Strategies to Remove Particulate Contamination from Ablator Capsule Surface

Presentation at Target Fabrication Meeting

Santa Fe, New Mexico May 24, 2012

Suhas Bhandarkar, J Reynolds, S, Baxamusa, S. Letts, N. Antipa,

E. Lindsey, K. Moreno

• CH capsule surfaces – Isolated contamination of the

outer surface (particles, etc) – Isolated defects on the outer

surface (bumps)

• Likewise, hydrogen ice layer – Roughness, isolated defects,

contaminant crystals, etc

Debris on the capsule surface can contribute to Mix

Suhas Bhandarkar 2

R-T instabilities are exacerbated by surface roughness of the

ablator layers

R-T instabilities: Colors indicate density profile after unstable acceleration

The numerous parameters that affect ignition have been grouped into four categories

8/21/2012

Capsules can be contaminated with particles with a range of sizes, shapes and compositions

3

1. Implementation of clean-rooms and clean handling procedures

1. Both at GA and LLNL

2. Metrology of the capsule surface with sub-μm resolution

3. Mitigation strategy: Capsule cleaning

Three efforts to diagnose and reduce particle contamination

Suhas Bhandarkar

Particle Volume (μm3)

Cube dimension

(μm)

Number Allowed

>30* >3.1 0

15 – 30 2.5 – 3.1 10

7.5 – 15 2 – 2.5 50

< 7.5

Cleaning needs to happen just before final assembly to minimize risk of further pick up of debris

4

Inspect

Clean

Inspect

Capsule placed in diagnostic band

Final assembly

Suhas Bhandarkar 8/21/2012

Fill-tube bond is extremely fragile: all cleaning techniques have to avoid any degradation of the bond

5

NIC scale capsule fill-tube bond fails in tension at ~0.03N (3 gmf) We have set the spec at 1 gmf for any cleaning operation

Delicate fill-tube bond (max 5pL) is the main challenge

Tensile pull tests of ignition-spec CFTAs:

Suhas Bhandarkar

0

1

2

3

4

0 100 200 300

forc

e (g

)

distance (um)

57 CH

0

1

2

3

4

Bon

d st

reng

th (g

)

8/21/2012

10 μm

5 μm

We need to overcome the particle – capsule interfacial bond strengths that can vary by orders of magnitude

• van der Waals – London dispersion: 2 induced dipoles (< 45 kJ/mol) – Debye: 1 induced/1 permanent dipole ( < 3) – Keesom : 2 permanent dipoles (< 25)

• Hydrogen bonding (< 55) • Electrostatic

– Localized regions on CH capsule surface can easily tribo-charge

• Liquid bridge – Capillary retained liquid (water)

• High temperature bonding – Say during mandrel pyrolysis

• Visco-elastic force – Particle embedded in a high viscosity fluid or the

capsule surface itself • Solid bridge

– Deposition/Precipitation at interface with capsule surface

Gravity: 1e-13 N

vdW: 5-10e-8 N

Capillary 3e-6 N

Viscous drag 1e-5 N

Glued particle 1e-3 N

Electrostatic 5e-7 N

6 Suhas Bhandarkar

Particles can be held by different mechanism with a wide range of forces

8/21/2012

Suhas Bhandarkar 7

Electrostatic – Induced charged on

a 25um metal wire

Surface energy – Particle would rather be in

a fluid than on the capsule

Sacrificial layer – Removable coating that

sheds just before assembly

Localized agitation – Thermal shock using

a liquid cryogen

Hydrodynamic drag – High velocity helium jet

from ~75um capillary

Increasing particle removal efficiency

Incr

easi

ng ri

sk to

cap

sule

or f

ill-tu

be

Strategies to remove particles from a CH capsule with an attached fill-tube

8/21/2012

0

200

400

600

0.E+00

2.E-02

4.E-02

6.E-02

0.0E+00 5.0E-06 1.0E-05

rate

of L

N e

vapo

ratio

n g/

mm

2.s

heat

flux

(J/m

m2)

Time (s)

heat flux

rate of evaporation of LN

Cleaning using liquid nitrogen (LN) relies on thermal shock that the very small fill-tube bond can withstand

Suhas Bhandarkar 8

Rapid evaporation at the surface concentrated pressure bursts

75

125

175

225

275

0 50 100 150 200

tem

pera

ture

T (K

)

Wall location (um)

)2( 22

drdT

rrT

dtdT

+∂∂

= κ

α1

α2

Bubble initiation:

Intense bubbling & any CTE (α) mis-match can generate powerful stresses on particle-capsule interfacial bond

1

2

3

4

Initial condition

Free-stream entrainment

Dislodge particle

Capsule surface

Foreign particle

Liquid nitrogen

t= 10 μs

t= 0.2 μs

Temperature profile in capsule wall vs time

Heat flux & LN evaporation rate

Outer surface

Inner surface

Increasing time

8/21/2012

Though no detriment to the C-FT bond, results from LN cleaning showed considerable variability

0 5

10 15 20 25

20 10 5

Num

ber

Particle 'dia' (um)

1st capsule before after

0

5

10

20 10 5

Num

ber

Particle 'dia' (um)

4th capsule before after

0

5

10

15

20

20 10 5

Num

ber

Particle 'dia' (um)

7th capsule before after new

Process not seen to be robust due to the variability in particle removal efficiency

0

20

40

60

80

100

20 10 5

Rem

oval

Effi

cien

cy (%

)

Particle diam (um)

Liq N2 Cleaning trials 1st

2nd

3rd

4th

5th

6th

7th

8th

Ave

9 Suhas Bhandarkar

Before

After LN clean

Capsule cleaned with LN flow

30μm

8/21/2012

LN cleaning has lower efficiency for particles with higher surface charge

20

40

60

80

100

20 10 5

Rem

oval

Effi

cien

cy (%

)

Particle size (um)

PP

Al

Cu1

Cu2

SS1

SS2

10 Suhas Bhandarkar

Typical after

• Non-polar LN can make only vdW bonds with particles • Need polar liquid to entrain the particles

Typical before

100um 50um

Polypropylene particles

SS Cu, Au Si, PP Teflon

Triboelectric series +ve

-ve

Skin, hair Glass Nylon Al SS

Expected particle charge vs material composition

Measured efficiency for different materials

Typical after

Increasing charge

8/21/2012

Low boiling point for fast drying; counter to high P and H

To sustain entrainment of charged particles

Another technique is to use surface energy principles to separate particles from capsule

To enable liquid-particle bonding

11 Suhas Bhandarkar

Low surface energy to wet the particle and help debond it: counter to high H Oxide~1000 Metals~500-1500 Plastics: 40-60

Hamaker constant A123 can be negative if the intervening εliq is intermediate, leading to repulsion RICH~1.55 RIPlastic~1.5-1.6

We want a solvent with: • γ < 40 and high n, ideally 1.55, (for bond breakage), •high P and H (for good entrainment); •B little over 30oC (to minimize impact on CH or adhesive)

Liquid Polarity

P Hyd bond

H Boiling Pt

B Surf Energy

γ RI n

(Hansen parameter) C dyne/cm methyl formate 6.5 12.5 32 25 1.36 methyl iodide 7.7 5.3 42 25.8 1.53 methylene chloride 6.3 6.1 40 27.2 1.42 methanol 12.3 22.3 65 22.6 1.33 diethyl ether 2.9 5.1 35 17 1.35 toluene 0 0 36 18 1.51 trichlorotrifluoro ethane 1.6 0 44 14 1.36 Liquid nitrogen 0 0 0 9 1.21 water 16 42.3 100 72 1.33 formamide 26.2 19 210 58 1.45 glycerol 12.1 29.3 290 63 1.47

8/21/2012

Of the four top solvents methyl formate provides the best combination of the attributes we are looking for

Liquid Polarity Hyd bond Boil Pt Surf

Energy RI methyl formate 6.5 12.5 32 25 1.36 methyl iodide 7.7 5.3 42 25.8 1.53 methyelne chloride 6.3 6.1 40 27.2 1.42 methanol 12.3 22.3 65 22.6 1.33

Suhas Bhandarkar 12

0

15

30

45

1 10 100

Stre

ngth

(MPa

)

time (min)

Methyl formate

methylene chloride

Strength of adhesive after solvent exposure

Halogenated solvents (such as methylene chloride & methyl iodide) are too aggressive

8/21/2012

Concept is to combine strong solvation forces with hydrodynamics to get most efficiency of removal

Suhas Bhandarkar 13

Technique: Methyl formate impinged on capsule at high velocity through a 100μm nozzle with limited total time duration

• Tip velocity: 2-5 m/s • Max force on capsule due to

momentum of jet: 0.02 g • Holding force of a 500μm

vacuum wand: 1g • Driving pressure for fully

developed flow of jet through the nozzle: 12 psi

Experimental station to test the solvent-jet technique Removal of the separated particles is mediated by a

high- velocity jet

Drain

Nozzle

CFTA

8/21/2012

Initial tests with intentional contamination with a cocktail of particles were very effective

14 Suhas Bhandarkar

Bef

ore

Afte

r 1st tr

ial

Afte

r 2nd

tria

l

SEM image of full capsule before & after solvent jet cleaning

50um

50um

High Magnification SEM image

Bef

ore

Afte

r

Surface has mixture of metal (Al, Cu), oxide (Al2O3) and polymer (nylon & polypropylene) particles

8/21/2012

This technique was also effective on ‘real’ debris on capsules

Suhas Bhandarkar 15

500um 100um

20um 20um

200um 200um

Bef

ore

Bef

ore

Bef

ore

After

After

After

SEM images of sections of real capsules

8/21/2012

Fill-tube joints on 8 capsule assemblies showed no signs of degradation & were leak-tight

Suhas Bhandarkar 16

100um

10um

10um

SEM images of the bonds connecting the fill-tubes to the capsule after 30 min of exposure to methyl formate show little change

8/21/2012

We have modeled the fluid velocity field on the capsule using mechanisms described by Watson*

Suhas Bhandarkar 17 * E.J. Watson, J Fluid Mech. 20, 481, 1964

r

u(r, y)

y

jet, flowrate Q kinematic viscosity ν

h(r) ∫=)(

0

2rh

udyrQ π

)()()(1

0

2 rhgrhdrdhg

yudyru

drd

r xyh

+−∂∂

−=∫ ν

Continuity:

Momentum:

For a sphere:

r= Rs(sinθ)

Radial spread of impinging jet on a sphere

jet

θ Rs

8/21/2012

Model indicates that the fluid spreads over the capsule as a film whose thickness is ~40μm, independent of the incoming jet velocity

Suhas Bhandarkar 18

Modeled results of film thickness compared to experiments

Model Capsule periphery Liquid layer

Video 0

100

200

300

400

500

0 2000 4000 6000 8000 10000

Max

siz

e of

dro

p fo

r 1 p

g re

sidu

e

contamination level in fluid (ppb)

Typical drop size

Max allowed contamination

Fluid cleanliness needed for a residue of no more than 1pg per drop

8/21/2012

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

0 50 100 150

Free

sur

face

vel

ocity

(m/s

)

angle (degrees)

Velocity, 2m/s

Velocity, 4m/s

Highest fluid velocities only occur close to the incoming jet stream, calling for rastering of the jet

Suhas Bhandarkar 19

We need to raster the jet stream over the capsule

Jet velocity 2m/s or 4m/s

Modeled results for free-surface velocity of the liquid film around the capsule

8/21/2012

Strategies to Remove Particulate Contamination from Ablator Capsule SurfaceDebris on the capsule surface can contribute to MixCapsules can be contaminated with particles with a range of sizes, shapes and compositionsCleaning needs to happen just before final assembly to minimize risk of further pick up of debrisFill-tube bond is extremely fragile: all cleaning techniques have to avoid any degradation of the bondWe need to overcome the particle – capsule interfacial bond strengths that can vary by orders of magnitude Strategies to remove particles from a CH capsule with an attached fill-tubeCleaning using liquid nitrogen (LN) relies on thermal shock that the very small fill-tube bond can withstandThough no detriment to the C-FT bond, results from LN cleaning showed considerable variabilityLN cleaning has lower efficiency for particles with higher surface chargeAnother technique is to use surface energy principles to separate particles from capsuleOf the four top solvents methyl formate provides the best combination of the attributes we are looking forConcept is to combine strong solvation forces with hydrodynamics to get most efficiency of removalInitial tests with intentional contamination with a cocktail of particles were very effectiveThis technique was also effective on ‘real’ debris on capsulesFill-tube joints on 8 capsule assemblies showed no signs of degradation & were leak-tightWe have modeled the fluid velocity field on the capsule using mechanisms described by Watson*Model indicates that the fluid spreads over the capsule as a film whose thickness is ~40μm, independent of the incoming jet velocityHighest fluid velocities only occur close to the incoming jet stream, calling for rastering of the jetSlide Number 20