Hybrid direct drive with a two-sided ultraviolet laser

Inertial confinement fusion (ICF) employs the energy of a high-velocity implosion to compress and heat (and ultimately confine) a variety of low-Z fusion fuels.^1–3 The most common mixture is deuterium and tritium (DT) as it has the lowest threshold for ignition (i.e., the highest reactivities) and the potential to generate 340 MJ per mg of fuel. Products include alpha particles at 3.5 MeV and neutrons at 14.1 MeV. Assuming a central hotspot (hs) can be made with the density and scale needed to stop alpha particles, or a $ρRhs∼0.3$ g/cm²—at temperatures of 4–5 keV—the propagating burn of the surrounding cold fuel (cf) can be initiated, as recently demonstrated on the National Ignition Facility (NIF).^4,5 ICF could lead to the commercial production of energy if the product of the driver efficiency and target gain G can be 7–10, with the gain defined as fusion yield over the incident energy. Assuming that the wall-plug (to on-target) efficiency is 5–15% accounting for input energy and other subsystems, it is likely that inertial fusion energy (IFE) will require target gains $≥$ 50. Important metrics include the areal density of the cold fuel at its initial radius, $ρRinitial$⁠, and also during burn, $ρRcf$⁠. The nuclear yield can be expected to scale as $mDTϕb$⁠, and the burn fraction $ϕb$ as $ρRcf/(ρRcf+6)≈(ρRcf/3)2/3/3$ for $ρRcf∼3$ g/cm². If the areal density of the stagnated DT scales with the ratio of the kinetic-to-internal energy at the peak implosion velocity as $ρRinitialEKE/EIN$ (see Ref. 6), and $EKE∼ηhydroElaser$ (accounting for laser absorption and target ablation), then the path(s) to high gain are straightforward and require us to maximize $G∼ηhydro(ρRinitial)2/3(EIN/mDT)−1(EKE/EIN)−1/3$⁠.

Most concepts in ICF need to put a strong emphasis on compressibility^7,8 due to limitations in the hydrodynamic efficiency and scale, and details in laser–target coupling. The internal energy of the DT fuel is a function of pressure and will commonly be defined by a term called the adiabat, $α$⁠. $EIN∼pV∼αpFV$⁠, where $pF$ is the pressure of the Fermi-degenerate DT and V is volume. The Fermi pressure $pF∼ρ5/3$⁠. The adiabat of the DT is determined by the power in the laser as a function of time, and low- $α$ designs (⁠ $≤$ 2) tend to have fuel layers that are particularly dense and thin and can be susceptible to a variety of hydrodynamic instabilities. For this region of parameter space, it is common for ignition to be inhibited by mix (seeded by high mode imperfections in the laser and target) and premature heating of the fuel (caused by laser–plasma interactions and high-energy electrons). We can expect low adiabat implosions to eventually reach higher gains,^9,10 but requirements for stockpile stewardship and/or inertial fusion energy might still be difficult to achieve. The obvious alternative is to focus on the hydrodynamic efficiency and scale—as done here—with implosions that can couple much more energy to the DT fusion fuel.^9,11–16

In this paper, we present a “hybrid” approach to direct drive ICF that can exploit a high energy laser with two opposed beams. The concept employs an ultraviolet (UV) gas laser at 248 nm that is being tested within the Milestone-Based Fusion Development Program,¹⁷ with the potential to reach the efficiencies quoted above, although it could also be compatible with light at other wavelengths. Rather than design for high levels of compression, the goal is to maximize the laser–target coupling in a regime that greatly exceeds the criteria to ignite.^18,19 The geometry of the system is two-sided to take advantage of a reactor with thick liquid walls (which can efficiently breed tritium and protect the structure from x-rays and neutrons) and also be compatible with a cylindrical hohlraum (to shield the cryogenic fusion fuel from the chamber environment).^20–22 All of these elements are useful for reducing cost. The target is termed a hybrid because it combines x-ray and laser drives in a manner that is complementary^23–28 and not because it uses the dimensions and pulse shapes of prior work.^29,30 In net, this concept couples 300 kJ (or more) to the imploding fuel and enables high-adiabat implosions with a factor of 12 (3) more mass (areal density) than contemporary experiments on the NIF. Common metrics for stability and gain are substantially improved and could help to relax requirements on the target. The paper begins by introducing the principles of hybrid direct drive (HDD) and the design features used to suppress laser imprint and cross-beam energy transfer. Next, it summarizes the tools needed to obtain near-1D implosions with a two-sided driver: (1) a highly absorbing plasma atmosphere, (2) electron-conduction smoothing, and (3) a ring-peaked laser profile. Finally, we describe the hybrid concept with detailed radiation-hydrodynamic calculations and discuss several measures of margin. The HDD target is found to benefit from efficiency and scale and high levels of coupled energy as expected, and shows promise for achieving breakeven and potentially more.

We provide a schematic of the UV laser system in Fig. 1 and the target in Fig. 2. Long pulse KrF amplifiers and schemes for Raman and Brillouin pulse compression are used to deliver multi-Megajoule energies in the spatial and temporal formats needed for the HDD (Ref. 17). This approach has almost no need for solid optics—which also helps to reduce cost—and can operate at very high fluences (⁠ $≫$ 10 J/cm²). By design, the system represents a departure from current practices but is particularly well-suited to IFE as upcoming papers on the laser and target fabrication will show. The energy from the driver is to be delivered in multiple stages. In the first stage, a short laser pulse (P1) burns through a thin window and heats a high or mid-Z lined cavity or hohlraum to 104 eV. Similar to experiments on the NIF, the window film and hohlraum protect the capsule from the chamber environment,³¹ and the first pulse (and associated x-rays) ablate and shock the capsule to set the adiabat of the cryogenic fuel.² The capsule itself is a deuterated plastic (CD) that is relatively easy to manufacture. The target borrows heavily from learning at the NIF^{9,12,14,16,32,33} but has otherwise been adapted for two-sided illumination. To limit the blow-off of plasma inside the hohlraum relative to conventional indirect drive,⁵ the laser energy intercepting the wall is reduced to a factor of 40 per unit area when integrated over the full pulse. There is no need to tamp hydrodynamic motion with a high-density fill gas (for example, helium). For the second stage of the implosion, a series of short laser pulses (P2 to P7) drive the capsule directly at pressures exceeding 250 Mbar. The mass ablated by the first pulse (P1) enables this mode of operation, as we will explain. To limit laser–plasma instabilities or LPI, the wavelength of the laser is short, 0.25  $μ$m, and the peak incoming intensity is relatively low (⁠ $∼$ 10¹⁵ W/cm²). Cross-beam energy transfer is avoided by virtue of the two-sided geometry and a laser beam profile that can follow or “zoom” with the implosion. (To maximize spectral bandwidth, the laser is capable of operating on multiple lines, but this is not planned for the first experiments.) By putting the capsule in a high-Z cavity and staging the drive, the HDD is meant to exploit features from both indirect and direct drive and avoid many of their weaknesses: the DT fusion fuel is protected from the hot chamber background; there is no risk of laser imprint (like indirect drive) and the hydrodynamic efficiency is quite high (like direct drive).

Two-sided drive has additional benefits, which will be described, but might be expected to make implosion symmetry a challenge. The typical direct-drive target on the OMEGA or NIF laser systems needs a large number of beams in $4π$ sr to generate an efficient implosion (e.g., 60 or many hundreds).^34,35 Each of these beams needs a clear path to the capsule and its own set of optics and control systems. [Spherical illumination might be preferred, but tradeoffs in systems engineering tend to motivate compromises in design.] The hybrid target has just two beamlines, but several novelties still allow for good implosion symmetry. First, and most importantly, is the use of the first pulse (P1) to generate x-rays that ablate the capsule and form a thick plasma atmosphere. Indirect drive is very capable in this regard and can be symmetric when the ratio of the hohlraum and capsule radii (or, case-to-capsule ratio) is the same value used here (⁠ $∼$ 3).^2,36 As a consequence, when the capsule is driven directly by pulses P2 to P7, the laser can be absorbed in a large volume of underdense plasma. This approach not only benefits from the coupling efficiency of direct drive for most of the laser pulse but also ensures that the heat flux to the ablation front will always be smoothed by thermal conduction. To provide insight(s), we find it useful to introduce a model that can generate quick estimates but is also insensitive to many details. We start by assuming that electron transport is predominantly diffusive when averaged over a single pulse (say, P2) and treat the conservation of energy as a Laplace equation. This formulation can be exact in some circumstances,³⁷ but mainly, we seek a solution that is easy to address in 3D. If $∇·(k∇T)=0$ and $k∼T5/2$⁠, then $∇2(T7/2)=0$⁠. The region of interest is assumed to be spherical (⁠ $≈$⁠) and extend from an inner radius of $r1$ (the ablation front) to an outer radius of $r2$ (near critical density). Finally, we assert that the temperature at $r1$ is small compared to the corona. The solution of the Laplace equation is known, and it is easy to show that asymmetries in heat flux at $r1$ can be smoothed relative to $r2$ by a factor of $∝(r1/r2)ℓ+1$ at a Legendre mode number of $ℓ$⁠. In detail, the resulting smoothing factor $S12$ is $(2ℓ+1)(r1/r2)ℓ+1/[ℓ+(ℓ+1)(r1/r2)2ℓ+1]$⁠, as shown in Fig. 3. We might have liked to account for advection, but this is unnecessary given the benefits of thermal conduction. For even simple implementations of hybrid drive, we find that $r1/r2$ will be $≤$0.9 and will smooth imperfections in modes $≥$50 by a factor of-order hundreds. To put this in perspective, the conventional direct drive target has no protection when the laser is turned on, and variations in the drive will be immediately imprinted on the ablation front (Refs. 38 and 39). In all simulations to-date, the hybrid approach has been found to significantly reduce the risk of hydrodynamic instabilities, as shown in Fig. 4. Furthermore, the standard direct drive target starts with no plasma around the target and is very cold, and this results in an initial transient that is very uncertain. This scenario is completely avoided when using hybrid drive. In any case, the HDD could ultimately allow for implosions with lower adiabats and might also be combined with other types of beam smoothing. A related method is to coat the capsule with a layer of metal and generate a thin plasma atmosphere with a laser prepulse.^23–28 In the future, a hybrid target could also use a thick subcritical outer layer to tune absorption further.^40–42 

The smoothing model also predicts benefits to mid modes (⁠ $ℓ∼$ tens) whenever $r1/r2$ is less than one. As a consequence, the hybrid implosion will be symmetric (or has the potential to be) with a scheme for tuning lower modes. Our calculations show that this can be done with a limited number of beams when the incident energy has a specific time-dependent profile vs radius. Normally, a two-sided laser would not be compatible with direct drive. The incoming beams would be normal to the capsule at its waist and would not be absorbed over a large solid angle. By contrast, the hybrid target has a plasma atmosphere that absorbs the driver volumetrically (as described above) and by also having laser spots that zoom with the implosion. To explain, we start by assuming that laser deposition must be uniform and spherical at any given time from radius $r2$ to $r3$⁠, with $r1<r2<r3$⁠. Two opposed beams can deliver the same energy profile vs impact parameter, r, but the incoming intensity vs radius $I(r)$ must be proportional to the plasma chord that needs to be heated. As a simple function of geometry, for a ray with impact parameter $r<r2$⁠, this requires $I(r)∼(r32–r2)1/2−(r22–r2)1/2$⁠. If $r≥r2$⁠, this requires $I(r)∼(r32–r2)1/2$⁠. For common values of $r2$ and $r3$⁠, we find that the intensity profile must be ring-peaked and have a maximum-to-minimum intensity of $∼$ 2:1. (Of course, this energy will also be smoothed by thermal conduction.) This solution is similar to findings by Schmitt⁴⁶ and shows there is more than one way to generate a symmetric implosion. In detail, the ideal profile will also depend on refraction, material properties, and the trajectory of the capsule vs time (see Fig. 1). These issues must be accounted for in 2D calculations (as we will show) or iterative tuning (in experiments). Implosions using conventional forms of indirect drive are subject to similar issues but can overcome variations in flux vs time of-order 20%.^5,29,30 The hybrid target will also require tuning but should achieve adequate levels of implosion symmetry, as shown in Figs. 5 and 6. To increase the likelihood of ignition further, the HDD target will also be overdriven to provide “margin”, which we will also discuss.

To provide additional control of low-mode symmetry, mostly as a corrective tool, the hybrid target can also have a fuel layer that is modestly thinner near the hohlraum wall. This type of asymmetry is commonly called a “shim”^47–52 and can be fabricated with a variety of techniques.^53,54 To the first order, a shim is used to add or remove mass where the laser deposition is high or low. The main purpose is to make the acceleration history of the capsule more uniform and improve the symmetry of the final assembly. To provide an example, we can consider an asymmetry in the incoming laser of-order 10%, most of which is in Legendre mode 2. If the ablation pressure is assumed to scale with the local intensity as $I1/2$⁠, then an asymmetry in the mass of-order 5% could be expected to compensate. Variations like these can be imposed by wetting a foam with the required shape or by adding a heat source that shifts isotherms within the capsule. For reasons that were outlined previously, this approach could prove useful for tuning modes $≤$ 4. The use of shims has been tested in both calculations and experiments rather extensively⁵⁵ and have been found to correct asymmetries in drive as high as 2:1.

The geometry of the hybrid target and laser system has another novel and somewhat unexpected benefit. For example, in most schemes for laser ICF, the energy directed at the target can be redistributed by time-dependent cross-beam energy transfer (CBET).⁵⁶ These phenomena are caused by the seeded Brillouin scattering and effectively generate gratings in plasma density that can redirect laser light and reduce predictability and have also been known to reduce laser–target coupling. CBET can be particularly troublesome if many beams overlap at high intensities in the low-density corona and are partly directed along local plasma flows. By virtue of the two-sided geometry and laser spots that get smaller with time (see Fig. 1), the hybrid target does not excite common resonances, and little or no CBET is expected. In net, a larg(er) fraction of the incoming laser is coupled to the target, and this helps to increase efficiency.⁵⁷ 

Finally, we find that the hybrid approach is particularly complementary to excimer lasers and any strategy based on scale. If laser light at wavelength $λ$ is incident on a planar plasma at an angle of $θ$⁠, the absorption fraction $AL$ is known to implicitly increase as $1−exp(−IL*/ALIL)$ for $IL*∼ZL cos (θ)5/λ4$ (Refs. 3 and 58). The atomic number in the corona is given by Z, and its density gradient length scale by L. Relative to data using the OMEGA laser, the hybrid has values for Z, L, and $λ$ that differ by factors of-order 2, 5, and 0.7, respectively. The metric $IL*$ is increased by a factor of $∼$ 40, and this leads to high levels of laser absorption (97%) and in-flight kinetic energy (300 kJ/4 MJ $≈$ 8%). For related reasons, we do not expect convective stimulated Raman or a two-plasmon–decay to pose undue problems. Most metrics for LPI risk (i.e., gain) can be assumed to increase as $L(Iλm)n$⁠.^59,60 Depending on the gain per speckle and plasma damping, respectively, the exponent m will have values from 2 to 3, and n will be 1/2 to 1. Tests of direct drive on the NIF with 5 mm capsules are the closest surrogates to the HDD at present and tend to report little LPI (⁠ $∼$ several %) at intensities of-order $1015$ W/cm² (Refs. 57, 61–63). The HDD target will have a similar value for L and I, and the wavelength $λ$ is reduced from 351 nm to 248 nm. LPI risks have the potential to decrease relative to the NIF, but as these systems are very different, this type of extrapolation is highly uncertain. Historically, the benefits of short(er) wavelength light have always been found to be significant. To prepare for experiments with KrF lasers at larger scales, a series of tests are in planning for existing facilities. We are also making plans to investigate the potential of ArF at 193 nm (Ref. 64), and other novel concepts for mitigating risk (Ref. 65). Changes to the ablator, including the use of materials at higher atomic number might be expected to reduce LPI at a small cost in ablation efficiency. To reduce the importance of hot electrons produced by Raman or two-plasmon–decay instabilities at temperatures $∼$ 50 to 60 keV⁶⁶—if preheat were to be a problem—the hybrid implosion is also thick (as a simple function of size) and starts at an already-high adiabat.

The dimensions of the hybrid hohlraum are provided in Fig. 2. The CD capsule has an outer radius of 2185  $μ$m and a thickness of 40  $μ$m. To increase robustness, and the potential for reaching higher fusion gains, this radius is a factor of 2 larger than experiments on the NIF. The fusion fuel is comprised of a DT wetted CD foam developed at General Atomics,⁶⁷ having a thickness of 270  $μ$m, and a total density of 0.30 g/cm³ = 0.25 g/cm³ of DT and 0.05 g/cm³ of CD. The carbon in the foam absorbs x-rays from the corona and would appear to increase adiabat shaping and stability.^68–70 A wetted foam should make high shot rates more practicable, as DT can be wicked into the porous foam as a liquid, and in a manner that simplifies layering and reduces the inventory of tritium.^71,72 Targets of this type might leverage other novelties in design, as discussed in Refs. 15, 73–75 but these have not been exploited here. The density of the plastic capsule is 1.09 g/cm³. The vapor at the center of the target is assumed to be in thermal equilibrium with the layer and to have D and T in equal ratios at a density of 0.3 mg/cm³. The mass of the DT fuel at stagnation is 3 mg and intentionally exceeds contemporary experiments by more than a factor of 10.

To demonstrate the HDD target, we have made use of the radiation-hydrodynamic code DRACO,^43,44 which is an advanced tool for designing and analyzing experiments in ICF. The mesh in these calculations is 2D, and the laser ray trace is 3D. The deposition of laser energy is a function of inverse bremsstrahlung absorption and has been averaged in azimuth. First-principles models are used for all material properties,⁷⁶ and the DT-wetted foam is assumed to behave as if it were uniform in composition, as found in prior studies.^77,78 These simulations include refraction, CBET,⁷⁹ and nonlocal heat conduction⁸⁰ and have been validated against a large number of focused experiments. DRACO has previously been used to predict and analyze experiments on laser imprinting with good success⁸¹ and provides a strict test of the design concepts introduced above. Consistent with all existing targets, the calculations in this paper include physical imperfections and flaws, but future work will consider all of the specifications that could plausibly relate to ignition and burn.⁸² 

To estimate the initial x-ray impulse, we have employed a view-factor model that only requires the initial geometry of the target.^2,12,16 The radiation source is introduced at the boundary of DRACO and in the same manner as high-resolution calculations for indirect drive.^10,51 Very conveniently, the albedo for Pb has been measured and is known to be very similar to Au.⁸³ To ensure the initial x-ray drive is highly symmetric, the first laser pulse (P1) has been designed to intersect thin baffles at the zeros of the Legendre mode P₄. The geometry of the system has also been tuned to minimize mode 2, and all higher modes are smoothed by radiation.² [The waist of the capsule sees the warm high-Z wall, whereas the pole sees the cold hohlraum entrance hole, which is balanced with a bright ring of illumination coming from the baffles.] In comparison to indirect drive experiments on the NIF, which have a single and fixed laser pointing, the initial impulse for the HDD is easy to design, as it is completely separable from the rest of the pulse. The inner diameter and location of the baffles can be adjusted, as well as their intersection with pulse P1. As the baffles and liner are thin, and do not need to be optimized for the later generation of x-rays, the resulting waste stream is also more manageable. Simulations of high-Z materials will be discussed in the work to follow and use codes specific to that purpose.

To optimize the hybrid target in simulations, we note it was convenient to fire and optimize pulse P1 individually, then P1 followed by P2, etc., to cumulatively build the drive. To find the best $I(r)$ per pulse, we simply iterated through various candidates and interpolated for a round hotspot. The need for tuning is common to indirect and direct drives and should be aided by facilities having high shot rates. For all of the results that follow, approximately 50 calculations were required per pulse (P2 to P7) for a total of 300 simulations in 2D. These calculations show that the HDD can implode symmetrically with ring-peaked laser profiles, and also be very stable, although the design is nonoptimal and will be improved in future work. Design variables including the DT adiabat were not discovered by a process but were instead selected to be conservative. These values will be quoted and discussed relative to contemporary data, but again, no scan or study has been done. Given the amount of energy coupled to the target, it would not be difficult to simulate higher gains, but that is not the purpose here. For similar reasons, we have not included a shim, but only to emphasize that it is not a requirement. If we did use a shim, it can be assumed that the incoming laser profiles would be more uniform. Other than demonstrate the overall approach and tune the laser profiles as discussed above, the hybrid design is otherwise unrefined. Similarly, we have yet to investigate a number of other features that could be useful. Obvious examples include an absorbing or “Saturn” ring⁸⁴ and graded capsule dopants. If aspects of refraction or absorption were to be problematic—or we were to pursue higher levels of performance—a large number of options remain.

The hotspot for the hybrid target is shown in Fig. 6, and performance metrics are provided in Table I. By design, the hybrid implosion reaches stagnation pressures similar to those on the NIF⁵ and exceeds the Lawson criterion^18,19 and metrics for alpha heating⁸⁵ by more than a factor of 2. The HDD target readily ignites and has a prompt thermonuclear yield of 256 MJ, which results in a gain of 65. Efficiency and scale have significant benefits and can be expected to relax requirements on symmetry and other concerns (see Figs. 5 and 6). These findings are intentionally at an in-flight adiabat of 6, which is understood to lower fusion performance but is also meant to reduce requirements on the target. In general, experiments in this regime approach expectations on both the OMEGA and the NIF.^12–14,16 The in-flight aspect ratio is 20 at a convergence ratio of 1.5, and this also compares favorably with contemporary data. Findings are consistent with NIF implosions that ignite, since the imploded mass and areal density for the HDD are greater by a factor of 12 and 3, respectively. Assuming the hybrid-type target can be tuned, it has the potential to meet requirements for IFE even if implosions have to operate at relatively high adiabats. It is also possible that the thermonuclear yield would be suitable for applications in national security (i.e., stockpile stewardship). A strategy for increasing the gain might include an intermediate adiabat or a target with increased mass and laser energy. If we use the formula provided in the introduction, with $G∼ηhydro(ρRinitial)2/3(EIN/mDT)−1(EKE/EIN)−1/3$⁠, we would expect an implosion at adiabat 3 and twice the $ρR$ to have prompt gains $∼65×(1)(2)2/3(0.5)−1(1)−1/3≈200$ at laser energies $∼$ 8 MJ. These figures exclude the production of energy in the surrounding cooling blanket, which is designed to boost the net gain by an additional factor of 1.2–1.3.

It is also possible for the gain to be further increased if the peak laser power and energy can be reduced. This strategy could seem counterintuitive, or even counterproductive, but should be expected for implosions at such large scales. The hybrid target has been designed to meet the criteria for ignition while still imploding, as implied in Table I. In other words, there are tradeoffs that might still be made in gain and the probability of ignition. At present, the hybrid target reproduces the ratio of kinetic to internal energy of experiments on the NIF despite an increase in the scale of $121/3≈2.3$⁠. The HDD ignites at a convergence ratio (CR) of $∼$ 20 when it might otherwise reach 30. If we assume the compression of the hotspot is adiabatic, and its internal energy $∼$ CR², then the work available to the hotspot should exceed requirements by approximately the same factor (2.3). This figure is one way to quantify margin and the degree to which the target has been overdriven. In calculations at reduced energies, the peak velocity and laser power are found to exceed the threshold to ignite by a factor of 1.27 and 1.54, respectively. If we were to make another comparison with the NIF, this would be equivalent to taking a diamond target that ignites at 2 MJ (Ref. 5) but instead applying 3 MJ. The hybrid target has been designed to be robust and to also be capable of very high gains. Future efforts will investigate requirements for ignition and burn and include tolerances on the capsule, its smoothness, the thickness of various layers, and other imperfections in the incoming laser. Preliminary calculations are very promising and suggest specifications on the capsule can be greatly relaxed. We will also consider adaptations for the OMEGA and the NIF, and several alternatives including fast ignition⁸⁶ and shock ignition.⁸⁷ Finally, as part of the milestone-based fusion development program, we report that tests of the Xcimer laser system have just begun and will proceed in parallel.¹⁷ 

This work was made possible by colleagues at LLE, LLNL, LANL, General Atomics, and Xcimer and the encouragement and support of C. Deeney, S. P. Regan, V. N. Goncharov, and T. J. B. Collins. This material is based upon work supported by the Department of Energy National Nuclear Security Administration under Awards No. DE-NA0004144 and No. DE-NA0003856, the University of Rochester, and the New York State Energy Research and Development Authority. This work was also supported by an INFUSE award with project number RFA2022a-61. The support of DOE does not constitute an endorsement of the views expressed in this paper. This report was prepared as an account of work sponsored by an agency of the U.S. Government. Neither the U.S. Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the U.S. Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the U.S. Government or any agency thereof.

The authors have no conflicts to disclose.

C. A. Thomas: Conceptualization (lead); Funding acquisition (lead); Writing – original draft (lead); Writing – review & editing (lead). M. Tabak: Conceptualization (equal); Writing – review & editing (equal). N. B. Alexander: Conceptualization (equal); Writing – review & editing (supporting). C. D. Galloway: Conceptualization (equal); Funding acquisition (lead); Writing – review & editing (supporting). E. M. Campbell: Conceptualization (supporting); Writing – review & editing (supporting). M. P. Farrell: Conceptualization (supporting); Writing – review & editing (supporting). J. L. Kline: Conceptualization (supporting); Writing – review & editing (supporting). D. S. Montgomery: Validation (supporting); Writing – review & editing (supporting). M. J. Schmitt: Conceptualization (supporting); Writing – review & editing (supporting). A. R. Christopherson: Conceptualization (supporting); Writing – review & editing (supporting). A. Valys: Conceptualization (supporting); Writing – review & editing (supporting).

The data that support the findings of this study are available from the corresponding author upon reasonable request.