Instability modeling for NIF ignition targets and Omega experiments - - PowerPoint PPT Presentation

Instability modeling for NIF ignition targets and Omega experiments

*This work was performed under the auspices of the U.S. Department of Energy by the University of California, Lawrence Livermore National Laboratory under contract No. W-7405-Eng-48.

S W Haan, T Dittrich, G Strobel, M Marinak, D Munro, G. Glendinning,

• P. Amendt, and R. Turner

X-division

Lawrence Livermore National Laboratory

University of California

Presented to: IWCTM 2001 Pasadena Dec 2001

SWH IWCTM_2001 2

Summary: We are continuing to explore hydro instability issues on NIF targets, and verifying modeling with Omega experiments

Specifications are being completed for a variety of indirect drive targets: Beryllium, polyimide, CH(Ge) ablators Drive temperatures 250 - 350 eV, spectra for gold or cocktail hohlraum Scales from 100 kJ to 600 kJ into capsule (NIF energy ~1.8 MJ) Details such as 3He buildup in the core are being analyzed Modeling of Omega planar polyimide Rayleigh-Taylor foils is close to experiments A new design for convergent Rayleigh-Taylor experiments on Omega will test other aspects of the modeling

SWH IWCTM_2001 3

Generically the ignition targets all look the same as for the last 10 years or so

• r U2Nb0.28AuTaDy

C22H10N2O4 1.085 mm 0.5 mg/cc

& &

SWH IWCTM_2001 4

Our current instability modeling is based entirely on explicit full simulations of perturbation growth and its impact on ignition and burn

• Single shell cryogenic capsules are ablatively stabilized on
• utside during acceleration, and on inside during

deceleration

• Simulations indicate that modes beyond about 120 do have

any appreciable amplitudes at any times of interest

• Experiments have generally been compatible with

simulations giving us confidence in them

• Modeling is done in 2D (LASNEX and Hydra) and 3D

(HYDRA) for single modes, and for multiple modes over various solid angles

• Biggest uncertainties are considered to be in the input:

spectrum of drive radiation, opacities, characterization of initial perturbations

SWH IWCTM_2001 5

There are three failure modes we see in our simulations

• Acceleration: Modes l ~100 grow and disrupt the shell

Especially a problem if shell is too thin

• Deceleration: Modes l ~15 create spikes that cool the hotspot

Especially a problem if shell is too thick

• Low modes: If there is much solid angle with ρ

ρ ρ ρr < 1 g/cm2, bubbles blow out and yield is reduced A successful target is optimized to trade off the first two issues, and has enough 1D ρ ρ ρ ρr to minimize the third. Requires power and energy to have room to trade them off!

SWH IWCTM_2001 6

This plot summarizes ablator-seeded Rayleigh- Taylor results for the different capsules

Old Dittrich 250eV result w/ graded dopant, 0.3 mg/cc

20 40 60 80 100 200 400 600 800

0.3 mg/cc DT gas 0.5 mg/cc Rms roughness for 50% YOC, nm All with gold hohlraum spectrum Capsule energy, kJ

120

New CH(Ge) 300eV, 0.5 mg/cc 300 eV Polyimid, both mg/cc 300 eV Be(Cu) 350 eV Be(Cu)

Be(Cu) is better, and higher TR helps a lot

PT

Different calculation details

SWH IWCTM_2001 7

600 kJ capsules might be constrained in foot length, at a significant energy price

Largest scale might have foot increased in order to keep total pulse length close to 20 ns

Time (ns) 250 200 100 20

350 kJ 190 kJ 600 kJ

Drive TR (eV) 25 15 10 5 150 50 300

If shock-crossing time is fixed, velocity ~ S1 foot level flux F ~ S2 Adiabat β β β β ~ S1.2 Margin ~ S3β β β β-1.5 ~ S1.2 ~ E0.4 instead of E1

SWH IWCTM_2001 8

1H,

0.3 mg/cc DT + 1H 0.5 mg/cc DT + 1H

3He:

0.3 mg/cc DT + 3He 0.5 mg/cc DT + 3He 2D simulations (ablator roughness for 50% yield, normalized to 65 nm, include 0.93 µm DT rms)

Surface roughness specifications are tighter if there is 1H or 3He in the central gas

• Both are “dead weight” w/ respect to hydro, ignition & burn
• Atom-for-atom, 3He is worse—more electrons and ion

charge, increases radiative and conductive losses

• But gram-for-gram, 1H is slightly worse—3x more atoms/g

0.2 0.4 0.6 0.8 1 1.2 0.1 0.2 Relative ablator roughness requirement (ablator roughness for 50% YOC, normalized)

3He or 1H density (mg/cc)

SWH IWCTM_2001 9

The calculated NIF cocktail spectrum is intermediate between Planckian and gold

Time (ns)

5 10 15

A (black) typical gold spectrum B (red) cocktail calculation (Pollaine) C (blue) Planckian w/ same flux

Need to do simulations

• f effect on Rayleigh-

Taylor of actual cocktail spectrum

SWH IWCTM_2001 10

With a Planckian drive, baseline polyimide NIF capsule shows 85% more Rayleigh-Taylor growth

Complicated interplay of growth on the various interfaces With doped ablators, may be able to reoptimize w/ cocktail wall Growth in 2D simulations, very small multi-mode pert on ablator initially Time (ns)

5 10 15 1 10 100 1000

Time - Ignition time (ps)

300 200 100 1000 ρ ρ ρ ρr rms ρ ρ ρ ρr avg Initial value 500 2000 4000 8000

Black Au Blue Planckian Red cocktail hohlraum wall Growth on ablation front Growth on DT/PI interface Deceleration growth

ρ ρ ρ ρr rms ρ ρ ρ ρr avg Initial value

See

• ther

plot for detail

SWH IWCTM_2001 11

We are doing Rayleigh-Taylor experiments on Omega to verify modeling of polyimide

Omega hohlraum Backlighter for face-on Rayleigh-Taylor growth measurement Backlighter for side-on trajectory measurement Rippled Polyimide foil View for Face-on

SWH IWCTM_2001 12

Peter Amendt has done hohlraum simulations that fit the Dante flux measurement

Post-process to simulate Dante: almost high enough to fit data (black curve compared to green). Simulated drive for package is red curve, about 10 eV lower

There’s a significant geometrical correction (like the old albedo correction, but now in the

• ther direction) that we

need to incorporate

Simulated Dante Simulated flux onto foil Shot 19010 Dante data shifted 320 ps

1 2 3 4 50 100 150 200

TR (eV) Time (ns)

SWH IWCTM_2001 13

Simulated drive extracted from Peter’s hohlraum calculations makes sideons very close to data

250 200 150 100 50 Posi t i

• n (

µm ) 5 4 3 2 1 Time (ns 19011 19013 19014

Simulation using Peter’s simulated drive Also shown shifted in time, improves fit Peter’s hohlraum simulations include a foil, its side-on motion agrees with my foil-only simulation

SWH IWCTM_2001 14

I have finished one case faceon and sideon from June 00 shots with the new source info

Source was Dante-25eV, with M-band adjusted (by factor of several) to match Dante M-band fraction

250 250 250 250 200 200 200 200 150 150 150 150 100 100 100 100 50 50 50 50 5 5 5 5 4 4 4 4 3 3 3 3 2 2 2 2 1 1 1 1 Ti m e ( ns) Ti m e ( ns) Ti m e ( ns) Ti m e ( ns) 20154 20154 20154 20154

Gail says this is the one reliable side-on from this series

0. 20 0. 20 0. 20 0. 20 0. 15 0. 15 0. 15 0. 15 0. 10 0. 10 0. 10 0. 10 0. 05 0. 05 0. 05 0. 05 0. 00 0. 00 0. 00 0. 00 4 4 4 4 3 3 3 3 2 2 2 2 1 1 1 1 Ti m e ( ns) Ti m e ( ns) Ti m e ( ns) Ti m e ( ns) λ λ λ λ = 30 µm, 2.0 µm amp Modulation (OD) 0. 25 0. 25 0. 25 0. 25

Face-on Old drive New drive This is late and slow, meaning we’ve overcorrected the drive, which is very good news

SWH IWCTM_2001 15

The simulations I’ve shown previously for the June 00 faceons used this drive profile

Dante: Black 19010, 1, 3 (Feb 00) (sideons we’ve been trying to fit) Red 20154 5 6 (June 00) (faceon shots) All Dante retimed to go through (1.2 ns ,120 eV) All plots are with CEA calibration Profile I used for old face-on work 19010 simulated source from Peter (aruguably fits sideons) Black solid to black dashed is geometry correction + ~10 eV that Dante is still high compared to

• simulations. (Arguably fits sideons)

Same correction to red curves would be “right” profile, compare to green curve. Red dash is face-on Dante -25eV, shifted 0.1ns to get good time 0 -- best guess at drive for faceons 20154-6. On old green profile, foot was too high, peak not bad

A A A B B C C C E E E E F F F G G G G

1 2 3 4 5 50 100 150 200

SWH IWCTM_2001 16

With that profile I had a decent fit, need to revisit now that sideons are more or less sorted out

_ Better simulations use opacity tables generated from OPAL code _ Increases growth slightly, improves agreement at 30 microns Simulations using XSN opacities, Dante drive, calculated spectrum (same as above) OPAL opacities, drive shown above and calculated spectrum OPAL opacities, Planckian spectrum 0. 20 0. 20 0. 20 0. 20 0. 15 0. 15 0. 15 0. 15 0. 10 0. 10 0. 10 0. 10 0. 05 0. 05 0. 05 0. 05 0. 00 0. 00 0. 00 0. 00 4 4 4 4 3 3 3 3 2 2 2 2 1 1 1 1 Ti m e ( ns) Ti m e ( ns) Ti m e ( ns) Ti m e ( ns) λ λ λ λ = 30 µm, 2.0 µm amp Modulation (OD) 4 4 4 4 3 3 3 3 2 2 2 2 1 1 1 1 Ti m e ( ns) Ti m e ( ns) Ti m e ( ns) Ti m e ( ns) λ λ λ λ = 70 µm, 1.9 µm amp λ λ λ λ = 50 µm, 1.8 µm amp 4 4 4 4 3 3 3 3 2 2 2 2 1 1 1 1 Ti m e ( ns) Ti m e ( ns) Ti m e ( ns) Ti m e ( ns) 0. 25 0. 25 0. 25 0. 25

SWH IWCTM_2001 17

Recent shots in cocktail hohlraum had this drive

250 200 150 100 50 TR ( eV) 5 4 3 2 1 Ti m e ( ns ) 23680 ( cock t ai l ) 23682 ( Au) 23683 ( cock t ai l ) 21885 ( Au,Dec 00)

Dante curve 23683, minus 25 eV from “typical” temperature for this series, shifted +200ps for Dante timing -- used as source for foil simulations

SWH IWCTM_2001 18

At 70 micron wavelength, we see good agreement between simulated and observed perturbation growth

0.4 0.3 0.2 0.1 0.0 M odul at i

• n (

OD) 5 4 3 2 1 Time (ns Cocktail (Oct) Au (Oct) Cocktail (Jun) Au (Dec)

λ λ λ λ=70 µm Nominally adjusted Dante, too high to fit side-ons 17eV lower, too low to fit side-ons At 70 µm wavelength there is no difference between Au and cocktail drives in modulation growth. Early shots seemed to show experimental difference, but not more recent data 70 µm happens to be the wavelength at which experiments have worked to date. Need to get data comparing Au and cocktails at smaller wavelengths!

SWH IWCTM_2001 19

We are also planning convergent Rayleigh-Taylor experiments with a mock fuel layer

• On NIF capsules, perturbation ends up growing on

interface between ablator and fuel, which becomes increasingly unstable as shells implode

• Converging geometry is a big part of the physics

determining densities, plus something we haven’t done enough with yet

CH Be stays at density > CH Doped with silver for diagnosis Similar to experiment calculated by Dittrich for 0.6-scale NIF noncryo Impose perturbations, view face-on Image here may give “side-on” growth measurement Image here will give “face-on” ρ ρ ρ ρr growth measurement

SWH IWCTM_2001 20

With a beryllium mock fuel layer we do a decent job

• f mocking up the interface instability

0.2 0.4 0.6 0.8 1 2 3 4

Fuel density Ablator density Interface radius (NIF, mm, scaled for Omega)

Black NIF Green CH over Be(Cu) Also tried CH(Ge) over CH(Ti) Ge= 0,1,2.5, and 4% (red curves)

SWH IWCTM_2001 21

We are working on optimizing this experiment

Current thinking: ramp pulse that pushered single shell program developed, they are verifying symmetry Capsule 210 µm outer radius, 23 µm CH / 4µm Be+0.5% Ag / 3 µm mandrel Gives good density profiles, and good images

Simulated image from 1 µm initial amp, 50 µm initial wavelength, 2 1/2 waves at waist cut into ablator