> Once it becomes self sustaining one should be able to add more fuel to the pellet to get more energy out for the same input energy.

That's not how ICF works. Plasma, being a gas-like state, will always expand to fill whatever volume is presented. With ignition, the rate of expansion is essentially lower than the rate of fusion, allowing you to fuse all of the fuel before the plasma dissipates and cools down.

In ICF as studied at NIF, you start with an extremely precisely machined piece of metal called a hohlraum, you put a solid pellet of fuel inside at an extremely precise location, then fire a laser with extremely precise alignment to heat the hohlraum until it generates X-Rays that heat the pellet just right so that its outer layer explodes, creating an equal implosion, generating two shockwaves inside the pellet; if the two shockwaves meet just right, at the center of their meeting place you get a fusion reaction, and you hope that that fusion reaction has enough time to heat up and cause more fusion reactions before the initial implosion loses speed and expansion happens.

That initial shock is the only thing containing the plasma - once it has lost its velocity, the plasma dissipates and cools down. If ignition was reached, the gas that cools down and dissipates should be 100% He, instead of a mix of He, D and T. However, there is no way to stop this dissipation, it is a fundamental part of ICF.

The only way to keep an ICF reactor going is to shoot one laser burst at one pellet, capture the energy of the fusion, and use that to power the next laser burst fired at the next pellet.

Of course, after each burst of laser heating the hohlraum so much that it radiates the heat as X rays, and then briefly containing a 1-10M kelvin burst of hot plasma, plus a neutron bombardment, the hohlraum is destroyed. Since machining the hohlraum to the precise shape required to achieve the shockwaves discussed above is never going to be a cheap process, it is impossible to imagine ICF would ever be even a tiny bit close to economical, even if it could in principle output more energy than it requires as input.

As such, ICF is strictly a scientific pursuit, mostly interesting for nuclear weapons research.

This report found that ICF could reach LCOE as low as $25/MWh "with optimistic but not obviously unrealistic inputs."[0] This does require hohlraums cost about $2 each and are fired every 20 seconds. With mass production and process optimization it may not be ridiculous to reduce hohlaum cost to this amount. However, the yield is about 5 gigajoules which is equivalent to about 1 ton of TNT.

Making equipment that can handle 1 ton of TNT exploding every 20 seconds is an interesting engineering challenge.

[0]https://royalsocietypublishing.org/doi/10.1098/rsta.2020.005...

"Optimistic but not obviously unrealistic inputs" include reducing the cost of hohlraums from the million dollar range to 10$ (not even sure if that accounts for the price of the gold itself), a reactor capable of resisting 50 million pulses before needing replacement, and a few others.

It also considers the price of a fusion power plant to be less than that of a fission power plant, based entirely on the observation that it would have less stringent safety requirements.

Overall this article may be right in principle if taken to refer to an arbitrarily far away future (hundreds of years away at least, if ITER and DEMO are to be taken as realistic examples of the pace of improvement of fusion power in general, even if they are MCF instead of ICF).

This paper was not well received on reddit.

https://www.reddit.com/r/fusion/comments/ja879n/nuclear_fusi...

Don’t you still need to spend energy on confinement? You don’t need to “reignite” the plasma but w/e confinement solution you chosen still has a cost and a non marginal one when it comes to magnetic confinement.

this uses inertial confinement rather than magnetic confinement.

You still need big ass lasers or particle accelerators for ICF too, these tend to be quite energy intensive too.

Yes, with ICF you do need to constantly reignite the plasma.

What is the goal of NIF? I've read repeatedly that fusion power isn't their end goal but rather to study inertial confinement. That's fine but why study inertial confinement if not to generate power? I've always been very confused about their goal. I'm a total layman when it comes to this stuff so there's some nuance I'm not understanding. Appreciate any clarification anyone can give.

> What is the goal of NIF?

Nuclear weapons, more specifically stockpile stewardship (what happens as weapons age) and verification of weapons codes/simulation software (can we make new weapons without full-scale testing).

Everything else is gravy. There's a reason it's at one of the weapons labs (vs. the unclassified work done at most other national laboratories).

The thing you are missing is that in addition to fusion power research (which is valuable and NIF has made major contributions to) there is also fusion weapons research. Inertal confinement is (kinda) close to the conditions inside a fusion bomb, and NIF also has a mandate to research those conditions. For that kind of research, a single pulse of fusion ignition is exactly the kind of data they need. Since we have a nuclear weapon test ban, and computer simulations need some kind of ground truth to be calibrated against, achieving fusion ignition in a lab is valuable to NIF for that reason alone.

In terms of the NIF's broader goal, as opposed to the specific goals for their ICF work, the NIF is meant to keep nuclear physicists fresh on research relevant to nuclear weapons design in the aftermath of the end of the Cold War and Comprehensive Nuclear Test Ban Treaty. [1]

[1] https://en.wikipedia.org/wiki/National_Ignition_Facility#NIF...

The NIF is a facility for conducting experiments. The goal for the field is fusion power, and these experiments may wind up contributing toward it, but it will never be anything more than a stepping stone. The primary purpose of the NIF is validating computer models for simulating nuclear reactions. These models are used both for the design of nuclear weapons and nuclear reactors. They also develop technologies to support their activities, such as new sensors and laser control methods. Compare this program to say a mars rover where we don't expect the rover itself to do anything of great practical utility, but the lessons learned along the way have many potential applications both directly for future missions, and indirectly for spinoff technologies.