A prospectus on laser-driven inertial fusion as an energy source

Producing energy by fusion has been a grand challenge for more than 60 years (Post, 1956). Many people (including this author) were attracted to this field because of the potential for harnessing fusion for producing electricity. Recent developments make fusion energy closer than ever before—although there are still challenges to be overcome in terms of both physics and technology needed. In this prospectus, we will discuss some recent results, lessons learned, and challenges still ahead for inertial fusion. It is important that we keep these lessons learned in mind as we move forward toward inertial fusion energy—particularly as we try to accelerate our progress.

Shortly after the invention of the laser, the concept of using high powered lasers to compress and heat deuterium-tritium (DT) fuel to fusion conditions was proposed (Nuckolls , 1972). This technique, called inertial confinement fusion (ICF), has been pursued as an alternative to magnetic confinement fusion (MFE). Recently, the first burning plasma (Kritcher , 2022a; and Zylstra , 2022a), the first plasma to ignite by the Lawson criteria (Abu-Shawareb , 2024;, Kritcher , 2022b; and Zylstra , 2022b), and the first target gain > 1 (target gain = fusion energy produced divided by laser energy into the target) (Abu-Shawareb , 2024) have been achieved using inertial confinement on the National Ignition Facility (NIF) laser (Moses , 2016) located at Lawrence Livermore National Laboratory. These achievements demonstrate the physics proof-of-principle for inertial fusion energy (IFE); we can ignite and produce propagating burn an inertial fusion capsule and produce more energy than the laser energy that we put into the target.

In the inertial confinement concept, the DT fuel is frozen into a solid layer inside a small, spherical capsule, made of a low atomic number “ablator” material such as plastic (CH), beryllium (Be), or high-density carbon (HDC). The outside of the capsule ablator is either heated directly by high power laser beams (“direct drive”) (Craxton , 2015) or by x-rays generated by high power laser beams or beams of ions (“indirect drive”) (Lindl, 1995). The outside of the capsule is rapidly heated and ablates outward—causing the inner part of the capsule, including the DT fuel, to be accelerated inward like a spherical rocket. The fuel reaches high velocity (∼400 km/s) and stagnates at the center—converting the kinetic energy to PdV work (pressure times change in volume) which heats a central hotspot. If the central hotspot reaches fusion conditions, then, fusion reactions occur (Atzeni and Meyer-ter-Vehn, 2004). One of the products of the fusion reaction is an energetic alpha particle (the other being an energetic neutron); if the areal density of the stagnated core is high enough, the alpha particle stops inside the hotspot and heats it further. If the conditions are right, the alpha particle heating causes additional fusion reactions to occur and results in a thermal instability which causes a burn wave to propagate through the compressed fuel layer. The fusion reactions stop when the shell explodes outward and cools.

One of the advantages of the inertial confinement concept for generating electricity is that the fusion system (the target which contains the DT fuel) is physically separated from the driver (a laser, in this case). For a fusion power plant, the neutrons generated by the fusion reaction are captured in a “blanket” and their energy converted to electricity via a conventional steam cycle. The blanket is also used to breed tritium from lithium to produce fuel for subsequent targets. Because the target is separated from the laser, it is easier to protect the laser from neutron damage—although the final optics are at risk. In addition, the amount of energy produced from each “target shot” can be controlled via different target designs to meet the electrical needs of the consumer.

In this paper, we will discuss the NIF results and what we have learned from the journey to ignition and energy gain on NIF. We will then discuss what is needed to go from the NIF result to inertial fusion energy and how the lesson's learned from NIF can help us achieve fusion energy. Finally, we will discuss the role of public/private collaborations in inertial fusion energy.

The NIF began target physics experiments in 2009, with the first layered experiments done in late 2010. NIF ICF experiments are primarily focused on indirect drive hotspot ignition, using a laser driven hohlraum (cylindrical cavity about 1 cm long) made of high atomic number material (such as gold or depleted uranium) to create an x-ray bath that drives the implosion of the DT filled capsule in the center of the hohlraum, as shown in Fig. 1. 96 laser beams enter the hohlraum through a “laser entrance hole” from the top of the chamber and 96 beams enter from the bottom. The laser pulse shape is designed to launch several shock waves through the capsule and gently accelerate the capsule while keeping the adiabat (or entropy) relatively low to allow for high compression of the fuel.

The two most important factors (largest exponents) are the implosion velocity (⁠ $vimp$⁠) and the size or scale of the capsule (S). This is because the kinetic energy of the implosion is the energy that is converted into hotspot energy when the implosion stagnates. Additionally, larger scale alters the surface to volume ratio in favor of more heating (going as volume) relative to electron conduction-limited cooling (related to surface area). Having a large amount of hotspot energy is key to getting ignition and yield. Of course, driving a large capsule or one with a high velocity requires more input energy—either from the laser itself or, in the case of indirect drive, from a more efficient hohlraum. Driving a larger capsule to high velocity (∼390 km/s) was key to getting ignition on NIF.

A second important factor in Eq. (1) is the inflight ablation pressure (⁠ $pif$⁠). This factor was not fully appreciated in early NIF experiments—however, it was found that keeping the drive (and hence, the ablation pressure) high until late in the implosion was important to getting high compression and neutron yield by increasing the mechanical power associated with the rate of conversion of kinetic energy into hotspot internal energy (Hurricane , 2017). Keeping the ablation pressure high is achieved by keeping the laser on until very late in the implosion and is often measured by the “coast time” which is the time between the end of the laser pulse and the time of peak compression. By keeping the “coast time” short, the drive on the capsule was kept high late into the implosion thereby delaying the time and decreasing the radius at which deceleration begins—essentially equivalent to keeping your foot on the gas and then rapidly braking your car. Again, the cost for keeping short coast time is additional input energy (at fixed power)—either through more laser energy or a more efficient hohlraum.

The neutron yield is inversely proportional to $αif$ the fuel adiabat inflight (related to the fuel entropy). This describes the compressibility of the fuel—lower adiabat allows higher compression and should, in theory and simulation, produce higher neutron yield. In experiments, this has not proven to follow the theoretical scaling—either for indirect drive (Clark , 2019; 2022) or direct drive (Goncharov , 2014) and work continues to try to develop a better understanding of high compression, low adiabat designs.

The final two factors in Eq. (1) describe degradation mechanisms that can prevent energy from coupling to the hotspot. The first, $[1−m.f.ηTτ]$⁠, where (m.f.) is the “emission mix fraction” defined by Pak (2020), as the radiative energy loss rate due to higher Z material relative to the radiative energy loss rate for pure DT, $η$ is a fitting parameter, T is temperature, and $τ$ is burn-width. This factor represents the degradation due to mixing of material with higher atomic number than hydrogen into the hotspot. If higher atomic number material—such as carbon from the ablator—is mixed into the hotspot, radiation losses are enhanced which cool the hotspot. This factor has been compared with data in the limit of low to moderate mix in reference (Pak , 2020).

The second factor, $1−RKEnorm$⁠, represents the loss of energy from the hotspot when there is incomplete conversion of kinetic energy (RKE = “residual kinetic energy”) into pdV work. Residual kinetic energy generally occurs when the implosion is asymmetric and parts of the implosion are still in motion at minimum volume (Kritcher , 2014). Asymmetry of the areal density of the shell of the implosion is the primary cause RKE, thus, it is important to minimize shell areal density variations (Hurricane , 2020; 2022).

Getting to ignition required control of all these parameters simultaneously. Several things frustrated progress toward ignition—control of hydrodynamic instabilities, control of low mode asymmetry, coupling enough energy into the capsule, and target quality. It was through a series of design iteration loops (see Fig. 2) that we would improve on the performance, then hit a limiter, identify that limiter, and redesign to mitigate the limiter that got us to ignition.

One of the advantages we have now that ignition has been achieved on NIF is a set of “lessons learned” that can help us accelerate progress toward fusion energy. Here, are some of the things that we learned from the NIF experience that can impact the strategy for accelerating inertial fusion energy.

The initial NIF point design (Haan , 2011), which is now often referred to as the “low foot” design, was intended to be a very high-performance design. As such, it pushed the target physics to (and beyond) the limit of our understanding. This resulted in several different things going wrong at the same time, which made it difficult to understand the results and fix the issues.

We started to make more rapid progress on performance when we backed off to a more conservative, lower performance design, called the “high foot” (Dittrich , 2014; and Park , 2014) which addressed one mode of failure—ablation front Rayleigh–Taylor instability. While the ideal, one-dimensional neutron yield of the high foot design was lower than that of the low foot design, the experimental results showed the opposite—the high foot outperformed the low foot.

Once a more understandable design had been established, we were able to navigate toward ignition by systematically identifying the issues that were limiting performance (Callahan , 2015; and Hurricane , 2014). We could then redesign the target or make improvements in the laser and target fabrication to fix these issues. While some felt that this methodology was too incremental, it kept us from getting lost because we never extrapolated too far; we could understand the gradients presented by the data, and, thus, could navigate a successful route to higher fusion performance. The extreme nonlinear nature of ignition made this incremental approach exceedingly important.

The lesson learned here is that extrapolating too far can result in getting lost in parameter space. Of course, how far is “too far” is a judgment call. Much like the tortoise and hare fable, while taking steps can appear slower, the experience of NIF suggests that it is the best way to navigate a very complex, multidimensional space to the ultimate goal.

While we knew going into the NIF experiments that the simulations would not be predictive enough to get to ignition, we assumed that the simulations would be close enough that small corrections to the simulations model, based on experiments, would adequate. Although the simulation codes are state of the art, they were further away from the data than anticipated and were often overly optimistic about what would ignite. At the end of the day, the experimental data were very important to find the correct direction to move to ignition. Post-shot simulations were eventually able to explain a significant number of experiments, but there are still mysteries even now.

One surprise was that simple models and physics based scalings were surprisingly good at explaining the data trends. Some examples are a model of the capsule as a set of pistons (Hurricane , 2020; 2022) and a model of the hohlraum based on the gold bubble dynamics (Callahan , 2018). The 2023 Ronald C. Davidson award was given for the publication of the gold bubble dynamics paper and this model was used heavily as a guide to controlling implosion asymmetry in the recent designs—including the designs that achieved ignition and gain.

While ICF targets are complex, multi-physics systems, if the dominant/key physics can be identified, then the data may be explained by very simple models. Moreover, these simple physics models can focus further design work because they identify the key parameters that must be adjusted or controlled.

The lesson learned here is that we need to rely on the three legs—experimental data, simulations, and simple models/theory. Getting enough experimental data to be able to study trends, compare with simulations and simpler models is very important for making rapid progress. This was challenging on NIF because of the limited shot rate and number of experiments.

Because the simulations are not perfectly predictive, understanding the design space and taking advantage of the separation between the target and the laser is important. Since the target is replaced on every shot, the target design can be changed based on the data. (Once we reach an operating power plant, the design will need to be kept fixed, but that is not the case in the early stages of development). Developing an understanding of the tradeoffs in design space will aid in rapidly finding an optimized design in the multidimensional parameter space and is preferable to narrowly focusing on one design, for example.

The exciting results from NIF are really a “proof of principle” for inertial fusion energy. While these important results demonstrate the basic concept of igniting a target in the laboratory and generating more fusion energy than was put into the target, there is significant additional work needed to make a commercially attractive, practical source of fusion energy. Here, we will review some of the most important requirements needed for fusion energy.

The target gain must be high enough that the energy produced is significantly higher than is required to power the laser and the rest of the power plant. Otherwise, the power plant would only produce enough energy to power itself and leave nothing left to sell to the customers. Figure 3 shows a simple diagram of the flow of energy—laser energy is amplified by the target gain G and is then converted to electricity using a thermal cycle with efficiency $ηt$⁠. A fraction, f, of that is returned to run the power plant leaving (1 – f) to deliver to customers.

The electrical energy needed to run the laser is the laser energy divided by the laser efficiency $(EL/ηL$⁠) and the electrical energy produced by the power plant is (⁠ $ηt G EL$⁠). (Here, we assume that the laser is the largest consumer of electricity in the power plant—more detailed studies, such as in Meier (2009) suggest the laser accounts for 85%–90% of the electricity used by the plant.) The fraction of the electrical energy that goes back into the plant is $f ∼1/(ηtηLG)$⁠. So, if we want to limit the fraction of energy going back into the plant to be less than f, then, $ηLG ∼ 1/(fηt)$⁠. For a 25% recirculating power fraction (f) and thermal efficiency (⁠ $ηt$⁠) of 40%, then the product of the target gain and the laser efficiency needs to be greater than 10. This means that for a 10% efficient laser, we need gain of 100 or more. This is considerably higher gain than currently achieved 1.5 on the NIF laser.

How do we get from gain 1.5 on NIF to gain ∼100 needed for a fusion power plant? While there may be several approaches, one approach is to move to a directly driven capsule rather than the hohlraum targets used on NIF. In a hohlraum target, the energy absorbed by the capsule is only about 10%–15% of the laser energy; the “capsule gain” (fusion energy produced/capsule absorbed energy) for NIF shot N221204, is 12 while the target gain is 1.5 (Abu-Shawareb , 2024). So removing the hohlraum and directly illuminating the capsule will increase the gain for the same laser energy by a factor of 5 or more.

Driving a thicker DT fuel layer increases the fusion gain in two ways. First, it can increase the areal density and burn faction, as discussed previously. Second is that there is simply more fuel to burn. For high gain, we would like to drive a thicker (more massive) fuel layer without increasing the laser energy—this can be achieved by driving a more massive layer with lower implosion velocity. For example, if we drive the fuel to an implosion velocity of 200 km/s rather than 400 km/s on NIF, then we can have 4 times the amount of fuel for the same total kinetic energy. With these slow implosions, we need an additional method to ignite the target, however.

The company that this author works for is pursuing proton fast ignition (Roth , 2001) and shock ignition (Betti , 2007) as methods to ignite the target for high gain with a slow, high DT mass target. In proton fast ignition, the compressed fuel is ignited by a beam of protons created by a short pulse laser. In shock ignition, the fuel is ignited by launching a strong shock after the fuel is compressed. Following the lessons learned on NIF, we are keeping the design space open to both approaches because it is unclear which will be optimal. Both are predicted to achieve high gain (>100) with moderate laser energy (∼1–1.5 MJ), but we still need to demonstrate this in the laboratory.

In addition to higher gain, a power plant will need to fire at a high repetition rate to produce the desired power plant output. In general, we will need a repetition rate of 10–15 Hz to support a power plant in the 1000 MWe range, depending on the fusion yield per target.

This means that we need high repetition rate, high efficiency lasers (see e.g., Bayramian , 2011 or Obenschain , 2020), are needed. One advantage of developing these high repetition rate lasers is that a large number of target physics experiments can be done in a relatively short time. One of our lesson's learned was that we need experimental data to help us navigate toward an optimal design. Having even one shot every few minutes will allow us to rapidly explore parameter space experimentally and find an optimal design.

Having high repetition rate lasers will also require high repetition rate diagnostics and automated analysis. Being able to do many experiments will also require a different approach to an experimental campaign. On NIF, with a limited number of shots per experimental campaign, each shot is heavily planned, simulated, and diagnosed. With a higher repetition rate, things like parameter scans can easily be done as a way to understand the design space. We anticipate that machine learning will also play a larger role when we have hundreds or thousands of experiments for a given design.

Having high repetition rate also means that we need to produce a high volume of low-cost targets produced each day, and automated system to get the targets into the chamber will also be needed. For a fusion reactor, we will need to inject the targets into the chamber. Previous IFE studies (Latkowski , 2011; Perlado , 2011; Petzoldt, 1998; and Schultz , 2001) have studied these issues; however, we will need to address these for a future fusion power plant.

So, while the proof of principle for inertial fusion has been established on NIF, there is still work to be done to produce designs that meet all the requirements for a fusion power plant. The results and lessons learned from NIF can be used to help accelerate the progress toward this goal.

Much of the research in fusion energy has historically been carried out in universities and National Laboratories funded by public money. In recent years, several fusion companies, funded by a combination of private investment and public money, have been established with the goal of getting to a fusion pilot plant as quickly as possible. Private companies and public institutions like National Labs and universities each bring different elements to advancing fusion research; like many others in our field, we feel that having a partnership between the publicly funded institutions and the private companies is the best way to get fusion energy on the grid quickly.

In inertial fusion, much of the research has been funded in the US by the National Nuclear Security Agency (NNSA) within the Department of Energy. The mission for this work is not energy, but national security. However, the scientific results from the NNSA work—from NIF and other facilities—has relevance to inertial fusion energy. The National Labs and universities have expertise in inertial fusion, as well as simulation codes/tools and experimental diagnostics in addition to facilities. Other National Labs have expertise in handling tritium or in materials for fusion blankets needed in a reactor.

The private companies have the opportunity to bring in private investments and can move more quickly—as taking technical risk is part of the DNA of startup companies. This should allow progress to be made on a shorter timescale and with less overhead—accelerating the pace of progress in fusion energy.

Partnerships between public institutions and private companies seems to be an ideal way to make progress. The public institutions bring decades of experience, knowledge, and tools to the table while the private companies bring the startup culture to move quickly, while accepting more technical risk for the sake of rapid progress. The US Department of Energy recognizes the benefits of this type of partnering and has setup several funding mechanisms to encourage these public/private partnerships.

In addition to technical progress, these public/private partnerships should be used strengthen and grow the workforce in fusion. Having private fusion companies opens additional training opportunities and jobs for people in fusion. Having staff move from public to private and vice versa also brings new perspectives into both our private companies and our public institutions. Getting those fresh ideas and perspectives brings a diversity of thought that will benefit fusion as a whole.

This is an incredibly exciting time for fusion. With the recent NIF results, the proof of principle for the target physics for inertial fusion is demonstrated. Getting ignition has been a grand challenge problem more than 50 years in the making and we can now say, that we have achieved that goal.

The next challenge is to take that result and build on it to turn fusion into a clean source of energy. While ignition has been demonstrated, many challenges remain—target gain of 100 or more; high repetition rate (10 Hz), high efficiency lasers (10%–15% efficient); rapid, inexpensive target manufacturing; target injection into the chamber; and a reactor chamber that can survive the neutron flux.

Oftentimes, the history gets lost because the “corporate knowledge” is not passed on or documented properly. Even over the 15 years of NIF experiments, results get forgotten—particularly when they are negative results. This author believes that we need to capture what we learned about the strategy that worked to get to ignition so that we can use that learning in making inertial fusion energy a reality.

While there are still challenges, fusion energy is much closer than ever before. The excitement in the community and new opportunities in public and private sector are better than ever. This means it is also a time where we will be expanding our community and growing our workforce. That gives us an opportunity to think about diversity, equity, inclusion, and accessibility, to question our processes and procedures, and to educate both our future workforce and the general public. This is a great time to be in fusion!

Debra Callahan: Conceptualization (equal); Writing – original draft (equal).

The authors have no conflicts to disclose.

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