PERSISTENT SURVEILLANCE FOR PIPELINE PROTECTION AND THREAT INTERDICTION

H. Huang, H.W. Xu, K.P. Youngblood, D.R. Wall, R.B. Stephens, K.A. Moreno, A. Nikroo of General Atomics

K.J. Wu, M. Wang of LLNL

2012 TFSM Conference Santa Fe, NM

May 21-24, 2012

Work supported by U.S. Department of Energy under Contract DE-AC52-06NA27279

Inhomogeneous Cu Diffusion in NIF Beryllium Ablator Capsules

IFT\P2011-037

Unlike GDP, beryllium is heterogeneous at micron scale

• Process/structure interactions critical to performance

Micro-structure

Target performance

Fabrication process

• This talk presents a Cu-Be case study • Same methodology applicable to other NIF dopants Introduction

NIF Be capsules have step-wise dopant profiles; diffusion during pyrolysis blurs the steps

• Interfaces become diffused after mandrel removal − Measured non-destructively by contact radiography

Cu/Be

• NIF design uses 5 distinctive layers to optimize shock timing

Need quantification to estimate the impact on implosion

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0 20 40 60 80

Cu (a

t%)

Be Thickness (um)

S6NIF165 Before and After Pyro

Before Pyro

After Pyro

SEM Backscattered Electron (BSE) contrast reveals dramatic difference in pre- and post-pyro samples

• Post-pyro interfaces become diffused and heterogeneous

• Heterogeneity length scale: 0.5-5 µm radial, 0.5-10 µm lateral − >10x larger than bulk diffusion behavior (Butrymowicz 1975)

Pre-pyro

Layer #3: 0.70 at% Cu

Layer #4: 0.35 at% Cu

Layer #5: 0.00 at% Cu

Post-pyro Higher Ar

5keV field emission SEM on ion polished Be cross-section

Energy-Dispersive Spectroscopy (EDS) unambiguously identified Cu variations as the cause of BSE contrast

EDS Cu Map BSE Image

Post-pyro: Diffuse Interface

Pre-pyro: Sharp Interface

Need to distinguish two scenarios: (1) polishing smear (2) Cu diffusion

~5 um

Electron interaction

volume

Sample surface

20 keVelectron beam

4um

BSE Detector

backscattered electrons

35°

~2 um

~0.05 um

secondary electrons

Measuring Cu from outside, through undoped Be layer, eliminates the possibility of polishing smear

5.3um

• 20keV e-beam probes <5 µm depth into bulk beryllium − Can’t see through 5.3 µm undoped buffer layer unless Cu diffuse

1 at% Cu

0 at% Cu Outside

e-beam probes outside surface

Test suggested by J. Edwards @ LLNL

Copper diffused into the undoped buffer layer in highly heterogeneous manner during pyrolysis

Line is where EDS scan was performed

• BSE contrast invisible on pre-pyro surface

• BSE contrast conspicuous on post-pyro surface

• Post-pyro contrast: “super nodules” on textured background

This nodule was FIB Lifted for TEM

EDS element line-scans provide further validation

0

200

400

600

800

1000

1200

0 20 40 60 80 100 120 140 160

Cu-L

X-ra

y (C

ount

)

Distance (um)

Post-pyro GA226-05 outer surface

Cu-L Count (20KeV e-beam)

Bsd Image Intensity

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0 20 40 60 80 100 120 140 160Co

pper

(at%

)Distance (um)

Post-pyro GA226-05 outer surface

Cu at% (20KV)Cu at% (15KV)Cu at% (10KV)Cu at% (5KV)

• BSE contrast agrees with Cu profile measured by EDS linescan − Establish BSE as quick semi-quantitative Cu mapping technique

• Varying e-beam energy proves Cu from an internal source − Higher energy probes deeper and measures higher Cu at%

Max depths: 20KeV: 5.0um 15KeV: 3.0um 10KeV: 1.5um 05KeV: 0.4um

This nodule was FIB Lifted for TEM

X

Y FIB Slice

Nodule center location On FIB slice

White ring, due to oxide, marks the “super nodule” growth column boundary

White dot due to enhanced Cu diffusion ~130 nm thin FIB slice

shows porous structures near nodule center

FIB slice was lifted-out from the “super nodule” on post-pyro sample to study structural-composition correlation

Top-down SEM image (BSE contrast)

Lots of voids near nodule center

Lack of voids outside nodule

Scale bar: 1 µm Composite TEM Image

TEM shows nodule center having finer, heterogeneous, and porous structures, which can enhance Cu diffusion

Nano-voids present mostly near nodule center

Grain sizes vary greatly: 50-200nm near nodule center 500-1000nm outside nodule

TEM Black represents disorder

~20 µm “super nodules” are growth columns induced by mandrel defects, terminating as surface bumps

Large bumps on outside surface

SEM image of a bump cross-section

Abnormal growth initiated by mandrel contaminant

FIB through the nodule

X-ray imaging confirms strong correlation between “super nodules” and mandrel surface contaminants

X-ray tomography slices

“Super nodules” strings

Mandrel surface defects

Two features parallel to each other

X-ray transmission

Inhomogeneous background Cu diffusion produces atomic Z contrast on both cross-section and OD

(1) Cross-section view (2) Top surface view

• Pattern length scales match between two BSE views of the same post-pyro sample

Background Cu diffusion produced ±0.06 at% 1-σ variations, which is unwanted by target designers

0

0.2

0.4

0.6

0.8

1

0 5 10 15 20 25 30 35 40 45 50 55

Cu (a

t%)

Lateral Distance (um)

Copper Distribution (Estimate from BSE Image Contrast)

D=-1.5um

D=-1.0um

D=-0.5um

D=0.0um

D=0.5um

D=1.0um

D=1.5um

“D” indicates distance to the original Cu interface

GA226-05 post-pyro BSE contrast line-outs

• Copper profiles near interface are anti-phased, expected if a conserved total quantity is redistributed

A phenomenological description of Cu diffusion in beryllium

85um

Sputtered beryllium grow in vertical columns, capped at the surface as nodules. Each column is a cluster of submicron columnar grains

Cu: 0.0 at%

Cu: 1.0 at%

There are ~25um “super” nodules embedded inside a sea of 5um regular nodules.

5.7um

25um 5um

Pre-pyro

Post-pyro Cu diffuses faster toward center of each nodule, even faster for “super” Nodules

See Poster P-28 by H. Huang for metrology technique details

Oxide provides an effective diffusion barrier

Thermal oxide stopped Cu diffusion

Strong Cu diffusion where no oxide was introduced

• Oxide idea evolved – ALD Al2O3 – Thermal oxide – In-situ oxide

See Oral Tue. AM1-4 by K. Youngblood

Cu diffuses inhomogeneously in sputtered beryllium

• Cu structure length scale varies: – 0.5 to 5 µm radially, – Minimum 10x larger than bulk prediction

• Two types of Cu structures: – background variations & “super nodules”

• Background Cu varies by ~+/-0.06 at% (1-σ) • “Super nodules” correlates to mandrel contamination

– Nodule centers are porous, with smaller grains which enhanced diffusion

• Cu diffusion, uniform or not, can be stopped by oxide barriers – Beryllium can remain to be an ablator choice

Summary

Same methodology can be applied to other dopants, Work in progress on W, Zr, Al, Si

Summary

• Dopant behaviors vary greatly − Uniform and stays uniform: W and Zr − Uniform and becomes non-uniform: Cu − Non-uniform but doesn’t further evolve: Si − Non-uniform and worsens upon pyro: Al

• Different physical processes in the play − Solubility => phase segregation − Mobility => diffusion

• Same measurement strategy apples