I think you're overestimating the amount of energy released in a single fusion event and underestimating the machinery needed to generate and focus the laser pulses on the hohlraum. Remember there is no fusion chain reaction, the thing would need to keep on pumping pellets into the "primer" while firing laser pulses to keep the energy flowing. Compared to a 'simple' fission primer with a neutron reflector I don't really see the gain in a laser-primed fusion weapon, at least not one in the line of current fusion bombs.
Actually, the tech that you are hinting at would guarantee clean fusion power. If you could generate clean fusion explosions, detonate underground and use the heat to generate power. Nuclear bombs are currently the only net energy positive fusion devices we have.
Most bomb designs are fission-fusion-fission - most of the energy comes from the fission of the casing of the secondary by neutrons from the fusion process in the secondary.
Edit: Mind you - there were tests like the huge Tsar Bomba that de-rated at 100+ Mt design to 50Mt be leaving out the final fission stage - the output being mostly from fusion.
My understanding is that what you're referring to as a "fusion" stage is really a fusion-enhanced fission stage. All the stages of normal "H-bombs" are fission stages, using a small fusion core to generate a flood of neutrons, which results in the fission fuel combusting more completely.
Pure fission bombs tend to disintegrate long before enough neutrons have been generated to achieve complete combustion. The inclusion of a small core of lithium deuteride results in a much more efficient fission bomb.
I suppose these NIF tests are tiny pure fusion explosions; but I'm not aware of any practical weapon that generated energy mainly from fusion.
The primary is fusion boosted fission, the core ("spark-plug") of the secondary is also fusion boosted fission surrounded by dry fusion fuel (lithium deuteride) and then a fissionable tamper.
My understanding is that X rays from the primary compress the entire secondary package - which ignites the spark-plug and then the fusion fuel is caught between the incoming tamper and the exploding spark-plug at its core and ignited.
So there does appear to be a real "fusion" stage - with a lot of other steps involving fission and fusion.
"Fast fission of the tamper and radiation case is the main contribution to the total yield and is the dominant process that produces radioactive fission product fallout."
The way I read that, even in a large H-bomb, little of the yield is the direct result of fusion reactions. Rather, the neutrons produced by the fusion reaction dramatically increase the efficiency of the fission reactions, which on their own would not produce enough neutrons to fission more than a small part of the fissile material before the device disintegrates.
There is still a considerable about directly from fusion e.g. in the Tsar Bomba the fusion stage produced about 47Mt of the 50Mt total (it had two smaller H-bombs as primaries for the fusion stage). The full design was 100+ Mt so about 50% from the final fission stage seems a rule of thumb, which is also mentioned in this article about "neutron bombs":
It might guarantee clean fusion power, but there's no guarantee it would be in any way economic.
Edward Teller proposed something along these lines with conventional two-stage nuclear explosives back in the 1970s, but it was rapidly abandoned as horribly inefficient and costly:
I know this stuff is fun to speculate about but the boring and unsexy option of more solar panels, wind turbines, and transmission lines is likely to be both cheaper and more politically saleable.
From what I understand, practical routes to fusion all produce a lot of fast neutrons, and tritium. The neutrons make the surrounding equipment radioactive, and create toxic elements. The tritium is itself toxic, and is hard to contain. And the high neutron flux means that such a machine could easily be used to create weapons-grade plutonium, by irradiating uranium. And come decommissioning time, the reactor itself is a pile of toxic waste that will need long-term storage.
I don't believe that fusion is inherently dirty; just that none of the technologies currently being investigated can be described as "clean". And they all run the risk of nuclear proliferation.
These issues apply to D-T fuel, whose output energy is 80% high-energy neutrons. Helion is using D-D/D-He3, whose output would be only 6% neutrons, at lower energy. If you do the D-D reactions in specialized fuel production facilities, you could sell the He3 to anyone with an He3 power plant and they'll have zero neutron radiation to deal with or misuse.
Even with D-T, the reactor waste would stop being dangerously radioactive in a few decades.
All of what you say is true. But on the plus side, you don't have the transuranic waste. Transuranic waste is the worst waste, the most dangerous and the most expensive and vital to store properly.
Sure, it's complex interaction. But where should our focus priorities be, what task should we do first...what should we publish most about??? Maybe my priorities are wrong...what are you saying?
Humans usually do the cheap thing. Historically making the right thing the cheap option has been much much more effective than trying to nag the invisible hand, which basically never works.
It's not the point, having cheaper energy is not going to magically reverse priorities. Focus on what's important. Anyway, you work it out. I'm not gonna waste my time having a discussion here.
Who cares what the priorities are as long as the right result happens? If cheap energy makes incentives align so the right outcome happens, why does it matter the wrong priority caused that outcome.
You choose: no electricity for 1 month, or no food for one month? Do your priorities matter now? It matters what we get to first. Unlimited energy won't matter if everyone starved. We have multiply redundant energy, but food is vulnerable. We can't grow our food with soil that's broken and without the right fertilizer (if you don't think it's a risk look into it more).
If we could afford doing the sustainable thing, but are doing the unsustainable thing anyway... that usually means we actually can't afford the right thing, or just can't do it.
Note: you have to account for costs of coordination and taking on risk - then you realize a lot of the "if only everyone would ${trivial thing}..." aren't done because they actually are prohibitively expensive or downright impossible.
I think you need to look at history more. The world is not driven by rational moral actors...We could have been farming sustainably, but haven't been. It's even cheaper to rotate crops in the right way, that preserves soil, but it wasn't done. There are engine designs that are more efficient, use less gas, but gas companies held them back. There are treatments that are better, save more lives, but pharma companies would lose out if there were less chronically sick people. "Can't afford" is a dirty word. Depends on perspective, encodes morals/values. Who can't afford? The greater good? Or the greedy? Who can't afford? The present generation? Or 3 clicks down the line when everything's dead because of choices made now?
The blocker to a better world is not some engineering impossibility, or market-rational overprice...not always. Often it's because someone with some power would lose out if we changed the way we did stuff, and they don't want that.
Human fucking nature. Can't cure that. Small bore example: Your engineering org would be better without the power-mad meeting-coordinator kahuna middle manager who holds hours long meetings for no purpose than to lord over proceedings and justify their own pretense of importance. The greater good would be served by demolishing said meetings, and removing kahuna from org. Does it happen? Often, no. Why? Because the world is not driven my moral, rational actors.
You don't think having a cheap way to power artificial light sources might be relavent to the food crisis in places not suitable to growing food the natural way?
Think about it. I don't think "those places" are going to be getting first dibs on fusion anyway, not de point. It's not that it ain't relevant. Of course interacts, but focus on food first.
Look everyone this is fun and all but I’m not going to waste my time going back and forth on this. Sort it out amongst yourselves. You can pretend you’ve won and everything, I’m cool with that.
The title doesn't match the contents. There is no actual justification of why this is a breakthrough.
IMHO it's a milestone, not a breakthrough. This is because it is only a marginal improvement on previous results and it is very unlikely to significantly change the state of either the science or industrial products without further work.
This specific marker is a breakthrough. We did not know until three weeks ago if laser ignited fusion reactions could be created that released more energy than was transmitted into the fuel.
Now we know it is possible, it has been done.
The next question is, can we create enough industrial efficiency to create a working motor generator set.
A hard problem to be sure, but proving that it is a problem that we have is itself a breakthrough.
I'm confused why it matters. If I heat something with 2J of energy and it released 1J of energy, it's like the matter has been heated with 3J of energy. If I can extract heat energy with 80% efficiency, I can extract 2.4J, more energy than was put in.
That is very small though, for a "breakthrough". That's my issue.
We already knew that laser ignited fusion worked (that it triggers fusion). And we still don't know it can produce more energy than it consumes overall (one of the big questions for this sort of fusion).
Really, we have gone from being 0.5% efficient to 1% efficient (I believe).
One way to describe that is doubling the output. And that is technically true. But it's quite misleading as it will have no actual impact in the real world.
Add to that that this was quite predictable. And that we will need at least 3 very big breakthroughs before this is viable (much more efficient lasers, a continuous method and either replacement of the fuel or a very easy supply of deuterium). I just don't see this as a big deal.
That's why I called it a milestone. It's progress. But it isn't huge, unexpected, change-in-the-real-world progress.
Most interesting fact from the article.
At the system level it took 300 million Joules to produce 3 Million a 1% return - so still a very long way to go!
With this kind of approach I can’t see how fusion can ever be cheap, the capital costs are enormous. Better to focus on standardised production lines for small fission reactors.
Engineered lasers are over 10x more efficient at emitting light than what they got, and it is very possible they could improve the fusion ignition process to generate more per laser needed, so I wouldn't rule this out yet.
For one, it's theoretically impossible to make lasers 102 times more efficient, especially if you consider the current industrial efficiency of up to 20% (30% under special conditions).
For example, the maximum efficiency of a laser is limited by the quantum efficiency of the laser material, which is the ratio of the number of photons emitted by the laser to the number of charge carriers (electrons and holes) injected into the laser material. The quantum efficiency of most laser materials is less than 100%, which means that there is a limit to the efficiency that can be achieved.
In addition, the efficiency of a laser is also limited by the thermodynamic laws of thermodynamics, which state that it is not possible to convert heat or any other form of energy into work with 100% efficiency. This means that there is always some amount of waste heat generated when a laser is operated, which limits its efficiency.
In other words, fusion is unlikely to be using lasers for the ignitions.
UPDATE: People here seem to believe in self-sustained fusion, but this has never been tried or proven to be possible - no one know how much energy is needed for repeated ignition.
The current plan is to re-ignite new pellets fed into the containment.
For one, it's theoretically impossible to make lasers 102 times more efficient, especially if you consider the current industrial efficiency of up to 20% (30% under special conditions).
That is not true. You can of course not make a 10 % efficient laser more than 10 times more efficient, but there is no limit to how much more efficient you can make a laser in general, you just have to start with an inefficient enough one.
And the NIF lasers are quite inefficient, according to Wikipedia - which might have slightly outdated numbers as someone pointed out to me - they turn 422 MJ stored in capacitors into 1.8 MJ UV laser light, that is an efficiency of 0.43 %. Bringing this up to the 20 % you mentioned would already almost be a factor of 50. Also not all of the gain has to come from more efficient lasers, the fusion process could also be made more efficient.
But you are of course correct that there is some general limit to laser efficiency, therefore the question is if the fusion gain can compensate this and all additional losses during the conversion into electricity further down the line.
I'm not sure if it's theoretically impossible. But it's certainly an incredible challenge to build a high efficiency laser system that can consistently generate many pulses per second without degrading either the laser head itself or any associated micro-focussing optics.
Can we stop pretending this isn't about nuclear weapons research and - basically - continued funding?
LLNL is at least a decade behind its initial predictions of ignition. The current lasers may be inefficient but they've been fine-tuned very precisely for optical quality. There's no guarantee more efficient lasers would have the same characteristics and could be focussed in the same way.
I'm sure LLNL know the maximum theoretical fusion gain for a realistic pellet design, and it's worth nothing that that number hasn't been mentioned anywhere. Clearly there's only so much energy available for each cycle. If that energy is on the wrong side of what's needed to release it even after efficiency improvements, the entire system isn't workable, no matter how it's re-engineered.
My point was purely about the claimed impossibility of a hundredfold efficiency increase which is only true if you are already about 1 % efficient. I have no real opinion on what the NIF is doing or not, at best my layman judgment would be that laser fusion is way further from becoming practical than tokamaks or stellerators.
You're right that we don't have enough information here, I've been reading a lot about it's theory and practicality throughout the years. There's also theories about ignition in parallell, by using multiple pellets. Since the first step has been taken, there's a lot of avenues for improvement. However, one also has to take into account the "chaos" created by the ignition in the first place, re-introducing an additional pellet or chain reaction has it's own challenges.
>Can we stop pretending this isn't about nuclear weapons research
The laser practicality issue that prevents this from directly becoming a power source would also be a major barrier to its application as a weapon. The laser fires in the UV-B (351 nm), which is scattered and attenuated by air, to say nothing of smoke or dust; it also requires incredibly high targeting precision (<2 mm target diameter) and consequently precise placement of a target weighing only milligrams. Additionally, the various optical components of the laser must be very well aligned, which is difficult to achieve in any battlefield conditions. And the whole thing is a very obvious and vulnerable target.
I ascribe a small possibility to its utility as a weapon in space, but practically zero on Earth without other major developments.
>I'm sure LLNL know the maximum theoretical fusion gain for a realistic pellet design, and it's worth nothing that that number hasn't been mentioned anywhere.
I wouldn't be so sure. Fusion is in general quantum chromodynamics, which is not so well characterized (being the subject of the famous YM mass gap conjecture). Even in this case it was stated that the yield exceeded expectations and damaged the sensors, which was probably not desired.
NIF isn't used for research into a new kind of weapon, in the context of nuclear weapons, it's used to generate similar conditions to an actual nuclear explosion, as the US government has signed treaties that agree to not test nuclear weapons anymore.
It's better to focus on standardised production lines for small fission reactors currently if your goal is to produce energy for the power grid right now, but that's not what cutting-edge research or science is about. Whoever launched the first firework probably didn't see how people could walk on the moon either.
Science is rarely a linear path. It's also rarely "I have this problem so I solved it this way." That's more engineering than science. It's really disappointing to me to see people, for lack of a better term, just shit on this accomplishment on HN when it's some seriously amazing stuff. It seems like a lot of people here don't understand science, they only understand engineering and then think about engineering mostly from a dull business and product oriented perspective.
Except it isn't the HN readership who choose to constantly frame these stories in terms of engineering. From the article:
"These and other scientific, technological and engineering hurdles will need to be overcome before fusion will produce electricity for your home. Work will also need to be done to bring the cost of a fusion power plant well down from the US$3.5 billion of the National Ignition Facility. These steps will require significant investment from both the federal government and private industry."
If fusion research scientists continue to insist that their research has any viable path to use in commercial power generation, and to demand large amounts on funding on that basis, then they should expect to be critiqued on that same basis.
What that that paragraph means is that things that do not exist will cost money to bring into existence. A significant amount. It's not a critique. This is not new. This is not a fusion problem. There is an initial cost and development to any new technology. This technology is particularly difficult, however. It's also sustainable real renewable energy. You have to worry about actually being able to invent something before you can worry about saving money on that thing. Look up how much the Manhattan Project cost! This shit is not software engineering; It's really really fucking complicated it will cost a load of money.
We don't really know what path fusion research will lead to. It's science. We don't know what we could find out tomorrow that could apply this research. But even the promises of commercial power generation alone should be enough to keep funding the project whether it will happen in 50 years or 100 years. It's not like they aren't making progress. You can't rush research and also under fund them.
The Manhattan Project cost $2 billion, equivalent to $23 billion in today’s dollars — for the entire project.
~$2 trillion was used to bailout big businesses. If money was properly accounted for, $2 trillion could fund approximately 87 Manhattan Projects simultaneously in today’s dollars.
Nuclear fusion is stuck into an "early MASER/LASER phase".
Laser was cool, but nobody knew what to do with it. These days we're at the stage where we're thinking: what can't we do with it??? But for about 3 decades (before CDs, basically) laser was a pop science laughing stock, more or less.
And fusion is much harder plus has been talked about and hyped for at least as long.
Fusion is different: we know what we want it to do, but it's proving very difficult. It's a problem in search of a solution, unlike lasers which were (initially) a solution in search of a problem.
> laser was a pop science laughing stock, more or less.
Yup:
In college English classes I took, they wanted the students to write term papers. Ah, sure, they expected some review of some case of belles lettres, maybe Medieval French romantic poetry!!!!
Instead, in one case, I picked the transistor and another, the laser.
For the laser, I had no idea of the applications. For the transistor, all I knew was, what it seemed was all Bell Labs had in mind -- replace their usage of vacuum tubes, that is, analog amplifiers and not digital, Nyquist sampling, etc., even though Shannon was at Bell Labs, etc.
The idea of a few billion transistors on a sheet of silicon about the size of a large postage stamp, 16 cores, 64 bit addressing, 4.0 GHz clock, etc. -- beyond all expectation or belief. The graphics processors -- still less belief! Lasers sending trillions of bits per second per hair-thin glass fiber -- not even ready for science fiction!
No doubt I picked the transistor and the laser out of media hype. So at least some people in the media expected something from those two. Here the media was not wrong, and in the long term the hype was way below the reality.
That's because this breakthrough is being sold - from a dull business and product oriented perspective - as an incredible etc etc which makes a new kind of product more likely.
No one is shitting on the research itself. But the PR around this story has carpet bombed sci-comm as a potential practical energy source. Not as basic research.
So it's perfectly reasonable to ask if there's a there from a commercial POV. And to note out that currently there really isn't.
If they really had announced a viable commercial product everyone here would be cheering.
I don't see one part of the article that sold this from a dull business and product oriented perspective or any unreasonable sci-fi nonsense. They only explained some of some people's future goals of fusion. Not necessarily even the goals of this program. You can't ask about a commercial POV of something that likely isn't even close to existing. Think past the damn business perspective! It's science! This is Hacker News not Shark Tank! It's like asking what is the commercial POV of lake houses on Titan. It's just ridiculous. One step at a time.
The comment I replied to stated "Better to focus on standardised production lines for small fission reactors." To you, is that not shitting on the research? It implies the only purpose for this research is for power generation. And that it is clearly inferior to small fission reactors for that purpose, when the technology doesn't even exist yet. It just diminishes the accomplishment as a whole. It's so short sighted from a group of people who's jobs only came to existence <100 years ago.
People can do multiple things at once...or rather, different people can do different things simultaneously, there is no need to "better focus on" x over y.
Well, early transistors were 3-5 orders of magnitude larger, slower and more dissipative than those we have today. So by the same argument, the computer revolution would have seemed impossible just 50 years ago.
Transistors? Computers before them used vacuum tubes. In 1943 Thomas Watson, chairman of IBM, said, "I think there's a world market for maybe five computers." Now, it would be easy for us to laugh at him, thanks to us being almost 80 years in the future, but on the topic of "what does the future hold for us", put yourself in Watson's place. You're reasonably smart and knowledgeable, and understand what goes into a computer, and thus you'd make the same claim.
Fusion is similarly incomprehensible to us here and now. There are untold advancements in materials sciences and engineering technologies that need to happen before it's possible so we have to invest in trying to so we can make those advancements in order to make it possible.
Just 75 years after Thomas Watson said that, we were already doing EUV lithography, a process whose description and commercial incentives would have been totally incomprehensible to him:
"In our laser-produced plasma (LPP) source, molten tin droplets of around 25 microns in diameter are ejected from a generator at 70 meters per second. As they fall, the droplets are hit first by a low-intensity laser pulse that flattens them into a pancake shape. Then a more powerful laser pulse vaporizes the flattened droplet to create a plasma that emits EUV light. To produce enough light to manufacture microchips, this process is repeated 50,000 times every second."
Transistors, microchips and lasers were not invented yet in 1943. How would one even guess that it could soon be a viable business to build a giant machine that shoots molten tin droplets at 50,000 Hz to produce a particular bandwidth of light so that you could create billions of tiny computers out of sand?
Hopefully fusion is on a similar path where the description of 2100's commercial reactor will sound similarly incomprehensible to us.
Thomas Watson's quote is misrepresented and it is more prescient than it appears. There are indeed maybe five computers on the world market: the BlackRock computer [1], the Google computer, the Tencent computer, and maybe a few others. A computer can be implemented in one CPU, or in millions of CPUs: if the control and the data ownership behind the CPUs is the same, it is still one computer. Watson's quote is more about psychology rather than technology, just look at the hullabaloo of crypto: a theoretically decentralized technology and the people still use centralized exchanges.
In the extreme you could say that today we have only one computer, the ASML computer [2], since they are the only ones making the machine that makes the machine.
> Watson's quote is more about psychology rather than technology, just look at the hullabaloo of crypto: a theoretically decentralized technology and the people still use centralized exchanges.
Well, that's easy. I forgot who mentioned it, but crypto has 0 support for actual clients, and the vast majority of day-to-day computing now happens on mobile end user devices (phones, tablets, laptops). So besides (what should be) this fatal flaw, the second thing crypto requires you is to manage your keys SUPER carefully yourself or a server.
Nobody[1]'s going to do that. I'm a techie and I don't want to maintain my own servers. What hope does Joe Locksmith have?
Energy production has always been an enormously capital intensive venture.
R&D labs aren't known for their industrial optimization unless that's their research goal. They proved this from the perspective of the fuel pellet - now they have to increase energy yield (larger fuel?) , reduce total system power, etc.
That said, they doubled energy yield in less than a year. The early exponential function on emerging technologies is always fun.
That's not the only big challenge. The fuel pellets are 1mm diameter hollow spheres machined out of diamond to extreme precision and filled with a tiny amount of deuterium-tritium gas inside. The pellets need to be carefully cooled to cryogenic temperatures so that the gas condenses evenly around the inside of the pellet, in order to maintain the near-perfect geometry needed for ignition. Once the pellet is hit with the x-rays produced by the lasers striking the interior of the chamber (hohlraum), it compresses and the fuel undergoes fusion, vaporizing the whole package.
So it takes an incredible amount of precision engineering to produce these tiny diamond pellets of fuel which then produce an even tinier amount of energy while being destroyed in the process. There is no indication whatsoever that the pellets can be scaled up due to the incredible difficulty of scaling the precision geometry involved.
So I remain incredibly skeptical that this is anything more than hype for a project that's really about maintaining nuclear weapons stockpiles. I sincerely doubt this approach will lead to a real production power plant without some major research into how to overcome the requirement for extreme precision geometry and a similar effort to scale up the size of the reaction. Then throw the laser efficiency issue into the mix!
One of the things about reaching this point is that we can reconsider just how much of that complex process is required - like getting the poc of some code working, then working backwards to strip away what isn't necessary.
Do you think they’ll drop the cubes into the reactor, The Expanse style, or have to move them in from outside between firings, while also keeping them cryogenically cooled? That sounds like a slow process, especially in a chamber that just had an artificial sun inside it and is running a power generator off of the heat produced.
Could they really fire this design more than twice a minute?
Solar panel and power cost were 100s of times more than they are today. Costs will come down that is actually not the problem having a sustainable fusion reaction is. How would you know how efficiently you can build something when you don't know what to build in the first place. If we can get sustainable fusion reaction then only should peope worry about how to bring down the costs.
Even if heat from fusion could be extracted at exactly zero cost, thermal fusion would still be uncompetitive. There is no way around it, wish as you might.
Aneutronic fusion could possibly have a future, but hardly anybody is working on that.
This is still fundamental research not commercial engineering. Like everything good before it.
Even in engineering problems, two people started a little electric car company in 2003. No, youknowwho wasn't involved yet. Would you have said "i can't see how electric cars can ever be workable" then?
Hmm .... Mars has nearly no atmosphere, and, thus, nearly no bad weather, clouds, fog, rain, etc. That is, it's sunshine all the time!
Sooo, for Mars, just put the solar panels near the equator and fairly densely all around the circumference. Then, wow, have solar power 100% of the time!
Do the same on Earth??? Ah, have some jungles, mountains, two major oceans, and a lot of bad weather. For the oceans, sure, have floating solar arrays. Right, need to think about how deep the power cables to the shore would be, 3-7 miles down and then back up?
But the Internet has some cables across oceans!! Sooo, we could also have power cables???
People don't believe the hype anymore so the solution is to not make 1 news article, but dozens of news articles. Maybe the people will start believing again!
I've seen as many articles that says it's not as major of a breakthrough. It's marginal at best.
Real Engineering published a video about Helion Energy company.
I don't have the tools to filter out commercial BS, still it seems to me to be most promising effort. The idea to get electricity directly from magnetic fields instead of boiling water is very intriguing.
I've also been quite surprised watching that one, it's completely novel approach to fusion and it may be the best way to mass adoption.
If they can make it work that is.
What most surprised me about NIF's ignition announcement was that this was even news. I'd assumed for years that most if not all tokamaks and stellerators could already achieve at least some brief fusion, just not to a useful degree. Turns out they're all just elaborate plasma heaters...
I'm very skeptical about those great announcements, even more when there's an announcement of the announcement. 300MJ in, 3MJ out, I wouldn't call it a breakthrough, maybe a milestone, nothing more.
I mean yeah it doesn't really change much in terms of usability, but it is the point from where onwards you can actually say you're doing fusion and not just heating stuff to high temps.
Someone said in another comment that they elected to use old lasers, and that state of the art are about 40x more efficient. That’s still a big gap, and I’m dubious about the tritium supply, and the amount of beryllium they’re using to breed the tritium. Also tritium is nuclear proliferation of fusion bombs.
They basically have no real achievements yet - the energy harvesting system isn't working, they are just fusing D+He3 - even though the concept sounds impressive. The video was cool but not cool enough to break through healthy fusion vaporware skepticism.
There is no energy harvesting system yet, next prototype will be include that. I like their "fast" (relatively speaking, of course) prototyping process.
Why not do a major announcement with the working prototype though, is what's ringing the alarm bells for me. Yeah sure they can promise the next prototype will do the magic thing but why not wait for that.
Because you can do reliable aneutronic fusion using usual methods? This seems explicitly commercial, therefore it needs to show value, which is in power generation, not making interesting gadgets.
As I understood it, reliably igniting aneutronic fusion efficiently was an open research problem. Whereas, they can consistently get to aneutronic burning plasmas.
He says in the video that they miscalculated a bit of physics regarding angular momentum and the Trenta design I suspect suffers from erosion of the throat (and power loss) because of that. The new gen is 25% bigger to avoid contact between the plasma and the walls.
> What most surprised me about NIF's ignition announcement was that this was even news. I'd assumed for years that most if not all tokamaks and stellerators could already achieve at least some brief fusion, just not to a useful degree.
I think they've done fusion in tokomaks and stellerators before, it's just that the energy applied to the plasma is greater than the additional energy that's created by fusion. It's not that no fusion is happening at all in the most powerful magnetic confinement devices yet built, it's that there isn't very much of it happening.
I'm not sure if it's true or not exactly, but I don't see how the NIF experiment would be any different from that if that were true, and in that case it wouldn't be news at all. It's not like they didn't use absurd amounts of laser power to get a tiny poof.
Going from a Qplasma of 0.9 to 1.5 for inertial confinement fusion seems like a step-change to me. For gains greater than one you have a positive feedback mechanism going for you -- the fusion you create makes more and bigger fusion. It is the difference between a tiny little bomb that works or doesn't work. It's this positive feedback that make scaling up potentially "easier". That's why it is a breakthrough, and concerns about lasers, etc. seem to be getting the cart before the horse. On the other hand, for magnetic confinement fusion, going from the same Q=0.9 to Q=1.5 is a much more "meh" event. I can see why some people are concerned about "where to draw the line" about total energy consumption, if it were MCF we were talking about. Maybe people have the expectation that ICF is further along the learning curve than it actually is? That's not to say that ICF is going to be viable for commercial fusion reactors in future. There are still many, many hurdles to cross. And maybe they will never be crossed for financial reasons.
The milestone is that they surpassed Qplasma > 1. This has not been achieved before even though fusion has been done in ICF and MCF machines.
Qplasma > 1 is significant because it now enters a regime where the burning fuel heats up unburnt fuel. In MCF machines this is referred to as a burning plasma. In ICF machines this is referred to as ignition. Ignition in MCF machines is Qplasma = infinity, where no external heating is used. This is on the table (maybe, theoretically), we just need to build machines with sufficiently high triple product and plasma control surfaces then learn how to do it. We'll have MCF burning plasma machines soon.
Helion has not talked about triple product numbers. I wish they would, because that's the key metric for understanding how far they are from D-He3 burning plasma. Also, having the triple product history for all of their machines would help show scaling laws. These are currently closely guarded secrets. Targeting D-He3 fuel means a much higher coulomb barrier needs to be overcome and a higher triple product is needed to reaching burning plasma.
Check out slide 40 for more details on fusion fuels.
Tokamaks do fuse a lot of atoms, just not yet enough to get more energy out than they put in. The record so far is energy output at 70% of the energy input.
The output scales with the square of plasma volume and fourth power of magnetic field strength, which is why net power is expected from ITER (which is huge) and SPARC (which uses new superconductors for especially powerful magnetic fields).
Everything about that was amazing. The design loos so simple and it's so small. The fuel used and the forward thinking of the fuel availability, separating fuel making from the reactor, how the fuel interacts with the reactor. Harnessing power from the magnetic field not by boiling water.
What made me super skeptical about the project is that I could not find out any numbers regarding economic viability:
1) How much energy per pulse is a commercial device expected to achieve?
2) How many pulses per second are realistic?
3) How many tons of copper/steel/capacitors per MW of power capacity are expected?
Because if you're gonna need a warehouse full of capacitors, several big turbogenerators worth of high quality steel and copper PLUS all the vacuum/plasma tech just to hit a few 100MW of continuous electrical power-output, then I frankly do not see how that would EVER be viable/attractive ANYWHERE, and NO amount of scientific progress might be able to change that...
But the concept to me at least looks more attractive than big magnetic-confinement plants, where the problems are even bigger (unaffordable plasmachamber + cryo-infra, super problematic neutron-flux, AND STILL needing all the heat => motion => electricity circus from a conventional plant).
In the video they’re trying to transition to 1/s from 1/10s, and targeting 10/s in a future generation.
Always hard to tell from publicity shots, but it looked like their existing capacitor bank was around the size of a shipping container. That probably goes up with the frequency though. I’m not sure you can do single or dual banks when aiming to fire every hundred milliseconds. Plus I think you need somewhere to send the produced power for dumping into the power grid.
But they’re already trying to produce their own capacitors to deal with that level of cycles per hour. I could see these guys spinning out a couple of companies that supply other designs or even industries. Especially if the money runs out.
I’m both intrigued and disappointed when people end up using ultra fast computers to solve these engineering problems. One, it’s a sign of progress and a potential solution. Two, it suggests a lie around the notion that we should have had an interplanetary civilization with flying cars, fusion reactors, and personal robots generations ago.
We didn’t have the compute power for any of this stuff in the 80’s. I did some reading a while back and discovered that there are elements of Computational Fluid Dynamics that became state of the art around 1990, so we are maybe two decades behind, not five.
Everything i have heard about this makes it sound more like a milestone than a breakthrough. I.e. its great it happened and an important step, but its not shocking that we eventually did it.
When i hear breakthrough i think of something more unexpected. I.e. figuring out how to solve a problem we had no idea how to solve.
Exactly.. The article fails to answer the question posed by its very title. What I've gathered, this was a very much expected result and NOT a "breakthrough". And with regards to making fusion a viable source for commercial, large-scale energy production .. this is only tangential at best to that effort.
The problem with celebrating "fake victories" (and being scolded for calling them out) is that when there really is a breakthrough (i.e. something so big it fundamentally changes our perception of the problem), people are going to think you are making just another over-hyped publicity stunt. I am definitely of the opinion that such media attention focused advertising of (otherwise perfectly valid) fundamental research is doing no one favors (not the researches nor the public.)
I've seen quite a few people arguing in (a European context) that we shouldn't build out nuclear in the "short term", "since we can just get fusion without any of the downsides of fission".
I've seen also people arguing that we should never ever build nuclear fission again because it is "bad".
What I want to say is that bad faith arguments will always exist, and calling this a breakthrough won't significantly increase the number of people using bad faith arguments, because they would just find another excuse instead of the fusion one.
I feel like HN is being especially pedantic and jaded about this specifically and I don't really get why.
Scientifically this is absolutely a breakthrough, just like the detection of the Higgs boson was despite them already having expected it or just like LIGO's detections of gravitational waves were despite that being entirely what it was built for or just as JWST's detections of galaxies much older than those seen from other telescopes was despite that being part of the entire point of dumping billions into it. Ingenuity's first flight was also hailed as a breakthrough despite that being exactly what it was designed to do.
In every case we had an idea of how to solve a problem and the outcome was generally expected. All that had changed was that the data had been collected of the solution working, just as it has been in this case.
To put it differently, it's a breakthrough because after decades of work, the National Ignition Facility can actually achieve the thing in its name.
I think after decades of fusion 'breakthroughs' people have become a bit jaded and would simply like to short-cut to the point where there is a fusion reactor in a box that produces more energy out than is put in.
Responding in a pedantic/jaded manner is one way to assert a form of intellectual dominance. For this audience I feel like that sort of dominance is unsurprising.
Agreed, but I think it has the side-effect of signaling "This technology advancement isn't worth excitement/praise/prestige" which I think in this case could have effects like demotivating people who might be considering going into the field.
No it's not a breakthrough. It is only net positive because the NIF has an exquisitely stupid and convenient place where they are defining "energy in" which let's them cheat how good it looks. If you could take other types of fusion and draw a clever line around where you start tallying energy in, then they would have had "breakthroughs" long ago.
The definition of Q that is being used was established by the National Academy of Sciences not NIF.
NIF previously tied to use the hot spot energy instead of target energy in 2013, which they were criticized for, and rightfully so. But this definition of Q is analogous to the definition of Q used by MCF.
With the Higgs, we got the mass, which was not predicted by any of the theoretical models. We were basically done at that point. With LIGO we got tons of tangible data of black holes colliding.
Fusion is not useful until Q>1, and for NIF to claim that when it obviously isn't true is, IMO, a bad look. They invented Qplasma where they arbitrarily get to ignore over 99% of the energy that actually went in, and don't need to worry about capturing any of the energy actually coming out. They got Qplasma>1, which is a great milestone, but the overwhelming narrative in the media is that they got more energy out than they put in, which is simply not true.
Someone, and almost certainly someone *ELSE* will get to Q>1 very soon. Lots of groups are targeting 2024-2025. That will be a major breakthrough, but in the public mind, NIF already did it, and I guarantee that's going to cause a lot of confusion.
It just feels like a marathon race where a bunch of competitors are competing honestly, and one participant jumps in a scooter, and blasts through the finish line, and the crowd is cheering, and they're doing interviews about how great it was to win the race, and people are going home, but the actual race is still happening, and getting super interesting, but no one cares anymore because that milestone was already claimed by someone who didn't even accomplish it.
Depends on the device. Q_plasms > 150 iirc is the estimate for MCF fusion. The NIF lasers take in way more energy than they put into the pellet, so that I think you want something like Q_plasma > 10,000 iirc.
Iirc most engineers use a different Q, which refers to all in energy, so Q > 1.5 ish is enough
Because we've been told for decades that commercial fusion plants are a decade away. So people are understandably cynical about it. Really, physics has had a problem of late over hyping things that don't impact people's lives. Maybe that's unfair to expect every discover will directly lead to noticeable changes in technology. But I would argue that researchers over estimate how much people care about their discoveries. Laymen don't understand or care about minor milestones if they don't directly impact them. Or to put it another way, people will be receptive if you have a big achievement at your job but will get annoyed if you tell them every detail about it.
So yeah, when people are told fusion is around the corner but all they hear are small improvements that to them mean nothing, its easy to see why. Now obviously not everyone is that way, there are communities of laymen and amateurs online who are interested and do care. But you don't need public press conferences to get to them. A press conference is for Joe Sixpack. And Joe Sixpack assumes when you come calling about Fusion its to say that it works, global warming has been solved and his electric bill will be next to nothing.
> Because we've been told for decades that commercial fusion plants are a decade away.
Even if you take the most breathless headlines about NIF at face value, we are obviously still decades away from a commercial fusion power plant. Let's say laser fusion is great and perfect, all we need to do is design a commercial facility that uses modern lasers and then construct it, with commercial turbines/etc, as well as commercial production of the fuel pellets. Just that will take decades. Even basic bitch coal power plants take several years to construct and there's nothing novel about those.
But the reality is that even with the best lasers available today, they wouldn't get enough energy out to make this commercially viable. Decades more development time is needed before they can even think about designing a commercial power plant.
Actually, I think the pedanticism is on the part of the people announcing this "breakthrough." The definition of "break even" used here (that is, Q>1) is highly technical, and but one factor in the actual "break even" equation. This experiment didn't prove you could get more energy out of a fusion burn than you put in (which is the notion they want to evoke), but rather that they could raise the ratio of energy absorbed in one part of the burn, from the NIF lasers, to that released in another, from 0.7, which they had achieved in August of 2021, to 1.5. That's it. That, and the fact that it's funding season in Washington, so not a bad time blow NIF's horn.
Milestone, yes. You have to get Q above one if you're starting at .01, and need to get to 100 and 1 is a great mile marker. But, But breakhrough - I don't think so.
I mean NIF has spent a majority of its life (11 years) seeing yields below 200 kj. Only in 2021 did it jump above 1 mj and in 2022 above 3 mj. If your definition of breakthrough is rapid improvement, I think this demonstrates it:
I realize this may be too complicated to explain in one comment, but as a layperson... can anyone summarize what they actually did to achieve this sudden gain in power? Like, what have they been doing for those 11 years? Upgrading the lasers, improving the optics, refining the design of the hohlraum, changing up the fuel? I know this is the cutting edge of physics and all, and maybe there's some DoE secrecy involved, but it's interesting that I haven't seen a single article yet attempt to summarize what changes were actually made in order to achieve the "breakthrough".
I don't think anyone specifically knows. If you ask the physics, they will point to changes (like pulse shape or component sizes) they made due to their increased understanding of the physics. If you ask the target design people, they will point to decrease fill-tube size and number of defects in the capsule. If you ask the facility, they will say their ability increase both power and control power delivery. Likely, it is a combination of everything.
The other point that has been mentioned to me is when you in the self heating regime, there are exponential returns on increasing "quality" of a shot.
What does "component sizes" and "fill-tube size" refer to, BTW? And the quality of a shot thing you mention in the last line? Sorry for all the questions, just curious
I think he meant _useful_ energy. Is it so surprising that HN is primarily interested about civilian applications and cynical when none are in foreseeable future?
The existence of stars already did. This is more like "it's plausible to do so in a reactor that won't destroy itself". The "that won't destroy itself" part is what this demonstrates, but annoyingly reporting often shortens it to the "is plausible to do so" part.
Except... correct me if I'm wrong here... The "reactor" did destroy itself.
> The fuel and canister get vaporized within a few billionths of a second during the experiment. Researchers then hope their equipment survived the heat and accurately measured the energy released by the fusion reaction.
It didn't necessarily blow up like a fusion bomb, but the energy was hardly produced in any sort of sustainable, ongoing form that we can harvest.
>The definition of "break even" used here (that is, Q>1) is highly technical, and but one factor in the actual "break even" equation.
Isn't the Qplasma > 1 the most important one here, by a very wide margin? Maybe the only important one at this point in time. That fusion begets more fusion, in a positive-feedback way. That's the breakthrough here. All the other efficiency factors are secondary in nature. That is step 1, and the next step is to scale this up so that eventually Qplasma >> 1. And only then the hand wringing about efficiency of the lasers becomes something to address. It seems like the baby was just born, and people are concerned about which colleges to apply to.
NIF achieved plasma combustion over a year ago, as documented in a Nature article earlier this year. If there is a threshold being crossed, that was it. The recent shot tuned that experiment to increase the yield. So now the the total released energy (not capturable, or usable, but released) is greater than the total non-fusion energy captured by the target (not put on the target, but actually captured and turned into useful heating and compression). But that's a completely arbitrary milestone.
And, most significantly to me, there is no argument being made anywhere that this milestone is on the path to anything other than a dead end, local optima. That is, that you can further optimized the shape, quality and size of the hohlraum, and shape of the laser pulse to get to a Q that actually has something to do with power generation. (I have no doubt that achieving combustion/ignition is useful for NIF's real purpose, which is to simulate H-bomb physics to aid in maintaining our stockpile of city-destroying weapons - they have in fact created a nano H-bomb).
I think the distinction is that with a breakthrough there is something new to learn, where a milestone you basically say "good job, keep up the good work". With a breakthrough there is more content to engage with because you can try and understand what is new. A milestone is great, but there isn't anything new to really engage with, so it has a shorter news cycle.
As a comparison - stable diffusion is a breakthrough - we are still trying to figure out what it all means and how it will or wont change society. its been months now and we are still talking about it. We wont be talking about this fusion thing months from now. That doesn't mean its less important only a different type of event.
Exactly. If there was a breakthrough at all, it in 2021, when NIF finally reached the point of "burning plasma."
But breakthrough, or milestone, I fear it may all be for naught. The improvements over the last two years are essentially all due to changes in the hohlraum geometry and preparation. All good, if your goal is to get burning plasmas to study. But NIF is orders of magnitude away from anything useful for power production. They have a Q of 1.5, but need something more like 100. They have a cycle time of, maybe, .1 shot/hour, and need something like 10,000/hour. Some real breakthroughs in the lasers (or maybe particle beams instead), coupling, and scale, not fine tuning of hohlraum shape, are needed.
Still unconvinced why this is such a major breakthrough - it may be the first net-positive controlled fusion, but the distinction between controlled and uncontrolled fusion is a bit academic in this case. In uncontrolled fusion (https://en.wikipedia.org/wiki/Thermonuclear_weapon), you are using a nuclear fission primary stage to generate enough energy to start the fusion reaction in the secondary stage. In the NIF, you are using lasers (that have to be aligned to trillionths of a meter and damage their own guiding optics everytime they fire) to start the fusion reaction in a smaller pellet (that costs hundreds of thousands of dollars). So the only real difference between the first and the second case is that the H-bomb is smaller, much more expensive per amount of energy released, and explodes inside a chamber. And it's obvious to pretty much everyone who is paying attention that transforming this setup into a working and cost-effective fusion power plant is a very tall order. Also, what somehow always gets left out of these articles is that the NIF is a military facility financed from the "(Nuclear) Stockpile Stewardship" budget. So its primary objective is studying fusion in a controlled way, but for building better bombs (and retaining the expertise necessary to build them), not for energy production.
With regards to NIF being a military facility, I'd very much think that energy production (cheap, safe, limitless?) is very much a military needs as much as a civilian need.
Even if said energy goes further to power laser weapons or other exotic power hungry systems; energy is energy, and something has to produce it.
The internet was a military need from what I remember.
The military said something like: "Telephone routing is point to point and inflexible, if the enemy cuts out 1-2 lines of communication, an entire section is completely cut off. We need something better." That something better turned out to be packed switching where you just throw stuff along a network and the network ensures that the packet reaches the destination, but you could theoretically have two packets going from Bucharest to Johannesburg, one through India and the second one through Canada.
Yeah compare broken water pipes with broken facebook/whatsapp etc. If the former is broken, people don't have water until the issue is fixed. With latter, at least they can use one of the alternatives (and they do, usage counts of alternatives go up during outages).
It started as a way to network air defense bases for AESA radars, NORAD and Nike missiles together across CONUS. Feature creep and increased scope grew it into a general purpose data network that linked not only military sites but also research institutes by the 80's.
Depends - if someday they manage to make the fusion generators compact enough (think of the coffee-machine-sized "Mr. Fusion" that Dr. Brown brought back from 2015 in Back to the Future where you could throw in any kind of junk and it would use it to provide energy, or even the shipping container-sized units that Lockheed Martin is envisaging), it might be interesting for military purposes. But it's very optimistic to believe that the NIF will seriously work on sustained fusion and extracting the generated energy in a usable way - after all they are called National Ignition Facility, so now they achieved that, mission accomplished. They can now run further ignition experiments with various pellets, laser configurations etc. and study the results, but they will probably say that a new facility (and more money) is needed for further studying sustained fusion.
NIF being a "military facility" has less to do with laser weapons or power-hungry systems than their type of research indirect drive laser-fusion, being the exact mechanism that takes place in multiple-stage nuclear weapons. The lasers at NIF first generate x-rays, which are then focused on the target to energize the pellet and create fusion.
If you look at the "shot history" of NIF -- the vast majority of their facility's energy is spent on actual DOD weapons research, not fusion power research that would be incidentally beneficial to the DOD:
You greatly underestimate the order of magnitude difference of produced energy in these two examples. You can't contain a bomb's hydrogen fusion reaction in a chamber for the simple reason that you need a nuclear fission warhead just to ignite it. So if you want to produce useful, controlled energy outside of killing things, the lasers are your only shot for inertial confinement fusion. Yes they still get damaged, but lasers have seen ridiculous advancements in precision and instantaneous power over the past few decades. This news shows that practical fusion is now only an engineering problem, not a fundamental physics one - like it still is for Tokamaks and Stellarotors. For them noone knows if it is even possible to e.g. control turbulence long enough to achieve net positive energy gain.
This is the real key. A LOT of people (and digging through their comment history, many of them) are MCF people who somehow poo-poo this result but then go on to extol how MCF schemes or companies are still a safer bet...even when none of them have even achieved any level of Q, heck they didn't even beat NIF's old result.
If this result is "useless" and "shouldn't be news," then the whole field of MCF and all the start-ups are even more "useless" and should be even less as news, because a small non-zero number is still larger than zero.
For those who were wondering what "MCF" is (like I was), the above author is likely referring to "magnetic confinement fusion"[1].
Contrast MCF with "inertial confinement fusion"[2] (ICF), a different (and apparently competing) approach taken by the National Ignition Facility in this latest announcement.
But isn't the 'net positive', itself only a factor in the very last stage of all of the input energy required for ignition? So they need to factor in all other components before looking at these orders of magnitude being useful?
No, it's not. If you get to arbitrarily define at what point you count "energy in" (as the NIF people are), you could also define a clever spot for MCF to be net positive.
As I understand it - correct me if I'm wrong - one of the consequences of the December 5 experiment is that the engineering uncertainty was brought into the realm of "this now fits the progress of engineering, so we expect to be able to engineer the thing given sufficient capital and time", rather than the previous "we have no idea if we'll ever be able to engineer this thing".
This is actually a significant step, not a tiny step. Achieving fusion ignition, ie including the 2021 shot, is THE non-linear scaling effect needed to make fusion a viable energy source. The other items needed to make this practical would’ve been premature to fund without this significant advance. https://physicstoday.scitation.org/do/10.1063/PT.6.2.2022121...
For sure, you can’t even consider the possibility of a fusion power plant without the core reaction producing power. That they figured out how to do the energy equivalent of getting the coal to actually light (finally!) in this context, is indeed useful.
But it is a tiny, tiny step towards an actual working fusion power plant.
And it does nothing to fix the economic issues here in competition with all the other energy sources, which matters.
Igniting that bit of ‘coal’ took near 100x the energy we got out of it. It was essentially putting it in a giant blowtorch, and then getting excited when it finally got lit and burned. That we previously only got half smoldering was indeed a problem! And that it finally lit is indeed cool.
it also took an extreme amount of thought, precision, investment, etc. to do it once.
Meanwhile, solar plants in the area were churning out billions of kwh.
I would consider the sparking of fire a big step. And that it used inefficient lasers is irrelevant because it wasn’t intended to be anything but a demonstration of ignition and scientific breakeven. The follow-ons which would use modern efficient solid state lasers were delayed because there’s no reason to build a power plant without first having the fire. Funding for them was premature, and it was right to focus on achieving ignition first: https://en.wikipedia.org/wiki/Laser_Inertial_Fusion_Energy
No one needs to talk about delaying solar deployment, obviously that would be stupid. No one should be arguing for that, I certainly am not, and trying to do this whole thing where we attack genuine advances just because they aren’t the thing we prefer is super counterproductive to progress. Shameful to do that, no matter how common it is on social media.
This was built as a machine to test fusion, not to generate electricity. It did that, and doing so is a significant advance.
This isn’t sparking fire. We already know how to do that. Fusion reactors have been doing that for decades, and we’ve been doing it overall for 80ish years now.
It’s sparking fire in a very specific synthetic scenario which maybe could be useful in making a specific type of fusion reactor.
It’s a critical step, yes, in development of that type of reactor, if we wanted to do so. Currently there is no reason to believe we’d find it worthwhile to do.
But there are literally hundreds more steps just as critical with no known solution, that are likely just as hard before we could actually have net positive energy even in a lab for such a reactor. Even if we found the economics would pan out.
That’s why I’m saying it’s a small step and we’ve got miles to go - because we do.
You’re pointing out some of the next small steps.
But those won’t get us to even having the energy balance positive on paper.
making lasers 100x more efficient for one being a notably hard problem, from a physics perspective. Probably even harder than IC fusion. The planned upgrades to lasers will help, but not enough to be close to a positive energy balance.
If the ‘first 90% takes 90% of the time’, we’re now at 2% of that first 90%. But everyone is hyping it up as we just completed it. Which is not true at all.
Honestly, it sounds like a Hail Mary funding push to try to get a new project funded or stop the lab from being shut down.
NO. The lasers are immaterial. You could use a hammer to cause the initial fusion (in fact ... [1]). The point of ignition is to have fusion powered by fusion, NOT by lasers. We want fusion of two atoms cause at least 2.000001 other atoms to fuse, and then raise that number. Not even by much 3 will definitely do it. If Q > 1, then this must have happened, there is no other explanation. That's the big advance. The point is to have the fire burn after we've lit the match.
You want to use lasers to initiate some fusion, you can do that using a tabletop laser lab (although magnetic confinement and big voltages is much simpler).
So this is the first machine where we're sure: it wasn't fusion caused by lasers, it was fusion caused by neutron pressure! Where did that neutron pressure come from? From other atoms fusing!
Note also that the sun does NOT do this. Fusion inside the sun is powered by falling atoms converting their falling energy first into heat then into singular fusion events, which essentially explode and throw a lot of atoms back up. Then the cycle begins again. Fusion inside the sun is a self-limiting reaction, which is why stars don't just blow up once, but "burn" over a long time. All the fusion inside the sun really does is slow down gravity for a little bit, then immediately stop. Gravity is producing the heat from the sun.
The hohlraum is destroyed almost instantly by the reaction.
The reaction is initiated by the lasers.
You can’t actually use a hammer to initiate it at any useful scale because hammers can’t create the appropriate pressures and temperatures to initiate the process at the scale needed to produce a useful energy output. Which is why the lasers.
> rather than the previous "we have no idea if we'll ever be able to engineer this thing".
We still have no idea if we'll ever be able to engineer this thing to produce commercially useful energy. And we probably never will, for all interesting values of "never", the challenges are simply so massive.
Not really because we have known for almost a century that nuclear fusion is a real thing that can release energy.
The problem is the engineering of how to actually do it in a useful way. They have made some progress towards that but it's still extremely unlikely that laser based fusion will ever work as a power source because: a) they're still two orders of magnitude from break-even from the input side, and b) they don't have a practical way of capturing the energy that's produced at all as far as I understand it.
Additionally, this is laser energy, but the lasers aren’t used directly. They produce X-rays which are the ones that actually drive the ablative implosion of the fuel pellet. If you used X-rays directly, the gain value would be MUCH higher… I think by that measure, they would’ve achieved Q=1 back in 2013?
Note that hey COULD use the lasers directly to drive implosion but haven’t yet. The Hohlraum/X-ray/indirect method I think produces more consistent lighting on the target, and is more analogous to the operation of an H-bomb, which is partly what this entire process is intended to replicate. They may do future experiments to try direct drive, in preparation for actually using this as a power production method and improving the efficiency dramatically… in combination with far higher efficiency modern lasers, this could enable Q high enough for net electricity production.
It should be noted that even if the lasers were 100% efficient, this result is not sufficient to achieve a practical fusion power plant. That target could produce maybe $0.02 worth of electrical energy. That's nowhere close to paying for the plant, even totally ignoring operating costs.
A real plant would need target yields 100 to 1000 times larger, with all the engineering issues that would entail.
Of course, because you learn to spark fire before you build a power plant. NIF is nothing like a power plant. It was designed for stockpile stewardship and for fusion ignition demonstration, and they intentionally cancelled efforts to solve the engineering challenges of making this into a power plant, because doing so was premature without the ignition breakthrough.
https://en.wikipedia.org/wiki/Laser_Inertial_Fusion_Energy
A log curve, rather than linear, would make the bars in the center taller and not give the same appearance of a massive change in the measurement. Not for me to say, but I'm guessing that, as is, it doesn't represent the work put into the effort as well as possible.
Yes, that's what a log scale does to a graph. But why do you feel a log scale is appropriate in this situation? To me, a linear scale is a much better choice. The underlying data isn't necessarily exponential, so there isn't a compelling reason to use a log scale.
No. Just because the sun exists, does not mean that it is necessarily physically possible to emulate controlled nuclear fusion without an _immense_ gravity well. Ignition in this context is a validation of our understanding of the physics. _Now_ it is just an engineering problem.
Ordinary hydrogen-hydrogen fusion in the sun only works because the sun is so freakin' huge.
It's so unlikely even under the sun's massive gravitational and heat that the power output is about 270W/cubic meter-- compare to human metabolism or an energetic compost pile, which is over 1000W/cubic meter.
We actually have to do a whole lot better than the sun, under far less favorable conditions-- to have practical fusion for energy generation on Earth. Inertial confinement fusion ignition (and to a lesser extent, thermonuclear weapons) strongly suggest this is possible.
The fact that black holes exist does not mean that we can spin one up to harvest it's hawking radiation for power. It's not just an engineering problem. That's not what that phrase means.
Say that to rob74, the starter of this thread, who essentially says "so what if it is the first time we've ever had a net positive fusion reaction, there is still work to be done and I don't like the people working on it" in order to wave it off.
My crazy idea would be to use the heat and the pressure of these to pump water uphill. It is of course a crazy idea and you'd have to figure out a way to manage the pressure peak
Pumped hydro has been a concept for a while now for grid storage of power; one of the limitations of it is just the rather unique geography required. Most places with adjacent reservoirs and enough vertical separation between them also have cottages, roads, and other infra that make it hard to be constantly draining and filling.
It’s less that (IMO) - cottages and roads are usually not as dense on steep hillsides and the like, though of course they can’t be ignored entirely. The pipes involved are usually not the big issue, and the pumping ops are usually not the big issue either.
The business model for pumped hydro requires relatively low cost to be economic and if it requires building a giant concrete water retention systems on top of large flat topped hills/mesas, near large drops in elevation, with water retention at the bottom of the large drop that is also really really expensive to build and requires specific geology to be economic.
Thank you for pointing me to Pacer, that was a fascinating read. They were apparently also considering digging a new panama canal with fusion bombs...
In 1972 they decided it wasn't economical based on fuel price compared to yellowcake, but fissile fuel isn't the primary cost of fission power and the engineering costs of pacer would seem to be much lower.
There's some old Asimov "Big Book of Science" sort of book I've got hanging around somewhere (dude definitely had aides/ghost-writers on some of that non-fic, right? Surely? At least researching and first-drafting for him to come through and punch up? Especially the more generic ones like that? I don't see how he'd have had the time to physically write that many words, otherwise, let alone also think about the words, considering his total output) that features the idea of nukes-as-tools pretty heavily, especially for canal digging and excavation generally.
Pacer took thinking about nuclear fuels in a lot of new and interesting directions. For the use case of interest, they really wanted to maximize the fusion yield fraction, minimize the radioactive products, and minimize the total yield. Had it ever been allowed to proceed to a commercial scale, these would have been very tight, very challenging, technical constraints.
If you relax one of those requirements, then it becomes instantly clear that limitless cheap energy is possible. For a somewhat absurd example, imagine that we build Pacer plants in space instead of on Earth. There's no longer any size constraint on the blast chamber. You can go to triple-stage (like Tzar Bomba) or more designs, at which point the fusion fraction gets extremely close to 100%, and the cost per unit energy becomes somewhat hard to fathom. But on Earth, it still would have been expensive.
>> This news shows that practical fusion is now only an engineering problem, not a fundamental physics one - like it still is for Tokamaks and Stellarotors.
Hahaha! That is the funniest BS I've heard about it yet! Fusion has always been "only an engineering problem" since the first H-bomb and this approach might actually be the farthest from useful even with is.
"This news shows that practical fusion is now only an engineering problem, not a fundamental physics one..."
Yes, I'd agree with that statement that it's only an engineering problem, but with a laser energy of 2 million joules and a fusion yield of only 3 million joules and an overall input of nearly 300 million joules to produce the lasers then I've no expectation of ever seeing a total net positive output in my lifetime let alone fully established working fusion plants providing power to ordinary customers.
Moreover, compare the overall 'negative' efficiency of this experiment with the actual positive efficiencies of existing generation systems: solar: 20+%, combustion: 20%-40%, mechanical/electrical turbine: >97% and so on and we see there's three or so orders of magnitude to catch up upon.
The task may not be impossible but don't hold your breath.
The 300 million joules to produce 2 million joules was almost entirely due to <1% efficient lasers.. even using “old” off the shelf solid state lasers with 15% efficiencies you’re already down to ~15 million joules. We’re much closer that the “300->2->3” implies.
Some labs are already working with 20% efficient lasers, so hypothetically, Fusion is closer to “10->2->3” which is just amazing.
There are different kinds of lasers. Last time I checked in 2018, Nd:glass lasers were the only choice for producing short, ultra-high power laser pulses with regular beam profile. Beam profile is very important for tight focusing. They may be 80's technology but these were produced in 2005. Efficient fiber optic lasers were already available but they are not suitable for the task. https://lasers.llnl.gov/about/how-nif-works/seven-wonders/la...
The second link has more details, but they show a 0.7% efficient NIF vs. a 7% efficient Mercury laser in the tiny little chart graphic -- which would reduce the energy for this experiment from 285MJ to 28MJ. This is way outside my area of expertise, but from what I understand, the same general tech that LLNL used to generate the HAPLS for Europe's laser research could be applied to the Mercury-like lasers and increase efficiency even further.
Thank you for interesting links. However, I see that the specified pulse energy of a single Mercury laser is 100J vs 18.7kJ for NIF so my point is still valid. There is no way to arrange 20k lasers all pointing at one pellet.
That was just a prototype laser to prove the direct drive approach with a high frequency solution.. They demo'ed a suitable laser operating at 10hz vs. 0.0001HZ with the existing NIF capability. Since the 100J Mercury laser was transferring more than 10x the energy to target compared to NIF (with an obvious path to 20x) - think of this as a 1/20th scale replacement for a single NIF beam with the ability to fire multiple shots per second.
Even if they couldn't increase the power past 100J (which they certainly can), using 2,000 beams instead of 192 doesn't seem too difficult given the much smaller overall energy and heat-removal requirements.
Can you point me to somewhere to read more about this? Also from wikipedia: "NIF hosts the world's most energetic laser."[0] Is that true and if so can many less powerful but more efficient lasers make up for it?
The total energy consumed isn't what these places are optimizing for. LLNL/NIF "invented" a lot of this research so their facility is the most energetic but also ~the least efficient, while being the most important in the world. When they constructed their laser facility in 1996, NIF used the surest-bet laser system acknowledging that hypothetical competing systems could be better suited (solid state lasers were just coming into play at the time due to semiconductor manufacturing). They've done a massive amount of work ramp up the power and get their <0.5% system to ~0.7% efficient (power from the wall:power hitting the target). The NIF facility uses "indirect drive" where the energy from the lasers are first converted to xrays, and then the fuel is energized with the xrays. Not coincidentally, that's the exact mechanism that modern nuclear warheads use to drive their fusion stage -- so NIF's selection here makes even more sense given their DOD funding and 'stockpile stewardship' mission statement.
Since a lot of modern laser research now overlaps with advanced semiconductor manufacturing, the competing systems are now much more mature. Here's a good rundown, but experts think that it's plausible to use even up to 20% efficient systems. Rather than using the laser to first generate xrays - these "direct drive" systems deposit the energy from the lasers directly on the pellets.
Do we care about performance when the purpose is proof-of-concept? Those numbers aren't set in stone. Efficiency won't be utilized when it's besides the point.
"This news shows that practical fusion is now only an engineering problem, not a fundamental physics one..."
We've been able to reliably trigger stable fusion in controlled experimental facilities since before I was born in 1966. People have literally been claiming fusion power is just an engineering problem longer than my entire lifetime, and I'm no spring chicken. Ring me when they reach engineering breakeven, although optimistically the chances are it'll be my grandkids that take the call.
> This news shows that practical fusion is now only an engineering problem, not a fundamental physics one - like it still is for Tokamaks and Stellarotors. For them no one knows if it is even possible to e.g. control turbulence long enough to achieve net positive energy gain.
Well, sort of. This is "theoretical breakeven" - the reaction generated more energy than they put in. "Technical breakeven", where you generate enough power to run the thing, is 2-3 orders of magnitude away. It's not at all clear that pulsed fusion like this is a viable power source even if it reaches technical breakeven.
The magnetic containment people are still struggling, but making progress. For a long time, everybody was looking for some clever magnet geometry that would yield stable plasmas, with limited success. Maybe active control will work. There are now people throwing machine learning at the problem.[1]
If we can get unit-costs on those pellets down, that sounds like a great starting point for high-speed spacecraft propulsion.
Not in the Expanse "pellets explode and produce electricity" way but in the Project Orion "explode nuclear bombs behind a shock absorber and ride the shockwave" way [1]. Just that this allows you to build much smaller and more manageable spacecraft (important since you probably still want to reach orbit with more traditional propulsion)
“So the only real difference between the first and the second case is that the H-bomb is smaller, much more expensive per amount of energy released, and explodes inside a chamber” - that sounds like a major technological breakthrough to me. They managed to concentrate so much energy inside a chamber in a controlled manner.
This comment sums up HN beautifully: "Here's an article from a subject matter expert on this topic, but I'm going to go ahead and ignore everything that was said because it doesn't fit my narrative."
You find us a subject matter expert in building production fusion power plants and we’ll get back to you.
What we have are a bunch of R&D people, some of which are using ingredients for which the entire world combined possesses enough fuel to run a single power plant for only two months, and then we’re out.
Too late to edit, but my source on the 2 months comment comes from an interview with Helion, who I think made the front page a couple of weeks ago, but in summary they are trying to smash two plasma balls together and then directly capture the magnetic flux created by the resulting fusion, rather than using a heat engine:
They talk about why ITER has no viable fuel source, and about how they (Helion) are going to have to run two configurations of fusion plants in order to source their own fuel. One that produces a 10% power surplus, undesirable neutrons and the fuel for their power plants. They didn’t say anything but I suspect they’ll end up coming up with a use for the neutrons, perhaps making fuels for other processes. Someone will likely eyeball that waste heat as well.
Major breakthrough is a subjective term so it’s hard to argue that precisely, but is it a major milestone on the way to commercial fusion, or “historic” as Dr. Kuranz says in the article? I can accept that. In 50 years, I’d guess it’ll be described in Wikipedia in such as way.
Secondly I’m not sure what you mean by saying this experiment has only academic differences with a thermonuclear weapon. It seems the only similarity is they both generate a fusion reaction. The latter is essentially multiple bombs strapped together which is quite different practically.
It’s practical proof that in a single pulse we can get more energy out than we put in. We do spend a shit ton of energy setting up the pulse, and we haven’t figure out how to get that energy out and make electrons move, but it’s been proven possible.
The “breakthrough” is the proof that it’s possible. Everything else is now optimisation and engineering, and needs mostly time, sweat and money.
Yea, and I wonder how much our need, at a cultural level, for progress will factor into this. If we feel like we need to progress and can't go "backward" with our energy technology, there may just be an irrational impetus to make it happen.
Solar and wind etc may just feel too "primitive" and nuclear too old and toxic and Chernobyl. Fusion though, once there's enough buy in and enough of a sense of it being doable, would fit our expectation of technological progress.
Culture bends to circumstances. When or where people can't cook or stay warm using the power sources they are accustomed to they go back in an instant to old ones. Logging and campfires, stoves, etc.
The practical reality is that solar and wind are still getting deployed en masse, people who want electricity don't really care what how science feels about progress etc. They just want their lights to turn on and for cheap. Since renewables are currently cheap, that's what we're deploying a decent amount of (depending on geography of course).
This milestone is liberating more fusion energy than photons used to start the reaction, but the energy source for the machine was electricity and it produced zero electricity. If you tried to convert the heat back to electricity you would be talking less than 0.1% of the electricity used being recovered at the other end.
That was already proven in weapons tests in the 1950s (e.g. https://en.m.wikipedia.org/wiki/Castle_Bravo). The hard part is always building the machine to produce net energy gain, which is still 2+ orders of magnitude away.
The reason it looks that far away is that NIF uses lasers from the 1990s with only 0.5% efficiency. Equivalent modern lasers are over 20% efficient, which would put overall net energy gain only one order of magnitude away.
> which would put overall net energy gain only one order of magnitude away.
With modern lasers they would be closer to Q=1 without the funny energy accounting. But still several orders of magnitude away from commercially viable, which requires vastly greater efficiency than Q=1.
Energy return on investment. Some oil extraction is currently at EROI=10, assuming EROI and Q can be made to be the same concept. Higher than EROI=50 is unheard of today, even 40 is pretty rare. EROI for this fusion method would be ~0.015.
The worst fusion methods scales with the cube of the radius of the area that can be made to fuse. So we "only" have to get a 10x bigger area to fuse, or have the same area fuse for 10x longer (and make the facility not blow up or melt down from doing that).
I wonder how the composition of the pellet was optimized. I'm sure the pellet was setup to make ignition as likely as possible. Now that they've accomplished that i wonder if parameters can be changed to up the efficiency of the reaction for energy output? Granted, optimizing for likeliness of ignition may well be the same as optimizing for energy output. Can they just up the mass of the pellet for a longer fusion reaction?
Isn't additional stages in fission/fusion bombs just upping the mass with the result being higher yields? "higher yield" being a bit of an understatement as the range is like 100KT to 100MT (and beyond).
Can a multi-stage pellet be formulated? Ignition of an outer pellet provides some energy to get a more powerful inner pellet ignited.
/not even close to being a layman, let alone expert, on the subject
>Can a multi-stage pellet be formulated? Ignition of an outer pellet provides some energy to get a more powerful inner pellet ignited.
My understanding is that that is the whole point of the announcement. The initial laser induced fusion caused further fusion of the interior. That's why the gain is greater than one. And that's why there is excitement, the possibility is there to scale things up using larger/more massive pellets. To me, that's the real announcement, the laser initiated fusion created additional fusion. The whole break-even energy seems more like a red-herring. Also, with inertial confinement fusion you have the possibility of doing the harder fusions, (like aneutronic fusion), that magnetic confinement fusion would have a much harder time with.
No. This was inertial confinement fusion, like in a bomb. The reaction could not have continued under any circumstance. The only hope would be to run this test repeatedly, like ten times a second, and somehow pull put power. But this machine has no hope of even attempting that.
> more than 5 million degrees Fahrenheit (3 million Celsius) – about 100 times hotter than the surface of the Sun
As my fear of Crichton's Gell-Mann Amnesia sets in, I'm being pedantic: either its about 545 times hotter than the surface of the sun (5.500°C) or they've reached "only" 0.55 million degrees Celsius.
Also: its only 519 times hotter than the surface of the sun when using absolute temperature (5.778 K).
I find it interesting that people talk in absolutes ("fusion will never be economical!") when the devices they're using were "impossible" only a few decades ago.
Also the difference between a 4:00:00 marathon and a 3:59:59 is less than a 0.007% improvement but it's a huge victory nonetheless. Once someone proves an arbitrary numerical barrier can be overcome humans tend to make a lot of progress shortly thereafter.
I can tell you that regular people took the announcements as a sign that nuclear fusion as a real useful thing is imminent, and couldn't be convinced otherwise!
But then if your announcements are completely rational you don't get as much funding. A lot of us nerds on HN don't like it when people do propaganda, but I assume we also like to win, right?
I recall reading an old 60s issue of 'Physics Today'; someone prophesized fusion ws 20 years away.
None of that in the (look, it's still around) recent article. Insightful details are present. And some honesty ... e.g.
"The agency toned down the ignition objective, emphasizing NIF’s ongoing experiments to investigate materials’ behavior under extreme densities and pressures in support of nuclear stockpile maintenance."
This type of nuclear fusion uses tritium, and since we can’t extract it from nature the only way to obtain it is very energy extensive. For this type of fusion to make sense producing tritium+lasers have to require less energy than the output, right?
Korean reactor reached 7 times hotter than sun, when I was reading it made me laugh that humans said fuck the dyson sphere and skipped a step on Kardashev scale yet still I find it ironic that this one built with military support.
Gravity confinement fusion with optical coupling. Really much, much more practical than the fanciful notion that NIF will miraculously lead to a commercial energy source that ingests diamond-coated nanopellets and has a duty cycle of 1:1e15.
Being able to do “ignition” reliably means you can do experiments in a regime where the energy from fusion starts to become a large part of the reaction. Dynamics in this regime are non-linear, so being able to actually test there is fairly important. Updating models and simulation inputs based on these results should make some further progress possible. This device is significantly different than a power reactor, but other efforts might see a benefit from a better understanding of reaction physics during “ignition”.
For a hint, the work was at the NIF, National Ignition Facility.
Even to start a wood fire, need an ignition facility. Then the fire continues on its own putting out much more energy than was used with the matches or whatever during the ignition. Then the energy from the fire ignites the wood that is not yet burning.
As we now know, once
the nuclear reaction is ignited, the reaction generates more energy than was used with the lasers for the ignition. Then, sure, we can expect that the energy of the reaction should, without more energy from lasers, have the reaction continue.
Right, I'm ignoring a long list of what are sometimes called engineering details, e.g., that this is inertial confinement fusion!
Ah, at one time I worked for KMS, that is the company of Kip M. Siegel, as at Google and Wikipedia:
"KMS Fusion was the first and only private sector company to pursue controlled thermonuclear fusion research through use of laser technology."
This article is obviously a PR piece, not an objective review, and the interviewee receives funding from the program. Looking around a bit, here's a discussion from 2018 that's a bit more revealing:
> "In the latest round of experiments, the capsule shells consisted of diamond doped with a thin layer of tungsten, and the hohlraums were made of depleted uranium. Earlier experiments had used plastic shells and gold hohlraums. Sebastien Le Pape, lead author of the paper, says the uranium hohlraum boosted the peak energy deposited on the capsule by 25 terawatts, for a total of about 450 TW."
Clearly the cost of these fuel capsules is prohibitive when it comes to any practical power production system, plus there's nothing like an energy capture-and-conversion-to-electricity system being developed. It certainly looks more like nuclear weapons research than power production research. The real goal is just preserving the budgetary outlay:
> "The new results were achieved before the Trump administration proposed cutting funding for NIF by $57 million, to $287 million in fiscal year 2019. That would have forced a 30% reduction in the number of experimental shots of the 192-beam laser. Lawmakers instead added to NIF’s funding next year: The final appropriation will likely end up between the $330 million included in a House-passed bill and $344 million included in a Senate measure."
Notably, DOE budgets for solar PV development remain entirely minuscule in comparison, which is one reason why it's China, not the USA, that has mastered efficient production of high-efficiency, high-durability monocrystalline silicon PV panels at scale. These of course compete directly with natural gas for power production, and the US wants to be the biggest natural gas exporter in the world. Not hard to understand what's really going on at the DOE/NNSA, whose original name was the Manhattan Project, then the AEC:
Long story short: it has never been done before and that's why it's a breakthrough. Never mind that it's totally impractical and far from something we could use to solve our energy problems TODAY.
We need to focus on building hundreds if not thousands of nuclear fission reactors and to reduce the population growth through birth control outside the Western world.
well, we have a serious energy addiction problem as large parts of the population got used to unsustainably high per-capital energy lifestyles (basically burning fossil fuel repositories like there is no tomorrow). so any paths to clean energy production are worth pursuing - provided they don't create an illusory sense of comfort and unsubstantiated hope that some fancy tech will solve everything.
fundamentally though, one would think that trying to create artificially solar interior conditions on Earth while the planet is literarily drenched in sunlight is not the shortest path to energy salvation. Life has invented already several ways to harness that energy and we have invented a few more in the meantime. I think the odds are that we will have more (as in: sufficiently) efficient harnessing of solar energy long before we create locally little suns to play with.
Does anyone know what the conversion mechanism from the laser to x-rays is? Is that just because the lasers heat up the inside of the hohlraum, and the x-rays are just the black-body radiation from that?
288 comments
[ 2.5 ms ] story [ 174 ms ] threadEdit: Mind you - there were tests like the huge Tsar Bomba that de-rated at 100+ Mt design to 50Mt be leaving out the final fission stage - the output being mostly from fusion.
https://en.wikipedia.org/wiki/Tsar_Bomba
Pure fission bombs tend to disintegrate long before enough neutrons have been generated to achieve complete combustion. The inclusion of a small core of lithium deuteride results in a much more efficient fission bomb.
I suppose these NIF tests are tiny pure fusion explosions; but I'm not aware of any practical weapon that generated energy mainly from fusion.
The primary is fusion boosted fission, the core ("spark-plug") of the secondary is also fusion boosted fission surrounded by dry fusion fuel (lithium deuteride) and then a fissionable tamper.
My understanding is that X rays from the primary compress the entire secondary package - which ignites the spark-plug and then the fusion fuel is caught between the incoming tamper and the exploding spark-plug at its core and ignited.
So there does appear to be a real "fusion" stage - with a lot of other steps involving fission and fusion.
https://en.wikipedia.org/wiki/Thermonuclear_weapon
"Fast fission of the tamper and radiation case is the main contribution to the total yield and is the dominant process that produces radioactive fission product fallout."
The way I read that, even in a large H-bomb, little of the yield is the direct result of fusion reactions. Rather, the neutrons produced by the fusion reaction dramatically increase the efficiency of the fission reactions, which on their own would not produce enough neutrons to fission more than a small part of the fissile material before the device disintegrates.
https://en.wikipedia.org/wiki/Neutron_bomb
Neutron bombs don't have fissile tampers on the secondary exposing the neutrons from fusion directly rather than using them to fission the tamper.
Edward Teller proposed something along these lines with conventional two-stage nuclear explosives back in the 1970s, but it was rapidly abandoned as horribly inefficient and costly:
https://en.wikipedia.org/wiki/Project_PACER
I know this stuff is fun to speculate about but the boring and unsexy option of more solar panels, wind turbines, and transmission lines is likely to be both cheaper and more politically saleable.
From what I understand, practical routes to fusion all produce a lot of fast neutrons, and tritium. The neutrons make the surrounding equipment radioactive, and create toxic elements. The tritium is itself toxic, and is hard to contain. And the high neutron flux means that such a machine could easily be used to create weapons-grade plutonium, by irradiating uranium. And come decommissioning time, the reactor itself is a pile of toxic waste that will need long-term storage.
I don't believe that fusion is inherently dirty; just that none of the technologies currently being investigated can be described as "clean". And they all run the risk of nuclear proliferation.
IANAP.
Even with D-T, the reactor waste would stop being dangerously radioactive in a few decades.
Because it produces no free neutrons, it doesn't have the (small) waste problem as this laser fusion approach.
This article is newer from the one that I remember.
https://news.ycombinator.com/item?id=30841080
Not to mention more practical and achievable in less wealthy countries.
And safe to use in politically unstable parts of the world.
It won’t fix these crisis directly, but it would reduce or remove a lot of the contributing factors.
Are all ultimately problems of energy being too scarce (and therefore too expensive).
Note: you have to account for costs of coordination and taking on risk - then you realize a lot of the "if only everyone would ${trivial thing}..." aren't done because they actually are prohibitively expensive or downright impossible.
The blocker to a better world is not some engineering impossibility, or market-rational overprice...not always. Often it's because someone with some power would lose out if we changed the way we did stuff, and they don't want that.
Human fucking nature. Can't cure that. Small bore example: Your engineering org would be better without the power-mad meeting-coordinator kahuna middle manager who holds hours long meetings for no purpose than to lord over proceedings and justify their own pretense of importance. The greater good would be served by demolishing said meetings, and removing kahuna from org. Does it happen? Often, no. Why? Because the world is not driven my moral, rational actors.
Food is a particular form of matter.
Therefore...
IMHO it's a milestone, not a breakthrough. This is because it is only a marginal improvement on previous results and it is very unlikely to significantly change the state of either the science or industrial products without further work.
Now we know it is possible, it has been done.
The next question is, can we create enough industrial efficiency to create a working motor generator set.
A hard problem to be sure, but proving that it is a problem that we have is itself a breakthrough.
We already knew that laser ignited fusion worked (that it triggers fusion). And we still don't know it can produce more energy than it consumes overall (one of the big questions for this sort of fusion).
Really, we have gone from being 0.5% efficient to 1% efficient (I believe).
One way to describe that is doubling the output. And that is technically true. But it's quite misleading as it will have no actual impact in the real world.
Add to that that this was quite predictable. And that we will need at least 3 very big breakthroughs before this is viable (much more efficient lasers, a continuous method and either replacement of the fuel or a very easy supply of deuterium). I just don't see this as a big deal.
That's why I called it a milestone. It's progress. But it isn't huge, unexpected, change-in-the-real-world progress.
For example, the maximum efficiency of a laser is limited by the quantum efficiency of the laser material, which is the ratio of the number of photons emitted by the laser to the number of charge carriers (electrons and holes) injected into the laser material. The quantum efficiency of most laser materials is less than 100%, which means that there is a limit to the efficiency that can be achieved.
In addition, the efficiency of a laser is also limited by the thermodynamic laws of thermodynamics, which state that it is not possible to convert heat or any other form of energy into work with 100% efficiency. This means that there is always some amount of waste heat generated when a laser is operated, which limits its efficiency.
In other words, fusion is unlikely to be using lasers for the ignitions.
UPDATE: People here seem to believe in self-sustained fusion, but this has never been tried or proven to be possible - no one know how much energy is needed for repeated ignition.
The current plan is to re-ignite new pellets fed into the containment.
That is not true. You can of course not make a 10 % efficient laser more than 10 times more efficient, but there is no limit to how much more efficient you can make a laser in general, you just have to start with an inefficient enough one.
And the NIF lasers are quite inefficient, according to Wikipedia - which might have slightly outdated numbers as someone pointed out to me - they turn 422 MJ stored in capacitors into 1.8 MJ UV laser light, that is an efficiency of 0.43 %. Bringing this up to the 20 % you mentioned would already almost be a factor of 50. Also not all of the gain has to come from more efficient lasers, the fusion process could also be made more efficient.
But you are of course correct that there is some general limit to laser efficiency, therefore the question is if the fusion gain can compensate this and all additional losses during the conversion into electricity further down the line.
Can we stop pretending this isn't about nuclear weapons research and - basically - continued funding?
LLNL is at least a decade behind its initial predictions of ignition. The current lasers may be inefficient but they've been fine-tuned very precisely for optical quality. There's no guarantee more efficient lasers would have the same characteristics and could be focussed in the same way.
I'm sure LLNL know the maximum theoretical fusion gain for a realistic pellet design, and it's worth nothing that that number hasn't been mentioned anywhere. Clearly there's only so much energy available for each cycle. If that energy is on the wrong side of what's needed to release it even after efficiency improvements, the entire system isn't workable, no matter how it's re-engineered.
The laser practicality issue that prevents this from directly becoming a power source would also be a major barrier to its application as a weapon. The laser fires in the UV-B (351 nm), which is scattered and attenuated by air, to say nothing of smoke or dust; it also requires incredibly high targeting precision (<2 mm target diameter) and consequently precise placement of a target weighing only milligrams. Additionally, the various optical components of the laser must be very well aligned, which is difficult to achieve in any battlefield conditions. And the whole thing is a very obvious and vulnerable target.
I ascribe a small possibility to its utility as a weapon in space, but practically zero on Earth without other major developments.
>I'm sure LLNL know the maximum theoretical fusion gain for a realistic pellet design, and it's worth nothing that that number hasn't been mentioned anywhere.
I wouldn't be so sure. Fusion is in general quantum chromodynamics, which is not so well characterized (being the subject of the famous YM mass gap conjecture). Even in this case it was stated that the yield exceeded expectations and damaged the sensors, which was probably not desired.
https://aip.scitation.org/doi/abs/10.1063/1.50451
Science is rarely a linear path. It's also rarely "I have this problem so I solved it this way." That's more engineering than science. It's really disappointing to me to see people, for lack of a better term, just shit on this accomplishment on HN when it's some seriously amazing stuff. It seems like a lot of people here don't understand science, they only understand engineering and then think about engineering mostly from a dull business and product oriented perspective.
"These and other scientific, technological and engineering hurdles will need to be overcome before fusion will produce electricity for your home. Work will also need to be done to bring the cost of a fusion power plant well down from the US$3.5 billion of the National Ignition Facility. These steps will require significant investment from both the federal government and private industry."
If fusion research scientists continue to insist that their research has any viable path to use in commercial power generation, and to demand large amounts on funding on that basis, then they should expect to be critiqued on that same basis.
We don't really know what path fusion research will lead to. It's science. We don't know what we could find out tomorrow that could apply this research. But even the promises of commercial power generation alone should be enough to keep funding the project whether it will happen in 50 years or 100 years. It's not like they aren't making progress. You can't rush research and also under fund them.
~$2 trillion was used to bailout big businesses. If money was properly accounted for, $2 trillion could fund approximately 87 Manhattan Projects simultaneously in today’s dollars.
Laser was cool, but nobody knew what to do with it. These days we're at the stage where we're thinking: what can't we do with it??? But for about 3 decades (before CDs, basically) laser was a pop science laughing stock, more or less.
And fusion is much harder plus has been talked about and hyped for at least as long.
Yup:
In college English classes I took, they wanted the students to write term papers. Ah, sure, they expected some review of some case of belles lettres, maybe Medieval French romantic poetry!!!!
Instead, in one case, I picked the transistor and another, the laser.
For the laser, I had no idea of the applications. For the transistor, all I knew was, what it seemed was all Bell Labs had in mind -- replace their usage of vacuum tubes, that is, analog amplifiers and not digital, Nyquist sampling, etc., even though Shannon was at Bell Labs, etc.
The idea of a few billion transistors on a sheet of silicon about the size of a large postage stamp, 16 cores, 64 bit addressing, 4.0 GHz clock, etc. -- beyond all expectation or belief. The graphics processors -- still less belief! Lasers sending trillions of bits per second per hair-thin glass fiber -- not even ready for science fiction!
No doubt I picked the transistor and the laser out of media hype. So at least some people in the media expected something from those two. Here the media was not wrong, and in the long term the hype was way below the reality.
No one is shitting on the research itself. But the PR around this story has carpet bombed sci-comm as a potential practical energy source. Not as basic research.
So it's perfectly reasonable to ask if there's a there from a commercial POV. And to note out that currently there really isn't.
If they really had announced a viable commercial product everyone here would be cheering.
The comment I replied to stated "Better to focus on standardised production lines for small fission reactors." To you, is that not shitting on the research? It implies the only purpose for this research is for power generation. And that it is clearly inferior to small fission reactors for that purpose, when the technology doesn't even exist yet. It just diminishes the accomplishment as a whole. It's so short sighted from a group of people who's jobs only came to existence <100 years ago.
Fusion is similarly incomprehensible to us here and now. There are untold advancements in materials sciences and engineering technologies that need to happen before it's possible so we have to invest in trying to so we can make those advancements in order to make it possible.
"In our laser-produced plasma (LPP) source, molten tin droplets of around 25 microns in diameter are ejected from a generator at 70 meters per second. As they fall, the droplets are hit first by a low-intensity laser pulse that flattens them into a pancake shape. Then a more powerful laser pulse vaporizes the flattened droplet to create a plasma that emits EUV light. To produce enough light to manufacture microchips, this process is repeated 50,000 times every second."
https://www.asml.com/en/technology/lithography-principles/li...
Transistors, microchips and lasers were not invented yet in 1943. How would one even guess that it could soon be a viable business to build a giant machine that shoots molten tin droplets at 50,000 Hz to produce a particular bandwidth of light so that you could create billions of tiny computers out of sand?
Hopefully fusion is on a similar path where the description of 2100's commercial reactor will sound similarly incomprehensible to us.
In the extreme you could say that today we have only one computer, the ASML computer [2], since they are the only ones making the machine that makes the machine.
[1] https://en.wikipedia.org/wiki/Aladdin_(BlackRock)
[2] ASML: TSMC's Critical Supplier, Explained https://www.youtube.com/watch?v=CFsn1CUyXWs
Well, that's easy. I forgot who mentioned it, but crypto has 0 support for actual clients, and the vast majority of day-to-day computing now happens on mobile end user devices (phones, tablets, laptops). So besides (what should be) this fatal flaw, the second thing crypto requires you is to manage your keys SUPER carefully yourself or a server.
Nobody[1]'s going to do that. I'm a techie and I don't want to maintain my own servers. What hope does Joe Locksmith have?
So in practice crypto will always centralize.
[1] 99% of the population.
It just needs to enter into state of self sustained reaction that is net positive.
Ignition itself can be horribly inefficient.
Nobody is designing fussion where ignition's output is directly used for the whole system to reignite again.
It would be equivalent to moving hand with match that fires a match, where energy from ignition is used to trigger another hand to move another match.
R&D labs aren't known for their industrial optimization unless that's their research goal. They proved this from the perspective of the fuel pellet - now they have to increase energy yield (larger fuel?) , reduce total system power, etc.
That said, they doubled energy yield in less than a year. The early exponential function on emerging technologies is always fun.
So it takes an incredible amount of precision engineering to produce these tiny diamond pellets of fuel which then produce an even tinier amount of energy while being destroyed in the process. There is no indication whatsoever that the pellets can be scaled up due to the incredible difficulty of scaling the precision geometry involved.
So I remain incredibly skeptical that this is anything more than hype for a project that's really about maintaining nuclear weapons stockpiles. I sincerely doubt this approach will lead to a real production power plant without some major research into how to overcome the requirement for extreme precision geometry and a similar effort to scale up the size of the reaction. Then throw the laser efficiency issue into the mix!
Could it be made from a metal that is in greater supply?
Could they really fire this design more than twice a minute?
Aneutronic fusion could possibly have a future, but hardly anybody is working on that.
Even in engineering problems, two people started a little electric car company in 2003. No, youknowwho wasn't involved yet. Would you have said "i can't see how electric cars can ever be workable" then?
Sooo, for Mars, just put the solar panels near the equator and fairly densely all around the circumference. Then, wow, have solar power 100% of the time!
Do the same on Earth??? Ah, have some jungles, mountains, two major oceans, and a lot of bad weather. For the oceans, sure, have floating solar arrays. Right, need to think about how deep the power cables to the shore would be, 3-7 miles down and then back up?
But the Internet has some cables across oceans!! Sooo, we could also have power cables???
Ah, sounds like only for Mars!
I've seen as many articles that says it's not as major of a breakthrough. It's marginal at best.
If they can make it work that is.
What most surprised me about NIF's ignition announcement was that this was even news. I'd assumed for years that most if not all tokamaks and stellerators could already achieve at least some brief fusion, just not to a useful degree. Turns out they're all just elaborate plasma heaters...
They knew all along that fusion energy would take decades to develop, and every year they need billions in funding to keep pushing for advancements.
They can't just go radio silent until they're like "we've done it", they need to regularly publish papers and make big media announcements etc.
Come to think of it, the LHC has been silent for a good while. Edit: It had a 3 year hiatus, started up again April this year.
https://physicstoday.scitation.org/do/10.1063/pt.6.2.2021102...
Why not announce that?
As I understood it, reliably igniting aneutronic fusion efficiently was an open research problem. Whereas, they can consistently get to aneutronic burning plasmas.
I’d love to read more about that.
https://youtube.com/watch?v=_bDXXWQxK38
I think they've done fusion in tokomaks and stellerators before, it's just that the energy applied to the plasma is greater than the additional energy that's created by fusion. It's not that no fusion is happening at all in the most powerful magnetic confinement devices yet built, it's that there isn't very much of it happening.
That's my layperson's understanding anyways.
Qplasma > 1 is significant because it now enters a regime where the burning fuel heats up unburnt fuel. In MCF machines this is referred to as a burning plasma. In ICF machines this is referred to as ignition. Ignition in MCF machines is Qplasma = infinity, where no external heating is used. This is on the table (maybe, theoretically), we just need to build machines with sufficiently high triple product and plasma control surfaces then learn how to do it. We'll have MCF burning plasma machines soon.
https://twitter.com/jb_fusion/status/1506964692627034118
Helion has not talked about triple product numbers. I wish they would, because that's the key metric for understanding how far they are from D-He3 burning plasma. Also, having the triple product history for all of their machines would help show scaling laws. These are currently closely guarded secrets. Targeting D-He3 fuel means a much higher coulomb barrier needs to be overcome and a higher triple product is needed to reaching burning plasma.
Check out slide 40 for more details on fusion fuels.
https://suli.pppl.gov/2022/course/IntroductionToFusionEnergy...
The output scales with the square of plasma volume and fourth power of magnetic field strength, which is why net power is expected from ITER (which is huge) and SPARC (which uses new superconductors for especially powerful magnetic fields).
https://www.youtube.com/watch?v=_bDXXWQxK38
1) How much energy per pulse is a commercial device expected to achieve?
2) How many pulses per second are realistic?
3) How many tons of copper/steel/capacitors per MW of power capacity are expected?
Because if you're gonna need a warehouse full of capacitors, several big turbogenerators worth of high quality steel and copper PLUS all the vacuum/plasma tech just to hit a few 100MW of continuous electrical power-output, then I frankly do not see how that would EVER be viable/attractive ANYWHERE, and NO amount of scientific progress might be able to change that...
But the concept to me at least looks more attractive than big magnetic-confinement plants, where the problems are even bigger (unaffordable plasmachamber + cryo-infra, super problematic neutron-flux, AND STILL needing all the heat => motion => electricity circus from a conventional plant).
Always hard to tell from publicity shots, but it looked like their existing capacitor bank was around the size of a shipping container. That probably goes up with the frequency though. I’m not sure you can do single or dual banks when aiming to fire every hundred milliseconds. Plus I think you need somewhere to send the produced power for dumping into the power grid.
But they’re already trying to produce their own capacitors to deal with that level of cycles per hour. I could see these guys spinning out a couple of companies that supply other designs or even industries. Especially if the money runs out.
We didn’t have the compute power for any of this stuff in the 80’s. I did some reading a while back and discovered that there are elements of Computational Fluid Dynamics that became state of the art around 1990, so we are maybe two decades behind, not five.
When i hear breakthrough i think of something more unexpected. I.e. figuring out how to solve a problem we had no idea how to solve.
The problem with celebrating "fake victories" (and being scolded for calling them out) is that when there really is a breakthrough (i.e. something so big it fundamentally changes our perception of the problem), people are going to think you are making just another over-hyped publicity stunt. I am definitely of the opinion that such media attention focused advertising of (otherwise perfectly valid) fundamental research is doing no one favors (not the researches nor the public.)
I've seen quite a few people arguing in (a European context) that we shouldn't build out nuclear in the "short term", "since we can just get fusion without any of the downsides of fission".
What I want to say is that bad faith arguments will always exist, and calling this a breakthrough won't significantly increase the number of people using bad faith arguments, because they would just find another excuse instead of the fusion one.
I share your skepticism towards commercial impact — but this is certainly a breakthrough in their technology. Just look at the chart.
https://en.m.wikipedia.org/wiki/National_Ignition_Facility#/...
Scientifically this is absolutely a breakthrough, just like the detection of the Higgs boson was despite them already having expected it or just like LIGO's detections of gravitational waves were despite that being entirely what it was built for or just as JWST's detections of galaxies much older than those seen from other telescopes was despite that being part of the entire point of dumping billions into it. Ingenuity's first flight was also hailed as a breakthrough despite that being exactly what it was designed to do.
In every case we had an idea of how to solve a problem and the outcome was generally expected. All that had changed was that the data had been collected of the solution working, just as it has been in this case.
To put it differently, it's a breakthrough because after decades of work, the National Ignition Facility can actually achieve the thing in its name.
NIF previously tied to use the hot spot energy instead of target energy in 2013, which they were criticized for, and rightfully so. But this definition of Q is analogous to the definition of Q used by MCF.
Fusion is not useful until Q>1, and for NIF to claim that when it obviously isn't true is, IMO, a bad look. They invented Qplasma where they arbitrarily get to ignore over 99% of the energy that actually went in, and don't need to worry about capturing any of the energy actually coming out. They got Qplasma>1, which is a great milestone, but the overwhelming narrative in the media is that they got more energy out than they put in, which is simply not true.
Someone, and almost certainly someone *ELSE* will get to Q>1 very soon. Lots of groups are targeting 2024-2025. That will be a major breakthrough, but in the public mind, NIF already did it, and I guarantee that's going to cause a lot of confusion.
It just feels like a marathon race where a bunch of competitors are competing honestly, and one participant jumps in a scooter, and blasts through the finish line, and the crowd is cheering, and they're doing interviews about how great it was to win the race, and people are going home, but the actual race is still happening, and getting super interesting, but no one cares anymore because that milestone was already claimed by someone who didn't even accomplish it.
Iirc most engineers use a different Q, which refers to all in energy, so Q > 1.5 ish is enough
So yeah, when people are told fusion is around the corner but all they hear are small improvements that to them mean nothing, its easy to see why. Now obviously not everyone is that way, there are communities of laymen and amateurs online who are interested and do care. But you don't need public press conferences to get to them. A press conference is for Joe Sixpack. And Joe Sixpack assumes when you come calling about Fusion its to say that it works, global warming has been solved and his electric bill will be next to nothing.
Even if you take the most breathless headlines about NIF at face value, we are obviously still decades away from a commercial fusion power plant. Let's say laser fusion is great and perfect, all we need to do is design a commercial facility that uses modern lasers and then construct it, with commercial turbines/etc, as well as commercial production of the fuel pellets. Just that will take decades. Even basic bitch coal power plants take several years to construct and there's nothing novel about those.
But the reality is that even with the best lasers available today, they wouldn't get enough energy out to make this commercially viable. Decades more development time is needed before they can even think about designing a commercial power plant.
No sooner than 50 years.
Milestone, yes. You have to get Q above one if you're starting at .01, and need to get to 100 and 1 is a great mile marker. But, But breakhrough - I don't think so.
See below for previous yields of NIF shots. https://en.wikipedia.org/wiki/National_Ignition_Facility#/me...
The other point that has been mentioned to me is when you in the self heating regime, there are exponential returns on increasing "quality" of a shot.
What does "component sizes" and "fill-tube size" refer to, BTW? And the quality of a shot thing you mention in the last line? Sorry for all the questions, just curious
https://www.orau.gov/support_files/2022ssap/presentations/Da...
Haven't H-bomb tests already proven that beyond all doubt?
> The fuel and canister get vaporized within a few billionths of a second during the experiment. Researchers then hope their equipment survived the heat and accurately measured the energy released by the fusion reaction.
It didn't necessarily blow up like a fusion bomb, but the energy was hardly produced in any sort of sustainable, ongoing form that we can harvest.
Worse - it's not even just the chamber that essentially destroys itself - the lasers that create the event also damage themselves: https://lasers.llnl.gov/news/controlling-backscatter-damage-...
Isn't the Qplasma > 1 the most important one here, by a very wide margin? Maybe the only important one at this point in time. That fusion begets more fusion, in a positive-feedback way. That's the breakthrough here. All the other efficiency factors are secondary in nature. That is step 1, and the next step is to scale this up so that eventually Qplasma >> 1. And only then the hand wringing about efficiency of the lasers becomes something to address. It seems like the baby was just born, and people are concerned about which colleges to apply to.
And, most significantly to me, there is no argument being made anywhere that this milestone is on the path to anything other than a dead end, local optima. That is, that you can further optimized the shape, quality and size of the hohlraum, and shape of the laser pulse to get to a Q that actually has something to do with power generation. (I have no doubt that achieving combustion/ignition is useful for NIF's real purpose, which is to simulate H-bomb physics to aid in maintaining our stockpile of city-destroying weapons - they have in fact created a nano H-bomb).
It reminds us how pointless our Javascript-based existence is.
I think the distinction is that with a breakthrough there is something new to learn, where a milestone you basically say "good job, keep up the good work". With a breakthrough there is more content to engage with because you can try and understand what is new. A milestone is great, but there isn't anything new to really engage with, so it has a shorter news cycle.
As a comparison - stable diffusion is a breakthrough - we are still trying to figure out what it all means and how it will or wont change society. its been months now and we are still talking about it. We wont be talking about this fusion thing months from now. That doesn't mean its less important only a different type of event.
I guess milestone assume you already know you will reach it.
I also think breakthrough as a lot of meaning and people would assume there is a few years before a real application.
In a sense it's probably a breakthrough within the scientific ring rather than from the people perspective.
1. a sudden advance especially in knowledge or technique
2. an act or instance of moving through or beyond an obstacle
Milestone:
1. an action or event marking a significant change or stage in development.
I'm failing to see why labeling it one or the other even matters.
But breakthrough, or milestone, I fear it may all be for naught. The improvements over the last two years are essentially all due to changes in the hohlraum geometry and preparation. All good, if your goal is to get burning plasmas to study. But NIF is orders of magnitude away from anything useful for power production. They have a Q of 1.5, but need something more like 100. They have a cycle time of, maybe, .1 shot/hour, and need something like 10,000/hour. Some real breakthroughs in the lasers (or maybe particle beams instead), coupling, and scale, not fine tuning of hohlraum shape, are needed.
The military said something like: "Telephone routing is point to point and inflexible, if the enemy cuts out 1-2 lines of communication, an entire section is completely cut off. We need something better." That something better turned out to be packed switching where you just throw stuff along a network and the network ensures that the packet reaches the destination, but you could theoretically have two packets going from Bucharest to Johannesburg, one through India and the second one through Canada.
It's much easier than the same operation for physical infrastructure.
If you look at the "shot history" of NIF -- the vast majority of their facility's energy is spent on actual DOD weapons research, not fusion power research that would be incidentally beneficial to the DOD:
https://lasers.llnl.gov/for-users/nif-target-shot-metrics
If this result is "useless" and "shouldn't be news," then the whole field of MCF and all the start-ups are even more "useless" and should be even less as news, because a small non-zero number is still larger than zero.
Contrast MCF with "inertial confinement fusion"[2] (ICF), a different (and apparently competing) approach taken by the National Ignition Facility in this latest announcement.
1. https://en.wikipedia.org/wiki/Magnetic_confinement_fusion
2. https://en.wikipedia.org/wiki/Inertial_confinement_fusion
The physics are known to work and have been demonstrated decades ago in a different engineering context (i.e. weapons).
A bomb? Already done.
A fusion reactor? We’re still light years away from anything of the sort. This is only one tiny tiny step there, and we’ve got miles to go.
EDIT: Being down-voted, why? Previous steps to actually develop a power plant around NIF-like inertial fusion were cancelled in order to focus on achieving ignition first: https://en.wikipedia.org/wiki/Laser_Inertial_Fusion_Energy
For sure, you can’t even consider the possibility of a fusion power plant without the core reaction producing power. That they figured out how to do the energy equivalent of getting the coal to actually light (finally!) in this context, is indeed useful.
But it is a tiny, tiny step towards an actual working fusion power plant.
And it does nothing to fix the economic issues here in competition with all the other energy sources, which matters.
Igniting that bit of ‘coal’ took near 100x the energy we got out of it. It was essentially putting it in a giant blowtorch, and then getting excited when it finally got lit and burned. That we previously only got half smoldering was indeed a problem! And that it finally lit is indeed cool.
it also took an extreme amount of thought, precision, investment, etc. to do it once.
Meanwhile, solar plants in the area were churning out billions of kwh.
No one needs to talk about delaying solar deployment, obviously that would be stupid. No one should be arguing for that, I certainly am not, and trying to do this whole thing where we attack genuine advances just because they aren’t the thing we prefer is super counterproductive to progress. Shameful to do that, no matter how common it is on social media.
This was built as a machine to test fusion, not to generate electricity. It did that, and doing so is a significant advance.
It’s sparking fire in a very specific synthetic scenario which maybe could be useful in making a specific type of fusion reactor.
It’s a critical step, yes, in development of that type of reactor, if we wanted to do so. Currently there is no reason to believe we’d find it worthwhile to do.
But there are literally hundreds more steps just as critical with no known solution, that are likely just as hard before we could actually have net positive energy even in a lab for such a reactor. Even if we found the economics would pan out.
That’s why I’m saying it’s a small step and we’ve got miles to go - because we do.
You’re pointing out some of the next small steps.
But those won’t get us to even having the energy balance positive on paper.
making lasers 100x more efficient for one being a notably hard problem, from a physics perspective. Probably even harder than IC fusion. The planned upgrades to lasers will help, but not enough to be close to a positive energy balance.
If the ‘first 90% takes 90% of the time’, we’re now at 2% of that first 90%. But everyone is hyping it up as we just completed it. Which is not true at all.
Honestly, it sounds like a Hail Mary funding push to try to get a new project funded or stop the lab from being shut down.
You want to use lasers to initiate some fusion, you can do that using a tabletop laser lab (although magnetic confinement and big voltages is much simpler).
So this is the first machine where we're sure: it wasn't fusion caused by lasers, it was fusion caused by neutron pressure! Where did that neutron pressure come from? From other atoms fusing!
Note also that the sun does NOT do this. Fusion inside the sun is powered by falling atoms converting their falling energy first into heat then into singular fusion events, which essentially explode and throw a lot of atoms back up. Then the cycle begins again. Fusion inside the sun is a self-limiting reaction, which is why stars don't just blow up once, but "burn" over a long time. All the fusion inside the sun really does is slow down gravity for a little bit, then immediately stop. Gravity is producing the heat from the sun.
[1] https://en.wikipedia.org/wiki/General_Fusion
The reaction is initiated by the lasers.
You can’t actually use a hammer to initiate it at any useful scale because hammers can’t create the appropriate pressures and temperatures to initiate the process at the scale needed to produce a useful energy output. Which is why the lasers.
So no, the lasers are not immaterial at all.
We still have no idea if we'll ever be able to engineer this thing to produce commercially useful energy. And we probably never will, for all interesting values of "never", the challenges are simply so massive.
The problem is the engineering of how to actually do it in a useful way. They have made some progress towards that but it's still extremely unlikely that laser based fusion will ever work as a power source because: a) they're still two orders of magnitude from break-even from the input side, and b) they don't have a practical way of capturing the energy that's produced at all as far as I understand it.
None of that has changed.
The way to capture the energy and make it into electricity was studied here: https://en.wikipedia.org/wiki/Laser_Inertial_Fusion_Energy
But it had been premature before they achieved ignition. You don’t build a boiler before you even have been able to spark a flame.
https://en.wikipedia.org/wiki/National_Ignition_Facility#/me...
Credit: Mark Herrmann/LLNL
From this article: https://physicstoday.scitation.org/do/10.1063/PT.6.2.2022121...
Clear qualitative difference.
Additionally, this is laser energy, but the lasers aren’t used directly. They produce X-rays which are the ones that actually drive the ablative implosion of the fuel pellet. If you used X-rays directly, the gain value would be MUCH higher… I think by that measure, they would’ve achieved Q=1 back in 2013?
Note that hey COULD use the lasers directly to drive implosion but haven’t yet. The Hohlraum/X-ray/indirect method I think produces more consistent lighting on the target, and is more analogous to the operation of an H-bomb, which is partly what this entire process is intended to replicate. They may do future experiments to try direct drive, in preparation for actually using this as a power production method and improving the efficiency dramatically… in combination with far higher efficiency modern lasers, this could enable Q high enough for net electricity production.
A real plant would need target yields 100 to 1000 times larger, with all the engineering issues that would entail.
It's so unlikely even under the sun's massive gravitational and heat that the power output is about 270W/cubic meter-- compare to human metabolism or an energetic compost pile, which is over 1000W/cubic meter.
We actually have to do a whole lot better than the sun, under far less favorable conditions-- to have practical fusion for energy generation on Earth. Inertial confinement fusion ignition (and to a lesser extent, thermonuclear weapons) strongly suggest this is possible.
The fact that black holes exist does not mean that we can spin one up to harvest it's hawking radiation for power. It's not just an engineering problem. That's not what that phrase means.
You certainly can, if you use a large enough chamber. Pacer was exactly this. You just have to accept high proliferation risk.
My crazy idea would be to use the heat and the pressure of these to pump water uphill. It is of course a crazy idea and you'd have to figure out a way to manage the pressure peak
But it would be very interesting
Hydro is the best, so good in fact that most of the good spots are already taken, because as you said, it requires a very specific geography.
The business model for pumped hydro requires relatively low cost to be economic and if it requires building a giant concrete water retention systems on top of large flat topped hills/mesas, near large drops in elevation, with water retention at the bottom of the large drop that is also really really expensive to build and requires specific geology to be economic.
And if something gets messed up, it’s a huge liability [https://en.m.wikipedia.org/wiki/Taum_Sauk_Hydroelectric_Powe...], as it can cause very damaging flooding and be expensive to repair/replace.
Worth doing when there is cheap power that can be stored and sold at higher rates later, but construction costs are really a huge part of it.
I'm fairly certain we have these multi level semi natural systems here in Norway, but I am not sure if we use them for pumped hydro.
(Og course, if one doesn't care about ecology, one of those levels can be the sea.)
This is a great idea, not crazy.
It ignores the steampunk future that awaits us all!
In 1972 they decided it wasn't economical based on fuel price compared to yellowcake, but fissile fuel isn't the primary cost of fission power and the engineering costs of pacer would seem to be much lower.
If you relax one of those requirements, then it becomes instantly clear that limitless cheap energy is possible. For a somewhat absurd example, imagine that we build Pacer plants in space instead of on Earth. There's no longer any size constraint on the blast chamber. You can go to triple-stage (like Tzar Bomba) or more designs, at which point the fusion fraction gets extremely close to 100%, and the cost per unit energy becomes somewhat hard to fathom. But on Earth, it still would have been expensive.
Hahaha! That is the funniest BS I've heard about it yet! Fusion has always been "only an engineering problem" since the first H-bomb and this approach might actually be the farthest from useful even with is.
Yes, I'd agree with that statement that it's only an engineering problem, but with a laser energy of 2 million joules and a fusion yield of only 3 million joules and an overall input of nearly 300 million joules to produce the lasers then I've no expectation of ever seeing a total net positive output in my lifetime let alone fully established working fusion plants providing power to ordinary customers.
Moreover, compare the overall 'negative' efficiency of this experiment with the actual positive efficiencies of existing generation systems: solar: 20+%, combustion: 20%-40%, mechanical/electrical turbine: >97% and so on and we see there's three or so orders of magnitude to catch up upon.
The task may not be impossible but don't hold your breath.
Some labs are already working with 20% efficient lasers, so hypothetically, Fusion is closer to “10->2->3” which is just amazing.
https://en.wikipedia.org/wiki/Mercury_laser
https://www.laserfocusworld.com/test-measurement/research/ar...
The second link has more details, but they show a 0.7% efficient NIF vs. a 7% efficient Mercury laser in the tiny little chart graphic -- which would reduce the energy for this experiment from 285MJ to 28MJ. This is way outside my area of expertise, but from what I understand, the same general tech that LLNL used to generate the HAPLS for Europe's laser research could be applied to the Mercury-like lasers and increase efficiency even further.
https://www.techbriefs.com/component/content/article/tb/insi...
Even if they couldn't increase the power past 100J (which they certainly can), using 2,000 beams instead of 192 doesn't seem too difficult given the much smaller overall energy and heat-removal requirements.
[0]https://en.wikipedia.org/wiki/National_Ignition_Facility#
Since a lot of modern laser research now overlaps with advanced semiconductor manufacturing, the competing systems are now much more mature. Here's a good rundown, but experts think that it's plausible to use even up to 20% efficient systems. Rather than using the laser to first generate xrays - these "direct drive" systems deposit the energy from the lasers directly on the pellets.
https://physicstoday.scitation.org/do/10.1063/pt.6.2.2021102...
We've been able to reliably trigger stable fusion in controlled experimental facilities since before I was born in 1966. People have literally been claiming fusion power is just an engineering problem longer than my entire lifetime, and I'm no spring chicken. Ring me when they reach engineering breakeven, although optimistically the chances are it'll be my grandkids that take the call.
Well, sort of. This is "theoretical breakeven" - the reaction generated more energy than they put in. "Technical breakeven", where you generate enough power to run the thing, is 2-3 orders of magnitude away. It's not at all clear that pulsed fusion like this is a viable power source even if it reaches technical breakeven.
The magnetic containment people are still struggling, but making progress. For a long time, everybody was looking for some clever magnet geometry that would yield stable plasmas, with limited success. Maybe active control will work. There are now people throwing machine learning at the problem.[1]
[1] https://news.mit.edu/2022/fusion-machine-learning-turbulence...
Not in the Expanse "pellets explode and produce electricity" way but in the Project Orion "explode nuclear bombs behind a shock absorber and ride the shockwave" way [1]. Just that this allows you to build much smaller and more manageable spacecraft (important since you probably still want to reach orbit with more traditional propulsion)
1: https://en.wikipedia.org/wiki/Project_Orion_(nuclear_propuls...
What we have are a bunch of R&D people, some of which are using ingredients for which the entire world combined possesses enough fuel to run a single power plant for only two months, and then we’re out.
https://youtube.com/watch?v=_bDXXWQxK38&t=1081s
They talk about why ITER has no viable fuel source, and about how they (Helion) are going to have to run two configurations of fusion plants in order to source their own fuel. One that produces a 10% power surplus, undesirable neutrons and the fuel for their power plants. They didn’t say anything but I suspect they’ll end up coming up with a use for the neutrons, perhaps making fuels for other processes. Someone will likely eyeball that waste heat as well.
Secondly I’m not sure what you mean by saying this experiment has only academic differences with a thermonuclear weapon. It seems the only similarity is they both generate a fusion reaction. The latter is essentially multiple bombs strapped together which is quite different practically.
The “breakthrough” is the proof that it’s possible. Everything else is now optimisation and engineering, and needs mostly time, sweat and money.
Solar and wind etc may just feel too "primitive" and nuclear too old and toxic and Chernobyl. Fusion though, once there's enough buy in and enough of a sense of it being doable, would fit our expectation of technological progress.
This milestone is liberating more fusion energy than photons used to start the reaction, but the energy source for the machine was electricity and it produced zero electricity. If you tried to convert the heat back to electricity you would be talking less than 0.1% of the electricity used being recovered at the other end.
https://physicstoday.scitation.org/do/10.1063/pt.6.2.2021102...
With modern lasers they would be closer to Q=1 without the funny energy accounting. But still several orders of magnitude away from commercially viable, which requires vastly greater efficiency than Q=1.
The worst fusion methods scales with the cube of the radius of the area that can be made to fuse. So we "only" have to get a 10x bigger area to fuse, or have the same area fuse for 10x longer (and make the facility not blow up or melt down from doing that).
Working fusion is surprisingly close.
Isn't additional stages in fission/fusion bombs just upping the mass with the result being higher yields? "higher yield" being a bit of an understatement as the range is like 100KT to 100MT (and beyond).
Can a multi-stage pellet be formulated? Ignition of an outer pellet provides some energy to get a more powerful inner pellet ignited.
/not even close to being a layman, let alone expert, on the subject
My understanding is that that is the whole point of the announcement. The initial laser induced fusion caused further fusion of the interior. That's why the gain is greater than one. And that's why there is excitement, the possibility is there to scale things up using larger/more massive pellets. To me, that's the real announcement, the laser initiated fusion created additional fusion. The whole break-even energy seems more like a red-herring. Also, with inertial confinement fusion you have the possibility of doing the harder fusions, (like aneutronic fusion), that magnetic confinement fusion would have a much harder time with.
I don't think the 1% system wide would matter if it just went on after the first few seconds.
A few seconds would be a major breakthrough indeed.
As my fear of Crichton's Gell-Mann Amnesia sets in, I'm being pedantic: either its about 545 times hotter than the surface of the sun (5.500°C) or they've reached "only" 0.55 million degrees Celsius.
Also: its only 519 times hotter than the surface of the sun when using absolute temperature (5.778 K).
Also the difference between a 4:00:00 marathon and a 3:59:59 is less than a 0.007% improvement but it's a huge victory nonetheless. Once someone proves an arbitrary numerical barrier can be overcome humans tend to make a lot of progress shortly thereafter.
But then if your announcements are completely rational you don't get as much funding. A lot of us nerds on HN don't like it when people do propaganda, but I assume we also like to win, right?
If your goal is to gain funding, the rational announcement IS to sensationalize.
None of that in the (look, it's still around) recent article. Insightful details are present. And some honesty ... e.g.
"The agency toned down the ignition objective, emphasizing NIF’s ongoing experiments to investigate materials’ behavior under extreme densities and pressures in support of nuclear stockpile maintenance."
[https://physicstoday.scitation.org/do/10.1063/PT.6.2.2022121...]
I may not speak for everyone but I certainly don’t find that comforting and know at least a few people who would agree.
https://www.sciencealert.com/koreas-fusion-reactor-ran-7-tim...
For a hint, the work was at the NIF, National Ignition Facility.
Even to start a wood fire, need an ignition facility. Then the fire continues on its own putting out much more energy than was used with the matches or whatever during the ignition. Then the energy from the fire ignites the wood that is not yet burning.
As we now know, once the nuclear reaction is ignited, the reaction generates more energy than was used with the lasers for the ignition. Then, sure, we can expect that the energy of the reaction should, without more energy from lasers, have the reaction continue.
Right, I'm ignoring a long list of what are sometimes called engineering details, e.g., that this is inertial confinement fusion!
Ah, at one time I worked for KMS, that is the company of Kip M. Siegel, as at Google and Wikipedia:
"KMS Fusion was the first and only private sector company to pursue controlled thermonuclear fusion research through use of laser technology."
https://en.wikipedia.org/wiki/Kip_Siegel
The report at the time was that KMS had achieved fusion neutrons.
Hmm ....
I didn't work on fusion but was hired because I knew quite a lot about the fast Fourier transform (FFT).
https://physicstoday.scitation.org/do/10.1063/pt.6.2.2018061...
> "In the latest round of experiments, the capsule shells consisted of diamond doped with a thin layer of tungsten, and the hohlraums were made of depleted uranium. Earlier experiments had used plastic shells and gold hohlraums. Sebastien Le Pape, lead author of the paper, says the uranium hohlraum boosted the peak energy deposited on the capsule by 25 terawatts, for a total of about 450 TW."
Clearly the cost of these fuel capsules is prohibitive when it comes to any practical power production system, plus there's nothing like an energy capture-and-conversion-to-electricity system being developed. It certainly looks more like nuclear weapons research than power production research. The real goal is just preserving the budgetary outlay:
> "The new results were achieved before the Trump administration proposed cutting funding for NIF by $57 million, to $287 million in fiscal year 2019. That would have forced a 30% reduction in the number of experimental shots of the 192-beam laser. Lawmakers instead added to NIF’s funding next year: The final appropriation will likely end up between the $330 million included in a House-passed bill and $344 million included in a Senate measure."
Notably, DOE budgets for solar PV development remain entirely minuscule in comparison, which is one reason why it's China, not the USA, that has mastered efficient production of high-efficiency, high-durability monocrystalline silicon PV panels at scale. These of course compete directly with natural gas for power production, and the US wants to be the biggest natural gas exporter in the world. Not hard to understand what's really going on at the DOE/NNSA, whose original name was the Manhattan Project, then the AEC:
https://en.wikipedia.org/wiki/United_States_Atomic_Energy_Co...
We need to focus on building hundreds if not thousands of nuclear fission reactors and to reduce the population growth through birth control outside the Western world.
fundamentally though, one would think that trying to create artificially solar interior conditions on Earth while the planet is literarily drenched in sunlight is not the shortest path to energy salvation. Life has invented already several ways to harness that energy and we have invented a few more in the meantime. I think the odds are that we will have more (as in: sufficiently) efficient harnessing of solar energy long before we create locally little suns to play with.