We should never forget who pays the bills here. Laser fusion, in the context of the NIF, means weapons research. It might seem unrelated, but using radiation/lasers to collapse metals onto hydrogen to initiate fusion is the literal core of modern nuclear weapons research. This is not about electrical energy production.
“Activities are about one thing and one thing only” is a fallacy.
It's good to remind people about weapons research interests. It's less good to completely dismiss the interests concerning energy production, or basic scientific research. They don't go away just “omgz military funding!”
There are no true interests concerning energy production related to inertial confinement fusion. It is completely impossible for ICF to ever have a cost/watt that would make it even remotely economical. A coal power plant burning diamond instead of coal would almost certainly have better economic fundamentals than ICF can hope to achieve.
So this is weapons research and basic physics research caked in the language of power generation.
You're basically right about laser-driven fusion. But ultrafast mode-locking lasers do have other uses, such as X-ray generation [1]. And MagLIF is a derivative technology that should have a significantly better energy in/out ratio [2]. The big problem with all existing ICF concepts is their tendency to damage the facility (boom!), though.
As far as I understand from the wiki page, the Z-machine still requires a hohlraum, which I assume also requires the kind of precise machining that the NIH laboratory does as well. As such, it is essentially burning an extremely expensive fuel, obtaining small bursts of energy from it - requiring vast amounts of such targets every hour in constant operation, which basically makes it a complete no-go from any kind of economic standpoint.
Note that I'm not saying that it isn't an interesting or useful scientific experiment - it's just not a promising power source.
We spend vast sums characterizing materials before they become cheap plentiful and profitable, this certainly could be an important step towards power generation without making a step in that direction directly. I imagine characterizing the likes of fusion it would be some comfort that it's inefficient. Question how do you avoid the great filter? Answer very carefully.
That’s assuming they ever even make it to net positive energy output. The article carefully avoids mentioning that the lasers they use are wildly inefficient. It is very hard to break even when you lose 90% of your input power right at the start
But the relevant question is, how efficient are equivalent modern lasers? NIF was built in the early 2000s and laser technology has advanced quite a bit in the past couple decades.
The lasers they use are extremely specialized, since they have to deliver outrageous amounts of power for a very short period of time. Advancements in normal lasers may not be of much use here
Edit: Wow, looks like they have improved considerably
Yes, and lasers that do exactly that have advanced very quickly. We're up to multi-petawatt picosecond lasers now. We're getting high repetition rates instead of the once or twice per day of NIF, by using laser diodes in place of flashlamps. There are dozens of extreme laser facilities operating or under construction all over the world. And if I'm not misinterpreting this, efficiency is as high as 60%:
The game played with inertial fusion is that when the energy people are skeptical, say it's about weapons, and when the weapons people are skeptical, say it's about energy.
Experiment is the arbiter of truth -- being able to verify that aging fissile materials continue to function as-expected is very important.
It really matters that nuclear weapons go bang when you want, but not when you don't.
An example of why one needs to care about such things in stockpile stewardship -- plutonium's own radioactivity means that helium is effectively being born within the crystal lattice, causing damage, distortion, and swelling. It isn't a guarantee that a bomb that worked 50 years ago will work today.
Oddly, the US Department of Energy's statutory responsibility for "the safety and reliability of the nuclear arsenal" is officially interpreted within the department as: safety, it goes bang when you push the button; and reliability, it makes as big a bang as it says on the tin.
Of course this "safety" has absolutely nothing to do with what the public naturally assumes is meant, which is typically that it won't go bang when you don't push the button, or leak, or be hijacked, or fall out of a plane carrying it. The DoE considers none of those its responsibility under the law.
One might guess that going bang when the button is not pressed would be extremely unlikely. But when a B-52 crew made a booboo and dropped a pair of hydrogen bombs in a farmer's field, investigation showed that all of the processes that would have set them off completed; just not all on the same bomb.
I am listening to Eric Schlosser’s book “Command and Control” at the moment and it has quite a bit of detail on this and other nearly disastrous incidents and the complete lack of serious command & control around nuclear material/weapons. We are very lucky not to have had several accidental or unwanted nuclear detonations to this point. I’ve not finished the book but I am not optimistic the situation is very much better today.
I'll admit it's just a hunch telling me you don't actually need experiments on extremely high temperature plasma to prevent a reaction that never naturally takes place on earth.
Intuitively, simply not building a bomb seems sufficient to prevent any accidental detonations.
Since it's not trivial to start a chain reaction that lasts long enough to create a sizeable explosion, unexpected (nuclear) explosions are a very unlikely failure mode for nuclear weapons.
And now you've made an abusive edit; your original comment here was mocking and pointing out that it was the GP "wishing" for a failure.
GP wished for a (moderately) likely failure: failure to detonate. GP obviously didn't wish for unlikely failures, like spontaneous detonation, ICBMs all flying to Washington DC, etc.
The US military has accidentally dropped dozens of nukes out of planes (this was on the front page a few days ago) and none of them accidentally went boom. They don't spontaneously ignite, they are designed with several levels of safeguards to prevent that.
You've misrepresented their point three times now, that's enough don't you think.
See 'the principle of charity' embedded in the guidelines as "Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith.".
Project Orion used fission bombs "the size of a soda can" to power its propulsion. Designing such bombs was a big part of the military aspect of the project and how it got government funding. Now the results are still classified, of course, but it appears they were successful from a design angle. At least the scientists involved seem to think it was possible. This source gives an estimate that a 1 kg fissile mass plutonium can yield about 100 tonnes of explosive power, which is a sphere of plutonium about 2 inches in diameter:
In principle there isn't really a limit to how small you can make it, just with decreasing efficiency and (with traditional implosive designs) a higher detinator-to-fissile material ratio. Things like laser implosion remove this design constraint, however, and would let you consider things like millimeter-sized pellets.
Project Orion did not use fission bombs the size of a soda can. It may have used bombs whose CORE was the size of a soda can, but to get that to explode requires much more high explosive (than in a typical nuclear weapon) to implode it more aggressively. That's fine, as the mass of all that high explosive acts as reaction mass.
This is a serious albatross round the neck of nuclear technologies; if they're dual-use, they'll get lots of funding, making them look expensive, while the true purpose and feature set is obscured, making them look nonsensical.
I wouldn't say boondoggle. The need for nuclear deterrent is back in focus, and so long as that exists then I'd much prefer we do anything but actual weapons tests.
What was the expected long term outcome of the nuclear test ban treaties? It seems unless there is some breakthrough in simulations and NIF type experiments, all countries with nuclear weapons will end up unsure of their weapons efficacy. Is this intentional? Is the hoped for outcome that no one has nuclear weapons they are sure will work and thus effectively disarming them? If not, are facilities like NIF what was envisioned by all the countries who have nuclear weapons?
I cannot speak for the people responsible for drafting those treaties, but it appears to have been a timely effort to apply multi-lateral pressure against proliferation. Put another way: for some nations to pull up the ladder. Not long-term thinking perhaps, but medium term at least.
In the longer term, who could have known then or even cared how to maintain the transition to a no-test regime? What the treaties' drafters might have been thinking is now moot. Those people are long gone, it's your problem now.
The University of Rochester's Laboratory for Laser Energetics was working on laser fusion back in the late 70s. Despite multiple upgrades, they never achieved the goal, eventually outgrew the facility and the DoE moved on to the NIF. It really underscores how difficult laser fusion is and how important it must be to merit pouring all that money into.
And even then, we are decades away from a commercial design due to the large number of engineering challenges needing to be solved. The next major milestone is getting a plasma gain of > 5, where helium-4 interactions become the dominant source of heating and dramatically change the behavior of the plasma. This has been simulated but never experimentally archived. Hopefully ITER will achieve this by 2030. Keep in mind that continuous operation with a Q plasma gain of at least 75 will be required for a commercially viable reactor. ITER will be lucky to hit 15 with upgrades through 2040.
Fusion would necessarily be a lot more expensive than fission. But fission is itself not competitive, and gets less so with each passing day. So, the longer it takes before fusion is actually, technically possible, the less value there is in it.
Fusion reactors are more complex in every possible way than fission reactors because of the sheer temperature differentials (super cool magnets next to hot neutron blanket).
Plus, they have much more parasitic power loss (running the superconducting magnets and cooling them to keep them superconducting, plus keeping the vaccum in the reactor chamber all require massive power, and they scale with the size of the reactor). You need some extra power to produce hydrogen through electrolysis as well, though that is probably lower. Note that making the magnets stronger to keep the vessel lower (so that everything is more easily cooled) will not get you much more, since we are already nearing the limits of material resistance for the supporting steel structures in the presence of the extreme magnetic forces trying to crush the reactor together - especially given embrittlement.
Not to mention, you will always need at least a small fission reactor to breed some tritium, since the fusion reactor will always have some losses.
Finally, the much much higher neutron flux that fusion expels (32x or more neutrons, and much higher energy per neutron as well) means that everything close to the reactor becomes brittle in 2-4 years. This means that things like support beams and the magnets themselves need to be constantly replaced. Even worse, they become medium level radioactive waste, which needs to be stored.
These are all intrinsic limitations of fusion that fission reactors just don't have.
> you will always need at least a small fission reactor to breed some tritium
The idea for future designs is to breed Tritium in the blanket surrounding the vessel. This obviates the need for external sources.
> everything close to the reactor becomes brittle in 2-4 years
I'm no expert on solid state physics, but that seems a little short? JET is more than ten times as old, and as we speak it's in its second run of D-T experiments.
> These are all intrinsic limitations of fusion that fission reactors just don't have
To be clear these are limitations of a specific kinds of fusion reaction and/or reactor design, primarily that of Deuterium-Tritium in a solid-walled tokamak. This may be a nitpick as it's currently far and away the most promising for energy production, and the alternatives are much further from any sort of workable prototype, but I have heard of them undergoing active research: tri-alpha (aneutronic reaction, meaning no activation or embrittlement of reactor components), liquid metal for the walls and divertor (in effect, continuously replacing neutron-bombarded material), etc.
All of your alternatives make it even more expensive to build and to operate, and so even less competitive. Regardless, extracting the few grams of 3H or 3He bred for fuel in the thousand tons of molten, radioactive lithium jacket, every day, is a problem no one has solved. Good luck with that.
It is conceivable that fusion could be made to work in outer solar system spaceship propulsion, where the constraints are very different. It will never generate commercial power on Earth. The billionaires pumping cash into fusion startups are being taken for a ride. They can afford it.
By contrast, Bill Gates got US taxpayers to pony up fully half of the scratch on his pet SMR project, without giving up any ownership. So, we are the ones taken for a ride, instead.
> The idea for future designs is to breed Tritium in the blanket surrounding the vessel. This obviates the need for external sources.
The problem is that, at best, you can create as much Tritium in the blanket as you put in as fuel (since every emitted neutron is coming from a Tritium atom). So, to be self-sustaining, every emitted neutron would have to be caught by the blanket to form a Tritium atom, and you would have to be able to extract every single atom of Tritium from the blanket back as usable fuel - and this is assuming 100% of the tritium you put in actually fuses, which is unlikely given how hard tritium is to contain (essentially every material is porous to Tritium). So, since there are losses at each of these levels, you need to inject new tritium into the cycle.
Also note that this entire blanket design is entirely theoretical at the moment: no fusion experiment has ever attempted to do anything with the fusion products other then measure the amount of heat generated.
> I'm no expert on solid state physics, but that seems a little short? JET is more than ten times as old, and as we speak it's in its second run of D-T experiments.
I'm no expert either, but these are the estimates I have read everywhere. JET is not in any way representative, as they do a handful of fusion events per year, for a few seconds - while a DEMO plant would be running continuously, 24/7. The amount of irradiation is incomparable.
> To be clear these are limitations of a specific kinds of fusion reaction and/or reactor design, primarily that of Deuterium-Tritium in a solid-walled tokamak.
These are all limitations of the only fusion electrical power-producing technology that is anywhere close to realistic.
All other fusion reactions require much, much higher temperatures and pressures to ignite, so they are many more decades away (regardless of what some snake oil start-ups are claiming).
All other magnetic confinement D-T fusion reactions have the same problems I discussed.
And inertial-confinement fusion approaches are much less likely to ever be economical given the huge costs of the actual fuel.
These are all excellent points—thank you for clarifying some of the parts I hadn't considered. In terms of the last and how practical it is, we're more-or-less saying the same thing; my hope was to be clear on what is a result of a tough engineering dilemma, as opposed to impossible or intrinsic, just so casual readers of this thread don't go away with the wrong idea.
Fusion blankets include lead or beryllium as neutron multipliers. When a nucleus of either is hit with a neutron, it releases two more. The resulting neutrons can still breed tritium from lithium.
CFS for example uses beryllium in a FLiBe salt. General Fusion and Zap Energy use lead.
Nice collection of very selective points. To respond to just a few:
So if something is complex we shouldn't or can't build it? OK. I hope you aren't an engineer.
Energy is needed to perform electrolysis to create hydrogen. What about the energy needed to mine uranium? I think it might require quite a lot of effort, particularly once supplies get harder to reach and extract.
The parts get radioactive and need maintenance and storage. The radioactivity is much more short term than fission. Fission fuel needs stored away from all life for 100,000 years. It will be a miracle if we manage to achieve that.
I think small scale modular fission is a good option, but your arguements against fusion aren't good.
> So if something is complex we shouldn't or can't build it?
We were discussing why fusion plants are necessarily more expensive than fission plants - not whether they can (or even should) in principle be built.
> Energy is needed to perform electrolysis to create hydrogen. What about the energy needed to mine uranium? I think it might require quite a lot of effort, particularly once supplies get harder to reach and extract.
I listed many other energy costs - including some uranium to breed tritium.
> The radioactivity is much more short term than fission.
Sure, but that still means decades for tritium and centuries for the neutron-bombarded materials - more than enough to make it as big of a problem in our lifetimes per kg, just with many more kg of waste from fusion.
He didn't list the worst part (for DT): the volumetric power density of the reactors will suck compared to fission reactors. So not only will the reactors be much more complex than fission reactors, they will also be far larger. How can this possibly be a cheaper source of heat?
Fusion's going to have to go with advanced fuels to have a chance. For that and various related reasons, I consider Helion the least dubious of the fusion efforts.
Isn't Helion claiming they can achieve non-DT fusion in a few years? That seems extremely suspect to me - as far as I know, non-DT fusion has barely ever been seen in a lab.
Yes, they do. An important point, though, is that the neutrons from DD fusion are much less energetic than those from DT fusion. This means they (like fission neutrons) are largely below the threshold where (n,alpha) and (n,2n) reactions occur, which eases the radiation damage problem.
So what? It's been a good idea for a hundred years. That doesn't mean it's simple to engineer in practice, but nor does it mean it's a poor idea that can't be achieved.
It's very funny to read that they couldn't even replicate the initial experiment, while there were scientific articles imagining how this could be a realistic power source given "modest assumptions" about how the hohlraums can be industrially cheaply built.
Fusion is hard. I thought of another use for that laser:
Take a small cylinder of pyrolytic graphite (it is diamagnetic, repelled by magnets) milled to be in the ideal axis and polystyrene layered in it. Then shoot it into a somewhat large ring of powerful magnets that have a trough in them to such a degree that it can trap the cylinder as it accelerates around the loop. The cylinder is propelled by successive shots of laser bursts that are not powerful enough to ablate (but close). One of the electromagnets can be quickly deactivated to allow for the cylinder to escape. Then it travels for a safe distance into very large lead target. Fusion could proceed if the speed is fast enough as the fuel is inertially kept together and friction should raise it’s temperature. If it produces neutrons, they will be absorbed by the lead. You can just collect the heat.
Edit. I think the lead target would be a no go. It would be no different than a bullet hitting a soft block of lead. Better would be a block of silicon carbide with some kind of borehole made of tungsten metal with fuel packed at the end of it. So the rice sized cylinder of nuclear fuel shoots right into the borehole and before the material can squeeze itself out, it has to overcome the enormous inertia keeping it going in one direction. Might be better though if the target itself was fuel. And it would be much easier to fuse lithium dueuteride than hydrogen and carbon.
I seem to recall a recent story on HN about something similar. So it isn't really a new idea. Perhaps making it run a circular course is different or that it doesn't use ablation for propulsion, it uses light pressure. I think the project was in the UK.
The interesting thing here is that publications like Nature continue to release this kind of nonsense. The PR campaign being run to keep these places operating must cost a much as the research
We should call him Doctor Hurricane. He must have heard this joke a million times already. He might be one suck joke away from becoming a supervillain.
Reading between the lines it seems the failure to replicate results is
down to a lack of precision engineering capability. Is anyone
who understands the experiment in detail able to comment?
Sounds like what goes on at the focal point of the laser requires
nanometers of accuracy in design, position and orientation. Perhaps
part of the problem is that we abandoned physical engineering and
manufacture in the west in favour of outsourcing it all?
This is a facility that is critical to US stockpile stewardship. None of the tech or engineering or manufacturing has been offshores/outsourced.
The engineering and fabrication side is just that precise.
In this facility they indirectly drive the target with the lasers. So instead of uniformly heating a sphere with lasers, they heat the inside of both ends of a gold tube (holhraum). This causes a bath of x-rays to uniformly heat a glass sphere that contains your fusion fuel. That glass sphere needs to be suspended in the center of the tube, but any inconsistency can cause huge knock on effects on yield. So they hold it between two <1um thick sheets of plastic. But if you can imagine, part of the sphere will be in contact with this plastic and part will not. That adds some reduction in how uniformly you heat the glass sphere.
But let's back up, how do they even make this perfect sphere of glass that is hollow? And how big is it? Well the sphere can be anywhere from 420um to 800+ um in diameter. From what I recall, the only way to make these is to drop small amounts of glass from a tall tower into water. Then do metrology on it, and toss away >90% of them right off the bat. Another issue is that some of the fuel leaks out of this hollow glass sphere over time. So timing from when you load the target from storage to actually firing the lasers can have non-zero effects on the yield.
Long story short. It's really hard to convey how close to the edge of precision ICF research is to the edge of micro fabrication. Now add the fact that some of these targets need to survive going from atmosphere to ultrahigh vacuum, and not break or tear... Remember that 1um thick plastic?
So yeah, no one is outsourcing this stuff. It's not being negatively effected by brain drain or lack of funding. It's just absurdly difficult to build to the designed tolerances the physics demands.
Fun fact, this is why a lot of physicists got behind the ITER project. The bigger you make something, the less important the minor fluctuations are to the overall yield. :D
I see. A similar problem to aligning the detonation wave in the early
plutonium bomb designs then, only with the speed of light as a
constraint. Tricky.
