Development of new platforms for hydrodynamic instability and asymmetry measurements in deceleration phase of indirectly driven implosions on NIF

Hydrodynamic instabilities^1–4 have been shown to impact neutron production in inertial confinement fusion (ICF)^5–9 implosions on the National Ignition Facility (NIF).¹⁰ The ICF scheme consists of a hollow spherical capsule of a low atomic number material (ablator) and a cryogenic layer of deuterium-tritium (DT) ice. An x-ray drive causes the capsule surface to ablate, propelling the inner surface towards the center of the capsule. The x-ray drive is generated from the walls of a cylindrical, high atomic number (e.g., Au or deuterated-uranium, DU) hohlraum driven by a shaped laser pulse. The implosion increases the density and temperature of the DT fuel, reaching peak compression and forming a hot spot where the required plasma conditions for fusion can occur. The implosion can be broken into two phases before peak compression, namely, the acceleration phase where the shell moves without resistance into the capsule cavity and the deceleration phase where the shell meets the returning shock reflected from the capsule center back into the fuel. Figure 1 (Refs. 11 and 12) schematically illustrates how a sinusoidal ripple on the surface of a capsule can grow through the acceleration phase of an implosion using a simulated portion of the capsule shell, shown on a density map. The acceleration phase of the implosion is dominated by Richtmyer-Meshkov^3,4 and Rayleigh-Taylor^1,2 instability growth, such that very small amplitude density perturbations have the potential to grow.¹³ The acceleration phase of the implosion ends when the shock launched into the internal gas of the capsule returns and meets the incoming shell. This forms a new instability regime where the existing growth at the end of the acceleration forms the seed for further growth in the deceleration phase. The most catastrophic growth can rip apart the shell, removing confinement and introducing a higher atomic number shell material into the DT hot spot. This is illustrated in Fig. 1 with a simulated x-ray self-emission image where an example perturbation on the top of the capsule has launched a jet of doped plastic ablator into the hot spot. As the shell consists of heavier elements, such contamination can impact the implosion through introduction of mass, radiation energy loss, and reduction of neutron producing volume.

We study the emergence and growth of perturbations through different stages of the ICF implosion with separate but complementary experimental methods. This approach is demanded as inferring the origin, and the impact of hydrodynamic growth at the stagnation stage of the implosion (where the fusion burn occurs and neutrons are produced) is problematic due to the large number of potential seeds that exist and their complex evolution through all stages of the implosion. The methods described focus on a convergent geometry although much recent investigation of hydrodynamic growth has been done in planer geometry.^14,15

Examination of ablation front seeds has been undertaken using a “hydrodynamic-growth radiography” (HGR) technique that highlights the areal density (ρR) variation in the shell through the early stages of the acceleration phase.^11,16,17 This was achieved through the introduction of a cone into the capsule which allows the use of an external x-ray backlighter to radiograph the shell. The presence of the cone inhibits the measurement into the deceleration phase, which occurs near 5–6× convergence of the capsule, as the opening of the cone perturbs the capsule convergence after this point.

Measurements of ρR perturbations in capsules at higher convergence are challenging and require a different approach from the HGR experiments to radiograph the shell. The first measurements at higher convergences, closer to stagnation, were acquired in a direct drive configuration at the OMEGA¹⁸ Laser Facility. At peak compression, the hot compressed core and the inner surface of the shell produce strong continuum x-ray emission. This emission is used to self-radiograph the outer colder shell.^19–22 These experiments used the K-edge, or 1s-2p absorption, of a dopant in the shell in combination with the continuum self-emission and differential imaging of the implosion “above” and “below” these absorption features to measure the shell ρR. However, measurements when the capsule is fully converged have further limitations. The Rayleigh-Taylor growth from both the acceleration and the deceleration, as well as low mode asymmetries from the drive, contributes to the ρR perturbations observed, making interpretation difficult. Furthermore, at peak convergence, the spatial resolution of available imaging systems limits the spatial mode numbers that can be resolved to <10, whereas it is expected that the amount an initial perturbation will increase in amplitude or density is most damaging to the implosion at mode numbers of 40–60.^23–25 

The first technique discussed in Sec. II of this paper improves on these previous radiography results by boosting the self-emission of the hot spot near peak velocity through the introduction of an argon dopant to the capsule gas fill.²⁶ Measurements of ρR perturbations near peak velocity (PV) have a significant advantage that they reflect the integrity of the capsule after the inward acceleration growth is complete, at a time when the capsule is still large enough to measure perturbations with spatial mode numbers near ∼40.

The second complementary technique discussed in Sec. III visualizes these most catastrophic perturbations and low Legendre-mode asymmetries using a preferentially doped inner few microns of the capsule ablator. Specifically, a small percentage of a high-atomic-number element is placed on the inner surface of the shell in contact with the forming hot spot, allowing the interface of the shell and gas to be imaged with the x-ray radiation produced through the deceleration phase and into stagnation.

On NIF, three ablators are being pursued, and each has different sensitivities to hydro-instabilities.²⁷ The techniques in this paper are tailored to the specific ablators through the dopant materials and their concentrations. The self-backlighting experiments described in Sec. II were performed in plastic ablators with low-adiabat, high compression drives where we expected the largest growth of both pre-imposed and native perturbations. The imaging techniques descried in Sec. III were applied in the highest neutron yield design which uses a high-density carbon (HDC) ablator where we highlight perturbations invisible in the high convergence DT experiments such as the capsule support tent and low mode asymmetries.

These new experimental platforms have highlighted the total impact of highly 3-D perturbations on the shell integrity late into the implosion. Section IV will discuss in more detail what we have learned and quantified with the new experiments. The conclusions are further summarized in Sec. V.

In order to radiograph the capsule shell at peak velocity and into the deceleration phase, we enhanced the self-emission of the hot internal hydrogen gas of the capsule through the addition of 1% Ar, forming a 69.3%He4 + 29.7%H + 1%Ar mix.^26,28,29 The technique is illustrated in Fig. 2(a) with a 2-D HYDRA³⁰ simulation showing the rebounding shock temperature profile and the cold outer shell with a pre-imposed modulation (a sinusoid with λ = 140 μm and a peak-to-valley amplitude of 220 nm). The radiograph shows the variation in the ρR of the shell, imaged using the hot-core x-ray emission as a backlighter to the surrounding cold shell material. The addition of the argon gas enhanced the continuum emission by an order of magnitude²⁸ near peak velocity, allowing the shell condition to be probed at this time. Figure 2(b) shows a shell pie diagram for the first Ar-self backlighting experiments. 1% of Cu dopant is added to the CH shell to increase the radiographic contrast. The experiments were performed in both a 4.69 mm diameter Au hohlraum with a 920 μm outer radius capsule and a 5.75 mm diameter Au hohlraum with a 1146 μm outer radius capsule using all 192 beams of NIF with a “low-foot” laser pulse shape.³¹ This low foot pulse was chosen because it showed evidence of hot spot mix and shell perturbations during the National Ignition Campaign (NIC).⁸ Figure 2(c) shows an example experimental x-ray image of the pre-imposed ripple “self-radiograph” during deceleration.²⁸ The attenuation from the pre-imposed ripples is visible as vertical stripes in the image.

We performed a number of experiments with pre-imposed ripples which allowed the technique to be proven, showing that we can measure the total growth of a known feature. Figure 3(a) shows a self-radiograph near peak velocity. By fitting a low-mode profile to infer the backlighter distribution, we can subtract the spatial variation of the backlighter and recover the amplitude of the perturbation, as shown in Fig. 3(b). (Without independent witness of the rebounding shock profile, we assume that the low-mode component of the image represents the backlighter, and this is described in more detail in the study by Pickworth et al.²⁸) The experiment with a known initial sinusoidal ripple allows us to easily infer the radius of the capsule at the time of the radiograph as the mode number of the ripple does not change with the convergence. The initial 2-D, 220 nm peak-to-valley perturbation has grown ∼7000× in ρR by the end of the measurement at peak compression. However, we found an unexpected difference in total growth between the equator and the pole of the capsule²⁸ which we attribute to asymmetry in Au M-band emission within the hohlraum. This asymmetry is an important indication of the imperfect laser drive that could affect the symmetry of compression and/or instability growth at the ablator-ice interface.³² These experiments also use a hohlraum gas density of 0.96 mg/cc and rely on cross-beam energy transfer³³ to tune the symmetry of the x-ray drive. Thus, the radiation drive between the waist and the pole could be changing as a function of time³⁴ and be responsible for the observed instability growth asymmetry. Figure 3(c) shows self-radiograph images with no pre-imposed ripple, indicating little variation in the optical depth in the absence of a pre-imposed perturbation.

This self-radiograph method allows us to investigate new capsule mounting techniques. The polar contact tent provides an alternate support mechanism to hold the capsule in the center of the hohlraum and is illustrated in Fig. 4(b) against a standard tent geometry shown in Fig. 4(a). In previous experiments, the tent^35–37 that drapes over and under the capsule, as shown in Fig. 4(a), has been used, departing from the capsule at a contact angle of ∼45° from the horizontal axis. The polar contact tent reduces the contact region of the support and capsule from ∼700 μm radius to a ∼200 μm radius ring at the pole of the capsule. This configuration aims to reduce the contact region of the tent feature and thus reduce the impact of the perturbation on the implosion over the tents previously employed. The contact region is sufficiently small that we can visualize its impact on the capsule ρR using the self-radiograph platform. For this experiment, we used a 5.75 mm diameter hohlraum with a 1146 μm outer radius capsule, which is a larger, hydro-dynamically scaled sister implosion than those previously described with pre-imposed ripples. The self-radiograph, in optical depth modulation, of the polar contact tent from the pole is shown in Figs. 5(a) and 5(b). Figure 5(c) shows an optical image of the polar tent contact region prior to the experiment. A complex radial absorption pattern is visible in the two images in Figs. 5(a) and 5(c); in contrast, a circular lift off region is observed in the optical image of the assembled target. The radial features are likely present in the material of the tent, a 48 nm thick polyimide film reinforced with 8 nm of carbon. The striking 3-D perturbations were measured in this experiment with amplitudes several times larger than expected at the locations of the polar tent contact on the capsule.³⁸ This result and its implications will be discussed in Sec. IV.

Areal density (ρR) perturbations observed in the shell near peak velocity represent the total growth from the acceleration phase. However, as the capsule moves into the deceleration phase, these perturbations undergo further growth. The self-radiography technique assumes that the shell material purely absorbs the self-emission from the radiative shock. Thus, the technique is blind to scenarios where the shell is punctured by catastrophic growth that launches an ablator material into the hot spot, which is heated and itself emits brightly. This can be confused for a low-density region of the shell in the absence of other information. To visualize and diagnose these scenarios, we develop a new deceleration phase imaging technique, using a localized high atomic number dopant in the inner layer of the capsule, initially in contact with the gas fuel. This made it possible to clearly visualize the temporal evolution of high-mode perturbations including the tents and fill tubes, as well as low-mode asymmetries that affect the overall shape of the implosion. This method additionally allowed systematic studies of neutron yield degradation due to high-mode perturbations, low-mode asymmetries, and x-ray radiation losses, using the “self-emission” imaging technique in these low-convergence, gas-filled implosions.

Figure 6(a) shows the shell composition for a “standard” HDC “symmetry” capsule, and Fig. 6(b) shows the shell composition for the “imaging” HDC capsule with a 0.17% tungsten dopant at the inner surface. In these experiments, DU cylindrical hohlraums were used to implode an HDC capsule supported by 45-nm thick membranes (“tents”). The hohlraum length was 10.2 mm, with a diameter of 5.75 mm, and a laser entrance hole (LEH) diameter of 3.37 mm. The hohlraum was driven with the 6-ns long laser pulse based on a 3-shock design with a nominal peak power of 400 TW and an energy of 990 kJ.^39–42 The capsules were filled with 4 mg/cc deuterium gas through a 10‐μm diameter glass fill tube. The HDC capsules were nominally 64‐μm thick with an outer radius of 909 μm. The baseline capsule had 18 μm thick, 0.2% atomic W-doped layers offset by a 6 μm thick undoped HDC layer from the inner capsule surface, as shown in Fig. 6(a), while the “imaging” capsules had 24‐μm thick, 0.17% atomic W-doped layers that extended all the way to the inner surface of the capsule, as shown in Fig. 6(b). These implosions do not have a DT ice layer, and the inner surface of the capsule is the interface between the shell and the deuterium gas fuel. The purpose of the W-doped inner layer in the “imaging” capsules was to enhance the x-ray emission of the inner shell surface near peak compression and to study the effects of the enhanced x-ray radiation and resulting radiative cooling on the implosion performance. The undoped 6 μm layer in the standard configuration is sufficient to inhibit the W doped ablator entering the hot spot when no perturbation is present; in this case, the hot spot self-emission dominates. The effect of placing the dopant in contact with the hot spot is schematically shown in Fig. 6(c), indicating the region of enhanced emission highlighting any penetrating perturbations of the shell material. The capsule is imaged with the Kirkpatrick-Baez microscope (KBM),^43–45 allowing the perturbations to be tracked through the implosion with high spatial resolution.

The x-ray images for one of the “imaging” capsule experiments show the evolution of the perturbations and asymmetries near peak compression [see Figs. 7(a) and 7(d)]. These images were captured with the KBM on a gated x-ray framing camera with a spatial resolution of ∼6 μm full-width-half-maximum, a temporal resolution of ∼100 ps, and an energy content in the image of 10.3 ± 1.5 keV. The perturbations corresponding to the fill tube and tents are marked in Fig. 7(b). The emission profiles of each frame are shown with horizontal lineouts through the fill-tube centers in Fig. 7(e) and with vertical lineouts through the image centers in Fig. 7(f). The emission profiles are displayed on a logarithmic scale to better show the change in image intensity through the implosion. The fill tube is clearly identified as the brightest feature in Figs. 7(a) and 7(b) although it becomes less visible near and after the bang time when the emission from the inner shell increases. Asymmetry in the emission is also visible and linked to a missing drive beam in this case. The tents are easily visible ∼200 ps before the peak compression or “bang time” (BT), although the upper tent perturbation (bright horizontal stripe) is brighter and more visible than the lower tent perturbation, in line with the overall asymmetry in the x-ray emission and attributed to the drive asymmetry.

An important feature of the “imaging” technique is its sensitivity to low-mode asymmetries. This sensitivity allows the correlation of their x-ray signatures with the bulk motion of the neutron emitting region of the hot spot that can be caused by low-mode asymmetries in the incident x-ray drive. If the x-ray drive is stronger from some direction of the capsule, it can cause the shell to accelerate along this direction, causing the bulk motion of the hot spot. This bulk motion during the burn can be detected with neutron time of flight (NTOF) detectors using a Doppler shift of the fusion neutrons. Simultaneously, the asymmetry of x-ray emission at the inner shell can be correlated with the bulk fuel velocity measured with the NTOF detectors. We observed several experiments with large up or down bulk fuel velocities.⁴⁶ These velocities are detailed for three imaging experiments in Table I along with the locations of four neutron detectors used to make these velocity measurements in spherical coordinates of the NIF target chamber. The three detectors in Table I are approximately orthogonal, allowing a velocity vector to be reconstructed. The fourth detector was used to check the consistency. Figure 8 shows time resolved x-ray self-emission near peak compression or bang time for three different “imaging” capsule experiments which exhibited a relationship between the x-ray emission and the vector of the bulk velocity reconstructed from the NTOF measurements. These equatorial images were captured with 10 μm pinholes at 9× magnification with an average image x-ray energy of ∼10 keV. These images were very similar to those observed with the narrower energy band, higher resolution KBM. We note a correlation between brightening in the x-ray image and the direction of the velocity on the neutron production as viewed from the equator.⁴² This brightening or asymmetry is more pronounced in the KBM images (see Fig. 7 which is the same experiment as the positive P1 image in Fig. 8), possibly due to the centering of the instrument on the L-shell line emission from the W dopant which is a dominant source of x-ray emission. The brightening of x-ray images and the resulting bulk velocity of the fuel can be used to detect and understand the origin and mitigate low-mode asymmetries of the implosions. The total magnitude of the velocity is very similar between the three experiments shown in Fig. 8, suggesting even when the large P1 (polar) velocity is mitigated, the M1 (equatorial) velocity can still be significant. In the individual cases of the large P1 velocity observed, we can attribute the origin to a missing set of beams (positive P1 case) or offset from nominal target alignment (negative P1 case). All three experiments have a significant M1 velocity of uncertain origin.

As an additional advantage of the “imaging” platform, the short-scale perturbations can also be better visualized using enhanced emission from the tungsten-doped inner surface of the shell. Figures 9 and 10 show time sequences of the x-ray images near peak compression measured in the polar direction for two shots. In the first shot, shown in Fig. 9, the high-mode perturbations were detected early in the implosion (at 260 ps before bang time). They coalesce as the capsule converges (at 10 ps before bang time) and merge in the final stages of the implosion during the rebound phase (at 110 ps after bang time). In the complementary implosion driven with the same drive (shown in Fig. 10), the level of perturbations is much reduced. The isolated high-mode perturbations are distinctly visible before peak compression (at 130 ps before bang time). They reach the center of the hot spot later in time (at 240 ps after bang time) in the rebound phase. It can also be seen from the images that the low-mode asymmetries are similar in these two shots and related to the position of the diagnostic windows (which in this cased are uncoated HDC). The correlation of the measured high-mode features with the neutron yield will be described later in this section, while the hypotheses about the origin of these perturbations will be discussed in Sec. IV.

Figures 11(a) and 11(b) show the simulated x-ray images from a 2-D HYDRA simulation of the tent and fill tube, respectively, for the imaging capsule experiment shown in Fig. 7. Good agreement between simulations and experiment was achieved when the up/down asymmetry observed in laser delivery was included in the simulations [compare Figs. 11(a) and 7(b)]. Figure 11(c) shows the experimentally measured separation of the two emission features attributed to the tent, as seen in the x-ray images in Fig. 7 (square markers), against the simulated separation of the two tent features (lines). The solid line represents the case with the asymmetric drive, which matched the laser delivery of the experiment, while the dashed line is for the symmetric case. Visible divergence of these features occurs close to stagnation; however, the penetration of the tent jet feature into the hot spot is captured well in the simulation. Figure 11(d) shows simulated density and temperature maps for the image in Fig. 11(a), showing the effect of the velocity on the shell uniformity and evolution of the jet from the tent feature. The variation in observed x-ray emission can be explained as a result of both a lower opacity region of the shell in the direction of the velocity vector and that this region is closer to the hotter regions of the compressed internal gas. Figure 11(e) shows the density map from a 2-D HYDRA simulation without tent or fill tube perturbations that replicate the overall low mode density perturbation, similar to that first reported by Spears et al.⁴⁶ The inclusion of the asymmetric drive allows the simulation to match well the experimental data. We also gain insight that the jets are being “swept” in the direction of the bulk velocity, following the internal movement of the hot spot.

The high-resolution images obtained by the KBM allow some more advanced image analysis to be applied to the fill-tube perturbation. Figures 12(a) and 12(b) show “zoomed in” regions shown above in Figs. 7(a) and 7(b). The images have been rotated such that the fill tube “moves” left to right (a small rotation from Fig. 7), along the horizontal axis, and the background of the shell emission has been removed. Figures 12(c) and 12(d) show slices through the “emission volume” reconstructed using a forward-fit iterative unfolding procedure, which allows left-right asymmetry in the reconstructed emissivity by including the asymmetry in the forward-propagated functions.³⁵ This analysis makes the following assumptions about the nature of the 3-D emitting volume: the emission is not self-opaque, the volume is following front-rear symmetry about the fill tube axis and along the line of sight, and the viewing angle is orthogonal. This reconstruction implies a hollow volume that fills in over 100 ps, separating the images, and the later image is 24% smaller in diameter and with increased peak emission density (∼2.6×). The spatial extent of the measured fill-tube perturbation is about 2 times wider than in the simulation shown in Fig. 11(b). The larger perturbation can be attributed to the effect of the x-ray shadowing that was recently discovered at earlier convergence of ∼2 using the HGR platform.⁴⁷ When the increased level of perturbations due to x-ray shadowing was included in the simulations, the inferred hollow volume was in agreement with the experiment.⁴⁸ This will be discussed in Sec. IV.

Through comparison of experiments with and without the inner doped surface, we can determine the role of radiation losses in these implosions, one of the major questions in ICF. The measure of the increased x-ray radiation above that expected from deuterium plasma was defined using the “x-ray enhancement factor.”⁹ To determine if enhanced loss of energy through x-ray radiation may account for the losses observed in neutron yield between the baseline (∼2.3 × 10¹³) and imaging capsule (∼1.2 × 10¹³), we form an experimentally based x-ray enhancement factor.⁴⁹ The total x-ray emission is given in the following equation:

where E_n is the energy in DD neutron yield in J, τ_E is the attenuation of the shell, the scientific constants are evaluated using the factor (C_X/C_N) = 1.895 × 10⁻⁴, E is the photon energy in keV, and T_i and T_e are ion and electron temperatures, respectively in keV. We assume local thermodynamic equilibrium (LTE), T_i = T_e, and use the measured T_i ∼ 3.7 keV for all the implosions. The emissivity from the compressed deuterium can be modeled as a free-free continuum, which is estimated from detailed configuration accounting, DCA⁵⁰ non-LTE at 1-GBar, and DD reactivity cross section (p. 45)⁵¹ and attenuated by the remaining shell. Equation (1) allows the total x-ray emission to be modeled with the experimental observables of DD neutron yield, DD ion temperature, and shell ρR.

To form the X-ray enhancement factor, we calculate the ratio of the observed emission on the image plate detector to the expected signal determined with Eq. (1) which differs from previous work⁵² by including lower energy x-ray emission. We compare the x-ray enhancement factor in Fig. 13(b) using two thin filter channels (∼11 keV and ∼14 keV mean energies) shown in Fig. 13(a) with the energy dependent response convolved using the estimated emission spectrum. The imaging capsule implosions (blue, Fig. 13) observed a factor of ∼2 lower measured neutron yield, and the x-ray enhancement factor is increased by an order of magnitude over the baseline implosions (red, Fig. 13). The enhancement factor is a factor of ∼3 higher in the ∼11 keV channel compared to the ∼14 keV channel. This simple analysis (described below in Sec. IV) indicates that radiative losses are a dominant loss mechanism for the imaging capsules⁴⁹ and that lower energy radiation contributes strongly to this energy sink. Figure 13(c) shows the contribution to the total observed emission as a function of time for the fill tube and tent in the experiment shown in Fig. 7. The fill tube dominates the emission in the early time images, over the tent. This will be discussed further in Sec. IV.

This article describes measurements of hydrodynamic instabilities in the deceleration phase of ICF implosions using two complementary platforms. While the “self-backlighting” platform is based on x-ray radiography of the perturbations, the “self-emission” platform takes advantage of the enhanced inner-shell x-ray emission to visualize the perturbations around peak compression at convergences of ∼5–10×. These two platforms complement the hydrodynamic growth radiography (HGR) platform previously developed for the acceleration phase of implosions used at lower convergences of ∼2 to 4×. The HGR results were critical for understanding of the hydrodynamic instabilities in indirectly driven implosions on NIF. The “growth-factor” measurements with pre-imposed 2-D sinusoidal perturbations established a correlation between the ablation-front stability and neutron-yield performance of the layered DT implosions.⁵³ The instability growth of 3-D perturbations revealed that perturbations from fill-tubes and capsule support “tent” were significantly larger than expected with both CH and HDC ablators.⁴⁷ The vorticity effects during initial stages of implosions increased the “tent” perturbations to the levels comparable to the shell thickness even at a relatively low convergence of ∼3× in CH-ablator implosions.⁵⁴ The x-ray shadowing during initial stages of implosions significantly increased the area of the “fill-tube” perturbations compromising the shell stability during implosions. In addition, the instability growth of 3-D “native roughness” perturbations was stronger than expected from “growth-factor” results measured with 2-D modulations.⁴⁷ One of the hypotheses to explain the results is based on the 3-D modulations of the oxygen content in the bulk of the capsule having a comparable effect on the overall growth of capsule perturbations as the outer-surface capsule roughness.⁵⁵ 

The first proof-of-principle argon self-backlighting experiments in the deceleration-phase, presented in Sec. II, were performed with pre-imposed 2-D sinusoidal perturbations. The measured instability growth factor in ρR of ∼7000× was observed at the beginning of the deceleration phase in these experiments, the largest ever observed in ICF implosions. This was performed in the CH ablator with a more unstable “low-foot” drive as this accentuated features and growth. The initial 2-D perturbation with Legendre mode 40 had a 220-nm peak-to-valley amplitude, the smallest amplitude that was possible to impose at the time of the experiment. At the time of observation, this perturbation was shown to be saturated in its growth due to nonlinear effects. In addition, it was discovered that the growth was different between the capsule equator and pole, possibly pointing to the different amount of pre-heat between these two directions. To date, we have applied this technique only in low-foot drives, although it in principle is compatible with all ICF configurations being investigated at NIF.

This platform was then applied to the observation of 3-D perturbations. The experiment tested the new “polar contact” tent feature, as a part of a tent mitigation effort. As in the earlier 2-D experiments, it was performed in the CH ablator with a more unstable “low-foot” drive to accentuate the hydrodynamic growth. It was expected that the perturbation would be dominated by the lift-off region of the tent, as seen in previous tent observations. However, measurements showed a complex 3-D pattern, not an idealized “small circle” on the surface of the imploded capsule. The detailed radial pattern indicated that the seed of the perturbation was present in the polar tent over the region in contact with the capsule, suggesting that the “stiff” tent bends or perhaps fractures into segments, although this was not observed in the pre-experiment metrology. Clearly, the seed present in the tent contact region is not an ideal circular pattern, and it is likely non-reproducible. The peak-average fractional ρR modulation in the experimental image is ∼150%, inferred from the experimentally observed optical depth and simulated average ρR at the time of measurement. The instability remains in the linear growth regime, and we infer that the initial seed has an “equivalent” δ (ρR)/ρR ∼ 0.02%–0.03% (for a linear growth factor of 1 × 10⁴ at mode number 50), five times larger than the simulations for the ideal circular contact. However, a recent layered DT implosion with the “polar contact” tent indicated improved neutron yield performance.³⁸ These preliminary results warranted further development of the polar contact tent as a mitigation strategy.

While backlighting techniques use absorption of the shell material to measure perturbations, the hot-shell emission is another complementary way to visualize the non-uniformities near peak compression. During the “CD mix” campaign on NIF, CH capsules were filled with tritium gas that had smaller contribution to overall hot-spot emission compared to the inner surface of the plastic shell.⁵⁶ In these conditions, the non-uniformities in the shell were measured and characterized. This technique was extended to the HDC shells with inner W-doped layers, as described in Sec. III

Observable early in the sequence of images (Fig. 7), the fill tube is the dominant emitting feature and the largest single perturbation. Initial 2-D HYDRA simulations expected a smaller feature [Fig. 11(b)]. The difference between the simulation and experiment can be understood with recent work, showing the extent and impact of shadows cast by the fill tube stalk itself early in the laser drive.⁴⁷ This causes an initial seed for hydrodynamic growth that is significantly larger than the tube, hole, and glue alone. This dramatically changes the dimensions of the feature (a wide jet that does not penetrate far in the capsule) from the thin jet that bisects the hot spot in simulation. In present low-convergence implosions, we did not expect the fill tube to have a significant effect on the yield reduction. For example, we estimated only ∼5% yield reduction due to reduction in the volume of fuel by 1%–2% and increased radiation cooling caused by the 10 μm fill tube. In the layered DT implosions with HDC ablators (which reach higher convergences than the imaging experiments), the impact of the fill tube has been shown to be stronger through comparison of 10 μm and 5 μm fill tubes. The smaller, 5 μm fill tube reduced the seed perturbation by a factor of 4× from the 10 μm fill tube and 16× in total seed mass and solid angle.⁵⁷ 

After the fill tube, the asymmetry in the x-ray emission was the most remarkable feature. We showed the correlation of this asymmetry with the observed fuel bulk velocity vector. This observation is also consistent with previous work by Spears et al.⁴⁶ In present experiments, the 3% drive asymmetry due to dropped laser beams resulted in the bulk fuel velocity of ∼100 km/s toward the direction of the reduced drive. The estimated yield reduction due to this drive asymmetry was not significant, only ∼10%. However, in the higher-convergence layered DT implosions, such drive asymmetries could significantly reduce the performance. Understanding the origins of these low-mode asymmetries is one of the critical issues for the ICF program. Visualizing the asymmetries using the “self-emission” technique can provide better understanding about the origins of these asymmetries.

The addition of 0.17% W at the inner surface of the shell increased the radiation loss from the capsule and impacted the neutron performance, even in current low-convergence implosions. We were able to relate this loss of radiation, ∼10× over similar implosions with no W at the inner surface, to the 2× loss in neutron yield over the same implosions. Separate from our experimental goal to investigate the penetration of hydrodynamic perturbations into the hot-spot, this opens an interesting avenue to investigate the role of radiation loss in ICF implosions. Emission from the W in the hot spot (enhanced continuum and L-shell) is visible to the imaging diagnostics as it is able to escape the strong absorption of the shell. Emission from the dopant M-shell is unable to escape the shell, but the presence of the L-shell enhancement in emission suggests that the M-sell emission is also present. In the case of different dopant materials or less catastrophic mix, this “dark” radiative mixed material, at low photon energy, could be overlooked as a source of performance degradation in the implosion, transferring energy out of the hotspot and cooling the fuel.

The imaging platform allows us to make estimates on the volumetric extent of features throughout the implosion. Estimation of the fill tube perturbative volume is ∼1%–2% of the hot spot. Interestingly, although less dramatic in radiative losses, the tents occupy a larger volume, ∼20% (two tent perturbations on the top and bottom), of the hot spot due to their large annulus-like perturbation. The fill tube and tent are volumetrically perturbing the hot spot formation, and it would be expected that these perturbations are similar in implosions with and without W at the inner surface. Some insight into the effect of the volume reduction and breakup of the hotspot is seen in the experiment with many high frequency perturbations observed, as shown in Fig. 9. Although the emissivity observed on the implosion was not enhanced more than the other imaging experiments, the yield was further reduced by a factor of two. We attribute this further dramatic impact of breakup of the hot spot to be consistent with a 25% reduction in hot-spot volume. This significant yield degradation due to high-mode perturbations in present low-convergence implosions is an important observation. Similar high-mode perturbations were routinely observed in some higher-convergence layered DT implosions where their effects deteriorating the performance were expected to be even stronger.

This paper has presented two new experimental systems to examine hydrodynamic growth in ICF capsules in the late stages of the implosion, from peak velocity through deceleration to stagnation. In the first innovative technique, self-emission from the hot spot was enhanced with an argon dopant to “self-backlight” the shell in-flight, imaging the perturbations in the shell near peak velocity. These experiments discovered unexpected 3-D structures originating from the capsule support structures. These new perturbations became a primary concern for an indirect drive ICF program requiring their origin to be understood and their impact mitigated. In a second complementary technique, the inner surface of the shell was visualized in implosions using x-ray emission of a high-Z dopant added to the inner surface of the capsule. With this technique, low mode asymmetry and high mode perturbations, including the gas fill tube and capsule support structure, were quantified near peak compression. Using this doping method, the role of perturbations and radiative losses from high atomic number materials on neutron yield was quantified.

This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract No. DE-AC52-07NA27344 and by General Atomics under Contract No. DE-NA0001808. LLNL-JRNL-750655.