Hacker News Reader

Comment by nutate | Hacker News Reader

The concern over startup energy is like saying we can't have internal combustion engines until the engine can start itself with gasoline. Turns out we've gotten along fine starting cars with electricity for about 100 years, including a long while where we had to hand crank them ourselves. Surplus is just a matter of runtime once you're getting more than you put in to maintain it.

View on HN

In this case, it's not clear how the analogy could apply, because you can't realistically deliver more fuel to the "reactor" while it's "running". An ultrafast laser pulse hits a target and it undergoes fusion for a few nanoseconds, maybe less, generating a huge explosion. You can't operate a fuel injector on that timeframe. So you need a big gain in terms of efficiency.

You can hopefully make a bigger fuel pellet, but that kind of scaling hasn't been demonstrated and isn't guaranteed, because it begins to disperse as soon as fusion initiates in this inertial confinement scheme. So "just" a matter of runtime is harder than it might seem at first.

I'm curious if you could power boats like this, though. It might not be economical for electricity. But the power-to-weight ratio is probably pretty good.

You do not need a continuous stream injecting at every nano second...you can operate much slower than that. If you watched the stream, they said a few hertz is enough. A milisecond is an eternity compared to a nanosecond, you remove the neutrons away from the interaction site well before the next start of the rep-rate when you inject the next pellet / holhraum. You just need to have something to capture the neutrons and convert that into heat to boil water, which will accumulate over various repetitions of course (water boils much slower than a GHz), and that thing need not be where the pellets are blasted.

That said, I'd need to think about it the design, I just don't think it's impossible.

From the PDF (page 15):

https://suli.pppl.gov/2018/course/Ma.pdf

typical confinement time for ICF is on the order of a tenth of a nanosecond. I don't expect they have made a factor-of-millions improvement on this. I generally avoid watching videos whenever possible, but I think you are referring to the frequency at which the fuel pellets can be repeatedly ignited by a laser — there are no plans to use the output of one fuel pellet to directly ignite the next. In fact not even the "magneto-inertial" techniques with putative confinement times in the microseconds have a roadmap to achieve this.

You don't use the ignition of one pellet for the next...each pellet ignites itself.

It isn't a nuclear fission reaction where it is a chain reaction between pellets. Each pellet interaction produces energy, and you capture that energy. It is ignition for the pellet, not other pellets in the machine. The boiling of water thankfully happens on a much longer timescale, being accumulationf of energy of many pellets over a few cycles.

The pellets produce a net energy release larger than the energy used by the laser. This energy gain is called "ignition" in fusion parlance. A fuel pellet does not ignite itself — otherwise, they would be destroyed as soon as they are produced, like the critical ball in a nuclear warhead. That's what I was saying from the beginning: the laser does not function like the spark plug on a car engine; it is required for every ignition. It's pretty clear if you go back and read the analogy I initially responded to.

"Getting more than you put in to maintain it" is the whole point for all these net power discussions.

In this particular case, it's a quick laser pulse and an exploding fuel pellet, so there's no lengthy runtime possible.

The next step is rep-rate, which a lot of the field has neglected, while some people (tooting my horn, my lab) were amongst the few pushing it. The thing is people couldn't even achieve the result with one shot, but now that it seems feasible now, hopefully scientists (and more importantly funders) will start to care about increasing rep-rate.

Yes, although they need to still leverage this ignition/scientific-breakeven achievement to get the higher energy gains needed for a practical power plant. Not just 1.5x but 25-50x. With ignition now being repeatable, that should be a doable proposition.

So, one of the things that I probably shouldn't say is the NIF design is not really optimized for fusion specifically (I won't say too much but you can guess why). There are schemes out there that can be more efficient, for example, not even bothering to convert to UV and then x-rays (start at UV in fact, or crank up the intensity of light, the light in NIF is very intense but modern high intensity systems are orders of magnitude more intense, they are however dramatically lower in pulse energy). That plus using modern laser tech when you realize NIF was designed in the 80s, we can definitely make it much more efficient eventually.

Don’t need to be too coy about the fact that NIF was built to recreate the processes inside an H-bomb after underground nuclear testing was banned. They literally use the same concept of a “Hohlraum” (here in miniature) where heat and light (from a fission trigger in the case on an H-bomb, lasers in the case of NIF) convert to x-rays on the outer casing and evenly illuminate the fuel pellet casing which ablates a bit to drive implosion which triggers the actual compression heating of the fuel pellet. Heck, some previous tests even used a uranium tamper (because in part, uranium has high inertia due its density thus prolonging the period of high pressure), just like a real H-bomb.

I don't see why not. Requirements are to inject the fuel (in pellet form) and ignite it with a laser pulse. Right now I believe the pellet is contained in an hohlraum that is held in place in order to perform the test but you could drop (or shoot) that assembly into the chamber and time the laser pulse to hit it when it's in the right location. That doesn't sound like it's beyond the realm of physical possibility.

Reply on news.ycombinator.com