Tracer Absorption Spectroscopy of ICF Ablator Materials An

Tracer Absorption Spectroscopy of ICF Ablator Materials: An OMEGA Experimental Campaign and Associated Modeling David Cohen, Swarthmore College with David Conners ('03) and Kate Penrose ('04), Swarthmore College and Joseph Mac. Farlane, Prism Computational Sciences Donald Haynes, University of Wisconsin Paul Jaanimagi, University of Rochester/LLE Otto Landen, Livermore National Laboratory We undertook an experimental campaign to study the effects of dopants on the radiation physics of ablators in indirectdrive hohlraum environments as part of the NLUF program on the OMEGA laser several years ago. Our primary goal was to demonstrate the delay in the Marshak wave propagation time to a specified depth in the interior of plastic ablator materials caused by the addition of Germanium dopant. The experimental campaign proved to be quite ambitious, necessitating the development of target fabrication techniques, new diagnostics (backlit tracer abstorption spectroscopy employing the LXS at OMEGA) used simultaneously with traditional ones, and detailed modeling of the hohlraum radiation field as well as the ablator physics in order to extract science results. While we have demonstrated the anticipated later turn-on times of the spectroscopic signal in the doped ablator samples, the results are muddied by the earlier than expected tracer turn-on times in both samples and the weaker than anticipated absorption signals. During the course of the campaign and with the associated computer modeling, we learned several things that may be of interest to the ablator physics and ICF communities apart from the specific results of our experiments.

OUTLINE 1. Scientific Context 2. Experiment Design 3. Overview of Results 4. Targets 5. Modeling 6. Data and Conclusions

Context Ablator dopants are used to control fuel pre-heat, but they also affect the radiation hydrodynamics of the interaction between the hohlraum radiation field and the capsule.

Ablator dopants affect the opacity and density, changing the manner in which energy is absorbed by the ablator. Controlling the process requires a means of diagnosing the properties of doped and undoped ablators in the hohlraum environment. Traditionally, samples in and on hohlraums have been evaluated spectroscopically via emission (in e. g. gas-filled capsules; and tracers in hohlraum walls) and using shock break-out measurements. Olson et al.

In a different context, Perry showed that absorption spectroscopy in multi-layered targets could diagnose radiation transport. And Chenais-Popovics et al. showed that Cl Ka absorption spectroscopy could diagnose material properties. Laser-produced Bi plasma provided the backlighter continuum source. From Chenais-Popovics 1989

We put these ideas together, and building on a previous effort to measure tracer emission spectra from aluminum witness plates with Tina Back, proposed an experimental campaign to use backlit Cl Ka absorption spectroscopy to diagnose radiation physics in the interior of ablator samples. We proposed to do this by placing thin tracer layers at specified depths in the interiors of ablator samples mounted on hohlraums. The spectroscopy monitors the ionization conditions in that layer, effectively diagnosing the timedependent radiation field properties at a specific location inside the sample. A time-delay in the turn-on time of the tracer signal between doped and undoped samples, for example, would allow us to determine the effects of dopants on the Marshak wave propagation.

We undertook these experiments under the auspices of the NLUF program at OMEGA in 1998. They continued until 2000. We encountered various challenges involving experiment design, target fabrication, and the interpretation of diagnostics. I will discuss the series of experiments -- but will focus on results from the April 2000 campaign in which we measured absorption spectra in two different shots, one with a (undoped) plastic witness plate and one with a 1. 75% Ge doped witness plate.

Our original plan was to make side-by-side measurements on doped and undoped ablators mounted on the same hohlraum We mounted them first on the outside of hohlraums, near the midplane; one spectrometer with two separate crystals was used A halfraum with two samples (two tracers, two backlighters) and two spectrometers

We never were able to successfully measure a tracer spectral signal on these experiments: • Lower-than expected drive temperatures (tracers deeper than they ought to have been) • “Cross-talk” between samples; emission seen by spectrometers not coming from line-of-sight through samples? • Problems with one spectrometer Other ambitious plans included use of wedge witness plates to make passive shock breakout measurements (VISAR with J. Oertel) simultaneously with tracer spectroscopy

Eventually, we settled on an approach involving single samples mounted on the ends of halfraums I will discuss mostly these experiments, carried out in April 2000, as they have the most straight forward target design and had the most positive results.

Targets were fabricated at General Atomics by Abbas Nikroo and assembled by Russ Wallace at LLNL ~5 m ~20 m We will be comparing shots 19526 and 19528 from April 2000: halfraum radiation #26: undoped, tracer depth 6. 3 m Na. Cl tracer <0. 5 m Ge-doped or undoped plastic (produced by GDP) #28: doped, tracer depth 4. 1 m Drive temperature history modeled and constrained by DANTE

Experimental Set-up Including schematics of diagnostic lines-of-sight Note: only one (blue) beam into the halfraum is shown here, for simplicity. All shots were carried out with 15 beams.

Quick Look at Main Result Shot #26: undoped Shot #28: doped Some absorption signal, apparently, on a noisy continuum: Turns on later in the doped sample (and tracer was even shallower in this sample); See progression through ionization states. But--in both cases--earlier turn-on than models predict

Experimental Configuration LEH facing P-7 (LXS in P-6) washer/aperture Gold Halfraums: L=1200 m, R=800 m positioning wires Pb-doped plastic mount Bi/Pb backligher foil TVS-X view

Targets and Target Fabrication SEM images (left) show that witness plates with KF are bumpy (not a problem with Na. Cl tracers)

SEM images of finished witness plates Leakage of KF tracer onto front of plate (right), but no similar problem in witness plates with Na. Cl tracer (left) GA produced these witness plates by first making the thicker plastic layer via glow-discharge polymerization (GDP); The salt layer was deposited on this plastic, and then the whole assembly was put back in the GDP chamber and an additional ~5 m of plastic was deposited on top.

SEM image of cross-section of plate Note the salt crystals…could the weak (or nonexistent tracer signals be due to this? )

Same target as previous slide; higher magnification

And higher still… If there are large gaps in the tracer layer, no amount of average areal mass will provide a strong signal.

Witness plates were mounted on the ends of halfraums; backlighter foils hung ~1. 5 mm from LEH TVS-Y view TVS-X view of plain foil

Target alignment in the chamber was non-trivial and relied on the creation of reticles, some keyed to the wires TVS-Y view (theta = 77. 28, phi = 19. 96)

TVS-X view (theta = 64. 44, phi = 270. 00)

Modeling 1. Viewfactor modeling of hohlraum drive, constrained by DANTE (and using measured beam profiles as input) 2. Hydrodynamic calculations for time-dependent witness plate properties (1 -D Lagrangian; DCA and UTA atomic and EOS models; short characteristics multi-group radiation transport) 3. CRE post-processing for spectral synthesis We use codes written by Joe Mac. Farlane at Prism Computational Sciences as well as some publicly available codes written at the U. Wisconsin Fusion Technology Institute.

Vis. Rad Viewfactor Modeling • 15 cone 2 and cone 3 beams into the halfraum • 1 ns square pulse • 3 beams onto the backlighter foil; also 1 ns square, but staggered in time for more even backlighter source. Note: not all beams are shown.

Constraining the viewfactor modeling • Beam powers and pointings are known • Temperature dependent albedo is modeled (separately) • X-ray conversion efficiency (of lasers) is a free parameter Radiation flux monitored on element at DANTE position

Shot #26

X-ray conversion efficiency was relatively low in order to match DANTE data: simulation on left used a constant XCE=0. 55, while the on the right is the same one from the previous slide, that reproduced the DANTE data.

DANTE temperature profiles (based on unfolds by Bob Turner and new calibration, April 2000) Are these temperatures low?

Calculated Drive Temperatures at Ablator • View factor calculations were performed to compute the flux incident on the ablator and the wall temperature seen by Dante. • The drive temperature at the ablator is slightly higher (up to ~ 5%) than the wall temperature seen by Dante. • Dante sees laser hot spots due to several beams. Only ~ 50% of the flux seen by Dante is due to wall reemission.

The Incident Flux at the Ablator Is Very Uniform View of entire halfraum 1200 mm diameter ablator region View from P 7 direction • The incident flux (= s TR 4) at the ablator sample is ~ 10 - 15 e. V higher than throughout the rest of the halfraum. • Using beams with uniform powers, the maximum change in flux (peak-tovalley) across the central 600 mm diameter of the ablator patch is only 1. 4% (0. 35% in TR). [Because of aperturing, only central ablator region is seen. ]

Effect of Beam Imbalance on Wall Temperature Seen by Dante Beams 25, 50, 45, & 69 are in the Dante field of view. 69 Approximately 50% of the flux seen by Dante is due to laser hot spots. 50 25 45 View from Dante (H 16) Figure (right) shows Shot 19526 beams energies were relatively low for beams 25 and 69. When taking into account actual beams powers: -- the flux at Dante decreases by 8. 7% -- inferred temperature decreases by ~ 3 - 4 e. V.

Effect of backlighter radiation on flux incident on witness plate

Once the hohlraum radiation field is modeled: Hydro simulations of the witness plate Electron Temperature Note: in Ge-doped sample, the peaks are narrower -- shock wave and radiation wave move slower

Same simulations: Radiation Temperature Slower Marshak wave velocity in the doped sample

Spectral synthesis: post-processing of hydro results (undoped sample) Note: Shouldn’t see Be-like until ~0. 6 ns.

Calculated tracer turn-on time vs. drive temperature

CONCLUSIONS Experiments were ambitious, but simpler single patch design employed in April 2000 generated time-resolved tracer absorption measurement in doped and undoped samples. In both types of targets, the signal turn-on time was earlier (by factor ~2) than expected, and lower ionization stages were not seen). But, signal turned on sooner in undoped sample than in doped sample. We looked at many causes for the weak signals and early turn-on times, including target fabrication issues; drive temperature may have been underestimated, but this cannot account for all of the discrepancy.

Extra slide following

In our Vis. Rad modeling, we looked at three cases: (1) fixed XCE=0. 55; (2) Modeled XCE (from G. Magelssen); (3) XCE adjusted to get match to DANTE temperature. Note: The XCE needed to fit the DANTE data is surprisingly low (above); and the witness plate temperatures are generally lower than the DANTE temperatures (right) in conflict with the slide from the IFSA meeting, prepared by Joe.