Establishment minimize the sources ofdata scatter. We now use three strain gaugesarranged 120° apart and average their results, asopposed to using a single extensometer.Minimizing data scatter will result in the need touse fewer test specimens.

To maximize the use of stockpile material, wehave been developing reduced-sized testspecimens for tension and compression tests andtwo new specimens for fracture and biaxialshear/compression tests. We believe the reduced-sized specimens are successful; however, we haveyet to determine the overall improvements fromactual warhead components. We have recentlysecured the first W62/LX-04 part that will becored for optimization.

We have successfully demonstrated the use ofthe biaxial shear/compression specimen withmock HE. To perform biaxial testing of PBX inLLNL’s HEAF, it will be necessary to develop atorsional actuator cell for one of the existing axialtest machines.

A schematic of the new test unit, the new mechanicaltest cell, and a plot of change in the tensilestress/strain response of LX-04 at 22°C for increasingstrain rates. The higher strain to failure as the strainrate increases is an unusual material response, one wehope to explain and ultimately model.


0

500

1000

1500

0 0.002 0.004 0.006

Ten

sile

str

ess,

psi

Strain, in./in.

3.37/sec

0.01/sec

0.0001/sec

1.25 × 10–5/sec


Organic Materials

Deliverables1. Complete parts retrieval from dismantled systems;expand parts acquisition to all extended stockpilesystems.

2. Run stress-relaxation tests for M-9750, S-5445,and S-5455 cushion materials.

3. Quantify thermal dependence of stress relaxationin silicone foams and elastomers and compare tosurveillance data.

4. Continue aging mechanism identification andcharacterization studies for additional prioritizedmaterials.

5. Produce a report describing the mechanicalbehavior of PBX-type explosives and other weaponsystem organic materials and how they are affectedby stockpile aging.

6. Apply nonlinear computational analysis to masterkinetic equations in PBX 9501 systems and refinethose master equations to include the new body ofEstane aging data as they become available.

7. Incorporate both diffusion of small species andspatial inhomogeneity into the above computationalanalysis.

8. Produce a description of an accelerated-agingmethodology for stockpile organics.

9. Perform quantum chemistry, molecular dynamics,and other more detailed calculations to answerotherwise unanswerable questions about Estanechemistry.

Pertinent TasksLA06 Polymeric Materials Aging ProgramLA24 Theory and Simulation of Polymer AgingKC07 Foams, Cushions, and SealsLL35 Compatibility and Predictive Modeling of

Weapon Organics and Polymers

28


Organic Materials

Focus AreaThis focus area addresses the assessment of

materials that are both nonenergetic and nonmetallicin character. The various elements that compose theOrganic Materials Focus Area provide for thedevelopment of baseline materials properties andcharacterization of stockpile-aged and accelerated-aged materials to support the development ofpredictive aging models. While the family of organicmaterials slated to remain in the enduring stockpile isquite broad in scope, emphasis in this focus area ison a limited number of polymers or polymercomposites that dominate organic material concernsand have been assessed to have a significant impacton weapon safety or reliability.


Polymeric Materials Aging ProgramThe aging of polymers and high explosives may well be

the limiting factor for the extension of the stockpile’slifetime. This task is looking at the 60 types of polymericmaterials and 200 polymeric components in the stockpile.The primary objective of our Polymeric Materials AgingProgram is to enhance the surveillance of polymeric weaponmaterials such that the viability of the stockpile can continueto be assessed with confidence. In order to meet thisobjective, we must understand how materials age and theimpact these changes have on the safety, performance, andreliability of the stockpile. Accomplishments in FY1998include experimental and theoretical efforts directed at theaging behavior of flexible polymeric foam. Some of thiswork has relevance to a broad variety of materials includingother polymers.

Two foam aging models were developed in FY1998. Thefirst model is derived from the statistical analysis of a nine-year aging study conducted at Los Alamos NationalLaboratory early in the 1970s. Silicone foam was compressedin fixtures and maintained at a constant temperature. Thefoam samples were removed from the compression fixturesevery year and compression load-deflection tests conducted.The sequence of aging under compression, unload, andrecovery followed by compression testing is similar to someaged parts in the stockpile that are tested in the existingsurveillance program. A model has been developed basedon data from the historical study that predicts the load-deflection response of the silicone foam material as afunction of material thickness, temperature, storedcompression, and time. The model also predicts theunrecovered thickness or compression set the material willexperience as it ages, which may be a life-limiting featurefor some parts. Predictions from the silicone foam agingmodel are being compared to surveillance data to indicatelong-term trends and identify inconsistentdata. As shown in the figure, the modelpredictions agree well with the data takenfrom stockpile aged parts.

The second foam aging model wasderived from stress relaxation studiesconducted on similar material. Whenpolymeric materials are strained or forcedinto a different shape, the resistive forceswithin the material decay with time. Theconstitutive properties of a polymericmaterial are likely to be different duringand after compression even when veryshort recovery times are allowed. Stressrelaxation studies are being conducted toimprove the understanding of howconstitutive properties change while undercompression. Preliminary results indicatethat data collected over short timeperiods can be very accurate in predictingbehavior over much longer periods of

time. Stress relaxation data collected over a 10-second period accurately predicted stressrelaxation data generated by an independentstudy conducted over a two-year period. Themethodology being developed has been appliedto a variety of materials in the open literature andtherefore may be applicable to other polymericmaterials that experience mechanical loading.

Although these two aging models are versatileand can be applied under a broad set ofconditions, they are only of value if the materialcan be quantitatively linked to the safety, reliability,and performance of the stockpile. To explore thisconnection, an ABAQUS-based constitutive modelis being developed to represent the material in aweapon configuration.

A variety of efforts are underway at AlliedSignal/Federal Manufacturing & Technology (AS/FM&T),Pantex, and Y-12 including material testing andaccrual of production process information.AS/FM&T personnel assembled crucial productioninformation on filled vinyl copolymer elastomer, orVCE, and are performing solvent swell experimentson filled silicone materials. Efforts at Pantex includetime-staged compression tests on stockpile agedparts. These tests will help establish a rational fortime limits on surveillance tests. Personnel at Y-12identified all canned subassembly polymericmaterials, are conducting tests to characterizeconstitutive properties, and supplied productionprocess information.

All of the information and data generated fromthese efforts are being compiled into material-specific handbooks to aid future users.

00

10 20 30 40 50 60

Load

ret

enti

on

Years

1.0

0.8

0.6

0.4

0.2

Load retention inference region calculated from the foam aging model. The lightgreen region represents the average change in load retent expected. Surveillancedata is shown in red.

Organic Materials • LA06 29


Theory and Simulation of Polymer AgingA detailed chemical model is being developed for the

aging kinetics of Estane 5703, which is used in the binderof the plastic bonded explosive (PBX) 9501. Whencomplete, the model will be able to reliably predict thechemical composition and molecular weight of Estane totimes longer than the current ages of any of the weapons.Its output will be used in other models being developed tocalculate and predict the changes in mechanical propertiesof the PBX with time.

Estane can degrade by hydrolysis, oxidation, free radicalattack, and thermal degradation. In accelerated-agingexperiments, any of these mechanisms can be made todominate. Our objective is to determine which dominatesin the mild conditions of weapon storage, and muchprogress has been made in FY1998.

We have observed that the stabilizer concentration inPBX 9501 appears to be slowly decreasing with time insamples stored in air but to show no net change with timein the stockpile. The differences between the stabilizerconcentration in the stockpile and its initial nominal valueappear to arise during initial processing. With the modeland the fact that no other component is present in smallconcentration, this implies the important conclusion thatthere is no catastrophe looming ahead; Estane in thestockpile will continue to age gracefully.

The acidities and the molecular weights of aging Estanelibrary samples have been measured at Pantex. Modelingthem shows that the rise in acidity is enough to account forall the molecular weight loss. That implies that hydrolysis isthe dominant mechanism for the degradation of Estane inair, even for conditions as mild as room temperature andambient humidity. The treatment of water sorption andhydrolysis of Estane in the kinetics model have beenimproved, and a single model now describes the aging ofEstane in air under a wide range of temperature andhumidity conditions.

Spectroscopic data of aged Estane wereobtained and analyzed by our collaborators atPantex, the University of Alabama, Los AlamosNational Laboratory, and Lawrence LivermoreNational Laboratory. These definitely identify theoligomeric (small fragment) products of theaging of Estane in moist air and also give theratio of the absolute molecular weights of theoligomers to their polystyrene equivalentmolecular weights obtained from gel permeationchromatography (GPC).

We have measured the sorption of water byPBX 9501 and have found that, in the form ofmolding powder, it absorbs a factor of two tothree times less water per gram than it does as apressed pellet and than neat Estane does (seeLA37). With the model, this lower sorptionaccounts for the fact that, when stored in airunder ambient conditions, Estane in PBXmolding powder degrades much more slowlythan neat Estane.

Finally, Pantex has analyzed Estane molecularweight (MW) data from surveillance samples tosee the effects of changes in GPCinstrumentation over time. Their results areplotted in the figure. The Waters 200 gavehigher values for the molecular weight for bothnew and aged Estane than do all the newerinstruments. Also, the MW of new Estane hasbeen remeasured and found to be higher thanpreviously thought. No instrument showed anyinitial rise in MW. Additional analysis indicatesthat some cross-linking can occur, but likely it isnot enough to make the weight average MWincrease. The calculated line in the figure showsour best current estimate for the actual behavior.

200000

180000

160000

140000

120000

100000

300200100Time, months

080000

Mo

lecu

lar

wei

gh

t

Micromeretics Waters 200 Waters 150 Waters 150/11-51 Model calculations

Estane molecular weight ofsurveillance samples versus time. Thedifferences between measurementswith different instruments should benoted, especially for new samples.

30 Organic Materials • LA24


Cellular Silicone Cushion Aging StudyCellular silicone cushions, such as those shown below,

serve several functions. They fill gaps between components,compensate for manufacturing tolerances of adjacentcomponents, and allow for thermal expansion. For thecushions to perform these jobs successfully, they arerequired to exert a specific compressive force atpredetermined maximum and minimum gaps. Because thecellular silicone cushions must fill these gaps for the life ofthe weapon, the long-term stress behavior of the cushionunder load is an ongoing concern.

AlliedSignal/Federal Manufacturing & Technologies hasconducted studies of various cushion materials in the past.At the end of those tests, the samples were left in the testconfiguration at specific compression and storagetemperatures. This task resumes testing of these existingaged parts, which are now up to 16 years old.

Nine different materials have been retested at least onceduring FY1998. These samples are representative ofmaterials in all weapons in the enduring stockpile and inseveral development formulations.

When the cushions were originally placed into the testingfixtures, the force loads required to close the fixtures wererecorded. At selected later time periods, these force loadswere remeasured to determine how much they hadreduced over time. The difference is reported as apercentage of the original fixture closure force, or percent

load retention (PLR). High PLR indicates thatnearly the same amount of force was required toclose the aged fixture as was initially required. Inpractical terms, high PLR indicates that there islittle compression set in the cushion, and that itsproperties have not significantly degraded duringfixture aging.

The results of tests on S-54 material show acontinuation of the slow decline in load retention.The material had been stored for 3,521 days(9.6 years) since last testing, for 15.6 years oftotal aging time, and still exhibited PLR thatfollowed the previous trend. Testing on M97-based materials also showed a slight decrease inload retention, at a lower rate than observed onS-54 materials. The material had been stored for3,609 days (9.9 years) since last testing, for15.3 years of total aging time, and still exhibitedPLR that followed the previous trend. The othermaterials included in this study have all beentested within shorter time periods, and all havePLR that are basically following anticipated trends.

The data were reported as plots in a topicalreport. Copies of the report have been sharedwith interested technical staff throughout theweapons complex.

100

23°C

50°C

70°C

90

80

70

60

50

400 800 1800 2400 3200 4000

Time, days

Perc

ent

load

ret

enti

on

Example of aging study results (left) and representative cellular silicone cushions.

Organic Materials • KC07 31


Compatibility and Predictive Modeling of Weapon Organics and Polymers

The goal of this new project, which began mid-yearFY1998, is to develop a complete and predictiveunderstanding of the long-term compatibility and agingphenomenology of nonenergetic organic components thataffect weapon performance. Our strategy is threefold:• We identify material properties that affect performanceand determine the sensitivity of performance to changes inthese materials. Based on this assessment, we developmaterial-unique roadmaps to identify, characterize, andmodel key aging signatures by characterizing stockpile-aged components of known pedigree. • We use knowledge gained in the above strategy toestablish appropriate experimental protocols to simulateand accelerate the aging of organics for validating models. • We address the long-term component performance issuesby collecting, analyzing, and preserving corporateknowledge of weapon organics; refining our understandingof each material’s unique aging environment; anddeveloping an effective strategy to retain and preserveorganic weapon components for future study.

We believe this strategy provides a focus on materialsissues important to weapons and a path for developingappropriate new surveillance tests that will provide muchearlier detection of changes in critical material properties.The strategy also improves understanding of the materialscience and chemistry of weapon organics as related toaging and compatibility phenomenology.

Traditional methods to evaluate the servicelife of organic weapon components are basedon the measurement of physical or mechanicalproperties of these materials over time or underaccelerated aging. This project develops asystematic methodology for identifying theunderlying chemical signature changes thataccompany stockpile-aged materials and isbased on the correlation of molecular chemicalalterations to the corresponding physical andmechanical property changes that can bedetected before gross mechanical changes haveoccurred. This approach addresses the short-termquestion of how a material is aging and thelonger-term issue of identifying predictivemodeling tools to determine when chemicalalterations in a material become substantialenough that macroscopic performance is inquestion. We are also addressing the potentialeffects of materials interactions that may occurover long times in the stockpile.

The techniques we use include: solid-phasemicroextraction and gas chromatography/massspectrometry to determine volatilization duringdecomposition or chemical structurerearrangement, positron annihilationspectroscopy to determine atomic free volume ordefect size distribution in the bulk material,infrared spectroscopy to determine changes inmolecular functionalities, electron spin resonanceto characterize free-radicals that might beassociated with long-term radiation damage, andscanning electron microscopy and atomic forcemicroscopy to determine any morphologicalchanges. AlliedSignal and Y-12 will supply agedsamples of materials and conduct selectedcomplementary experiments on these materials.

FY1998 progress included retrieving stockpile-aged organics from surveillance units, identifyingcomplex-wide expertise to preserve corporateknowledge, identifying design engineers whocan identify material properties affectingperformance, defining plant and lab tasks forfuture work, and evaluating experimentaltechniques. Techniques evaluated include low-level oxidation, solid state cross polarizationnuclear magnetic resonance, and material-specific solvent swelling.

0

100

200

300

0 20 40 60Time, days

Toluene only

Toluene + NH4OH

Tolu

ene

swel

ling

, p

erce

nt

wei

gh

t g

ain

Characterization of the degradation of cellular silicone isconfounded by the effects of the reinforcing filler fumed silica.The exposure of swollen polymer to ammonia eliminates thereinforcing effect (disrupts hydrogen bonds) and thus allowscharacterization of the polymer for evidence of degradation.

32 Organic Materials • LL35

Dynamics

Deliverables1. Fiber-optic velocity sensors.

2. Picosecond microscopy.

3. Laser holography.

4. A screening test to rapidly and inexpensively assessaging effects on high-explosive performance, andintroduction of this test into the routine surveillanceprogram.

5. Bigplate and smallplate tests on new and agedmain-charge high explosives.

6. Hydrodynamic tests using improved diagnosticsfor used, stockpile-aged components obtained duringdisassembly.

Pertinent TasksKC02 Fiber-Optic Velocity SensorsLL15 Development of New Hydrodynamic

Measurement and Diagnostic ToolsLL16 Aging Effects on Explosive Performance

34


Dynamics

Focus AreaThis focus area assesses the effects of aging on the

integrated hydrodynamic performance of nuclearweapon components. The work consists of thedevelopment of new diagnostic techniques to betterassess test results, precision characterization ofexplosive performance in high-explosive components,and improved utilization of hydrodynamic testing.The new diagnostic techniques and improvedmethods for hydrodynamic testing will be introducedinto the routine surveillance testing program, whilethe precision characterization is part of the assessmentof component lifetimes.

Work on new instrumentation techniques includesdevelopment of fiber-optic velocity sensors (pindomes) to better characterize the shock-front velocityand shape, multibeam Fabry-Perot velocimetry tobetter characterize behavior of large high-explosivecomponents, and picosecond microscopy to bettercharacterize behavior of small components. Thisadvanced instrumentation will be utilized inperformance tests of both new and aged high-explosive components, varying widely in size.Methods that allow hydrodynamic testing using themaximum possible number of aged components willbe developed, demonstrated, and introduced into theroutine surveillance testing program.


Fiber-Optic Velocity SensorsWe are developing a diagnostic tool using Doppler-

shifted light in optical fibers to measure the velocity andlocation-versus-time of high-velocity objects and shockfronts. The fiber-optic velocity sensor and recording systemwill augment diagnostics currently used to assess theperformance of explosively driven, imploding sphericalsystems in experiments conducted by Los Alamos andLawrence Livermore national laboratories. This sensorsystem will provide continuous velocity and locationmeasurements of a point on the imploding surface (insteadof the location at one instant in time as given by theelectrical pin currently used). The sensor does not rely onthe finish or reflectivity of the imploding surface, so it canbe used with materials known to provide poor light returnunder experimental conditions.

The system consists of two components: a velocityinterferometer and a fiber-optic sensor. The velocityinterferometer is constructed with optical fibers and micro-optics. We completed a prototype unit and are using it tocollect data on explosive experiments at Los Alamos’ DXexplosive facilities. The micro-optics components are beingreplaced with fiber components as they becomecommercially available. The fiber components will improvesensitivity and permit the use of smaller lasers. An alphaversion of the analysis software was written using IDLlanguage. The GUI-based software is customized for datafrom the interferometer and fiber sensor.

Numerous sensor geometries andoptical fibers have been tested. Themost promising geometry is achievedby encasing the fiber in a thin metaltube. Repeatable signals are obtainedfrom this geometry; however, thesignals are very complex and difficultto separate because they originatefrom different locations within andoutside the fiber. Current researchfocuses on improving signal returnfrom the end of the fiber andeliminating or reducing the sources ofthe unwanted signals. To do this, weare modifying the index of refractionprofile of the fibers and material usedin construction of the fibers.Computer calculations, as shown inthe figure, are used to assist inunderstanding the data and guidefuture experiments.

The contour map (left half of the figure)illustrates the pressure profiles in a fiber 300 nsafter it is struck with an aluminum plate at avelocity of 1,200 m/s. The fiber begins to“mushroom” on the end as it penetrates theplate at 570 m/s. The end of the fiber is crushingat 630 m/s. A high-pressure pulse is launcheddown the fiber at about 4,100 m/s. Both theunwanted pressure pulses and the crushingsurface return Doppler-shifted light. Theillustration (right half of figure) shows how thecrushing of the fiber affects its properties as adielectric waveguide. As the fiber diameterexpands, more transmission modes are createdthat dissipate the laser light, reducing the returnsignal. Modifying the core profile, changing thefiber material, and encasing the fiber in metaltubes improve the light return and reduceinterference from unwanted signals.

This task is a collaboration between LosAlamos National Laboratory and AlliedSignal/Federal Manufacturing & Technologies. Thefabrication facilities at the Kansas City Plant and the explosive experimental facilities,computational capabilities, and resources at Los Alamos are used to support the research and development of this sensor system.

–0.2 0.45 1.10 1.75 2.0Pressure, GPa

Aluminum

Acrylate

Silica

Higher modes

Singlemode

Middlemodes

Pressure profiles (left) andoptical modes (above) indeformed optical fibers.

Dynamics • KC02 35


Aging Effects on Explosive PerformanceThis task uses bigplate experiments to assess the effects

of aging on high explosive performance. It is theexperimental analogue of LL06, Material DegradationModel Development, which models and predicts the effectsof aging; this work is used to calibrate the codes beingdeveloped in that task.

A bigplate consists of a disk of pressed high explosive(100-mm radius, 20- to 40-mm thickness) with a 0.5-mm-thick copper or tantalum plate bonded to its front side.A point detonator on axis at its back is fired to create aspherical shock wave that interacts with the copperstraight-ahead on the axis and tangentially at the edge.Five Fabry laser beams measure the velocity on the frontedge of the copper at different positions. The figure showsa snapshot of bigplate blowing outward 14 µs aftercommencement of detonation.

Shots with Fabry laser velocimetry were made of twoLX-14 and five LX-04 sensitive high explosives and three

LX-17 insensitive high explosives. They measuredthe velocities to an excellent 0.066 mm/µs, or1%. The explosive causes the metal to suddenlyaccelerate in velocity, and these jump-offs weremeasured to 0.01 µs, the smallest number everachieved.

Our most unusual finding was that out of10 shots conducted in the series, 7 showedunexplained 0.1-µs lags between the jump-off at0 mm and at 80 mm. We hypothesized that thedetonation wave is slowed down by frictionalong the inner face of the plate as it movesoutward to 80 mm. This effect is not currentlyrepresented in our codes.

All samples showed unexpectedly large metalspall (i.e., internal breakage), which extended theon-axis first plateau time by 0.1 µs and created a“rebound” bump on the second plateau (seefigure). The spall diminished to near zero at theplate edge. Spall is thought not to affect explosiveenergy flow; our current codes do not treat it.

The fired explosives span the range of nearlyideal (LX-14 reacts almost instantly) to non-ideal(LX-17 reacts slowly), with LX-04 lying inbetween. The traditional explosive descriptionsused in the codes fit the ideal explosive well andthe non-ideal explosive badly. Code runs for LX-17 always show a low initial velocity becauseof this explosive’s long reaction time. TheCheetah Kinetic code, developed in Task LL06, isintended to provide code input to fix this time-dependence problem.

Bigplate is unusual in having fairly simplegeometry but with the shock wave crossingzones at various angles on the diagonal. Thismakes the code runs more sensitive than thebetter known cylinder, 1D plate and rib tests,where the shock wave always hits parallel to thezone edge. Thus, explosive code packages thatcalculate well for other experiments fail onbigplate.

The same 22-year-old batch of LX-04 was shotwith recently pressed and old samples, and nodifference was seen within the 0.05 mm/µs bandshown in the figure. Additional cylinder workshowed no change in detonation velocity ordetonation front curvature between recentlypressed and aged LX-04. This suggests that anaccelerated aging methodology will be neededbefore degradation effects can be seen inexplosives.

LX-04/Cu bigplate at 14 µs(right). The copper plate ismoving outward at 3 mm/µstoward an aluminum barthat will soon be blownaway. The center of thecopper is bowed because thedetonation commenced “on-axis.” The explosion hasreached the edge and sootygas from the explosive iscreeping to the front. Withinerror, no difference is seenbetween the samples fromthe same batch of LX-04, twosamples pressed recently andtwo pressed 22 years ago.

1.50 1 2

2

2.5

3

Vel

oci

ty,

mm

/µs

Time, µs

Aged LX-04

Freshly pressed LX-04

36 Dynamics • LL16

Diagnostics

38


Diagnostics

Focus AreaThis focus area includes the development of a new

suite of diagnostic methods and instruments that willaid in the early identification of aging problems inweapon system components and materials. New andmore sensitive techniques are necessary to measurematerials and component properties and the effects ofby-products, such as out-gassing of organic materials.The feasibility of real-time monitoring and self-awaresystem measurements is also being evaluated forincorporation into shelf-life testing programs.

The motivation is to improve the ability to detectpotential problems that may require refurbishment orrelated action. New techniques and methodologiesthat are developed will be made available to theroutine surveillance testing program.

This focus area is divided into the following elements:• Ultrasonics• Tomography• Mechanical testing and sensor development• In-situ diagnostics• Characterization and analysis• Pit surface analysis and monitoring

Advanced ultrasonic inspection methods are beingdeveloped for assessing the quality of bonds inweapons. These tools provide improved capability to

resolve interfacial features for early detection ofpotential defects such as corrosion and cracking.Neutron radiography and high-definition x-raytomography inspection tools are being developed forinspecting entire assemblies. These tools will providemuch-improved capability to resolve fine features.Vibration and shock testing, as well as other newpromising techniques for detecting materialsproperties, will diagnose damage and structuraldefects in weapon components.

This focus area also includes in-situ monitoring forshelf-life testing. Engineering feasibility studies ofthese strategies will determine their practicality, defineappropriate detection concepts, develop neededtechnology, and evaluate prototypes to support animplementation decision.

Characterization tools and analytical methodologiesare another element. This set of tools will allow theidentification of early signs of degradation of organiccomponents and the potential effect on organic andmetallic components. The suite of tools will be appliedto nondestructive examination of the gasses withinthe warhead space of stockpiled weapons and will beapplied in a similar fashion to stockpile cannedsubassemblies (CSAs).

Pertinent TasksKC10 Sensor Technology DevelopmentLA16 High-Energy Computed Tomography Detector ArrayLA18 Bayesian Inference TomographyLA19 Advanced Ultrasonic Methods for the StockpileLA20 Low-Level Vibration and Advanced Signal ProcessingLA21 Development of SQUID Microscope for SurveillanceLA26 Analysis of High-Resolution Vibration for Damage IdentificationLA28 Residual Stress PredictionLL13 Solid-Phase MicroextractionLL19 Enhanced Ultrasonic Characterization of AssembliesLL26 High-Resolution X-Ray Tomography

39


Deliverables1. Ultrasonic inspection tool with better resolutionthan currently available.

2. Operational neutron imaging system capable of detecting millimeter-sized voids or structural defectsin low-Z materials.

3. High-definition x-ray tomography system with at least 1-mil resolution for detection of defects inpartial and complete weapon assemblies.

4. Set of vibration and shock test procedures optimal for diagnosing structural defects.

5. Toolbox of signal analysis techniques fordetermining structural characteristics from test data.

6. Superconducting quantum interference device(SQUID) microscope to measure and map extremelyweak magnetic fields associated with defects orcorrosion.

7. Damage-detection system using high-resolution vibration analysis.

8. Extended state-of-the-art welding process models benchmarked against experimental data.

9. Laser-based technique for residual stress measurement in reservoirs.

10. Suite of fiber-optic sensors for monitoring of weapons materials and stockpile weapon systems.

11. Optochemical sensors with neural networks that measure, record, and broadcast the presence of various chemicals.

12. Miniature, nonintrusive gas monitor, which canbe incorporated in rebuilds or surgically implanted in existing stockpile units.

13. Solid-phase microextraction sampling method foranalysis of organic compound decomposition and resultant effects in weapon headspace atmospheres and CSAs.

14. Fiber-optic-coupled laser for analyzing high-explosive decomposition products.

15. Prototype infrared gas analyzer to supplement residual gas analyzer.

16. Laser drilling and welding process for sampling of CSAs.

OR04 Infrared Surface and Gas AnalysisOR09 Evaluation of Hydrogen Balance MethodsOR10 Materials InteractionsOR18 Laser Penetration and Rewelding for Gas SamplingOR19 Radiographic Parameter OptimizationSN13 Systems Engineering of an Arming, Firing, and Fuzing Device TestbedSN14 Weapon State-of-Health MonitorSR02 Residual Stress Measurement of ReservoirsSR05 Enhanced Inspection of Reservoir Girth WeldsSR06 Imaging Light Elements by Gamma Resonance AbsorptionSR07 Defect Detection by TV Holography and Speckle Interferometry


High-Resolution X-Ray Tomographyinspection data and other sources. Thisinformation can be used to guide the imageacquisition strategy and to constrain thereconstruction.

We performed simulations to guide thesystem design. Some are simple ray-tracingcalculations; others are full Monte Carlo transportcalculations with secondary-particle following.The simulations show that the spread ofsecondary radiation in the scintillator limitsresolution and the spread increases sharply withscintillator thickness. Image-acquisition time isnearly inversely proportional to scintillatorthickness, so a trade-off occurs between scantime and scan resolution. Experimentally, wedemonstrated a resolution of 0.005 in., based on the measured full width at half the maximumline-spread function.

Using a prototype system, we demonstratedCT of a complete weapon pit showing featuresnot possible with conventional film radiography.In addition, we are imaging test objectscontaining welds. The objects have beenfabricated to represent small intentional“defects” in the weld, but contain no specialnuclear material.

Using simulated and actual data, we havealso improved the algorithm for reconstructingan early test object made of depleted uranium

and steel but no weld. This project led to several

important innovations. Becausefocusing the lens using x rays wastedious and imprecise, we devised an insertable focusing camera usingvisible light. It is cofocal with the mainCCD, allowing focus on a resolutionchart at the scintillator using visiblelight. We also introduced an x-rayfiducial that is fixed with respect to the scintillator. Now we cancomputationally correct for any imageposition drift that occurs during dataacquisition.

We plan to complete thetechnology demonstrations, optimizethe basic design, and then work with aprivate-sector company to construct areliable, user-friendly system for use atPantex.

Computed tomography (CT) is a method ofcomputationally reconstructing an extensive set ofprojection (transmission) measurements into a three-dimensional volumetric data set. The so-called CAT scan in diagnostic medicine is one type of CT system. Thisapproach permits inspection and characterization of three-dimensional objects without the confusion fromsuperposition of structures that occurs with simpleprojection radiography. Once it is developed, CT willprovide an important tool for full inspection andcharacterization of welds in pits.

We are developing an x-ray CT system using acommercial electron linac with 9-MeV electron energy thatproduces an appropriate x-ray spectrum and gathers datain a reasonably short time. The x rays have a substantialpopulation in the 1–4 MeV range. X rays with this highenergy are extremely difficult to image with high efficiencyand clarity.

We adopted an approach for using a scintillator toconvert the x-ray image into visible light, then using opticsto transfer the light to a charge-coupled-device (CCD)camera. A schematic of the system we have developed isshown in the figure. To protect the camera from scatteredradiation, we employ two turning mirrors in the opticalchain. The large number of pixels in the CCD enables us tosimultaneously measure x-ray transmission for more than amillion rays through the object. We typically employ athousand different rotation angles (views of the object).

The second component of this work is todevelop computer algorithms thatemploy prior data to improve thereconstruction. Much is knownabout each pit from

Turning mirror

Optical path 63–165 cmScintillator Turning

mirror

3-m source detector distance

Object

Imaging lens

Axiom AX-2 digital camera 1536 × 1024 pixels 9 µm pixel size

Axis of column array

Center line of array

Axis of rotation Schematic of the data

acquisition system used forcomputed tomography ofweapon pits.

40 Diagnostics • LL26


Advanced Ultrasonic TomographyThis project supports research, development, and

modeling of novel ultrasonic inspection techniques andenhanced imaging and reconstruction methods to assessstockpile condition. The tools under development areproviding important new insights into the analysis andinterpretation of complex ultrasonic data and will ultimatelyprove crucial for nondestructively detecting earlymanifestations of component deterioration. The work isdirectly applicable to pits and gas reservoirs.

Ultrasonic imaging offers unparalleled sensitivity to smalldefects and density variations. For instance, digitalradiographic methods are incapable of measuring densityvariations of less than 0.5%, whereas ultrasonic techniquescan detect 0.01% variations. Thus, ultrasonic techniquesoffer the possibility of both measuring and imaging subtlechanges in materials due to aging that cannot be assessedradiographically.

Ultrasonic imaging methods have evolved substantiallyfrom the techniques used in the 60’s and 70’s when thebulk of the stockpile was created. One of the mostsignificant advances has been the ability to automaticallyacquire vast amounts of data with computerized high-frequency equipment that was not available even a fewyears ago. Because parameters inherent in the fullwaveform are often indicators of subtle material conditions,if not of overt flaws, full waveform-capture ultrasonicscanning is an important tool for rapid, inexpensiveevaluation of stockpile status.

Our goals are to perfect ultra-sensitive inspectiontechniques that capitalize on vastly increased hardwarecapacity and to reconstruct the enormous quantities ofmulti-dimensional raw data into standardized forms thatcan be easily interpreted, quantified, and compared againstone other. Problems of interest include quantifying residualstress, weld inspections, bond quality assessment, and grainsize measurements. Data reduction methods underdevelopment—specifically 3D interactive visualization,velocity tomography, and time reversal imaging—have thepotential to serve as powerful enhancements toconventional ultrasonic testing.

Significant progress has been made in several areas:• A new signal processing technique for enhancingsignal to noise in highly attenuating media such as

polymer potting compounds. The multiplecomparative energy loss method identifiesdifficult-to-detect voids with contrast andresolution far superior to infrared thermographand other competing methods.• A unique ultra-sensitive capacitative pressuretransducer for measuring third-order elasticconstants. Variations in such parameters mayprovide new insight into polymer bonddegradation and aging. The capacitativetransducer is also useful for calibrating ultrasonictransducers, superconducting quantuminterference devices, and quantum sensors.• An interactive, 3D visualization and imageanalysis package for processing ultrasonicinspection data as well as data from othercommon imaging modalities (e.g., x-ray andneutron computed tomography). The softwareimports, renders, and manipulates full waveformdata from various commercial systems incommon use within the nuclear weaponscomplex. The package includes a variety of signalprocessing, image enhancement, and displayfeatures, including contouring, edge sharpening,data warping, animations, and isosurfacerendering. To support multiple imagingmodalities, we are moving toward data fusioncapabilities for enhanced defect identificationand characterization.• Initial use of newly released ultrasonicmodeling codes to simulate data acquisition fromotherwise difficult materials and conditions todetermine algorithm performance and providequantitative inverse solutions.

From left to right (scan orientation not aligned withphotograph): Phantom test piece with fabricated voiddefects; typical infrared thermography results; normalultrasonic tomography scan using interface echotechnique; post processing of the same part imagedusing comparative multiple energy loss technique toenhance signal-to-noise ratio. The absence of silastic,shown in red, is clearly visible using this method.

Diagnostics • LA19 41


Enhanced Ultrasonic Characterization of AssembliesA team from Lawrence Livermore National Laboratory

(LLNL), Y-12, AlliedSignal, and Savannah River is developingultrasonic technologies for the nondestructive evaluation ofbonds in weapons components. Advanced ultrasonics willallow the quality and strength of autoclave, solid state, andadhesive bond lines to be assessed, thereby contributinginformation on how a weapon is aging and whatremaining life it has. The technology developed for this taskincludes specialized ultrasonic scanning systems and signalprocessing and pattern recognition algorithms thatadvance our capabilities to detect and characterize bonddefects. This work will result in a demonstratedmeasurement capability that provides high-resolutionimages of bond lines that display quality variations in aweapon component. This capability will be suitable forroutine surveillance testing at DOE production sites.

Because there are four participants in this project, thefirst step was to determine the baseline ultrasoniccapabilities existing at the four sites and decide on thestandardized hardware and software needed by eachparticipant for cooperating on the work. Each site acquiredthe necessary standardized ultrasonic instrumentation anddata acquisition systems. Next, the team agreed on thecharacteristics in bond specimens that would berepresentative of potential problems in the stockpile. Bondspecimens with those characteristics are currently beingmanufactured at LLNL and other sites. These bond

specimens will be used for testing anddiagnosing new ultrasonic hardware andsoftware.

Lawrence Livermore is developing andimplementing classification techniques forcorrelating information contained in ultrasonicbond line signals with the quality of the bond.Sets of bond specimens with known pedigreeshave been acquired, and ultrasonic dataacquisition systems are being developed tocollect data about the specimens. Signalprocessing algorithms are being written andimplemented to process the ultrasonic data andreject noise. Feature extraction and selectionsoftware are also in development to identify thepertinent characteristics of ultrasonic signals forbond quality determination. The classificationsoftware is based on statistical patternrecognition and probabilistic neural networks.

During the classification algorithm trainingphase, the best classification process will bedetermined by analyzing the ultrasonic signals ofspecimens with known bond quality; that is, thebonds have been measured by strength testingand microstructural analysis. When this trainingphase is complete, the selected classificationprocess will be verified by processing a secondset of ultrasonic signals from bonds of unknownquality. Once proven, classification procedureswill be implemented into ultrasonic inspectionsystems, along with advanced methods fordisplaying the results. For instance, variations inbond quality may be displayed as differentcolors, and the displays could be either two-dimensional or three-dimensional images.

The design and fabrication of an ultrasonicscanning system for weapon components beganthis year and went through three designiterations. The figure shows a three-dimensionalrendering of the system. The ultrasonicinstrumentation has been specified andpurchased. The control computer has beenacquired. Some of the data acquisition andmotion control software has been written. Thebond classification algorithms developed by thisproject will be implemented in this scanningsystem to provide images of bond quality.Production versions of this system will be builtand installed at the appropriate sites forsurveillance testing.

Ultrasonic scanner for assessing bond quality.



Enhanced Inspection of Reservoir Girth WeldsA unique ultrasonic inspection method has been

developed and demonstrated for inspection of tritiumreservoir girth welds. This method is capable of detectingintergranular cracks in the weld heat-affected zone as wellas weld flaws such as porosity and lack of fusion in theweld bead. We developed a separate type of inspection fortwo different reservoir configurations. The LF- and M-typereservoirs have a mid-body belly band girth weld, and theK-type reservoirs have a circumferential cap girth weld.Ultrasonic inspection has the potential to monitor in-servicedegradation in the Shelf Storage and Reservoir Surveillanceprograms and detect welding defects during fabrication ofnew reservoirs.

A conceptual drawing of the ultrasonic inspectionmethod for the LF- and M-type reservoir girth weldinspection shows that a water immersion method is usedto provide coupling of the ultrasonic signal to the reservoir.A computer-controlled inspection device simultaneouslyrotates the reservoir around its axis and indexes theultrasonic transducer in discrete increments to interrogatethe weld heat-affected zone. A mechanical gimbalmaintains a constant incidence angle of the inspectionultrasound beam to the surface of the reservoir. Ultrasonicsignal feedback is collected and correlated to positionalinformation by the computer-controlled inspection deviceand the data acquisition system. For complete inspectionof the girth weld full coverage, scans are conducted aboveand below the girth weld and in the girth weld bead. A standard 10-MHz ultrasonictransducer is used for inspection of the girth weld. The dataacquisition and processing system is capable of rendering color graphic representations of theinspection results.

The inspection method developedfor the K-type reservoir cap girthweld is a direct contact method thatcan be applied manually in the fieldor laboratory. It employs essentiallyidentical ultrasonic inspection anddata acquisition and processingequipment as the immersion method

described for the LF- and M-type girth weld, butit does not require a sophisticated mechanicalinspection device.

Accomplishments in this program to dateinclude the following: (1) The contact inspectionmethod for the K-type reservoir girth weld hasbeen fully developed and demonstrated in thelaboratory on calibration standards. Fieldinspection procedures have been written andinspection technicians have been trained andcertified in the inspection method. Five K-typereservoirs have been sampled from the stockpilefor inspection in a field laboratory. (2) Calibrationstandards for the LF- and M-type reservoirs havebeen machined, and the immersion inspectionmethod has been successfully demonstratedmanually in the laboratory. (3) A programmableimmersion inspection device with computercontrols has been located at the Savannah Riverultrasonic inspection development laboratory,and software and hardware upgrades required tofully automate the inspection system have beenidentified and cost estimates obtained.

This project is a collaboration with the LosAlamos National Laboratory. Funding is beingleveraged with the Reservoir SurveillanceOperations and Shelf Storage Program conductedat the Savannah River Technology Center.

Water

Ultrasonic reference

block

Ultrasonic transducer

Inde

x an

d gi

mba

l Schematic of ultrasonicinspection method forLF- and M-type reservoirgirth welds.

Diagnostics • SR05 43


High-Energy Computed Tomography Detector ArrayThis year, much progress has been made in developing

a high-energy computed tomography (CT) detector arrayand capability. Specific accomplishments include:• Acquisition of components.• Integration of hardware with prototype control/dataacquisition software.• Design and fabrication of necessary cables, boards,boxes, etc.• Evaluation of several collimator designs and simulation oftheir performance.• Selection of final collimator design and fabrication of asingle channel for actual testing.• Development of data acquisition and control softwaredesign and testing of prototype software.• Development and testing of reconstruction software inconcert with other programs. The software was found tobe compatible with all our detectors and sources.

The last four items deserve additional comments andexplanation.

We carefully considered the design of the collimators toyield the best performance possible. Three basic designswere proposed and individually evaluated. Some of thepoints considered were the feasibility and cost ofmanufacture, collimator performance, and ease of use.The “ideal” design, a 0.3-millimeter air bore with a solid

tungsten collimator, was dismissed after severalmachine shops across the country reported thatthey could not fabricate the parts required. Theselected design is a 0.3-millimeter beryllium borewith a collimator made of both powdered andsolid tungsten. Performance simulations on thisdesign show it to be nearly ideal and not alimiting factor in the performance of thedetector. This design was fabricated and tested.

Data acquisition and control software hasbeen designed to be interoperative with ourreconstruction and rendering software, thusreducing the cost and effort associated with thiscapability. The user interface for all our CTsystems have been unified in this effort. Theagreed upon data standard has already provedto be very robust and flexible in that it hasaccommodated several additions andimprovements without the need to redesign thedata and file structure.

The reconstruction software necessary toprocess the data from the detector array hasbeen designed to be common to all of ourdetectors and sources, and its development hasbeen supported by other programs. Thissoftware is PC-based, fast, easily amenable todistributed and/or parallel processing, and canexecute in parallel with the data acquisition. It isalso available to run, essentially without change,on a variety of platforms. It can easily beexecuted on workstations, larger computers, andeven teraflops machines, as well as on the PCscurrently in use. The figure shows the kind ofreconstruction possible with this system. Thedata in this image are from a low-energy sourceand are of a resolution target developed byAlliedSignal/Federal Manufacturing &Technologies.

In the final analysis, this high-energy detectorarray shows great promise as an accurate, high-resolution inspection tool for use with pits,canned secondary subassemblies, re-entryvehicles, other weapons parts, and non-weapons parts.

Preliminaryreconstructionof a resolutiontarget fromAlliedSignal/FederalManufacturing& Technology.The small wiresare 0.5 mm indiameter. Thebend in one ofthe small wiresis real.

44 Diagnostics • LA16


Development of SQUID Microscope for SurveillanceSince the first data was acquired in late August 1997,

numerous changes have been made to the superconductingquantum interference device, or SQUID, microscope systemto improve induction coil centering and to implement anautomated motion control and data acquisition for precisescanning. The SQUID microscope control and dataacquisition software package has been written in LABVIEWintegrating SQUID device control, motion control, and dataacquisition. We confirmed very high sensitivity for bothextremely small and deeply buried defects (beneathmillimeters to centimeters of intervening materials).

A stress fracture in a Ti-W surrogate was clearlyobserved, demonstrating the severe lattice defects werenot necessarily evidence of a physical separation. A seriesof measurements were made using aluminum plates withvarious cracks and holes. The SQUID microscope detectedeven the smallest cracks beneath several millimeters ofintervening conductor. Similar results were obtained using1-cm-thick plates of titanium (having similar conductivityto U and Pu).

A second induction coil design was tested and found to significantly decrease sample edge effects. Thisobservation led us to initiate the modeling effort using afinite-element electromagnetic code to better understandthis and other effects.

A systematic study of seam widths using the titaniumplates showed a clear trend in signal amplitude (a similartrend was observed for the phase) as a function of seamwidth. Seam-width resolution better than 0.005 in. caneasily be attained with this algorithm. A computer code to analytically calculate seam width is currently in progressand should enhance theresolution dramatically.The greatest challenge isdeveloping the requiredsensitivity within ourcomputationalcapabilities.

We began usingplates with a 100-µmcopper layer to testsensitivity to scratchesand determine spatialresolution. We recentlyscanned a platescratched with “P-21”and found we couldresolve the image fromraw data.

Currently our workfocuses on refining thefinite-element model to

the required sensitivity and resolution andvalidating the results. Work continues on theanalysis codes to further quantify seam widths intitanium samples. We are designing optimalinduction coils to test the ability to target featuresat specific depths in the sample. We havedesigned an experiment to identify cracks in athin layer of conductor buried under other layersof conducting and nonconducting materials.Calibration plates are being fabricated to helpquantify the effects of feature width and depth.

AlliedSignal/Federal Manufacturing &Technologies completed a brass-board flux-locked loop electronics package for controllingSQUIDs. The package was integrated with aSQUID, supplied by Los Alamos NationalLaboratory, similar to that used in the currentSQUID microscope. Slew-rate performance wasbenchmarked at >3 × 106 Φ0/s, exceeding thecurrent published world record (<2 × 106 Φ0/s)for DC SQUID performance. Significantimprovements will be realized by optimizing thecircuit and HF-packaging techniques. The hugeslew rate enables us to use SQUIDs overpreviously unrealized dynamic range and handletransients that would typically unlock a SQUID.This capability is important for the surveillanceprograms because it will enable the SQUIDmicroscope to be more easily incorporated intothe very noisy electromagnetic environments ofthe plants.

SQUID microscope prototype (above).Upper plot shows raw data for seamsof various widths in a 1-cm-thicktitanium plate. An unflawed titaniumplate of the same thickness covers theseamed plate. Lower plot shows thewidth of a fit to the data as a functionof seam width.

0

–0.01

–0.02

–0.03

–0.04

27

26

25

B,a

.u. a

t 63

3Hz

5 million 20 million 40 million 60 million

0 10 20 30 40 50 60

0 10 20 30 40 50 60

Position, mm

Wid

th

Gap width, mils



Solid-Phase MicroextractionUsing microcollection and analysis of chemicals in the

weapon backfill gases, we are developing a simple yetunique method to nonintrusively establish weaponenvironment and integrity of materials. The targetedchemicals outgas from the high explosives as well as fromother organic materials in the weapon assembly. Thistechnique detects changes in these materials because ofaging as well as defects. This valuable tool will remotelydetect and assess the aging of organics in weapons and willidentify potential compatibility issues (materials interactions)that should be more carefully monitored during surveillanceteardowns.

Microextraction is implemented within the weaponscomplex because it is safe, transparent to currentsurveillance activities, and provides a wealth of chemicalinformation. Collection, which is typically in the low-nanogram range, does not interfere with the weaponheadspace either by depletion or contamination. The gasvolume collected is limited to the dead volume of themicroextraction collection vessel that holds the fiberassembly. The current design, as shown below, has a deadvolume of 2.5 ml. This device, tested online with the newmoisture sensor, which is part of the Phoenix gas-samplingmanifold, has no effect on measurements down to thelowest tested point of 10 ppm.

The collection process uses a fiber, 1 cm long by 100 µmin diameter, coated with an absorbent that accumulates

chemicals without bulk collection of thebackfill gas. Following collection, the fiberis hermetically sealed and analyzed on a

Micrometer-sized fiber

B83coringtool

bench-top gas chromatograph/mass spectrometer(GC/MS).

The advantage of this collection technique isits high delivery efficiency for GC analysis.Compounds collected by the fiber absorbent canbe desorbed directly at the tip of a GC columnwith little loss or dilution. This results in highsensitivity, even for polymer fragments. The massspectrometer provides molecular weight andstructure information needed for positiveidentification in the low- and sub-ppb range.With this approach, we can identify unknowncompounds as they exist in a complex mixtureat trace levels.

In this application, we have found thatmicroextraction is particularly useful oncompounds having high surface activity (sticky)and those that are difficult to collect and transferto an instrument for analysis. Significant losses ofeven volatile compounds occur when sampling isdone at the gas bottle port through a gassampling manifold (Blue Goose) that is currentlyused for collecting permanent gases. Whensampling is moved upstream to the purge valve,we can collect and identify for the first time knownpolymer degradation fragments from silicone andpolysulfide materials. This information is beingcompiled into a living database that will be usedto help pinpoint potential problems.

During FY1998, a significant number ofweapons and compatibility test assemblies weresampled—nine B83s and 16 aging/compatibilitytest units from the W68, B83, and W84 programs.The W68 and W84 compatibility test units arerelevant because they use materials found in theenduring stockpile weapons. To assist ininterpretating the weapon signatures, data fromthese compatibility test units are combined withsignatures obtained from material and chemical

standards used in the weapon componentsand during manufacturing.

With Sandia, we are completinginitial evaluations of materialsused in the B83 weaponprimary and fireset. A surveyof materials used in the W87physics package wascompleted, and a report wasgenerated on formulationinformation, which is critical

for differentiatingdecomposition products.

5 10 15Time, minutes

20 25 30

30 31

O

OS

S

In coring tool At 50 cm At gas bottle port

Silicon decomposition productsPolysulfide decomposition products

Comparison of chromatogram signatures for theB83 weapon taken at three sampling locations:the weapon purge valve, 50-cm downstream, andthe gas bottle port (2-m downstream). At thelatter two downstream, less volatile compoundsare not detected at elution times of around 20min. (Note: Those peaks not labeled have beenidentified as either material synthesis precursors,aids, or byproducts, or manufacturing aids. Most of the signatures seen at 30 min arebackground from the fiber septum seal).



Sensor Technology DevelopmentFiber optic-based sensors for selected environmental

conditions and selected chemical species are beingprototyped and evaluated for weapons aging studies.Several different hydrogen sensors have been tested andportable systems for them have been prototyped. Threedifferent hydrogen sensors—palladium, yttrium/palladiumand palladium Bragg grating—have been fabricated. Theyare currently being evaluated in materials aging andcomponent studies.

The palladium fiber optic sensor and spectrometersystem fabricated at Savannah River Technology Center(SRTC) has been evaluated by AlliedSignal/FederalManufacturing & Technologies (AS/FM&T). The firstdeployment of this environmental hydrogen sensor isanticipated for a materials aging and compatibility testscheduled to begin in FY1999. Aging canister sensorheads utilizing glass-to-metal fiber optic seals (see figure,below left), which were developed in task LL32, werefabricated at the Kansas City Plant and sent to the Y-12Plant. These sensor heads were equipped with theSavannah River hydrogen sensor lens, Lawrence LivermoreNational Laboratory (LLNL) hydrogen sensor, and Entranpressure transducer. Core stack kinetic and aging studiesutilizing these sensor heads began at Y-12 in June 1998,over a year ahead of schedule.

The AlliedSignal yttrium/palladium fiber optic hydrogensensor has a potential increase in sensitivity greater than10 times the palladium sensor. Initial adhesion problemswith yttrium to glass have been corrected by adding athin coating of chromium to the glass before the yttrium/palladium coatings are applied. This new process was used

to coat lenses that are interchangeable with theSavannah River sensor lenses. Initial tests of thesecoated lenses in sensor heads indicate noadhesion problems.

The prototype Bragg grating sensor system(see figure, below right) uses a fiber laser withfiber optic Bragg gratings for hydrogenconcentration and temperature measurement.The fiber optic hydrogen sensor is 0.4 mm indiameter and 3.81 cm long. The fiber optictemperature sensor is 1.5 mm in diameter and3.05 cm long. Both sensors are totally opticaland introduce no electrical energy into the testvehicle. The prototype system was built into analuminum case for portability. A laptopcomputer provides data acquisition, reduction,and display.

A materials science study to improve theresponse time of the Bragg grating hydrogensensor is ongoing. It involves the introduction ofporosity into the palladium film by alternativefabrication techniques, such as electroplatingand plasma flame spraying. Although the currenthydrogen sensor is useful for enhancedsurveillance studies, a faster response time wouldmake it more desirable for kinetic studies.

AS/FM&T, SRTC, LLNL, and Oak RidgeNational Laboratory are working closely todevelop multi-use applications for real-timeenvironmental sensors for aging and materialscompatibility tests.

Prototype hydrogen sensor installed in aging canister cover and readout (left) and prototype Bragg grating hydrogen sensorand associated opto-electronics (right).

Diagnostics • KC10 47


Microsensors for Evaluation of Materials Degradationand Corrosion in Weapon Systems

The photograph shows integrated hydrogensensors and a schematic of the setup. Thesetests have been running with no apparentsignificant deterioration in sensor performance.Simultaneous pressure measurements in theretorts are also being recorded. Threeindependent test assemblies are beingcontinuously monitored. Using our currentsystem multiplexer, which switches betweensensors, monitoring is possible for up to 16 assemblies. Data acquisition and analysis is still in progress by the end of FY1998.

Prior to starting the tests, we determined fromcalibration curves that the sensors would showresponse to at least 0.1% H2 at 760 torr. We areworking to improve responses at lower pressuresto obtain better sensitivity limits and stability.Improvements in multiplexer and in the sensordesign (including adding an internal reference)are underway.

Our current effort focuses on developing fiber-optic sensors. The key element of these sensors isa chemically selective thin-film coating depositedon the end of the fiber. Any gas-phasecomponent that exhibits a high affinity for orreacts with the coating is collected on the end ofthe fiber and detected spectroscopically.

As a first step, we have chosen to demonstratea sensor for hydrogen. The hydrogen sensor thathas shown the greatest promise was developedat SRTC. The design, shown in the upperschematic, includes a thin palladium alloy filmdeposited on the end of the quartz lens. Twooptical fibers are affixed to the lens. Light from atungsten halogen lamp is launched down onefiber and collimated by the lens. The lightreflected from the metal film surface istransported back to a spectrometer by thesecond fiber. When hydrogen is present, it isabsorbed into the palladium alloy film, changingits optical (reflectivity) properties. Using thissensing concept, we have demonstrated areversible response to hydrogen in the range ofless than 0.1% to several percent. Our sensor isan improvement of an earlier micromirror design,first proposed by Butler and coworkers at theSandia National Laboratories.

This project’s long-term objective is to develop a suite ofchemical sensors for real-time monitoring of changes instockpile materials. In the near term, the sensors willprovide kinetic data never before accessible from CoreStack Compatibility Tests (a suite of tests that use a stackedarray of materials to evaluate materials compatibility andcompare various weapon design options). These sensorswill provide the capability to verify component andweapon integrity by analyzing their dead volume gases.The presence of certain volatile chemicals in the monitoredenvironment indicates material degradation. These sensorswill be incorporated in a protocol for componentinspection, refurbishment, and retirement. This year’ssuccess is the result of an extremely productivecollaboration involving AlliedSignal/Federal Manufacturing& Technology, Savannah River Technology Center (SRTC),Y-12, and Lawrence Livermore National Laboratory.

The most significant accomplishment this year is thesuccessful initial attempts to integrate chemical sensors forhydrogen into Core Stack Compatibility Tests at Y-12. Forthe first time ever, these sensors offer the potential forobtaining real-time kinetic information. In addition,progress has been made in developing chemical sensors forother gas-phase species.

Photo of Core Stack Tests with integrated hydrogen sensors andschematics of hydrogen sensor design and core stack assemblies.

Fiber-opticsensors

O-ringseal

To spectrometer

From lamp

Fiber-opticjunction

Modified swagelock reducer(quartz lens epoxied inplace)

Palladiumalloy film

Lens

0.25 in.



Systems Engineering of an Arming, Firing, and Fuzing Testbed

The purpose of this project is to understand the systemsengineering issues associated with incorporating chemicalsensors in a weapon subsystem. Systems issues of safety,reliability, volume constraints, sensor accuracy and drift,power demand, and communications are beingconsidered. Our focus at this stage is placing existinginstrumentation into an arming, firing, and fuzing (AF&F)unit taken from the stockpile to assess the state of the artand guide the evolution of embedded sensors. Ultimately,chemical sensors will be used in a suite of in-situ diagnosticsand built-in test features where they will enhancesurveillance testability and augment the traditionaldisassembly and inspection methods of determiningsubsystem state of health.

Based on past and present investigations, an initial set of sensors for hydrogen, water vapor, pressure, andtemperature were selected. The table lists the instrumentmeasurement ranges available for these sensors.

We are using an integrated nuclear materials monitoring(INuMM) electronic sensor package that contains each ofthese sensors. The INuMM package, which was originallydeveloped to monitor stored pits, will be installed into aW68 AF&F that has been modified to provide the necessaryvolume for the sensor package. The figure shows a W68AF&F and hydrogen sensor calibration data.

Two W68 AF&F units, retired from the active stockpile,are currently under modification at the Kansas City Plant(KCP). These units were selected because of their ready

availability and the recently developed databaseon the composition of the internal atmosphereof the W68 AF&Fs.

Testing will commence early in FY1999.During testing, gas samples will be withdrawnperiodically for gas chromatography/massspectroscopy analysis and compared against thedata telemetered from the self-powered INuMMpackage/hydrogen sensor sealed within thetestbed. The data will also be compared withthe periodic solid-phase microextraction datarecorded for similar W68 AF&F units sampledat KCP.

An improved water-vapor sensor that useschemiresistor technology and has a measurementrange of 200 to 50,000 ppm will be incorporatedin the INuMM package in FY1999. This sensorwill also detect a number of organic vapors.

W68 AF&F (above) and calibration data forthe hydrogen sensor in the INuMMelectronic sensor package.

6:00 9:00 12:00 15:00 18:00 21:00Time, June 23 and June 24, 1998

0:00 3:00 6:00 9:00 12:00

250

200

150

100

50

0

Hyd

rog

en s

enso

r o

utp

ut, c

oun

ts

100 ppm

1000 ppm H2 5% O2

0 ppm (sensor package turned off)

1000 ppm

1000 ppm

10,000 ppm

Chamber opened to turn on

sensor package

Sensor ranges for preliminary testbed

Sensor Measurement rangeHydrogen 10 to 10,000 ppmWater vapor 5% to 90% RHPressure 0 to 15 psiaTemperature –15 to 60°C

Diagnostics • SN13 49


Weapon State-of-Health MonitoringWe are developing chemometric tools based on

advanced statistical analysis and pattern recognitionmethods to extract time/age information from the gasanalysis of weapon internal atmospheres. The fundamentalhypothesis underlying this project is that the weapon’sintegrated time/temperature history is encoded in thecomposition of its internal atmosphere. Application of theadvanced modeling and pattern recognition techniquesprovide a general indicator of an individual weapon’s stateof health. Non-intrusive gas sampling coupled withchemometric data analysis extracts more aging informationfrom current surveillance gas samples and data than haspreviously been available.

Our research shows that each weapon type has a uniquegas signature resulting from the outgassing and aging ofmaterials in the weapon. As a weapon ages, its gassignature changes slightly. This change can be detectedand modeled using chemometrics. The chemometricmodel extracts age-related information imbedded in thegas signature, flags peaks (i.e., compounds), and predictsan age for the unit. Discrepancies between the predictedage and the actual age can signal the presence of a unitwhose signature is distinctly different than the rest of theweapons of the same type and of the same age. The unitcan then be studied further to determine the reason forthe discrepancy.

An example of extracting age-relatedinformation from a weapon’s internalatmosphere is shown in the figure. These datawere taken from routine surveillance for the B83and illustrate how age-related information can beobtained from the analysis of organic gas data.Although the data set is limited, the predictivecapability of the model is seen to be very good.A much larger data set has recently becomeavailable for the W80 from the field samplingeffort (see SN17, Enhanced Gas Analysis forDiagnostics and Surveillance). A cursory analysisof this data set (~100 samples) shows that age-related information is also contained in the W80 data.

The sampling of weapon and componentatmospheres has recently been enhanced byorders of magnitude with the introduction ofsolid-phase microextraction (SPME). The gassampling station in the figure is used to collectpressure and moisture data, as well as to collectgas samples both with and without the use ofthe SPME technique.

This technique extends the sensitivity ofmethods such as gas chromatography/massspectroscopy so that the number of species

detected is an order ofmagnitude higher (dependingon the system) and the signal-to-noise ratio of thechromatograph is more thanan order of magnitude higherthan that for a gas phasesample. With this richness ofdata, chemometric techniquesof data analysis become allthe more important.

This work is a collaborativeeffort between AlliedSignal/Federal Manufacturing &Technology, Pantex, LawrenceLivermore NationalLaboratory, and the SandiaNational Laboratories.

10.0

9.5

9.0

8.5

8.0

7.5

7.0

6.56.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0

Actual age, years

Pred

icte

d a

ge,

yea

rs

Model data set Test data set

Typical organic signature (gaschromatography raw data) from eightB83s is shown above; results frommodeling data obtained from twenty-twoB83 units with a prediction for eight moreunits are shown on the above right, and agas sampling station (right) for sealedcomponents such as arming, fuzing, andfiring systems at the Kansas City Plant.

No

rmal

ized

sig

nal

GC retention time, arbitrary units

50 Diagnostics • SN14


Identifying Damage with High-Resolution Vibration The goal of this work is to develop a nondestructive

weapon monitoring and damage-detection system usinghigh-resolution vibration analysis. The final product willaugment existing surveillance tools for enhanced assessmentof the stockpile. Detecting damage internal to cannedsubassemblies (CSAs) is the primary focus.

The basic concept of vibration testing is that dynamicparameters (acceleration time histories, resonantfrequencies, mode shapes, and modal damping) are afunction of the physical properties (stiffness, mass,damping, and boundary conditions) of the component or system. Thus, changes in physical properties of acomponent or system, such as its stiffness, cause changes in dynamic response.

A wealth of damage identification algorithms exists fordetecting the presence of damage and location in structuraland mechanical systems. Algorithms have been appliedwith varying degrees of success to detecting damage inbridges, the NASA Space Shuttle, and other structures thatundergo flexural vibration. However, the simplified CSAmodels, developed earlier as a part of this program,undergo axial or membrane vibrations rather than bending(flexure). Damage-identification algorithms had not beenapplied to this type of vibration; therefore, two promisingmethods, damage-index and change-in-flexibility, weremodified early in this program to specifically treat vibrationsexperienced by CSAs. Preliminary numerical simulationsperformed in FY1997 indicated that both these methodscould detect as little as a 10 percent reduction in stiffness.

During 1998, construction of a vibration testbed forverification of these modified damage-ID algorithms wascompleted (see figure), and experiments were performed.The generic vibration model consists of a series of massesconnected by springs, all free to vibrate on the mountingrod. The system was instrumented withaccelerometers mounted on each mass. Inaddition, a force transducer was mounted at the point of vibration excitation so thatfrequency response functions could beconstructed to provide both naturalfrequency and mode shape information.Excitation was accomplished either by anelectromechanical shaker or impacthammer.

Damage was introduced in the system bysubstituting a spring of lower spring constantat selected locations. Damage levels of 7, 14,and 24 percent were introduced sequentiallyat three locations in the system. After datawere taken for damaged and undamagedsystems and reduced in terms of naturalfrequencies and mode shapes, the modifieddamage-index and change-in-flexibility

methods were applied using the data. Thedamage-index method consistently detected thelocation of the damage in the generic vibrationmodel at all locations, even at the 7 percentdamage level. Results using the change-in-flexibility method were inconsistent.

Two favorable attributes of the damage-indexmethod: it indicates the presence and location ofdamage. Two significant drawbacks: the methodis appropriate only for detecting linear forms ofdamage and spatial information is required.

Because of the mixed results of the twomethods, another damage-ID method has beendeveloped, based on principles of statisticalpattern recognition (previously applied to speechrecognition). Although the method does notpredict damage location, it does predict thepresence of damage for linear and nonlinearsystems. Further, only external measurements on the CSA would be required. Using the samespring-mass data as above, this new method hasbeen successful in predicting the presence ofdamage in all cases.

Detection of impact-type damage is beinginvestigated. This problem is fundamentallynonlinear, so a simple analytical model for thecomponents has been developed. Testing usingthe vibration testbed, but with an impact rodinserted between two of the masses to simulate a different kind of damage, is under way now.Preliminary data are being evaluated using thestatistical pattern recognition. Soon, this methodwill be applied to external dynamic responsemeasurements from an actual CSA.

Experimental testbedused for vibration-based damageidentification.



Low-Level Vibration and Advanced Signal ProcessingWe have combined many of the computer analysis tools

developed in recent years into a toolbox known as SignalProcessing and Modeling or SPAM. SPAM is a graphical userinterface running in a MATLAB environment that combinescommon user interfaces for data management andpreprocessing with the modeling and analysis tools.Modeling and analysis tools incorporated into SPAMinclude:• ARMA and bilinear models• Bendat zero-memory nonlinear model• Canonical variate analysis (CVA) models• Neural networks• Probability density estimation routines• Bootstrap statistics• Pre- and post-processing tools including integration,differentiation, filtering, resampling, and typical spectralanalysis tools

This interface makes the analysis capabilities available toa wide range of users at the Los Alamos NationalLaboratory and Y-12. The startup user interface is illustratedin the figure.

Any signal-processing task selects certain importantfeatures of the test data. A feature of interest may be, forexample, mode shape or modal frequency. For stockpilesurveillance units, we have developed a trial set of criteria.Each test unit will be ranked according to the followingtrial criteria:• Resonant frequency and damping values indicated by thefrequency response function and/or ARMA coefficients

• Degree of nonlinear response (CVA, coherencemeasures, and dimension of the state space)• Deviation of probability density function fromGaussian• Decision statistics for Hinich’s Gaussianity andlinearity tests (bispectral statistics)• Generation of high-frequency signals in thepower spectrum• Air-bearing test results

Testing experience shows that thesemeasurements provide useful indications of unitcondition. Our ongoing air-bearing and vibrationtests have produced an extensive time seriesdatabase set, which is archived. Toolboxalgorithms are applied to the data, and reducedset quantitative discriminating features areextracted based on the above criteria. Weanticipate that these features will vary stronglywith different aspects of unit condition. Inconcert with engineering analysis, we aredeveloping statistical methods and criteria forcategorizing the units in our test database.

Low-level vibration and air-bearing testingcontinues at Los Alamos and Y-12. These dataadd to our test database for each type of unit. Fora number of systems, we have a statisticallysignificant set of test data, both for air-bearingand low-level vibration tests. Pantex is currentlydeveloping a capability for air-bearing testing.

We are continually improving the algorithmsused for the analysis of test data. Signalprocessing is a rapidly changing field, especiallywith regard to nonlinear systems. Systematicanalysis of the data from recent tests, combinedwith research into algorithms, should add to thecapabilities in our toolbox.

The startup user interface of SPAM, the Signal Processing andModeling toolbox.

52 Diagnostics • LA20


Laser Penetration and Rewelding for Gas Sampling

The purpose of this task is to develop a laser drilling andrewelding process for sampling and resealing the cannedsubassemblies (CSAs) so that new surveillance diagnosticsmay be incorporated in the CSA. Such a technique wouldprovide the ability to sample gas within a unit and,depending on the results of the analysis, to reseal the unitfor its return to the stockpile. Our goal is to demonstratethe feasibility of the hole drilling and resealing process andalso implement a laser gas sampling and real-time analysissystem in routine surveillance operations. The gas analysissystem incorporates techniques being developed throughother tasks, including infrared analysis in OR04 and solid-phase microextraction in LL13.

During FY1998, we demonstrated process feasibility onrepresentative geometries. We demonstrated the feasibilityof using a laser to drill a hole to sample the gas in a CSAand subsequently reseal the hole using a Nd:YAG andNd:Glass laser on stainless steel flat plates and outgassingtubes. Several tubes were resistance crimp welded usingproduction welding procedures to simulate secondarygeometries. Process feasibility on representative geometrieswas initiated as shown in the figure below. Tooling wasfabricated to orient the tubes at a 30° angle from theincident laser beam.

We also designed, fabricated, and tested alaser gas sampling stack/chamber. Thischamber was sealed onto a CSA, providing awindow for the laser beam and a port toobtain the gas sample. Leak and function testswere performed on the chamber assemblyrevealing that the system was leak-tight,although it exhibited an unacceptable leak ratewhen isolated from the vacuum source.Alternative sealing materials are beingidentified. A new gas sampling chamber withalternative sealing materials was designed andis scheduled for fabrication.

We evaluated process control parametersfor the drilling and welding operation. Holeswere drilled and welded in 0.010- and 0.030-in.-thick, type-304L stainless steel sheet usingOak Ridge Y-12 Development Division’sNd:YAG laser and the Y-12 surveillancelaboratory’s Nd:Glass laser. We drilledadditional stainless steel samples at variousparameter settings to determine optimum holesize for gas flow and fractionation. On the basisof these results, we are developing models ofgas flow through these holes.

We also began work on the implementationplan for the laser drilling and welding process.We modified the laser operating procedure inthe Y-12 surveillance laboratory to incorporatethe welding process for resealing. Drilling andwelding procedure specifications were written asseparate, stand-alone documents. Thesespecifications will contain all the necessaryparameters to perform the drilling and weldingoperations and will be used when referenced bythe laser operating procedure. A test of the laserdrilling and welding process on a war-reserve-quality CSA was performed in the Y-12surveillance laboratory. Drilling of the CSAadjacent to the resistance crimp weld was asuccess. The resealing and welding processrevealed that the laser being used in routinesurveillance operations is inadequate forresealing a CSA because of inherent alignmentproblems with the defocused laser beam. A newlaser, modeled after the laser used at Los AlamosNational Laboratory’s plutonium facility, is beingprocured and should resolve this problem.

Laser drilling and rewelding process tested on a representative geometry.

Diagnostics • OR18 53


Infrared Surface and Gas AnalysisWe have developed two nondestructive mid-infrared

inspection tools for evaluating contaminants that may be present on surfaces and in gases in returned stockpileunits. The identification of the presence and quantity ofcontaminants provides information that is critical forassessing the effects of aging in these units.

The diffuse reflectance infrared fourier transform (DRIFT) surface inspection tool, patterned after a Los Alamos National Laboratory design variation, has beendeveloped. DRIFT interrogates millimeter-sized spots bydiffusely scattering light off the surface, from which aninfrared spectrum is obtained that contains explicit chemicalinformation about the surface. A motion stage (not shown)is used to manipulate the specimen and enable DRIFT todisplay the inspected surface as a three-dimensional surfacemap. Figure 2 demonstrates the 3D spectral imaging

capability ofthis technologyby imaging the carbonylspecies of the

adhesive on a Post-it note. A visualizationattachment for this system, a recent Los Alamosdevelopment, was manufactured and is beingprocured so that the visual image of the areabeing inspected can be captured andcoordinated with the spectral image.

The nondestructive infrared gas analysis(NDIGA) inspection tool analyzes the trace gasesin a returned stockpile unit. NDIGA uses aninfrared spectrometer with a gas cell that has asensitivity to most simple, polyatomic gases.Figure 3 demonstrates the capability of NDIGAto simultaneously detect and distinguish betweengaseous (sharp lines) and liquid water (broadbands). The NDIGA is part of a largerinterrogation system involving technologiesdeveloped under Tasks OR18 and LL13.

During FY1998, we developed DRIFT,procured it, and prototyped it in a productionenvironment. The NDIGA system was alsodeveloped and tested by using returned weapon

materials. Both of these technologieswill be transferred to the

Surveillance Program in FY1999.

0.2

0.4

0.6

0

0

00.5

1.0

1.5

2.0

1.5

2.0

1.0

0.5

2.0

1.5

1.0

0.5

0

4 3 2Wavenumbers, thousands

Ab

sorb

ance

1

Fig. 2. A 3Dspectral imageof the spatialdistribution ofthe adhesivecoatings on aPost-it note.

Fig. 1. The DRIFT surface inspection instrument(top) configured for installation into the glovebox. NDIGA (bottom) can simultaneouslydistinguish between gaseous and liquid water.

Fig. 3. TheNDIGA systemas presentlyconfigured forgas analysis.

54 Diagnostics • OR04


Material InteractionsAlthough they are drier than many organics,

cellular silicone cushions inherently containmoisture. A primary concern about their use inan assembly is that the moisture in the cushionwill contribute to increased aging of theassembly. Key components of this task are theanalysis of the moisture content of cellularsilicone cushions and the effect that themoisture could have on assembly aging.

We selected a residual gas analyzer (RGA) asthe moisture analysis method. A schematic ofthe moisture analyzer is shown in the firstfigure. Prior to analysis, the cushion sample isexposed to a known moisture level for aspecified time. The cushion sample is heatedunder vacuum, and the RGA is used tomonitor mass 18 versus time. The data areintegrated to determine the total quantity ofmoisture evolved.

The second figure shows the analysisresults of the cushion moisture content. Theaverage initial moisture content of cushions atambient conditions, shown at time = 0, is0.1 wt%. When placed in a dry environment,the cushions rapidly lose moisture to anaverage level of 0.03 wt%.

Cushions were also aged with a substrateof interest under accelerated aging conditions.As shown in the third figure, the cushionsincreased the rate of aging of the substrate,dramatically at first, then at a more modestrate with increasing elapsed time. Overall,the increased rate of aging was withinacceptable limits.

It is expected that cellular silicone systems,with proper processing, are suitable for storagein fielded units.

Moisture incushions afterexposure toknownmoisturelevel innitrogen. Thecushionsrapidly losemoisturewhen placedin a dryenvironment.

Rel

ativ

e in

ten

sity

, 10–4

Time, h

Initial moisture content of cushions at ambient

2

4

6

8

10

12

00 4 8 12 16 20 24

Sub

stra

te a

gin

g r

ate,

arb

itra

ry u

nit

s

0.02

0.04

0.06

0.08

0

Time, h0 12 24 36 48 60

Substrate + cushion

Substrate

Heater

Sample

Roughpump

APCvalve

RGA

Standardleak

TM pump

Effect ofcushions onsubstrateaging.

Schematic of themoisture analyzer.



Radiographic Parameter Optimization

Radiography is used to measure gap and otherdimensional measurements on most parts andsubassemblies manufactured at the Oak Ridge Y-12 Plant.Current radiography methods are based primarily onexperience, and we do not know of any comprehensive,systematic effort to optimize radiographic parameters,hardware, or procedures. Our efforts under this task aredirected toward improving Y-12’s ability to identifystockpile problems, especially with regard to components,through optimization of the radiography process.Optimization will be aided by computer simulation of theradiographic process and by the development ofcomputer-based tools that will enable radiographers toselect appropriate radiographic exposure parameters forefficient production of the best image possible. Theproject’s approach is to survey any prior optimizationwork, develop a computer model of the radiographyprocess, validate the model through experimentation,and implement the model as an advisory system thatradiographers can use in operations.

A survey of radiography personnelthroughout the nuclear weapons complexrevealed that no concerted efforts have beenmade to optimize radiography processes. Wefound that capabilities were experiential ratherthan science-based, just as Y-12’s are. Standardradiography texts and handbooks proved to be the most helpful sources, and we obtainedsimulation codes. TART97, from LawrenceLivermore National Laboratory, is a radiationtransport code that enables estimation ofradiation intensity under defined material andgeometric conditions. A second code, XRSIM,written by Iowa State University, will be used to generate virtual radiographs to verify theacceptability of real exposures.

We are preparing a radiographic facility toconduct model validation experiments in a low-energy (<420 KeV) range. The 9-MeV linearaccelerator at Y-12 will be used for high-energyvalidation experiments. We are also fabricatingtest rings that simulate a canned subassembly’sgeometry and materials for model validationexperiments.

Modeling of the radiographic process isalso underway, and we have developed amethod for selecting a source. The methodinvolves determining the intersection of theappropriate beam and build-up curves vsphoton energy, as shown in the figure. Theenergy at the intersection becomes the basisfor the selection of the appropriate x-raymachine and source intensity.

0.12

0.10

0.08

Rel

ativ

e in

ten

sity

Photon energy (MeV)

Intensity (transmitted primary beam, without scatter)

Build-up

Bui

ld-u

p

0.06

0.04

0.02

0

24,000

20,000

16,000

12,000

8,000

4,000

00 1 2 3 4 5 6 7 8 9 10

X-ray energy optimization (monoenergetic photons). The energy at theintersection becomes the basis for selecting the appropriate x-ray machineand source intensity.

56 Diagnostics • OR19


Evaluation of Hydrogen Balance MethodsHydrogen corrosion causes weapons to age. Through

the surveillance program, the Oak Ridge Y-12 Plantsupports routine surveillance tests on weapons to establishthe balance between hydrogen sources and hydrogensinks. The hydrogen inventory is a basic tool forsurveillance, and its test results are used to predict theaging and reliability of fielded weapons systems, planfacility requirements to recertify or refurbish weapons,and assess the significance of observed anomalies.

Current methods have served the nuclear weaponscomplex well, but the push to extend weapon lifetimesbeyond their design lives requires re-examination of thehydrogen inventory tools. As shown in the figure, thesetools provide hard data for performing weapon agingassessments. The objectives of this task are to assess theaccuracy, precision, and predictive value of each hydrogenbalance test and to develop and implement methods withenhanced accuracy, precision, and predictive value whereneeded. The deliverables will be data sets that can be inputinto weapon aging models and enhanced interrogationtechniques that will be implemented into routinesurveillance operations.

We have surveyed hydrogen balance data forweapons on a specific system. We retrieveddisassembly records for 22 weapons that wereevaluated as a part of the surveillance program.Systematic review identified eight qualityparameters that clearly are increasing with age. Asimilar number of parameters were notablebecause they clearly do not change with age. Apublished review provides objective data againstwhich weapon aging models can be developedand validated.

We also analyzed organic hydrogen getters.Routine surveillance results to date show thatorganic hydrogen getters provide almostcomplete protection against corrosion. Most of aweapon’s aging history will be recorded by thehydrogen getter. Two fundamental flaws were

identified in current analysis methods: (1) all gasuptake is not measured, and (2) one analysis issubtly but decidedly flawed. To resolve theseissues, specific analytical methods will berecommended from a “round robin” experimentthat is in progress involving four DOE sites.

We combined resources with Task OR02 todevelop the low-temperature thermaldecomposition (LTTD) system. LTTD usestemperature-programmed desorption to quantifycomponent chemical characteristics that mightbe weapon aging indicators. A prototype systembased on bench scale work performed in TaskLL23 was designed and built in FY1998 and willbe transferred in FY1999.

Enhancements to the bake-out operation willinclude programmed temperature control withdata acquisition to monitor kinetics as a functionof time, temperature, and isotopic composition.Enhanced bake-out will be ready for introductioninto routine surveillance during FY1999.

Hard data for

aging assessments

Moisture outgassing

Weapon gas analysis

Recertification criteria

Stockpile assuranceLTTD

Bake-out

Facility requirements

Getter analysis

Hydrogen inventory toolsassess weapon aging andstockpile reliability.



Development of High-Energy Neutron Radiography

We are developing high-energy neutron radiographyand tomography for use as a nondestructive evaluation(NDE) tool for stockpile-to-target sequence (STS) deviceengineering tests and routine stockpile surveillance atDOE production and storage plants. Our goal is todevelop an operational neutron imaging system capableof detecting cubic-mm-scale voids and other structuraldefects in heavily shielded low-Z materials (e.g., lithiumsalts) within a nuclear device. The imaging system will berelatively compact (suitable for NDE facilities at Y-12 andPantex) and capable of capturing a full tomographicimage of a weaponized secondary assembly. It willconsist of a high-yield, accelerator-driven neutron sourceoperating in the 10–15 MeV energy range (a nominaloutput ≈ 1012/4π n/sec/sr in the forward direction andan effective spot size ≈ 1 mm in diameter will berequired), a low-mass rotation stage to support andmanipulate the device under inspection (allowing fortomographic imaging), and the imaging detector itself.The detector will be based on technology proven inunderground nuclear testing. It will consist of a simpleplastic scintillator viewed indirectly by one or more high-resolution, LN2-cooled imaging cameras.

During FY1998, we built a prototype of the imagingdetector using an LN2-cooled camera with a fast lens andtested it by imaging a set of phantom targets at the OhioUniversity Accelerator Laboratory (OUAL) in Athens,Ohio. OUAL was selected because its physical layout isgenerally similar to that envisioned in an actual neutron-imaging facility. The first imaging experiments in October1997 included radiographic imaging of a set of nine step

wedges and a polyethylene slab with 10-, 8-, 6-,4-, and 2-mm-diam holes shielded by up to 4 in.of Pb (max. thickness ≈ 120 g/cm2). The secondexperiments in January 1998 included edge-resolution measurements using a machined Cublock and radiographic imaging of a lithiumdeuteride slab shielded by depleted uranium(max. thickness ≈ 100 g/cm2). Tomographicimaging data consisting of up to 64 views of a4-in.-diam, right-circular Pb cylinder with a 2-in.-diam polyethylene insert with the same set ofholes were also taken during this run. Imagingtimes varied from 10 min to more than 1 hrdepending on the object under inspection andthe accelerator beam current available. Resultswere excellent in all cases. The figure showsclearly visible simulated defects and associatedlineouts.

We also made marked progress in developinga high-yield neutron source during FY1998. Wehave worked with vendors to find smallaccelerator systems suitable for use in an actualimaging system and are currently collaboratingwith Richard Lanza of MIT to develop anoperational prototype of a high-yield target thatcan be coupled to the accelerator.

Finally, we have established implementationteams with Y-12 to determine overall systemperformance requirements and expedite theinstallation of a neutron imaging system at that facility.

(a) A 10-MeV neutron radiograph of a LiD slab shielded by D-38. The image shows defects and fractured edges (simulated by 10-, 8-, 6-,4- and 2-mm-diam. holes drilled 1 and 0.5 in. deep). (b and c) Associated lineouts.

4000

4500

5000

5500

6000

6500

200 400 600 800 1000

(b) Lineout through full-thickness holes

CC

D c

amer

a d

igit

al le

vel,

coun

ts

4000

4500

5000

5500

6000

6500

200 400 600 800 1000

(c) Lineout through half-thickness holes

Image column Image column

(a)



Bayesian Inference TomographyWe have recently augmented the capabilities of our

Bayesian inference engine (BIE) to include 3D geometricmodels. The BIE is a unique tool that we have developedover the last few years for analyzing radiographic images.One of the unique features incorporated in the BIE is theuse of geometric models to describe the boundaries of theobject under investigation.

The BIE is already being used to interpret dynamicradiographs from hydrotests in terms of models of two-dimensional or axially symmetric, three-dimensionalobjects. We will adapt the BIE to analyze stockpilesurveillance radiographs to determine whether thedimensions of a warhead deviate from nominal. Weanticipate that this may be done with good accuracy fromseveral radiographs. Our new 3D models will enable us tomake the best inferences about 3D effects from theradiographs that we will obtain from the Dual-AxisRadiographic Hydro-Test (DARHT) facility, presently underconstruction at Los Alamos.

The 3D geometric models are based on representing theboundaries between various regions with a finelytriangulated surface. A number of challenges wereencountered in the development of these models,including computation of the projection of the surfacemesh onto a 3D rectangular grid and controlling thesmoothness of the surface model. Given the uniqueness ofthese techniques, we sought collaborators and peerfeedback from experts in medical imaging, which is thefield that appears to be the most advanced in its use of the3D image model.

We formed a collaboration with a team from theUniversity of Arizona, headed by Prof. Harrison Barrett, whois renown for his contributions to medical imaging. TheArizona team supplied us with nuclear-medicine data takenof a beating, CardioWest, artificial heart. Their study wasconducted to demonstrate the feasibility of a dynamicheart-imaging system that will track the time-dependent

shape of the heart during a few heartbeats. Theultimate purpose was to improve a cardiologist’sability to assess heart function. A standardradiopharmaceutical, based on technicium-99,was injected into the vein feeding the rightventricle. The sequential images recorded in 24 pinhole cameras situated around the heartphantom were collected in 50-ms intervals.Only around 2,000 x rays were detected in all24 detectors for each snapshot. The figure showsa typical x-ray image.

We reconstructed the ventricle using two typesof models. The first and simplest modelrepresents the surface geometrically, assuming auniform interior density. The second modelfurther includes a slowly varying density model totake into account mixing of the radioactive boluswith the simulated blood. Both of thesereconstructions, shown in the figure, areremarkably good, considering the limited numberof views and the extremely noisy nature of thedata. The second model seems to provide abetter representation of the distribution ofradioactivity.

To summarize, we have demonstrated a newand unique capability for 3D reconstruction. Welook forward to applying this technique tostockpile surveillance radiographs. We are usingour 3D reconstruction technique to help answerthe question of how many views would benecessary for the DARHT facility to certify thestockpile. We intend to develop time-dependent3D models (4D models), which will be essentialfor analyzing multi-pulse DARHT radiographs.Fortunately, the University of Arizona data consistof a sequence of 100 frames. These will give us anopportunity to test our 4D models in the future.

On the left, a snapshot of a beating artificial heart taken by one x-ray camera. This image is extremely noisy because less than100 x rays were detected in the 50-ms exposure period. The surface of the bolus of radioactivity reconstructed from 24comparable views is shown in the center. On the right is a projection through a 3D density reconstruction based on combiningthe geometric model for the boundary with variable interior density. This image clearly displays the hemispherical shape of theheart, as well as its values.


-

3

50

1

0

4

3

4

2

4

2

t

t

AD9850

DDS Sin

AD9850

DDS Cos

Max125

A/D

A/D

Transmit select

Data

Control

Gain

Receive select

60-Mhz clock

16 MHz

Systems

62


Systems

Focus AreaSeveral aspects of the Enhanced Surveillance Program

require integration across nuclear and non-nuclearcomponents, their surveillance methods, associatedpredictive assessments, and other tools in order toprovide results optimized at the systems-level for use by the DOE and the DoD. The Systems Focus Areacollects the outcome of those tasks in coherent groupsand incorporates them into the appropriate level ofsystem integration.

Our foremost goal is to implement a full-spectrumreliability assessment methodology, integratingadvances in evaluating the reliability of the nuclearexplosive package with development of predictivereliability methods for non-nuclear components.Another goal is to perform system-level flight testingwith the highest fidelity consistent with diagnosticcapability for key features of both non-nuclear andnuclear explosive package performance. The final goal is a surveillance information capability that addressesdeficiencies in the current information managementsystem, including transforming data to information and then to knowledge and assuring needed electronicaccess, with priorities determined by the surveillancecommunity.

Presently three major classes of activities are includedin this focus area: enhanced-fidelity instrumentation,database management, and reliability assessment.Enhanced surveillance efforts are supporting flight

testing through the development of diagnostics andadvances in miniaturization and data handling. Inaddition to providing data, these developments mustminimize intrusion and include the integration ofaccountabilities: nuclear, non-nuclear, system, andthose for the DoD. Many existing databases needmodernization, but much of the body of data is not yetin electronic form. ESP is defining the requirements andpriorities for data access by the surveillance community.Activities include prototyping a cost-effective, intuitive,desktop capability to access surveillance-relevantweapon information in both classified and unclassifiedenvironments and to cost-effectively assess trends inweapons surveillance databases.

Finally, this focus area is enabling a betterunderstanding of stockpile reliability, including currentreliability assessment and predictive reliabilityassessment. A critical understanding of systemperformance related to weapon stockpile reliabilityprovides important data in assessing the integrity of thenuclear deterrent. Current reliability estimates for thenuclear package have historic roots that may not beentirely appropriate for a significantly longer weaponlifetime. Non-nuclear reliability tools presently lack theability to incorporate age-aware performance models,performance trends, and early indication ofunacceptable changes to provide predictive reliabilityassessments past the originally protected period.

Pertinent TasksKC11 Enhanced Surveillance Data AccessLA03 Advanced Diagnostics Development for Flight

Test Scoring [a and b]LA14 Material Property and Surveillance DatabaseLA17 Stockpile Information Group: Information

ManagementLA23 Enhanced Reliability Methodology Test ProgramLA39 Sampling RationaleLL27 Surveillance Information Group: Database

Management SystemsLL28 Nuclear Explosive Package InstrumentationOR01 Historical Certification Data AnalysisSN15 Data SystemsSN20 Advanced Telemetry

Deliverables1. Develop new flight instrumentation to characterizevehicle flight dynamics and evaluate functionalperformance of nuclear and non-nuclear components.

2. Specify information requirements for surveillanceactivities and coordinate these requirements with otherprograms.

3. Gain a better understanding of the impact ofweapon aging on stockpile reliability.


Historical Data Analysis at the Y-12 PlantTo reduce the aging rate of weapons, a getter is

incorporated in its design. Right now, there is nonondestructive method to determine the health of thisgetter. However, a review of the historical data of severalweapon families has unveiled a correlation between ameasurement taken during surveillance before the weaponis destructively evaluated and the health of the getterinside it. This discovery could be of enormous value to asignificant fraction of the stockpile: methods could bedeveloped to routinely monitor stockpile weapons in anondestructive manner for signs of unanticipated andexcessive aging.

A common thread linking the weapon families appearsto exist, as shown in the figure. This common behaviorbetween different weapon types suggests that the samebasic physical processes must be occurring in all of them.

The actual discovery of this correlation was made over ayear ago. Since then, data from additional disassembledweapons were incorporated into the updated plot shown

in the figure. These new data are consistent withthe behavior observed over a year ago, addingcredibility to the hypothesis that a realrelationship—not just a statistical correlation—exists between the two variables shown.Particularly intriguing is the specific quantitativevalue for the slope shown in the figure. Its valuewould not have been anticipated given ourcurrent understanding of weapon aging.

Furthermore, when other historical data setswere reviewed during this fiscal year, the samevalue for this slope was discovered. It wasreported in a compatibility study conducted inthe early 1980s, and it was discovered insurveillance and certification records of aparticular weapon type recently reviewed. Allthese findings add credibility to the notion thata physical relationship, not just a statisticalcorrelation, exists between these two variables.

A correlation between measurements made atsurveillance. The different symbols representdifferent weapon families.

Systems • OR01 63

Surveillance measurement

Get

ter

con

dit

ion

Slope


Advanced Diagnostics Development for Flight Test Scoring

One of our most significant accomplishments was anagreement between Los Alamos National Laboratory(LANL) and Sandia National Laboratories (SNL) on aschedule to design, build, and fly development test unitsfor both the W80 JTA5 and JTA6 configurations. This workis scheduled to be completed in FY2002 or 2003. We havea list of tasks and a timetable for the development, design,and testing of instrumentation, telemetry systems, and thecomplete prototype flight test units. For the first time, LosAlamos and Sandia tasks for realizing telemetered flighttest systems are coordinated and linked in a self-consistentway, so the results of one set of tasks flow smoothly andlogically into the initiation of another. Also for the firsttime, we have a coherent plan for the development andflight of the JTA6 (late telemetry) type of flight test units.This type of flight test unit monitors the performance ofthe W80 primary and thus is of great interest to LANL.Minimal SNL component instrumentation is involved inthe JTA6.

The W80 SLEP (Stockpile Life ExtensionProgram) may require that LANL and SNLcomponents in the W80 be modified. For thatreason, the flight test units that are the endproducts of the schedule are not designatedDJTA (Development JTA) 5 and 6, but are ratherdesignated EFI1 and EFI2. This name changeindicates that if the SLEP results in significantchanges to the W80, additional DJTAs will beneeded before production JTAs for the W80 canbe realized. However, if the SLEP does not causesignificant W80 changes, EFI1 and 2 can beconsidered the Development JTAs for the earlyand late W80 flight tests (JTA5 and 6),respectively.

Roger Bracht of LANL has developed aminiature, ultrasonic transducer controller circuitsuitable for inclusion in the JTA5 telemetrysystem. This circuit, shown in the figure, consists

-

-

-

-

.

3

50

6

5

3

2

3

6

5

1

1

1

0

4

3

4

2

4

2

t

t

AD9850

DDS Sin

AD9850

DDS Cos

MUX

Maxim Max308

Transm

it

MUX

Maxim Max308

Receive

Max125

A/D

A/D

Transmit select

Data

Control

Gain

Receive select

60-Mhz clock

16 MHz

A miniature ultrasonic transducer controller circuit for the JTA5 telemetry system consists of an ultrasonic transducerdriver, or transmitter module (top portion of the circuit) and a transducer receiver module.

64 Systems • LA03

Systems • LA03 65


of an ultrasonic transducer driver, or transmitter module(top portion of the circuit) and a transducer receivermodule. The transmitter module uses signals from sweptsine and cosine generators (the AD 9850 chips) to driveup to eight transducers at frequencies ranging from 1 kHz to 10 MHz. The specific transducer being drivenat any given time over a specified time period is selectedby a controller (not shown on the diagram) operatingthe MUX switch in the top portion of the circuit. Thereceiver module also selects the transducer to bemonitored by means of the controller operating theMUX switch in the bottom portion of the circuit. Thesignals from the transducers are amplified and filteredand then demodulated, using the HA2444 circuitsconnected to the sine/cosine generators, so theamplitude/phase information in each transducer outputis reduced to a data rate of about 6 kHz. In this manner,megahertz ultrasonic information can be transmitted ata much lower bandwidth. The demodulated ultrasonicdata are then converted to a digital signal fortelemetering to a ground station.

Savannah River Technology Center (SRTC) is developinga radiation source/detector gap measuring system. Here, asmall radioactive source with the strength of a smokedetector’s source is positioned beneath two surfaces incontact, with block radiation emitted from the source. Asthe surfaces move apart, radiation from the source streamsbetween the surfaces to a detector positioned above thesurfaces. The amount of radiation that gets through thegap in the surfaces is proportional to the magnitude of thisgap. SRCT has successfully developed a source/detector/counting system that can measure several spectra persecond and detect a 0.004-inch change in gap magnitudebetween two surfaces with a 1-minute count window. Thedetector is small enough to fit in the available space in theflight-test unit. The next step in this development process isto interface with the telemetry system and to begin testswith gaps between actual warhead components.

High Explosive Radio TelemetryHERT (High explosive radio telemetry) is designed to

provide reliable information on the implosion process as itoccurs in a mock nuclear explosive package in a deliveryvehicle. Measurement of the implosion process during theterminal stages of flight provides an integrated, systems-level assessment of many components. HERT allows adiagnostic screening that is otherwise unattainable inground-based or component-based testing.

In a terminal-stage measurement of implosionperformance, time is a critical component. The time fromour earliest measurement to the time when the telemetry

system is vaporized is on the order of tens ofmicroseconds. Conventional telemetry encodingtransmits one bit at a time. This allows aseparation, similar to a 180° phase reversal,between the two characters (1 or 0) so that theyare easily identified. If we choose to transmit acharacter that represents more than a single bit,we would gain time to get the data out but wewould lose separation between the characters,resulting in more sensitivity to noise. HERT willtransmit characters that represent 4 binary bits ora hexadecimal character, which results in a factorof 4 in time saving. The tradeoff is considerablymore sensitivity to noise and radio frequency(rf) transmission properties. We have chosen a22.5° phase separation and a 1.6 amplitude ratiobetween adjacent characters for the currentversion of HERT (in a scheme known asquadrature amplitude modulation, or QAM).Extensive testing and simulation have shown thatthe increased noise sensitivity is acceptable. Onesignificant advantage of the short time availableto transmit data is the short burst of rf. Thisallows us to digitize the intermediate frequencyfrom the receiver and store the digital data in aconventional computer. A software program canbe used to demodulate the signal and providetiming data as an output well after thetransmission has ceased, eliminating the needfor real-time demodulation.

This shows an actual screen image of the LabView demodulationprogram used on an early QAM constellation. The bit assignmentabove the figure shows how a 20-bit block of data is transmitted as5 hexadecimal characters. The bit assignments and constellation arenot the current design.


Nuclear Explosive Package InstrumentationThe Lawrence Livermore National Laboratory (LLNL) is

developing diagnostics, electronic and electro-optic circuitsand systems, and high-density electronics packagingtechnologies to enable high-fidelity weapon flight testswith on-board instrumentation and telemetry. There is an urgent need for on-board measurement of the nuclear explosive package (NEP) and NEP-reentry vehicle interactions and flight dynamics in high-fidelityconfigurations and for observing NEP detonationperformance under flight test conditions. Technologydeveloped at LLNL and Sandia enabled the developmentand successful flight of the first instrumented, high-fidelityflight test vehicle, the Mk21/W87.

The next generation of joint test assemblies (JTAs) forstockpile surveillance must include a representative NEPwith a secondary and a functioning primary with fullhigh-explosive charge. For these JTAs, survivableinstrumentation and telemetry must be developed toobserve and report weapon functions during detonation.New diagnostics are needed for flight dynamics andNEP performance measurements and for observingdetonation events. One of our objectives is to developand demonstrate on-board mass memory for recordingtelemetry data for later transmission and to enablelong-duration surveillance missions.

We have developed new capabilities to instrument ahigh-fidelity weapon and reentry vehicle configuration tomeasure flight dynamics. This includes a system of sensorsto observe weapon tip-off from the bus, exo-atmosphericflight, and endo-atmospheric flight dynamics. Sensors aremounted in and around the NEP and reentry vehicle with

particular attention at the interfaces. LLNL andSandia have specified, evaluated, and developedflight dynamic sensors, including accelerometers,vibrometers, acoustic sensors, displacementsensors, temperature and pressure sensors,magnetometers, and rate gyros.

We enabled and supported development andflight testing of FTU12, a Mk21/W87configuration and the first instrumented high-fidelity flight test vehicle. In May of 1998, thePeacekeeper Glory Trip 27 mission was launchedfrom Vandenberg Air Force Base. On board wasFTU12 along with other JTAs. Flight dynamicsand NEP performance data were acquired fromon-board sensors, telemetered to ground stationsin real time, recorded, and are being analyzed(see figure).

We have developed new diagnostics toobserve terminal events. The engineering designand test of a novel, very fast shock arrivaldiagnostic has been accomplished. An LLNL-developed laser Doppler technique was used tomeasure shock arrival with 1-ns resolution in avery small package. We further investigated anew, very simple technique for observingneutron flux based on radiation effects indielectric materials. These diagnostics and otherswere demonstrated and evaluated in anexplosive ground test in August 1998.

We have extended the LLNL-developed laserpantography technology for packaging denseelectronics for telemetry. We developed theability to fabricate 3D stacks of unpackagedintegrated circuit dies. We designed, developed,fabricated, and tested high-density memorystacks for telemetry applications, and developedand demonstrated a novel technique formounting unpackaged integrated circuits to FR4fiberglass-composite printed circuit boards—along-time industry goal. This newly developedinterconnect system compensates for differencesin thermal expansion and enables easy mountingon the printed circuit board. We developed a0.9 G bit mass memory module for flight testtelemetry with memory control and data flowcircuits. The developed mass memory moduleprovides on-board data storage for down-rangetransmission, reducing battery power andvolume requirements and deferring telemetryheat buildup in confined volumes. This representsa new capability enabling broad-ocean areamissions for strategic weapon JTAs and long-duration JTA flight tests on cruise missiles.

Flight test reentry data from sensors aboard FTU12. Data fromacoustic sensors, accelerometers, displacement sensors, and rategyros observe lateral loads at various locations within the nuclearexplosive package (NEP) and between the NEP and the reentryvehicle. Reentry events are identified and analyzed. The magnitudeand phase relationships between these signals are used to developload transfer functions and refine flight dynamic models.

66 Systems • LL28


Advanced TelemetryThe Advanced Telemetry project is developing

technology for the high-fidelity flight test vehicles that willbe needed in a reduced-flight-test environment. These testflights collect data to certify the reliability of stockpilesystems for both the Department of Defense (DoD) andDepartment of Energy (DOE).

A high-fidelity joint test assembly (JTA) will be virtuallyindistinguishable from a war reserve (WR) weapon system(i.e., replacing few WR components and minimally changingmass properties), but it will monitor all required DoD andDOE functions, including implosion diagnostics. In the nearterm, advanced telemetry will affect development of testassemblies for the W87 enhanced-fidelity instrumentation(EFI), W80 JTA, W76 EFI, and the SLBM Warhead ProtectionProgram. Eventually, every JTA telemeter will be designed,or redesigned, for high-fidelity instrumentation.

New sensors, enhanced measurements, andminiaturization of instrumentation systems were our FY1997goals. Earlier this year, the EFI-2 instrumentation systemprovided a W87 flight test (FTU-12) with the smallesttelemeter ever to be fully contained within the nuclearassembly. The successful flight of EFI-2 provided enhancedmeasurements of reentry vehicle dynamics and nuclearpackage dynamics, and it was the first telemeter to includeacoustic and displacement sensors. The W76 EFI, scheduledfor a flight test in January 1999, is another new miniaturizedtelemetry system made possible by the technologiesdeveloped by the ESP Program.

In FY1998, miniaturization efforts were concentrated on development of a new distributed architecture for telemetrysystems that allows the instrumentation of flight tests witha full, live explosive main charge. This architecture isa key element of enhanced-fidelity instrumentationin future JTA systems. In contrast to conventionalcentralized telemeters which, despite sizereductions, invariably displace WRcomponents, a distributed telemetry (DTM)architecture places components throughoutthe weapon.

The EFI-3 instrumentation system (nowdesignated DTM1) being developed for theW87 is similar to EFI-2 in size, andincorporates a DTM subsystem. DTM1 is ahybrid system architecture for validatingDTM concepts while providing a maturetelemeter design for most instrumentation.Several new technologies were developed oradapted for DTM1, including a new serialdata link and communications protocol (tolink the distributed modules), miniatureconnectors, and data compression within ahigh-resolution, analog-to-digital converter.Pioneering efforts by AlliedSignal/Federal

Manufacturing & Technologies produced a new flex-circuit cable for the new digitalcommunications link that incorporates extremelysmall connectors. These new cables will eventuallyreplace the large “horsetail” of wires to analogsensors in conventional telemetry systems. In aclose facsimile of DTM1 (photo), several dataacquisition modules are connected to the centraltelemetry package by a single cable with onlynine conductors. An average module, designedwith discrete components, is 1.5 in. by 1.5 in. by0.5 in. in size. An application-specific integratedcircuit is currently being developed at SandiaNational Laboratories/New Mexico that willcombine the functions of several componentsinto one, further reducing the module size by atleast half.

Micro-electro-mechanical technologies are alsounder development to make measurements fromwithin the nuclear explosive package (NEP) thatwere never before possible. Magnetically excitedflexural plate wave devices are extremely sensitivestrain gauges, the size of integrated circuits, thatcan detect surface cracks when mounted tosurfaces inside the NEP. Because wires are notallowed to penetrate the hermetically sealed shellof the canned subassembly, new wirelesscommunications are also being designed to getdata from sensors inside the shell to the telemetry

package outside the shell.

The W87 DTM1 and W76 EFI are thesmallest telemetrypackages everfielded. The W87DTM1 demonstrationof a distributedtelemetry subsystemwill validate the DTMconcept for fullydistributed systemsof the future.

Systems • SN20 67


Stockpile Information Group—Database Management Systems

This task began by forming the Surveillance InformationGroup (SIG) to provide, throughout the nuclear weaponscomplex (NWC), desktop access to legacy, current, andfuture weapon component information. We continue tomake progress, adding content to our classified web sitethrough a combination of scanning (for legacy documents)and electronic file submission (for current information). Ourform-based process allows one to many files (includingwhole structures) to be uploaded to our server, withaccompanying metadata providing a rich description ofcollection content. The information is divided into logicalgroups to facilitate storage and retrieval through a platform-independent web interface. Information is found either byquerying the fielded metadata or by locating words andphrases within the optically recognized or electronic text ofthe document, using one of several commercial searchengines. This task has demonstrated the feasibility ofscanning and digitally storing entire vaults of classifieddocuments at reasonable cost. We have recently added acollection of 19,000 scanned aperature cards (courtesy ofSandia National Laboratories/California). This collectionallows the user to view Lawrence Livermore NationalLaboratory generated drawings online, which complimentsour ability to view NWC-wide unclassified drawingscontained in SNL’s Image Management System.

At Y-12, as a first step, we transformed the surveillancecycle reporting system from paper to electronic format. Inthe process, we greatly enhanced the content of the reportto include all pre- and post-disassembly procedures,laboratory reports, video clips, photos, and radiographs.

All previous cycle reports and the unit’s build bookare also included, and delivery time for the reporthas been shortened from 18 months to 1 month.In a cooperative effort with the AcceleratedStrategic Computing Initiative and AdvancedDesign and Production Technology (ADaPT)programs, Y-12 has installed a product datamanagement (PDM) system, designed to containall future surveillance cycle reports, an archive ofinformation on secondaries built for the nucleartest program, and electronic data capture forcurrent and future life extension programs. Y-12has recently accredited a web server to deliver thisinformation to the user. We have demonstrateddesktop access to this system using a web browserover SecureNet, as shown in the figure.

The collection of information at the Kansas CityPlant (KCP) resides predominately on a network ofsensitive unclassified computers. NetU has beenimplemented as the communications infrastructureto provide the capability for desktop access to thisand other unclassified resources. The focus of ourefforts with regard to providing information fromKCP data sources to consumers has been ondeveloping a distributed, platform-independent,object-centric interface to part-related informationstored in various computer systems throughoutthe site. This approach relieves informationconsumers from needing to know what data isstored where, in what format, and how to get it.Instead, they can think in terms of real-worldobjects that are familiar to them, such as parts ordrawings. Queries are formulated using theseobjects rather than database-oriented knowledge.The technical approach is based on a three-tieredarchitectual model and object technologies. Usingthis approach, KCP has developed a prototype PartObject and web-based user interface. Plans are tohave seven major surveillance-related data sourcesavailable online via the Part Object approach bythe end of FY1998.

At Pantex, we are making progress under theADaPT extended data capture and electronicprocedure pilots. The data capture pilot seeks tocapture W87 assembly and inspection records inan electronic data warehouse for access from thedesktop over SecureNet. Savannah River alsoparticipates in the SIG effort and is close toaccrediting both its SecureNet and NetU networks.

Secure web site at Y-12 providesdesktop access to test programarchives, surveillance, and currentinspection and certificationinformation.

PhotoData

Video

68 Systems • LL27


Online Material Property DatabasesWe are assessing the feasibility and utility of making

relevant engineering material property databases easilyaccessible on nuclear-weapons-complex (NWC) computernetworks and then to implement access in cases deemedbeneficial. Ultimately, we also plan to assess the quality ofexisting databases and promote needed upgrades.

We have acquired the hardware and software requiredto create databases for those production facilities that didnot previously have adequate equipment. Theseacquisitions have allowed us to generate and partiallypopulate at least one relevant prototype electronicdatabase at every nuclear weapons complex productionfacility. This has improved the internal availability ofimportant material property information and helped ensureits preservation. The emphasis to date has been onunclassified information so that it may be easily used to testcommunications protocols. With these pilot databasesavailable in-plant, local demonstrations to visitors havebeen possible, all of which have been well received. Mostwho view these pilots would like to be able to access themfrom their own desktops.

We are now attempting to make our prototypedatabases available at the desktop computers of weaponengineers. Unfortunately, we have yet to overcome thetechnical and administrative hurdles to implementing suchcommunications except for the Kansas City Plant (KCP)materials standards database that can be accessed byanyone in the NWC via the open Internet through the pathshown in the figure. Since the principal benefit tosurveillance-related activities from this task is easy access forweapons engineers to materials data, our attention nowcenters on surmounting these hurdles to communication.Originally, we had planned on demonstrations using theopen Internet with later transition to SecureNet. However,using the open Internet even for unclassified data raisedconcerns in several cases, while plants with largelyunclassified data preferred not to be burdened with theextra security issues surrounding SecureNet for theirunclassified databases. To address both of these issues,near-term efforts at KCP and Savannah River (SR) will nowfocus on use of the DOE NetU system for sensitiveunclassified information rather than a combination of opensystems and SecureNet.

The unclassified electronic version of the KCP materialsstandards database originally made available on the openInternet last year with limited data was well received andhas now been almost fully populated. Y-12 spent much ofthe last year populating databases with materials-relatedinformation, reviewing and correcting it during entry.Textual data from all of the unclassified Y-12 technical data

sheets have been captured, and scanned imagesfrom the reports are being added to theseelectronic files. Pantex completed the design ofan electronic high explosive surveillance databaseand has populated it with some of the higherquality unclassified data available from existingsources. (The final product will be an editeddatabase, not merely an archive). Equipmentnow exists at Savannah River to scan and digitizereports and micrographs, although an effort isunderway to upgrade its quality to bettercapture micrograph information. An SR materialproperty database home page has beendeveloped and will be put online via NetU assoon as administratively possible. Widely used (orspecifically requested) SR documents on materialproperties in hydrogen isotope environments willbe accessible electronically from this page.

Table of contents:Organized listing of (unclassified) materials and

associated numeric designators

Material ####### standard:

Specifications for ordering and testing

material. Contacts at KCP and design

labs. Etc.

Material properties

KCP home page

Technologies and

capabilities

http://www.kcp.com

GOAL: similar systems on SecureNet and NetU

Material standards

2000000

#######

contact ph

contact ph

List of numeric designators of

all KCP materials standards

Text search engine addition being explored

Not available yet, but an ultimate goal

username

password

Easy access to unclassified Kansas City Plant materialstandards database from a desktop computer.

Systems • LA14 69


Enhanced Surveillance Information Management Program The purpose of this task is to improve access to all

necessary information so that surveillance engineers andscientists can answer questions concerning the safety andreliability of stockpile weapons in a timely and efficientmanner. Our goal is to implement a better system forinformation acquisition, storage, and retrieval. This task isbeing undertaken in conjunction with Task LL27 and theSurveillance Information Group (SIG).

In addition to identifying specific information needed toanalyze surveillance units, we have mapped the currentprocess and identified limitations and roadblocks inretrieving surveillance information and in delivering it to thenuclear design agencies. We demonstrated communicationsprotocols to bring weapon part or surveillance data to thedesktop of engineers across the nuclear weapons complex

(NWC). In conjunction with other initiatives, suchas the Accelerated Strategic Computing Initiative,the Advanced Design and ProductionTechnologies, and the Nuclear WeaponsInformation Project, we have also initiated pilotprojects to improve access to information.

Classified communication throughout most ofthe NWC is now possible via classified email. Wehave also established connectivity that usesSecureNet to access the Y-12 web site. To improveelectronic information dissemination, we alsohelped with the design of a CD-ROM systemcontaining dismantlement information from Y-12.All future dismantlement information will beprovided in this manner.

Historical product manufacturing informationas well as current electronic data from tests werecaptured electronically and converted to a format(portable document format, or PDF) that can beaccessed and searched across different computerplatforms. For example, we helped incorporate aCD information system into the W80 cyclereporting system from Pantex. The goal is toimprove surveillance analysis by providing thenecessary tools for surveillance engineers toanalyze trends and do statistical analysis of eachweapon system for cycle reporting.

We have developed a prototype SurveillanceQuality Evaluation Tracking System using MicrosoftAccess for tracking surveillance components in thestockpile. The quality evaluation tracking (QET)points of contact (POCs) are beginning to use theMicrosoft Access database system. We areexploring the cost effectiveness of upgrading allQET POCs to the latest version of MicrosoftAccess. We designed, developed, and presentedseveral QET reports to DOE.

We have developed and implemented data-collection process flow throughout the complex

for two weapon systems, theW88 and the B83. Thisinformation was reported ona CD and included record-of-assembly data, test data,radiographic reports, etc. Theexercise defined many pitfallsto be corrected in the data-collection process throughoutthe NWC. The first figureoutlines the flow ofinformation at the highestlevel used in this process.The second figure is SandiaNational Laboratories’process flow.

Customer transfers data file to host

computer for processing

See Pantex data flow

See LANL data flow

*Record of Assembly

Customer requests ROA* data retrieval by

weapon

Dept. 9781 enters retrieval requests & starts program

Dept. 9781 transfers data file

Program builds data file

ROA data file on disk

ROA data file on disk

LANL procedure (bomb book)

Radiograph report

SRP procedure for retrieval

KCP procedure for retrieval

KCP

Sandia

Weapon

S/N

Y-12

ROA

Pantex

SRP LANL

Chemistry reportSXR reportVAR report

See Sandia process overview

LANL

LLNL

Network storage system

Assembly tractability database

Process flow for the Surveillance Information Group.

Process flow at Sandia National Laboratories.

70 Systems • LA17


Enhanced Reliability Methodology ProgramThe objective of this task is a validated methodology,

based on expert judgment, for assessing the reliability ofthe stockpile. Reliability is defined as the probability of acomponent or system achieving some performancemeasure (e.g., the reliability of a nuclear package would beits probability of achieving the certified yield).

Expert judgment is used to assess available knowledgeand interpret and synthesize it to estimate performance,namely, reliability and its associated uncertainty. Whenexperts deem that there are sufficient test data, reliabilitycan be calculated through classical statistics. However,when data are insufficient, experts may have to givesubjective estimates of performance based on theirsynthesis and interpretation of limited nuclear and non-nuclear tests, numerical simulation, material properties,communications with colleagues, and/or production andsurveillance data.

The methodology includes several elicitation techniquesthat may be used, depending on circumstances or theexpert’s needs. For example, the rule-based performanceprediction technique has been used when quantitative datawere lacking. This technique involves interviewing anexpert on the component or subsystem being evaluated tolearn which conditions would lead to degraded yield. Itelicits the expert’s membership functions and fuzzy ruleson these conditions (e.g., if x condition is high and ycondition is low, then yield is low); has the expert assignnumbers to these rules; and generates plots of yield as afunction of the conditions with the associateduncertainties. The technique provides a means forforecasting reliability as a function of the components’expected conditions. Forecasts have been obtained usinghypothetical component conditions.

The methodology includes a knowledge base, logicmodels, and elicitation techniques. Users can electronicallybrowse the information underlying the reliabilities anduncertainties, as well as the experts’ interpretation of thisinformation. The methodology is expected to offer asystematic and quantitative way to evaluate the effect onreliability of such actions as replacing a component, orconducting additional studies/testing, such as for stockpile

life extension. The test bed for the methodology isthe W76 nuclear package.

The methodology is being validated in theautomotive industry, where test data are moreplentiful. The aim is to check engineers’ earlyreliability estimates for systems such as fuel pumpsagainst subsequent test results and later “roadwarranty” data.

Three early validation results have beenreceived from the automotive environment as ofJuly 1998.

Reliability and uncertainty estimates by autoexperts on two systems predicted successfulperformance, later confirmed by test data. For athird system, the experts predicted a 95% chanceof failure to meet specifications and being recalled.Unlike the other two, this system was totally new,so its initial predictions had to be based solely onsubjective expert judgment. Independent tests ofprototypes showed that all of the 40 testedsystems failed, and in the way predicted by theexperts. This result has increased our confidencein the methodology and in the accuracy ofexperts’ subjective assessments, especially wheretest data are sparse; it has positive ramificationsfor estimating nuclear physics weapon reliability.

The prototyped software is available fordemonstration on a small, secure networkedsystem. It is being adapted and fielded at LosAlamos National Laboratory (LANL).

We incorporated physics-related chemical/engineering information into the knowledge base.Such information was obtained through ourinteraction with the Global Weapons InformationSystem (GWIS) project and through ourcollaboration with the surveillance program.

We investigated ways of integrating physicsperformance with the larger reliability framework. We have linked to other secure web pages atLANL as part of our plan to participate in thearchive integration effort.

Weapon designEngineering

Dynamic experimentation

Logic models of weapon divisions/groups

Expert judgment-based reliabilities and uncertainties. Underlying information data assumptions, interpretation.

Expert judgment is needed in ourmethodology to assess availableknowledge to estimate reliabilityand its associated uncertainty.

Systems • LA23 71


The Stockpile Surveillance/Data Management SystemThe Stockpile Surveillance/Data Management System

(SS/DMS) is a comprehensive database that serves as therecord copy of the data required by the Department ofEnergy for the Stockpile Surveillance Program. Existingstockpile surveillance data sets were converted into amodern, structured query language-compliant format.The SS/DMS currently contains the data from new materialand stockpile surveillance laboratory tests. That set of dataconsists of over two million attributes of data entries fromover 43,000 data sets collected from approximately19,000 sampled weapons.

The analytical toolset developed as a result of this taskprovides an automated, mathematically rigorous capabilitywith which to examine existing and future data for trendsfrom any cause. Because component production dates arepart of the SS/DMS test data sets, component performanceas a function of age is readily observable.

For example, the plot in the figure below was generateddirectly from a desktop computer linked to the SS/DMSserver through a classified local area network (data shownhere are unclassified data resident in the SS/DMSdatabase). The system evaluation engineer first submitted arequest to plot the values of the regulated high voltage atfiring as a function of component age. The SS/DMS queryengine collected the appropriate data (from the over twomillion data entries) and passed the selected data to theanalytical toolset, which generated the screen shown in thefigure. The elapsed time from request submittal to screendisplay was 30 seconds, with no further human interaction.The screen display shows all the data collected, a least-squares fit linear-regression plot (estimating equation), a

95% confidence interval for the estimatingequation, the parameters for the estimatingequation, and the value for the coefficient ofdetermination (R-squared). These results showboth visually and mathematically that there isessentially no significant mathematicalrelationship between the regulated high voltageat firing and the age of the component. Datavalues outside the confidence interval bands aredue strictly to natural variation, not age. Giventhe number of data points collected for thisquery, the analysis demonstrates no discernibletrend of degradation due to component aging.

Without the availability of the analyticaltoolset, obtaining this figure and theunderstanding gained from it would haverequired many hours of transcribing individualdata points into a statistical data analysis packagewith the potential of transcription errors. Withthe analytical toolset embedded in the SS/DMS,the entire data collection and analysis processtook 30 seconds, all under computer controlwith no further human interaction. Thissignificant saving in time allows the systemevaluation engineer to evaluate and interpret thedata rather than struggle with data transcription.The incorporation of the analytical toolset fundedby this task has contributed substantially to ourability to manipulate the millions of test datavalues collected over the last four decades by theStockpile Surveillance Program.

SS/DMS plot ofregulated highvoltage at firing as afunction ofcomponent age.

Age, years

Reg

ulat

ed h

igh

vo

ltag

e at

fir

ing

2.4

2.2

2.6

0 5 10 15

72 Systems • SN15

Secondaries

74


Secondaries

Focus AreaAssuming non-nuclear firing devices and the

primary portion of a nuclear weapon function asdesigned, the yield of the weapon depends upon theperformance of the secondary portion. The secondaryis a complex device with several potential avenues forperformance variation because it is composed of manymaterials and components assembled in a variety ofways. The assembled unit must be geometricallyprecise, and the materials included must remain intheir as-designed, as-installed condition to meetperformance expectations. Historical surveillance datahave revealed that materials interactions may affectperformance of the secondary in the aging stockpile.The Secondary-Specific Focus Area is benchmarked bythese observations and by investigation of otherpotential areas of concern.

This focus area includes a mix of materials science,supportive experiments, data analysis (historical and

new data), and modeling. These secondary-specificactivities integrate closely with related activities inother program elements, such as diagnostics. Theoverall objective is to develop understanding of theaging of materials, components, and systems to enableassessment and prediction of performance for decadesto come. Past knowledge and experience have beenbounded by relatively short stockpile lifetimes becauseweapons were replaced frequently with new designs.Modeling and prediction based on several decades ofperformance were adequate. Performance models areconstantly being improved and benchmarked againstnew data generated by ESP tasks. By approximatelyFY2000, performance models for the secondary will becomplete. Recognizing that prediction is not 100%accurate, new stockpile surveillance techniques are alsobeing developed to give early indication of any adversechanges in components.

Deliverables1. Results of experiments that characterize thechemical, physical, and mechanical changes that mayoccur as a result of aging.

2. Reports that describe the changes that might beexpected as a result of aging in constituentcomponents.

3. Predictive models that include aging mechanisms ofthe actual secondary configurations and can be usedto assess the expected life of secondary components.

Pertinent TasksLA10 Material Study on Special MaterialLA15 Uranium Corrosion and AgingLA22 Aging of Special MaterialLA42 Integrated Canned Subassembly ModelingLL20 Experimental Interaction and Degradation of

Materials of InterestLL22 Modeling of Materials CompatibilityLL23 Atomic Site-Specific Chemistry and PhysicsLL24 Development of High-Energy Neutron

RadiographyLL25 Development of Tests for Non-Traditional

ComponentsLL32 Microsensors for Real-Time Evaluation of

Materials Degradation and Corrosion in Weapon Systems

LL42 Sensitivity AnalysisOR02 Interrogation of Disassembled UnitsOR07 Hydrolysis Reaction on Special MaterialOR14 Materials Properties Evaluation of Uranium

Components


Effects of Solid–Gas Interactions on Component Aging We are combining a series of experimental and

theoretical tools to generate an improved understandingof canned subassembly (CSA) aging and a correspondingpredictive computer model. Three interrelated projectsinclude study and experiments in (1) materials interactionand degradation, (2) material compatibility modeling, and(3) site-specific atomic chemistry and physics. The goal isto blend together studies of the fundamental nature ofimportant gas-solid interactions with macroscopic modelsand experiments on the systems in relevant geometries.Understanding CSA components will also help us tounderstand the effects of any aging changes on structuralcomponents.

Because a CSA is large, expensive, and complicated,small-scale experimental test units have been prepared.By placing small amounts of materials in various simplecontrolled geometries, we can investigate the gas andsurface chemistry using techniques developed under otherESP tasks. These techniques then can be used as a test bedfor the macroscopic aging code.

The secondary is a sealed assembly containing a varietyof solid materials. The components can be characterized asnuclear or structural. Components that often includeuranium and lithium compound are the nuclear elementsof the thermonuclear device. The structural componentsinclude organic materials (among others) for maintainingthe correct geometry. The important distinction betweenthe two classes of materials is that, in the absence ofnuclear testing, it is very unlikely that the system’s nuclearcomponents can be changed without affecting itsperformance, whereas there is considerably more leewayon changes to the structural parts.

If the CSA contains uranium, the most likely destructiveevent in a CSA is the hydriding of uranium by molecularhydrogen. Despite efforts to remove H2 introduced duringthe manufacturing process, a small amount of moisture isalways present. The water in the CSA can react to formuranium hydride. These ESP tasks are a combined effort toexamine competing reactions and required gas transport ata molecular and macroscopic level.

Our primary long-term goal is a predictivecomputer model using both realistic CSAgeometries and accurate detailed chemistry andtransport. In the first step toward that goal, weare performing detailed three-dimensional“zoning” of the gas transport pathways througha representative CSA. Using Lawrence LivermoreNational Laboratory’s considerable expertise intransport codes, such as TOPAZ and ALE-3D, weare modifying these codes to include solid-gasreactions. In 1998, we developed a 3D meshand preliminary mass transfer calculation for H2migration in a stockpile system using the ALE-3D code.

One possible CSA reaction, the LiH/H2Oreaction, is being studied with molecular beamtechniques and quantum chemical modeling.Both theory and experiment show that the initialreaction, LiH + H2O —> LiOH + H2 is very rapidon a fresh LiH surface. It could be regarded asinstantaneous on the timescale of themacroscopic simulations. However, as this is agas-surface reaction, the resulting hydroxideslows continued reaction considerably. There arethen two other reactions, LiH + LiOH —> Li2O +H2O and 2LiOH —> Li2O + H2O, which aremuch slower. These reactions are also beingstudied experimentally and theoretically in orderto develop a rate model for the predictivecapability.

A similar effort is underway for developinguranium hydriding and organic hydridingmodels. Experiments indicate that the reaction ishighly localized on the surface. A model hasbeen developed that incorporates nonlinearkinetics. The presence of preferred sites showsbehavior similar to that seen in molecular beamexperiments. After the model is refined andparameterized, it will be incorporated into ourtransport code.

“Snapshot” from a quantumchemical study of the reaction ofwater with a lithium compound. Awater molecule attaches rapidly tothe surface and then rapidly reacts.

Secondaries • LL20/LL22/LL23 75


Interrogation of Disassembled UnitsBecause of the nuclear test ban and the need for weapon

life extension, information about the aging of weapons andpredicting lifetimes of the enduring U.S. nuclear stockpile ismore crucial than ever. Many of these weapons have beenin service far longer than their designed service lives, andthey are being removed from the stockpile fordismantlement because of the reduction in the U.S. nucleararsenal. This dismantlement effort provides a window ofopportunity to gather data on real-time-aged parts fromdisassembled canned subassemblies (CSAs).

The main purpose of this effort is to secure material,components, and information from returned weaponassemblies for aging assessment and life prediction tests.This task also implements enhancements to the currentinterrogation techniques in the surveillance program. Thecomponents and information recovered under this task arebeing used in other Enhanced Surveillance Program (ESP)

tasks to establish benchmarks and predict theaging of weapons in the enduring stockpile.

During FY1998, we transferred more than 100 components to ESP task leaders for agingprediction and life extension tests. An additional616 components have already been identifiedfor transfer.

We combined resources with Task LL27 toretrieve historical weapons assembly anddisassembly reports. All of the disassembly recordspublished for the Los Alamos and LawrenceLivermore national laboratories weapons wereconverted to digital media. We also worked withTask OR18 to assess the procurement ofsurveillance instruments.

We incorporated new interrogation techniquesinto routine surveillance. The first figure shows thevideo imaging system that was implemented inFY1998 to characterize, map, and digitally archiveanomalies in weapon components. Prototypes fortwo other techniques have been designed andbuilt for integration into routine surveillance inFY1999. These techniques include the low-temperature thermal decomposition (LTTD)system and the diffuse reflectance infrared fouriertransform spectrometer. Both of these technologiesprovide real-time, online diagnostic capabilities forunderstanding weapon aging. LTTD, shown inthe second figure, uses temperature-programmeddesorption to characterize component chemicalcharacteristics that might be weapon agingindicators.

Video imaging systemcharacterizes weaponanomalies (top). LTTDsystem studies agedcomponent chemicalcharacteristics (right).

76 Secondaries • OR02


Uranium Corrosion and AgingThis multifaceted project is working to understand the

microstructure and chemistry of unalloyed uranium instockpile components as they influence corrosion. This year,we also focused on the CSA aging modeling effort and, incollaboration with the U.K. Atomic Weapons Establishment,on experiments and modeling of uranium corrosion.

Microstructural studies have focused on the effects ofcarbon (the most common impurity in uranium) in theevolution of the uranium microstructure during acceleratedaging tests. Data have not yet been fully analyzed,although some very interesting results have been observed.For example, at constant carbon levels, we observedsignificant variations in hardness, grain refinement, andcarbide morphology. Because variations in mechanicalproperties clearly correlate with gross variations in thenucleation and growth kinetics of uranium hydride, thesestudies are leading us toward a better understanding ofhow physical metallurgy of the material affects thecorrosion mechanism.

We are also studying how the local nucleation of hydrideaffects the structure of the surface oxide by using atomicforce microscopy (AFM), scanning tunneling microscopy(STM), and Auger electron spectroscopy (AES) in inert gasat ultrahigh vacuum. Our studies show that the oxidizedsurface is very complex, exhibiting wide variations inthickness and topography between samples, betweengrains in the same sample, and within individual grains.However, relating this fine-scale variability to the relativelysparse hydride initiation sites is very difficult. The surfaceoxide layer does not appear to control the sites wherehydride attack is initiated, although it must play a role inthe induction period prior to hydride initiation. This work isdiscussed in greater detail in the Proceedings of the MaterialsResearch Society Symposium on Hydrogen in Metals.

Our studies on in situ observations of hydride nucleationand growth have, over the past two years, led us toconclude that variations in the chemistry of the near-surfaceregion of the sample strongly affect the local hydrideinitiation phenomenon. Numerous experiments with

uranium samples of various carbon contents,thermal histories, and levels of cold work showthat hydride nucleation sites are in the regionof second-phase particles (inclusions) in theuranium. We postulate that the region aroundinclusions, while depleted in the elements thatform the inclusions themselves (C, Si, etc.), isalso enriched in elements that are insoluble inthe precipitate. This gradient in impurityconcentration around inclusions thereforecontrols the locations where the hydridenucleates. In the figure below, apparent hydridenuclei have formed in rings around an inclusion.The hydride appears to precipitate below thesurface of the metal. This behavior is consistentwith published models of hydride growth buthas never been previously documented. Wesuspect that some impurities may act as catalystsfor the formation of uranium hydride. Forexample, we have determined using AES thatsulfur, calcium, and chlorine segregate to theuranium surface during a brief anneal at 200°C.

We are working to isolate the effects ofindividual impurities. Collaborators at Y-12have researched historical documentation of thechemistry of uranium parts produced there andidentified a stock of aged, depleted uranium thatappears to closely parallel the processing historyand chemistry of similar-vintage enricheduranium. We have taken advantage of newcapabilities for analyzing traces of uranium. Weare also preparing to produce test specimenswith tightly controlled levels of various impuritiesof interest. This effort is made possible by therecent acquisition of very-high-purityelectrorefined uranium produced by researchersat Argonne National Laboratory and materialpurified by researchers at Los Alamos NationalLaboratory.

These images show the surface of a cylinderof uranium containing nominally 300 ppmof C, which was exposed to hydrogen for 10 min at 70°C. The hydriding appears toinitiate adjacent to inclusions. The pit (leftimage) is adjacent to an inclusion in thecenter. The right image shows a ring ofwhat apparently are hydride blisters formedaround a pair of inclusions.

Secondaries • LA15 77


Material Properties Evaluation of Uranium Components

This task was begun in 1998 to develop anunderstanding of the aging process of components madefrom uranium-6% niobium alloy. The primary concern isthe varied appearance of binary material after aging andhow this may relate to changes in bulk properties. Thesurface condition varies from a very thin, light scale to aheavy black oxide. A limited number of previous studieshave shown that the varied surface scales may beassociated with local variations in Nb content. However,this relationship has not been quantified nor is there anybasis upon which to relate the observable surfacecondition with the bulk properties of the metal.

To correlate surface scales with variations in Nb content,we need an understanding of the oxidation mechanismand measurements of reaction kinetics. Although thedevelopment of oxidation-resistant U-Nb alloys was startedin the early 1950s, no studies published to date haveexplained how niobium imparts oxidation resistance touranium. One of the most commonly repeated theories isthat the dissolution of Nb in the UO2 lattice modifies theelectronic structure of the oxide, resulting in sloweroxygen transport through the lattice. If this were true, theoxidation rate would be reduced relative to that of pure U,but over the long term oxidation would still proceedunabated. An alternate theory is that the Nb forms a verythin second-phase oxide (Nb2O5) that protects theunderlying metal by acting as an oxygen diffusion barrier.Therefore our initial research has concentrated on

determining the corrosion mechanism of U-Nb alloys.

In this study, the initial stages of corrosion ofa U alloy with 6 wt% Nb by O2 or H2O wereexamined using the surface-specific techniquesof x-ray photoelectron spectroscopy (XPS),thermal programmed desorption, staticsecondary-ion mass spectroscopy, and sputteredneutrals mass spectroscopy (SNMS). X-rayphotoelectron spectroscopy studies of the U6Nbsurface following oxidation using O2 at 300 Kindicate production of a thin oxide overlayer ofstoichiometric UO2.0 intermixed with Nb2O5.While the same stoichiometry is exhibited foruranium when the oxide is prepared at 500 Kwith O2, niobium is much less oxidizedexhibiting a mixture of NbO and Nb. Contraryto previous XPS literature, the SNMS depthprofiling studies reveal that oxidation by O2 ismuch greater (as judged by oxide layerthicknesses) than that exhibited by H2O. Oxidelayer thicknesses of less than 20 Å are createdusing H2O as an oxidant at 300 K withexposures of <5000 L (oxide layer thicknessescreated from O2 exposures are approximately anorder of magnitude higher). Formation of acritical density of Nb2O5 is suggested to beresponsible for the enhanced corrosionresistance by preventing diffusion of O– (O2–) orOH– into the oxide/metal interface region. Thiswork is consistent with studies of materialintentionally oxidized in air.

It is therefore apparent that the addition ofniobium to uranium results in oxidationresistance solely by the formation of a protectiveNb2O5 network within the UO2 scale. At lowerniobium levels, the oxide must grow to a greaterthickness before this network is established. Atstill lower levels or in metal where the niobium isnot uniformly distributed, the oxidationproceeds identically to that of pure U. We arenow proceeding to demonstrate the relationshipbetween oxide thickness and Nb content incontrolled ingots produced at Los AlamosNational Laboratory and with actual stockpilecomponents obtained from disassembly. InFY1999, we plan to expand our studies toinclude mechanical properties characterizationin order to correlate variations in Nb contentand possible environmental embrittlement ofthe metal.

Niobiumoxide at 300 and 500 K on U6Nb

Nb 3d

Inte

nsi

ty

215 210 205Binding energy, eV

200

500 L O2 at 300 k

500 L O2 at 500 k

Clean U6Nb

1000 cps

The niobium 3dcore levels for U6Nbexposed to 500 L O2at 300 K (bottomtrace), U6Nbexposed to 500 L O2at 500 K (middletrace), and cleanU6Nb (top trace).

78 Secondaries • OR14

Materials-Aging Models

80


Materials-Aging models

Focus AreaThe reliability and surety of non-nuclear components

and subsystems in the enduring stockpile will change withtime as a result of the age-related degradation of theirconstituent materials. Existing stockpile assessmentmethods typically address materials-aging and reliabilityproblems after they occur, rather than anticipating andavoiding future problems or failure mechanisms. This focusarea promotes efforts to improve our ability to anticipatesurety and performance-reliability problems caused bymaterials aging.

In order to predict and thus avoid or correct reliabilityproblems in the surety and performance of weapons,significant improvement must be gained in modeling andsimulating the behavior of materials as they age in thestockpile environment. Detailed understanding of materialsencompasses the signatures of aging and degradationphenomena and projecting these phenomena out toeventual failure. This focus area supports investigations toimprove the fundamental understanding and modeling/simulation of materials-aging behavior in stockpileenvironments. Individual tasks emphasize the prediction offailures that affect the surety or reliable performance ofweapons. Supported investigations focus on quantifyingthe difference between “performance requirements” and“performance limits” and how this changes with time. Thisunderstanding is necessary to extend the lifetime of weaponsin a safe and dependable manner. The relevant integrationof materials understanding, materials modeling, andperformance simulations is the hallmark of this focus area.

Five principal categories of time-dependent changes in stockpile materials are being addressed. They include (1) energetic materials models, (2) organic materialsmodels, (3) solid-state materials models, (4) corrosionmodels, and (5) interface models. Our selection of thesecategories is the result of a risk assessment of stockpilematerials, which identified materials with the highestlikelihood of age-related changes and with the highestconsequence of material changes. Likelihood of aging is based upon a combination of data from the stockpilesurveillance history and engineering judgement fromour present understanding of the kinetics and thermo-dynamics of material degradation mechanisms instockpile environments. Evaluation of the consequenceof aging is based on our understanding of the designintent of the materials in the weapons, the impact ofdegradation of safety and reliability, and of otherrelated factors.

We will develop and validate age-aware materialsmodels that will be used in quantifying the useful servicelifetimes of non-nuclear components and subsystems inthe enduring nuclear weapon stockpile. We will alsoaccount for confidence bounds related to production orenvironmental variability. Where time-dependent changesin the structure and properties of critical weapon materialscannot be modeled with confidence, appropriatematerials characterization tools and surveillance methodsthat provide early signatures of aging will be identified.Models developed will be incorporated into higher-levelelectrical and mechanical models for component andsubsystem performance (developed in the Non-nuclearComponents Focus Area) to support the prediction ofoverall changes in system reliability with time.

Deliverables1. Develop model of the degradation of solid film lubricantfor application in the dynamic performance model of theMC2969 stronglink.2. Apply the PACER gas analysis surveillance method to quantify the leakage or permeation of outside air into thesealed weapon environment.3. Develop aging model for T-111 base and weld metal for inclusion in the pressure burst simulation of radioisotopic thermoelectric generator, or RTG, power sources.4. Develop aging model for PBX9501 in lens isolator.5. Integrate atmospheric corrosion model into electronic device performance model.6. Apply the “wear out” remaining life prediction technique to polymers from returned weapon hardware.

Pertinent TasksSN01 Atmospheric Corrosion of Weapon ElectronicsSN02 Aging of T-111 Welds in Radioisotope

Thermoelectric GeneratorsSN03 Stress Voiding of Integrated CircuitsSN04 Energetic Materials Predictive CapabilitySN05 Predicting and Verifying Elastomer LifetimesSN06 Ceramic Materials Lifetime Predictive CapabilitySN07 Degradation of Materials in Electromechanical

Devices


Atmospheric Corrosion of Weapon ElectronicsCorrosion caused by the atmosphere within nuclear

weapons is the leading age-related degradation modeobserved in the stockpile and forms the single largestSignificant Finding Investigation category. Traditionally, theseverity and potential consequences of corrosion problemswere addressed using engineering judgment, and whennecessary, materials of construction and the localenvironment were changed. Now, faced with an agingstockpile and active interest in weapon lifetime extension,our high standards for surety demand that we effectivelyanticipate any effect of corrosion processes on weaponreliability and nuclear safety.

Two general corrosion modes are relevant to weapon-system degradation: (1) product-layer growth (e.g.,sulfidation of electrical contacts) and (2) metal dissolution(e.g., corrosion of microelectronic aluminum metallization).Although structural and mechanical components alsocorrode, our main concern is corrosion in electroniccomponents because of the direct consequence on systemfunction and therefore reliability.

During 1996, a relevant example of Type I corrosion wasobserved on SA1388 diodes. During a 5-year storageperiod for war reserve product, a gaseous sulfur speciesemitted from rubber bands caused the growth of a copper-sulfide layer across the surface of a significant number ofdiodes, increasing reverse-bias leakage current (see figure).Once a corrosion product layer bridges the two leads, aconductive path is formed and conductance increases asthe film spreads and thickens. This degradation process canproduce subsystem and systemfailure that could occur in anactual weapon when anundesirable sulfur-containingenvironment exists (e.g.,polysulfide staking material orSO2-containing micro-balloons).

Our corrosion study combinescontinuum/phenomenological-level modeling with age-awareelectrical subsystem models,addressing both uncertainty andthe structure of the end-userinterface. Many types ofuncertainty are applicable,including extrinsic factors(environment), intrinsic factors(material properties, stochasticnature of corrosion initiation),modeling simulation (processsimplification), and experimentalmeasurement error (bias,unknowns). To ensure that a

proper end-use focus is part of every aspect ofthe integrated program and to allow time forunique computational or database requirements,this task is proceeding in parallel with thedevelopment of physically based corrosionprocess models.

We took a primarily empirical approach. Agoverning equation, based on the solution of themass balance (continuity) equation for theproduction and consumption of gaseous sulfurspecies, was derived to describe the rate ofproduction of copper sulfide on the diode. Whileseveral rate constants are deterministic, epoxyporosity, initial sulfide source concentration, andeventual film-growth characteristics arestochastic. Traditional computational reliabilitytechniques (Monte Carlo, pseudo Monte Carlo)were applied to treat the uncertainty and solvethe equation. Because of the stochastic nature ofthe corrosion process and environmentalvariability, the specific output is a distribution ofshort-circuit electrical resistance (conductance)values as a function of time. This materials-degradation information is input into age-aware,sub-system-level electrical models of thecomponents of interest to predict when thecomponents will no longer meet performancerequirements.

Diode sulfidation is shown in (a) andin cross section in(b); modeleddistributions ofreverse-biasconductance aregiven in (c) for threeinitial conditions:clean diodes, diodesalready sulfidizedduring 5-yearstorage, and cleandiodes encapsulatedin polyurethanefoam.

Corrosionproducts

Epoxy

15 µm, 2000 x

(a) (b)

0

2

4

6

8

10

12

14

0 0.1 0.2 0.3 0.4Conductance, µ-mho

Perc

ent

of

po

pul

atio

n

5 years in stores

FoamedClean

(c)

Materials Aging • SN01 81


Welds in Radioisotope Thermoelectric GeneratorsRadioisotope thermoelectric generators (RTGs), such as

the MC2730 and MC3500, are electrical power sourcesthat essentially consist of a radioisotope heat source incontact with a thermocouple pile. The heat sources used inRTGs, such as the MC2893, MC2893A, and MC3599, areplutonium oxide fuel encased within a welded tantalumalloy (T-111) capsule. During stockpile life, decay productsfrom the plutonium oxide increase the pressure within thecapsule, which—in order to avoid plutonium dispersal—must be contained during both normal operation andaccidental fire exposure. We are making an integratedassessment of RTG integrity as these heat sources exceedtheir original design life.

Capsule survival in a fire is the most demandingperformance requirement because high temperatures cancause creep in the tantalum weld, and the capsule couldleak. Finite-element models using base metal propertiespredict that the highest stresses occur in the dome of thecapsule, whereas failure often occurs in the weld region(see figure). This indicates that welding changes theproperties of the alloy and results in an unexpected failurelocation during simulated fire testing.

Our current surveillance activities include periodic,pressure burst-testing of inert capsules. Because ofsignificant variability in qualification and surveillance data,an accurate predictive model has not been developed,which explicitly incorporates effects of materials aging. To overcome these difficulties, we initiated an enhancedsurveillance program to improve understanding of creep and aging of T-111 welds. The program includescharacterization of creep deformation and failure modes

and development of weldment creep data for T-111 welds. Potential aging mechanisms werealso evaluated, improved thermal and mechanicalresponse models were developed, and resultswere correlated with qualification andconventional surveillance tests.

Our completed critical assessment of availablecreep data and behavior for T-111 translated basemetal data into a uniform and usable format andshowed that very little data for welds in this alloyare available. A welding system that reproducesthe procedures used in the original war reserve(WR) fabrication has been constructed tofabricate test welds with microstructural featuresas similar as possible to those in the WRcomponents. The welds will be used to developcreep properties data for T-111 welds and toassess the effects of aging on these properties. Thespecialized facilities for conducting the creep testshave been designed and are being assembled.

In parallel, we conducted metallographicanalysis of surveillance pressure burst testsamples. Initial indications show there issubstantial microstructural variation in capsulesdepending on material lot, fabrication site, failuretime, and possibly aging conditions. Hafniumoxide precipitates (hafnium is a constituent of thealloy) are present along the grain boundaries inthe welds as well as within the grain interiors;they may affect the failure mode and represent amechanism of aging. A first generation finite-element model of the capsule was constructed,and a plan for model characterization of theeffects of weld properties on heat source pressureburst response was developed. We exercised themodel using tentative values for creep strain-to-failure for weld and base metal over a range ofpressure burst test temperatures and pressures.The results appear to correlate, at leastqualitatively, with available qualification andsurveillance experiments. Experimental weldmentproperties data are currently being generatedthat will be used to validate the currentcomputational model and pinpoint areas whererefinements are needed.

These activities are coordinated withcomponent and weapon systems personnel tomaximize the utility and impact of the work.Further, the efforts are complementary withconventional surveillance activities and, whencompleted, should provide an integrated andpredictive surveillance capability for RTG heatsource reliability.

Finite-element analysis using currently available properties datapredicts that highest stresses would cause failure in the dome. Testfailures typically occur in or near weld, with burst severity and time-to-failure highly variable.

82 Materials Aging • SN02


Energetic Materials Predictive CapabilitySandia National Laboratories and Pantex are

collaborating in an effort to qualify changes in agedpentaerythritol tetranitrate (PETN). Understanding thesephenomena will give us a stronger technical basis formaking recommendations regarding replacement,rebuilding, or reacceptance of components containingthis material.

PETN and a silicone binder are components in theexplosive composite (XTX8003) used in the MC3028firing set. The PETN incorporated into XTX8003 is non-equilibrium material and, with aging, can changethermodynamically into new morphological forms.Changes in crystal-size distribution can alter detonationcharacteristics and potentially alter componentperformance. More importantly, the growth of crystals at afree surface represents the migration of energetic materialfrom the bulk composite, which could produce non-detonable regions depending on the nature of thediffusion. We have found that accelerated aging ofXTX8003 causes PETN to recrystallize into perfected crystalsat free surfaces. This redistribution of crystal sizes is alsobeing followed in bulk PETN samples. To avoid suchchanges, PETN is typically stabilized before long-term use,but stabilization is not used for the PETN in XTX8003.Because performance of the firing set depends on bothenergetic output and simultaneity, we intend to rigorouslyquantify the extent and effects of these aging phenomena.

In previous years, we developed the analytical tools toidentify numerous decomposition products of PETN. Theseproducts are not seen in naturally aged stockpile samples.The chemical stability of PETN in XTX8003 appears to beexcellent when the weapon environment is considered.Accelerated aging of PETN under conditions outside thestockpile-to-target sequence (STS), however, producestraces of decomposition gases, acetone from the

recrystallization of the explosive, and smallamounts of ionic species.

Detailed studies of the diffusion of PETN withinthe silicone binder demonstrate anomalousbehavior. Though not yet quantified under STSconditions, diffusion is rapid above 50°C andapparently is influenced by mobile, low-molecular-weight species within the cured binder matrixthat are known to influence aging and/or tracesof acetone from the PETN.

We have also examined the mobility and vaporpressure of PETN homologues and the effects theymight have on the vapor pressure and mobility ofPETN. These data will allow us to understand theeffects of homologues on PETN composites.

Ongoing velocity-of-detonation (VOD)experiments on aged XTX8003 in thin troughshave shown both increased and decreased VODdepending on aging conditions. The relationshipbetween XTX8003 composition changes andcomponent performance is spectacularlycomplex, but we are quantifying the nature andextent of ongoing diffusion phenomena. As werefine our quantitative understanding of theseactions, we will provide input to an evolvingexplosive/mechanical finite-element model of thefiring set so that we can understand resultingeffects on the function of the whole component.

Importantly, we have not identified any issuesthat would lead us to lose confidence in theperformance of the MC3028 in its currentconfiguration. Rather, we have identified agingphenomena that we must quantitativelyunderstand in order to confidently recommendlifetime extensions for the stockpile.

Explosive composite XTX8003(a) before and (b) after agingshows the growth of crystalson free surfaces; the size of thelarge crystal is 20 × 160 µm.Diffused PETN (c) precipitatesfrom silicone binder in minutesupon cooling in 20- to 40-µmcrystal clusters.

(a) (b) (c)

Materials Aging • SN04 83


Predicting and Verifying Elastomer LifetimesWe are developing unique methods and models for

better predicting and verifying elastomer lifetimes andapplying these advances to other polymeric weaponcomponents and materials. We have previously shown theimportance of diffusion-limited oxidation (DLO) during air-aging and developed theoretical models and uniqueexperimental approaches for predicting and verifying theimportance of such effects. We also derived theultrasensitive oxygen consumption approach, the first andonly quantitative method for testing extrapolations fromaccelerated aging data. This approach was applied to W88EPDM (ethylene-propylene rubber) O-rings, resulting in aconfident prediction of lifetimes exceeding a hundred years.

Environmental O-ring seals are an important polymericcomponent used on all weapon systems to protect theweapon interior from oxygen and water vapor leakage. Thetwo most common measurements for followingdegradation of seal materials are compression stress-relaxation (CSR) and compression set (CS). CSR, whichfollows the decay in force between the seal and its metalmating surface, is the ideal method to use for laboratorysimulations, modeling, and development of extrapolatedpredictions. However, because CSR cannot be easily appliedto real weapons, surveillance activities have used CSmeasurements as an immediate indicator of following thecondition of field-aged O-rings. The results from these CSmeasurements have often been quite interesting, asevidenced by a puzzling bimodal distribution of data forbutyl O-rings used on the W76. Unfortunately, there hasbeen a trend towards eliminating such measurementsbecause they have been difficult to perform using thehistoric hand-held caliper approach. Working with Pantex,we successfully transferred a superior CS method based ona laser micrometer. In addition, working with AlliedSignal/Federal Manufacturing & Technology (AS/FM&T), we havenow determined that the W76 bimodal CS behavior is due

to manufacturer differences, with Vendor Asupplying better butyl O-rings than Vendor B.

Although ideally correlated to the failurecriterion of a seal material, historic CSRmeasurement procedures were found to belacking. We recently published our improvedprocedures, which include new methods forseparating physical and chemical relaxationphenomena and finite-element modeling of DLOfor the disk-shaped samples used for CSRexperiments. We applied these new ideas to butylO-rings from Vendor A that are used asenvironmental seals for the W76, W78, W80, andW87. Modeling and experiments showed thatimportant DLO effects for standard CSR samples(15-mm-diam disks) led to meaningless results.To solve this problem, we developed a novelapproach that eliminated DLO anomalies bycompressing 50 mini-disks (2-mm diam) inparallel. This approach shows oxidative scissionprocesses dominate the CSR force decay,suggesting that our unique, ultrasensitive oxygenconsumption approach might allow us toconfidently extrapolate CSR data to ambientweapon temperatures. We are currently obtainingoxygen consumption measurements down toroom temperature to further test this approach.In addition, due to the recent discovery of thecause of bimodal CS behavior, we initiated a seriesof experiments on butyl material (W76, W78) fromVendor B. Another important question concerningCSR results is the force level that constitutes failure(leakage) of the seal. Working with AS/FM&T, wemodified a CSR device so the sealing force of anO-ring can be correlated to the leakage ratebetween the O-ring and its metal mating surface.

103

102

101

100

β

α'

10–1 100 101 102

10%

20%

30%40%50%60%70%80%

90%

Time at 125°C, days

Forc

e n

orm

aliz

ed t

o t

ime

= 0

0 20 40 60 80

1.00

0.10

0.01

Standard 15-mm disk important DLO effects

2-mm parallel mini-disks

84 Materials Aging • SN05

Theory for DLO of disk-shaped samples explainslarge errors in force decayfor standard CSRexperiments: (left)integrated relativeoxidation (%) as functionsof DLO theory parametersα′ and β for a disk ofradius a, where α′ is adimensionless parameterproportional to a2 and β isa dimensionless parameterproportional to the partialpressure of oxygen;(right) the large effect onforce decay caused bydiffusion-limitedoxidation.

Non-nuclear Components


86

Non-nuclear Components

Focus AreaThe fundamental objective of the Non-nuclear

Components Focus Area is to provide improvedinformation and tools to support stockpile managementdecisions for the enduring stockpile. Historically,surveillance of non-nuclear components and subsystemshas been based upon system-level functional testing.While system-level testing will remain as the principalmethod of non-nuclear surveillance, such performancetesting is forensic rather than predictive in nature—failures are found, but early indicators of unacceptablechanges are generally not observable. This focus area isdesigned to increase our ability to monitor, anticipate,and manage time-dependent changes in non-nuclearcomponents that would otherwise pose high risks tostockpile management. It will link together informationabout the aging of weapon materials withunderstanding of the performance margins of non-nuclear components to enable a predictiveunderstanding of the future evolution of non-nuclearcomponents in the stockpile. This program is intendedto identify the time scale to loss-of-confidence in non-nuclear components and subsystems and to providecomponent-level surveillance methods andrecommendations geared to discovery of changesbefore unacceptable performance degradation occurs.

There are more than 200 major non-nuclearcomponents and subsystems and literally thousands ofsupporting parts. These components (and subsystems)have been prioritized in order to assess and manage riskacross the stockpile. The highest-priority components forenhanced surveillance studies were selected on the basisof their pervasiveness in the stockpile, their effect onreliability and safety, and the likelihood that problemswith their materials would occur. The initialprioritization, performed in FY1996, will be reviewedand revised on a regular basis as increased informationon component-level risks is available. There are presentlythree tiers of non-nuclear component priorities. Tier Ihas 14 components that pose the highest risks to thestockpile, Tier II contains 35 components withsubstantial risks, and Tier III lists the remaining 150+components. It should be noted that meaningful life-extension decision of Tier III components would stillrequire some R&D.

There are three major classes of non-nuclearcomponents and subsystems: electronic componentsand subsystems, electromechanical components, andcomponents and subsystems containing energeticmaterials. Work performed within the EnhancedSurveillance Program is organized into three groups of

Pertinent TasksOR11 Arming, Fuzing, and Firing ShieldsSN23 Age-Induced Degradation of Nuclear SafetySN24 Component Prioritization and Stockpile Life Extension Program IntegrationSN25 Electromechanical Component—StronglinksSN26 Electronic Components and Subsystems—Capacitive Discharge Unit (CDU) Firing SetSN27 Energetic Components and Subsystems—Slim-Loop Ferroelectric (SFE) Firing Set


87

tasks, each representing one of these major classes ofcomponents. Objectives of these tasks are to assessexisting component degradation information, developpredictive component-degradation models that can beused to predict the system reliability over time,recommend and develop additional component-levelsurveillance methods, and guide future stockpiledecisions regarding component replacement.

This focus area has three principal objectives:

1. To identify methods for augmenting the existingstockpile surveillance program. These methods offer thereal possibility of substantially improved insights intoage-induced degradation processes.

2. To develop validated age-aware component andsubsystem models that link time-dependent materialchanges (characterized within the Materials-Aging FocusArea) to changes in system performance, with anemphasis on system reliability and nuclear safety.Validation of these models is accomplished through acombination of materials characterizations, evaluations ofstockpile-returned hardware, and stockpile surveillanceobservations (likely to play a greater role as newmethods are developed and deployed).

3. To provide much of the technical basis for StockpileLife Extension Program refurbishment decisions throughcharacterizations of age-induced degradations that likelyaffect system performance during future deploymentintervals.

Deliverables1. Support W80 decision milestones for CDU firingsets, intent stronglink, trajectory stronglink, andtrajectory sensing signal generators.

2. Develop age-aware performance for rolamites.

3. Furnish refined aging and low-dose-rate radiationmodels for critical electronic components for W80subsystems.

4. Develop a component level model capable ofincorporating aging effects for stronglinkenvironmental sensing devices for the W76.

5. Complete validated age-aware SFE firing set modelfor the W76 and W78.

6. Develop age-aware performance model for radaron the W76.

7. Complete aging and nuclear safety evaluation forthe W78.

8. Estimate safe life of radioisotope thermoelectricgenerators.


Age-Induced Degradation of Nuclear Safety This project explores the effects of material aging on

nuclear safety through a materials-based evaluation ofcomponents whose performance is critical to nuclear safety.The framework for identifying, prioritizing, and evaluatingmaterials issues of potential importance to nuclear safety isidentical to that used for original design and certification,emphasizing the principles of isolation, incompatibility, andinoperability to ensure a predictably safe response acrossthe full spectrum of credible accident threats. In addition toproviding initial aging evaluations, this project provides atechnical basis for guiding future assessments of field-returned hardware as well as a material basis for makingStockpile Life Extension Program (SLEP) refurbishmentdecisions.

Materials used in the manufacture of the componentscomprising the nuclear safety implementation for specificweapon systems were chosen for specific roles in integrity,form, and function. Aging could threaten the safety of aweapon by compromising the integrity and function ofcomponents made up of those materials.

This project has focused on the following: (1)identification of nuclear safety critical components, theirmaterials of manufacture, and the potential time-dependentmaterial degradations that could affect the nuclear safetyfunction of the components; (2) prioritization of materialissues and qualitative placement of issues into categoriesbased on importance to the weapon nuclear safety theme;(3) for each identified potential degradation, qualitativecategorization of the impact severity to the nuclear safetyfunction of the component; (4) development of a processfor potential materials degradation information to becommunicated and acted upon by the system andcomponent engineers in their SLEP, dual revalidation, andstockpile surveillance activities; and (5) incorporation of the

potential materials degradation information intothe nuclear safety assessment activities of specificweapon systems.

The loss of the internal silicone fluid of astronglink is an example of a plausible time-dependent degradation that would adverselyaffect weapon nuclear safety. Such a fluid lossmay allow the stronglink to respond to non-unique forms of energy (i.e., to forms of energynot unique to the intended use environments),thus violating the safety principle ofincompatibility.

To date, this project has examined fourweapon systems and hundreds of materials andmaterial combinations with respect to theirnuclear safety impacts. There is a process in placeto systematically work through the openmaterials issues, and field-returned hardware isbeing examined for evidence of the potentialdegradations. The potential degradations are alsobeing communicated to the appropriate systemengineers to be incorporated into the ongoingSLEP and revalidation efforts in progress.

To facilitate a rapid and managed retrieval ofspecific weapon system, component, materials,and safety information, the entire EnhancedSurveillance Report on each system in this projectwill be incorporated into the new format forfuture safety Sandia National Laboratories reportson weapon systems. Future nuclear safetyassessment reports will be a five-volume series.Volume 5 will be the ESP project results from theAge-Induced Degradation of Nuclear Safety task.This change is a direct result of this project.

Enhanced nucleardetonation safety is basedon the principles of isolation,incompatibility, inoperability,and independence(suggested by the two inputsignals) to assure a safe,predictable response acrossa wide spectrum of credibleaccident threats.

Stronglinks Weak link

Exclusion region

Nuclear explosive package

Non-catastrophicaccident

Incompatibility

Direct electrical threat

Isolation

Catastrophic accident

Inoperability

88 Non-nuclear Components • SN23


Component Prioritization and Stockpile Life Extension Program Integration

The Sandia National Laboratories, AlliedSignal/FederalManufacturing & Technologies (AS/FM&T), and Pantexhave been investigating since early FY1997 the agingcharacterizations of non-nuclear components whoseperformance is critical to sustaining the nuclear safety andreliability of the enduring stockpile. These activities, whichbuild on materials research whose origins pre-date theEnhanced Surveillance Program (ESP), grew under ESPsponsorship to encompass component-level characterizationswith a continuation of relevant materials support. Mostrecently, ESP has tried to enlist Sandia weapon systemorganizations as equal partners to assess the system impactof preliminary component characterizations and to championESP objectives and contribute to meaningful outcomes.

Natural partnerships with weapon system organizationswere impeded by their more immediate commitments anda lack of resources to provide the necessary coordination.The W76 Dual Revalidation (DR) program, conductedconcurrently with ESP, was an important exception to thesegeneralizations. At both Sandia and AS/FM&T, ESP was wellunderway in characterizing selected components andmaterials from the W76. Dual revalidation investments atboth facilities broadened the investigation of W76components well beyond the original ESP portfolio, and DRwas instrumental in acquiring substantial stockpile hardwareto aid in validation of postulated aging mechanisms. TheDR experience was a vital exploration of how system andcomponent organizations can work together to ensureuseful, achievable, and timely ESP outcomes.

Nevertheless, it was not the DR model but the emergenceof the Stockpile Life Extension Program (SLEP) that drovepartnerships with other weapon systems. Through SLEP,the project groups began to specify dates for systemrefurbishment decisions. The W76 and W80 systems, whichhave the more immediate decision milestones, are enteringinto phase 6.2/6.2A studies to evaluate refurbishmentoptions and costs. The W80 program has already identifieda minimum set of components for replacement. Othercomponents are less obvious and represent the greatestopportunity for ESP to influence the decision process. While

the W76 study is less certain to result in near-termstockpile refurbishments, component aging willweigh heavily in refurbishment decisions forW80/W76 systems as we explore the risks ofstockpile components that are well beyond theiroriginal expected life. The most valuable ESPoutcome would be high-confidence predictions ofcomponent lifetimes, but the real challenge ismaximizing the utility of lower-confidencepredictions. The W76 system, through its DRexperience, leads in ESP capabilities and how theyaffect refurbishment decisions. The W80/ESPcoordination only began this fiscal year and is stillevaluating how to prudently define ESP’s role inthe 6.2/6.2A study timeframe.

Some component characterizations wereredirected more specifically to the W80 system inearly FY1998. Today, the entire ESP componentportfolio is providing characterizations to supportidentified SLEP decision milestones, and FY1999program growth is aimed at recent candidates for“double lifetimes” on the W76 or W80. We willfollow a greater degree of formality in negotiatingand monitoring ESP milestones to ensureusefulness of the content and schedule of agingcharacterizations. Frequent ESP reviews areplanned with annual reports beginning in August1999 to document the characterizationconclusions for both the W76 and W80 systems.

It is important to note that, while the programhas adopted a strong emphasis on supportingSLEP refurbishment decisions, it has not sacrificedits initial emphasis on broad stockpile applicability.The current component portfolio still includesperformance critical components that are used inmultiple systems together with components whosedesign features, materials, and environmentaldrivers provide meaningful aging insights intolarger families of components throughout thestockpile.

FY1999FY1998 FY2000 FY2001 FY2002 FY2003

W80 Life Extension Program

ESP Life Assessment Reports

Components and subsystems

6.1 6.2/6.2A 6.3 6.4

Preliminary Update Update Update Final

Aging/Nuclear Safety Evaluation CDU Firing Set (MC3276) TSSG (MC3268, MC3269)

ICCB (MC 3289A) Rolamites (MC3150, MC3301)

Intent Stronglink (MC2969) Trajectory Stronglink (MC2935)

MCCS (MC2907)NWC approval

FY1997 A complex-wide ESP plan has beendeveloped to support W80 decisionmilestones as part of the Phase 6.2Life Extension Study for non-nuclearcomponents and subsystems. In theabove chart, 6.1 represents conceptassessment, 6.2 feasibility & optionselect, 6.2A definition, planning &costing, 6.3 design and development,and 6.4 production readiness (phases6.5, first production, and 6.6, full-scale refurbishment, are not shownon this timeline).

Non-nuclear Components • SN24 89


Electromechanical Components—StronglinksMany war reserve electromechanical components (e.g.,

environmental sensing devices and stronglinks) in thestockpile play critical roles in weapon nuclear safety andreliability. We have made significant progress assessing theuseful lifetimes of some components that are approachingor have already exceeded their intended design life.

We now have dynamic performance models of criticalfunctions for the MC2969 stronglink, the MC2854 fluidintegrating environmental sensing device (ESD), theMC2897 verge escapement ESD, and rolamite ESDs.Recently, using DADS (dynamic analysis and designsystem), we developed a highly realistic and detaileddynamic performance model of the MC2969 stronglinktimer. Its features include: inertial properties of all themoving parts, contact between the star wheel and pallet,contact between the cam follower and the drive cam,spring force on the drive cam, torque transmission throughthe gear train, spring force on the follower, and friction onthe pallet and star wheel journal bearings. Modelparameters that evaluate aging effects include contactfriction (see box, top) at the star wheel/pallet interface,contact friction at the drive cam/follower interface, andjournal bearing friction.

Changes in friction and other variables can alter thetimer beat rate, resulting in the misinterpretation of thepulse code that the mechanism discriminates. Component,subsystem, and weapon system failures occur when thewell-defined limits to the beat rate are exceeded. Thepredicted beat rate of the timer from the MC2969 timermodel (below, left) has been compared to the beat rate ofseveral of the timers assembled and tested as part of thesolid lubrication oxidation study. Model results comparevery favorably with experiment. Hardware testingconducted in the solid lubrication study will play an

important part in the validation of this model.Stockpile-aged MC2969s were obtained fromdismantled weapons and are being tested anddisassembled for materials studies.

High-speed video and acoustic emission datahave been gathered from the stockpile-agedunits as well as from artificially aged timerassemblies. Significant progress has been madetoward digital capture of the acoustic emissionsof aged units, and we are developing a methodto more accurately analyze the analog acousticemission data of the MC2969 originalproduction units for comparison to the digitaldata from aged units. The contributions ofAlliedSignal/Federal Manufacturing &Technologies have been crucial to our successesin collecting acoustic emission data, obtaininghigh-speed video data, defining componentperformance margins, and recovering hardware.The validation process for the MC2969 timermodel will help increase confidence in DADSperformance models of other electromechanicalcomponents (e.g., the MC2897) now beingstudied.

A problem that has long plagued our effortsto observe stockpile degradation is the potentialfor damaging the MC2969 during removal fromthe next higher assembly. Damage inflictedduring disassembly often cannot bedistinguished from aging effects. A newapproach to unit testing and residual gas analysishas reduced this damage; it requires only partialremoval of foam and potting material fromaround the unit and significantly reduces the

time needed for the lengthydepotting process.

Other ongoing areas of studyfor the MC2969 will increase thecertainty of lifetime extensionassessments. These include theobserved degradation of ceramichermetic seals for electricalfeedthroughs (see box, middle)and the measured effects of age onelectrical contact resistance duringswitch closure (see box, bottom).High-voltage holdoff testing is alsounderway to characterize theeffects of abnormal environments(600°C) on the performance of thisnuclear safety critical component.

Time

Star

wh

eel a

ng

ular

dis

pla

cem

ent


Predicted beat rate (heavy black line) and experimentally observed beat rates (lightlines) from the MC2969 timer show good agreement. The solid model of the timerassembly used to generate the dynamic analysis model is shown on the right.


Effect of Friction ChangesReactions between the internal atmospheric

environment and the solid lubricant used instronglinks can produce compounds with differentfrictional properties than the starting material. Wehave shown that oxidation of molybdenumdisulfide under conditions present in the stockpilecan increase its starting friction as well as steady-state friction coefficients. The effect of thesefriction changes on stronglink performance is nowbeing evaluated. The run-in data for artificiallyaged subassemblies show the influence ofoxidation on device timing and are being used tovalidate a dynamic simulation of the device thatincludes an age-aware friction model.

Ceramic Electrical FeedthroughsCeramic of 94% alumina is used to make

hermetic electrical feedthroughs in MC2969stronglinks. Failure of these hermetic seals allowsambient atmosphere to react with the solidlubricants and electrical contacts. The time-dependent aging mechanism that degrades theceramic is subcritical crack growth (SCG). Ourexperiments quantify the SCG susceptibility ofalumina. Examination of returned hardware isused to validate the materials aging model. Thisproject is also developing the methodology forpredicting lifetimes of other componentscontaining ceramic or glass.

Electrical Contact ResistanceAlloy aging processes, surface reactions, and

contaminant adsorption can all increase contactresistance, preventing circuit completion. Usingfield-returned hardware, we are measuring andcorrelating local contact resistance with surfacechemical composition, component-level contactloop resistance, and ambient gas composition.Local contact resistance has been shown to be afunction of surface chemical composition, appliedcontact force, and resistance probe voltage level.Measurements on the aged contact surfaces withand without cleaning by dynamic wiping will helpestablish higher confidence lifetime assessments.

Crack growthhas beenobserved inceramics usedin the MC2969electricalfeedthroughs.

00.0

0.2

0.4

EndsCenter

0.6

0.8

1.0

10

Co

nta

ct r

esis

tan

ce, Ω

20 30 40Load, g

50

Electrical contact resistance in MC2969 switches wasmeasured as a function of contact force and shows alarge variability, which must be explained by componentperformance modeling.

Measured effect of run-in and oxidation on device timing. At 40operational sequences and then again at 80, the test was stopped andthe samples oxidized.

95047% oxide, 14% sulfate 48% oxide, 46% sulfate 88% oxide, 77% sulfate

900

850

800

750

7000 20

Number of operational sequences40 60 80 100 120

Tim

e to

res

et, m

s

Crack

Alumina



Electronic Components and Subsystems—Capacitive Discharge Unit Firing Set

We are developing a new set of analytical and evaluationtools for making informed decisions about whether toreplace, rebuild, or reaccept weapon electronic subsystems.These tools will be used for early identification ofperformance trends and potential concerns with specificdevices and materials; they include advanced circuitsimulation, improved physics-based device models,statistically derived sensitivity analysis, accelerated agingtests of electronic devices, and improved surveillance dataacquisition and trend analysis.

The MC3276, a capacitive discharge unit (CDU) firingset, is representative of several CDU firing sets in theenduring stockpile. This project will evaluate stockpile-agedmaterials and components to identify potential material,mechanical, and electronic device degradation. The projectwill also develop (1) advanced circuit simulation, devicemodels, and analytical tools; (2) accelerated aging tests toapply controlled aging stress to electronic devices; and (3) improved surveillance data acquisition and trend analysis.In collaboration with Pantex and AlliedSignal/FederalManufacturing & Technologies, we have developed benignprocesses for recovering stockpile-aged firing-set materialsand components from dismantled stockpile weapons. Thesecomponents are used for materials investigations thatinclude the soldered connections on one of the oldest flat-pack integrated circuits in the stockpile (see box). Byapplying advanced circuit simulation techniques, weidentified the components that most significantly affect theperformance of the MC3276 type of CDU firing set. Thisinformation can be applied to several firing sets currently inthe stockpile, and the project can concentrate testing andanalytical resources on the few most critical componentsand conduct meaningful accelerated-aging studies.

Our improved accelerated-aging testmethodologies include a screening test toevaluate the susceptibility of electronic devicesto low-dose-rate radiation degradation and amethod for the automated acquisition of deviceparameters during test. Using thesemethodologies with the quick-look screen, weidentified the semiconductor devices that aremost likely to exhibit enhanced low dose-ratesusceptibility. More than 200 B61 and W80devices (10 part-types) were screened, and twoof them demonstrated enhanced low-dose-ratesusceptibility—the SA1889 NPN transistor andLM185 voltage reference. Second, the projectidentified, through circuit simulation, more than300 W80 and B61 critical devices (three part-types) that are installed on circuit boards,characterized, and ready for the long-term study.Third, we developed a generic age-aware modelfor a ceramic capacitor that uses Z5U dielectric.Fourth, we characterized potential age-relateddegradation of four semiconductor device typescurrently undergoing a thermal aging study.Accelerated aging tests will extend the agingprocess of electronic devices beyond currentlyavailable hardware, identify life expectancy, andprovide the data needed to develop accuratedevice models for physics-based circuitsimulation.

The project is developing an advanceddiagnostic test for use in surveillance. Typically,there is only one internal test point on warreserve firing set circuits that measures thecharge on the main CDU capacitor. However, asignificant amount of relevant circuit data arecontained within the basic waveform. Theobjective is to determine if the performancecharacteristics of internal and physicallyinaccessible devices (e.g., transistors andoperational amplifiers) can be derived solely fromthe circuit output waveform. Circuit analysis ofthis complexity has never been performed beforeand requires the use of a new 64-processorcomputer at Sandia National Laboratories. Theresults of a proof-of-principle calculation showthat the values of an internal resistor andcapacitor can be predicted based solely on the output waveform of the circuit.

15

10

5

0

0

Time, milliseconds

Vo

lts

5 10 15 20 25 30 35 40 45 50

–5

–10

Significant changes in the CDU charging waveform result from smallchanges in values of a selected resistor and capacitor.


The SA2371 integrated circuit (IC)represents one of the oldest flat-packelectrical components in the stockpile. As asurface mount package, the tin–lead (Sn-Pb)solder joints are particularly sensitive tofatigue due to fluctuations in thetemperature. Those solder joints are alsosubjected to additional stresses caused bydissimilar elastic properties and thermalexpansion of the foam used to encapsulatethe printed circuit board. An extensiveinvestigation is underway to assess the agingcharacteristics of the SA2371 solder joints.Measurement of the thickness of the solder/copper intermetalliccompound layer thickness, together with the kinetics model oflayer growth rate, provides a measure of the effective agingtemperature (35–40°C). A second series of measurementsevaluates the Pb-rich phase size distribution in the soldermicrostructure. These measurements are affected by two agingfactors: (1) the temperature and (2) residual stresses imposed onthe solder joints. These are coupled into a constitutive modelthat computes thermal mechanical fatigue damage to the solder.

Both the intermetallic layer thickness and Pb-rich phase sizeare being used, not only to compute the degree of aging by thestockpile environments, but also to predict future reliability underextended lifetime conditions. Concurrently, we are documentingthe conditions of solder joints in returned field units and thensubjecting the joints to accelerated thermal aging. These resultswill be used to validate the computational predictions andbenchmark the performance of the SA2371.

Aging Evaluation of Flat-Pack Solder Joints

Lead

Solder

Laminate

0.5 mm

Gull wing solder joint on the SA2371 flat-packoperational amplifier.



Energetic Components and Subsystems—Slim-Loop Ferroelectric Firing Set

The MC3028 slim-loop ferroelectric (SFE) firing set firesthe main weapon detonators and the neutron generator onthe W76 and W78 weapons. Large quantities of a similarfiring set, the W68 MC2370, have been made available asthe result of the W68 dismantlement activities at thePantex and Y-12 plants. Using this hardware, we createda database of firing set operational characteristics. Ourgoal is to (1) model and replicate the electrical output ofthe MC3028 as a function of multiple input parameters,(2) reflect the effects of aging of internal firing setsubcomponents, and (3) validate these effects withhardware characterization and analyses.

An apparent problem was discovered early in theinvestigation while evaluating dismantlement (MC2370)and shelf hardware (MC3028). Approximately 90% ofmeasured capacitance (Cap) and dissipation factor (DF)values were out of range. This was construed to indicatedegradation within the ceramic-ferroelectric-conductivelayer stack, and low electrical output was thus expected.However, when MC3028 and MC2370 units werefunctionally tested, all MC3028 hardware (which had notbeen in the field) performed properly and only 4 of 24MC2370 units had low electrical functional output. Thefailed MC2370 units were from dismantled weapons andconsequently had extremely short electrical leads, whichmay have been responsible for the failures. Concern withapparent interface degradation (low Cap and DF) led to arelated materials investigation to determine the nature ofaging within the stack interfaces (see box).

Pentaerythritol tetranitrate (PETN) in the XTX explosivein the lens has been shown to be chemically stable withinthe weapon environment (see SN04, Energetic Materials

Predictive Capability). However, thenonequilibrium PETN used in the XTX explosivehas shown potential for significant changes incrystal size distribution. Despite the potential formorphological change within stockpile-to-targetsequence conditions, thermally aged units haveperformed within specifications even after thesevere aging of 1 to 1.5 years at 60 and 80°C.These units were already 25 years old at thebeginning of the high-temperature aging. Thisperformance within specifications demonstratesa considerable insensitivity of the firing set tothermally activated aging mechanisms that maybe occurring in the explosive.

Our progress in the modeling area includessignificantly increasing the realism of theunderlying physics-based model, incorporatingadditional critical subcomponent parts into themodel, and using a larger number of smaller-sized cells to describe the component. The timeover which the transient behavior of thefunctioning firing set is modeled has also beenincreased. Highlights are: (1) nearly full-devicesimulation using 7 million cells at 200-µmresolution incorporates all subcomponentfeatures; (2) physical model of only one-quarterof the stack (device is nearly symmetrical)generates current output that represents thefunction of the firing set; (3) 45-million-cellcomputer simulation of the full stack offerroelectric, ceramic, and conductive layersshows XTX explosive pre-ignition in anunexpected area; and (4) explosive timingvalidation experiments have shown that theEMMA (electromagnetic model)/HVRB (explosiveburn model)/XTX (explosive) model burns 50%slower than experiment. This indicates the needto refine the HVRB model.

To date, no significant functional degradationhas been observed in the SFE firing set; however,a validated SFE computational model willaccelerate hypothesis testing in assessing theconsequences of material changes. Finally, theSFE firing set modeling—characterized by anextremely large cell count, multiple physicalprocesses, complex visualizations, and realopportunities for experimental validation—hasresulted in significant computational advanceslikely to be of benefit to a wide range of futurecomponents and systems.

A visualization of the output from the computer simulation of theexplosive burning in a quarter-cell representation of the track plateof the MC3028 SFE firing set. This calculation uses 5 million cells at200-µm resolution.



The electrode interfaces in the firing set arelaminate structures with an adhesive layersandwiched between copper and gold electrodes.These interfaces are extracted from the transducerusing cryogenic fracture techniques. Electricaltesting and acoustic imaging allow a specificinterface to be targeted. The exposed electrode isthen subjected to composition, structure, andmorphological characterization. The interfacesage as (1) copper is distributed throughout theadhesive, explaining observed electricalconductivity; (2) the electrode primer diffuses intothe adhesive with time, resulting in a structurallyaltered interface; (3) accelerated aged units(100°C) show severe electrode oxidation, withcryogenic fracture producing adhesivedelamination; and (4) the time-dependent natureof the interface most likely explains variableelectrical properties between units.

We are trying to correlate the void density andstructure with aging in an effort to understandhow materials aging might impact firing setperformance.

Scanning electron microscopy of the adhesive at an aged interface shows arange of microstructures, including voids and adhesive filler agglomeration.

Degradation of Slim-Loop Ferroelectric Interfaces



20

15

10

5

00 10 20

Age30 40

Rat

e

LINAC data Plot 1 regression SPHINX data Plot 2 regression

Routine Surveillance

97

98


Routine Surveillance

Focus AreaThe nuclear weapons complex (NWC) laboratories

and production agencies support their stockpilestewardship through the integration of enhancedtechniques into surveillance testing. New techniquesand technologies will result in the NWC’s ability toidentify failure mechanisms that are of interest in anever-aging stockpile. As such, tasks related to thisactivity are specific to affected areas and providecoherence only through their augmentation of theroutine surveillance program.

The purpose of the tasks in this focus area is tomeet the needs of the individual agencies whiledeveloping ongoing capabilities and resources withinthe NWC. Affected surveillance information andsystems developed in this effort provide enhancedunderstanding of the nuclear weapon stockpile.

Pertinent TasksPX07 Air-Bearing Technology Transfer to PlantsSN16 Enhanced Test MethodologiesSN17 Enhanced Gas Analysis for Diagnostics and

Surveillance

Deliverables1. Integrate new testing methodologies thatsignificantly enhance the capability for earlydetection of aging degradation into the routinesurveillance program.

2. Implement reliable measurements of dew pointand multi-gas analysis to evaluate the gas-phasesignatures of weapon aging.


Enhanced Gas Analysis for Diagnostics and SurveillanceThe end-of-life for the desiccant in the W80 was of

particular concern because it was thought to drive theStockpile Life Extension Program (SLEP) schedule, includingproduction and/or delivery of several components (neutrongenerator, type 1K valve, interconnecting control box)planned for replacement at the same time as the desiccant.An analysis of desiccant weight gains obtained duringsurveillance activities at Pantex Plant had suggested thatthe desiccant could reach its end-of-life within the next5–10 years. The surveillance data were inadequate,however, to yield predictions with enough statisticalprecision to unambiguously establish the end-of-life.Subsequent field analyses of W80 weapon atmospheres,described here, have uncovered significant new desiccantproperties that apply not only to the W80 but to mostweapon systems.

To develop the data, we set out in FY1997 and FY1998to more accurately and precisely determine the projectedend-of-life of the W80 desiccant. We designed, developed,fabricated, tested, qualified, and fielded two war-reserve-certified field portable gas analyzers and used them toobtain dew point measurements on 94 W80 warheads attwo service locations. More than 50 people from theSandia National Laboratories (SNL), Los Alamos NationalLaboratory, Pantex Plant, DOE Albuquerque (DOE/AL),and AlliedSignal/Federal Manufacturing & Technologiescontributed to this large project. The diverse set ofcustomers includes DOE/AL/WQD, DOE/AL/WPD, SNLW80 Program, SNL Military Liaison Department, SNLMechanical Measurements Team, U.S. Navy, and U.S. AirForce. The project was performed to DOE QC-1requirements and fully documented through the SNLengineering drawing system, where full details on thedesign requirements, product specifications, operatingprocedures, safety analyses, and equipment qualificationcan be located. Two DOE-approved (“diamond-stamped”)analyzers were delivered on December 15, 1997, and ourfield activities were successfully carried out betweenDecember 1997 and February 1998.

Results of field measurements (see figure) exposed alarge discrepancy between the internal dew points of theW80 and those expected on the basis of the desiccantweight gain measurements. Specifically, the internalatmospheres of the warheads were much drier thanexpected. More than half of the warheads sampled haddew points below –60°C, whereas the prediction based onsurveillance results was for a dew point of approximately–45°C. Further laboratory studies uncovered someimportant desiccant properties that quantitativelyaccounted for this discrepancy. The lab resultsdemonstrated conclusively that the desiccant used in theW80 (and most other weapon systems) can adsorbsubstantial quantities of carbon dioxide, a possibility

previously unrecognized in the weaponscommunity. In fact, we currently estimate that>50% of the desiccant weight gain observed inthe W80 can be attributed to CO2 adsorption.Additional experiments indicated that CO2 isdisplaced by water as water becomes available,so the desiccant should perform as intended.Our new understanding of competitiveadsorption enabled us to significantly extendthe W80 desiccant end-of-life estimates, anddesiccant replacement no longer drives theSLEP schedule.

This revolution in our understanding ofdesiccant performance impacts all desiccatedweapon systems in the enduring stockpile (B61,W62, W76 AF&F [arming, fuzing, and firing],W78, W80, B83 trajectory sensing signalgenerator and radar, W84, W88 case, and AF&F)as well as desiccated shipping/storage containers(AL2100). The budgetary impact is estimated tobe in the tens of millions of dollars, primarilythrough relaxation of schedule and budgetpressures on the W80 SLEP and on neutrongenerator development and production. Further,the capacities of the desiccants combined withthe measured rate at which they are adsorbingwater in the stockpile has raised the possibility ofperforming Alt 346 (W80 refurbishment) in thefield. This would provide tremendous additionalsavings as well as contributing to work loadleveling at Pantex.

A W80 engineer measures the dew point of a warheadusing Sandia’s W80 field portable gas analyzer whileU.S. Air Force and U.S. Navy personnel monitor theactivity. A second analyzer (background) is also in use.

Routine Surveillance • SN17 99


Enhanced Test MethodologiesTwo important contributions to the Stockpile

Surveillance Program were made possible by this task: (1) significant improvement in measuring the compressionset in W76 environmental O-ring seals and (2) increasedunderstanding of the aging behavior of W76 radars.

Compression set in O-ring sealsEnvironmental O-ring seals are an important component

used on all weapon systems to protect the weapon interiorfrom oxygen and water vapor intake. The two mostcommon measurements to monitor the degradation ofseal materials are compression stress relaxation andcompression set. Compression stress relaxation is the idealmethod to use for laboratory simulations and modeling.But because compression stress relaxation cannot be easilyapplied to real weapons, surveillance activities have usedcompression set as the method to monitor the conditionof field-aged O-rings.

In the past, the validity of the surveillance compressionset measurements for the W76 has been questionedbecause of method inconsistency. The past methodemployed hand-held calipers. This was highly operatordependent and lacked a standard technique, and thus theresults could not be standardized. This precluded statisticalanalyses of the results and the development of inferencesregarding aging degradation.

We introduced the use of a laser micrometer into theroutine measurements as part of the data collection forthe routine surveillance program. The Sandia systemsevaluation engineer and the Pantex program engineer for the W76 modified the disassembly procedures toincorporate the use of the laser micrometer to measure

O-ring compression set during weapondisassembly and inspection. The lasermicrometer method is operator-independentbecause standardized fixtures are used and nohuman handling is involved during themeasurement process. This enhancementprovides reliable measurements that aresystematic and consistent. As a result, the datacollected now can be used to develop a betterunderstanding of the viability of theenvironmental seals used on the W76.

Aging behavior of radarsIn December 1997, tests were conducted

on two W76 radars using the Sphinx flash x-raymachine at the Sandia National Laboratories.The Sphinx results shown in the figure indicate a possible aging effect in the protection circuitry,i.e., the increase in the y-axis parameter (Rate) isundesirable. The components organization thatconducted the test presented the results to aW76 working group that had been formed toidentify near-term enhancements to surveillancetesting. The working group agreed that insightscould be gained regarding whether or not anaging effect is occurring by conducting furthertests; results from two radars were deemedinsufficient to make any inferences.

Funds were provided, and another series oftests were conducted to further investigate thepossibility of aging effects. Six W76 radars ofdifferent ages were tested at a linear accelerator(LINAC) machine operated by the BoeingCompany. The results of that test series are alsoshown in the figure. Although the best fit linearregression line for the LINAC data does not haveas steep a slope as does the regression line forSphinx data, the LINAC data do show a definitetrend in increase with age.

These results warrant further investigation.Current plans are to disassemble the radars andextract the circuits and devices of interest. Wewill look for possible effects of aging and alsoexamine non-aging effects, such as differencesin devices between early production units andthose assembled later.

Test resultsfor radiationeffects onfield-agedW76 radars:Sphinx (red)and LINAC(black).

20

15

10

5

00 10 20

Age30 40

Rat

e

LINAC data Plot 1 regression SPHINX data Plot 2 regression

100 Routine Surveillance • SN16

U.S. Government Printing Office: 1998/783-046-80004

ucrl-lr-132042 enhanced surveillance program/67531/metadc784933/m2/1/high_res... · today. the...

Documents