First Light (Oxford University spinout) achieves nuclear fusion
firstlightfusion.com
firstlightfusion.com
https://makezine.com/projects/nuclear-fusor/
And just like everyone else, you'll suck up a lot more energy than you'll produce
Yes, i meant to say power density rather than energy. My bad.
The only power I can extract from a fusor is the current generated inside a Geiger counter. But in my head, I know there's all kinds of cool fusion reactions happening. Gamma rays, helium, tritium, neutrons... that's all going on and it's just cool to know you're making it happen.
It is a tiny sun!
(Well, sorta kinda)
No apology necessary! I'm well informed on the content of the US mainstream media, and there I've learned that without any doubt all of
force, power, energy, volts, amperes, Watts, Kilowatts, Kilowatt hours, and current
are all just the same things, just synonyms for just the same things!
When I was studying physics, I thought that there were important differences, but now with a lot of exposure to the mainstream media I've learned that I must have been badly wrong!
Any neutron generator that emits 14MeV neutrons is doing it with D-T fusion.
A little engineering later:
https://www.youtube.com/watch?v=sf4qRY3h_eo
Fusion is a harder problem than that but we have no physical reason to believe it is not possible and the surrounding technology like compact higher temperature superconductors has advanced significantly since the 1960s and 1970s.
I am typing this on a computer with a ~5nm feature size CPU. Hard things can be done. It takes time, focus, and funding.
One of the easiest (laziest?) positions one can hold is simply to be dismissive of anything that hasn't happened yet, and which appears to be moderately difficult or harder. Fusion, AGI, etc... Just dismiss those things as ridiculous and you position yourself as wise, informed, erudite, whatever - to most people.
In fact, just yesterday had a conversation with a friend and realized at current prices we jointly spent several hundred thousand dollars on VPN services. So yeah, I was there before there was a there.
Suppose they get this working, and able to produce, what, 300 MW worth of hot neutrons. They have to capture the neutrons and turn them into heat to boil water to drive a turbine to get out 150 MW. Thus, handle the, what, 1000 tons? 10,000 tons? of lithium needed to capture all those neutrons. And, I guess, sieve it for tritium? Maybe chemically separate micrograms of Li-3H from the thousand tons of pure, molten, radioactive lithium? And, every year replace all the pipes the lithium runs in, weakened by neutron bombardment. By remote control, because strongly radioactive.
This is clearly a bigger job than what needs to be done for a fission plant, where all you need to handle is water and fuel rods. (If you think a 1000 tons of molten radioactive lithium won't need containment, allow me to disabuse you.) But fission is already not competitive with solar/wind + storage. In 10 years, fission will be even less competitive than today. There is no scenario where this ends up economically useful.
The history, measured by the fusion triple product, is exponential progress on par with Moore's law [1], despite abysmal funding [2].
[1] Figure 1, https://www.scipedia.com/public/Sanchez_2014a
[2] https://hardware.slashdot.org/story/12/04/11/0435231/mit-fus...
The current funding level, given the abysmal prospects for any return, is too high. It was even higher before. We'll never get any of that back.
It sounds like your issue is with the PR, not the technology. Is there something faulty or misleading with the progress made in the triple-product score?
> The current funding level, given the abysmal prospects for any return, is too high. It was even higher before. We'll never get any of that back.
30 years of fusion research is what a single Nimitz aircraft carrier costs. The "even higher" level was one Nimitz carrier per decade. And it only lasted one decade. Eyeballing the funding graph, the US has spent a total of 3 aircraft carrier's worth of funding for fusion in total, since research began.
Innumerable commercial entities are building out solar and wind farms as fast as they can scare up capital. Literally not a single purely commercially-backed nuke plant has ever been built in 70+ years. Not one. Capital did build and operate coal plants, at a profit. But nobody is building new coal plants, anymore. Even operating an existing coal plant is not competitive any more; coal plants are being shut down with no plan ever to re-open, exactly as fast as solar and wind come on line.
Nukes are made out of steel, concrete, plumbing, and pumps. None of those are getting cheaper. They produce power by blasting steam through enormous turbines, that need regular expensive maintenance, not getting cheaper. Mining and refining uranium is expensive and not getting cheaper. Solar and wind generation cost have been declining at an exponential rate for two decades, and are still falling. Can you even conceive of an exponentially declining cost not crossing any given constant cost?
Suppose you figured out a way to get power from nukes at half the cost, and that was less than renewables just now. How long would it be before they undercut that, again? Would you be able to finish building one, in that amount of time?
There is no future for fission, and even less for fusion.
Sure, let's check wikipedia to see how this is not true : https://en.wikipedia.org/wiki/List_of_commercial_nuclear_rea...
Also, after googling it: yes indeed, in some countries renewable energies are more competitive than fission. But that's not the case everywhere : https://www.iea.org/reports/projected-costs-of-generating-el... Or maybe you were just talking about the US, in this case your statement might be true.
I hope renewable energies cost will keep decreasing as much as it did those last ten years but I'm not as optimistic as you are.
And, mining and refining fuel has been massively subsidized.
HOWEVER, it was also understood that the cost per ounce of the resulting gold was orders of magnitude higher than the cost of gold obtained via lower-tech methods.
So there were no serious attempts to scale up the original process. Nor to improve it. Nor to develop "new and better" lead-to-gold conversion processes. Nor to otherwise squander vast sums and resources chasing the "but it IS possible..." dream of making real gold from mere lead.
Those were flying in the 1980s.
For instance, the M1 has a density of 171 million/sq mm and claims 5 nm 2D equivalent, while Willow Cove from IBM looks much worse at 10nm but its density is 100.
Saying this not to "cast shade" on certain chip fabricators, just saying that it doesn't actually mean that they have taken the same style transistor fab processes from ~10 years ago and shrunk them down to 5nm. I mean, by the time an actual 2D feature would reach ~2nm, we'd be talking about features that were only about 10 atoms across.
We have excellent physical reasons to think DT fusion will not be practical. The power density will be terrible, so it would be more expensive than fission at boiling water -- and fission isn't competitive these days.
(There's that one quote from a startup founder, "we did it because we didn't realize how hard it was", or something like that. From a pg essay maybe?)
So the headline is very precise: they achieved fusion, but not power production using fusion. And as far as I can tell, achieving fusion is not the hard part.
This is just a press release, but an unusually informative one. For more information on what they have and have not done, see here: https://twitter.com/FLF_Nick/status/1511374600575365122
(Link from apendleton: https://news.ycombinator.com/item?id=30923527)
They have good evidence of deuterium-deuterium fusion. IIRC this is harder to achieve than deuterium-tritium fusion, but as tritium is radioactive with a half-life of ~12 years, it is much harder to acquire and work with. In their concept of a power reactor, they would apparently use the deuterium-tritium reaction, which with the latter being created by bombarding a lithium blanket with the neutrons produced by the reaction, something that is envisaged in most other fusion power concepts.
More power to them! We need to test a myriad of approaches. This could well turn out to be the best one.
>> For the first fifty years after 1945, every published description and drawing of the Little Boy mechanism assumed that a small, solid projectile was fired into the center of a larger, stationary target.[31] However, critical mass considerations dictated that in Little Boy the larger, hollow piece would be the projectile.
I had never heard this before and was in denial reading the part above that. So either this key detail was kept secret for 50 years, or somehow history has been changed to confuse would-be bomb makers. I wonder how this detail came to light.
Edit: Following the reference is was this guy: https://en.wikipedia.org/wiki/John_Coster-Mullen
Making a nuclear bomb has nothing to do with the knowledge of its' construction. Detailed plans are freely available to anyone who is interested. The reason you can't make a nuke is that the enrichment process of a suitable amount of fissile material requires nation-state level of industrial output. It is physically impossible for a small rogue actor to make a bomb from scratch. Germany during WWII, for example, was far advanced toward a bomb years before the Manhattan project, but their industrial capacity was simply never sufficient to build it.
Is this still true in the 2020s?
The laws of physics haven't changed in 80 years. Even if you started with yellow cake uranium (~70% enriched, and itself already nearly impossible for a non-state actor to acquire), to reach weapons grade at >95% you'd need hundreds of tons of it, and massive industrial scale chemical facilities to convert that into uranium hexafluoride and pull out the U235 isotopes [0], where the ratio of U235 to U238 (the non-fissile isotope) is 99:1. The vast majority of the Manhattan Project was in the engineering challenges required to do this, not really in the construction of the bomb. So far, only 5 countries in the world have been able to do it.
[0] https://web.evs.anl.gov/uranium/guide/uf6/index.cfm#:~:text=....
I think you’re a tad optimistic there - sadly we have many more than 5 nuclear powers today.
There’s the list of “known” meaning publicly acknowledged countries with operating enrichment facilities, more than 5.
Notice it doesn’t include Israel which more or less everyone believes has nuclear weapons meaning enrichment facilities. There are several other countries which are very reasonably believed to have or have had enrichment programs.
You still need thousands of kilograms of natural metallic uranium, which is a rare heavy metal with few commercial uses and one really salient illegal one, or thousands of tons of uranium ore, which, see previous. It's hard to buy this stuff secretly, because not many people have it, and people who do have it have a big, salient, vested interests in not selling it to people who want to make nukes! It's not like selling drones to Saudi Arabia and saying "oh well guess it'll suck to be a Houthi". You might get nuked!
Nation-states can pull this off, since they have both land and state security apparatuses. Empirically, private groups or terrorist cells haven't been able to do it. The rumors I've read and the impression I've received is that AQA spent decades trying, lost a lot of lieutenants and trucks full of cash to CIA honeypots, and eventually gave up. Too hard, too expensive, too many dead ends and entrapment schemes.
By default things are legal.
I've also spoken with scientists who do research that, at times, abuts nuclear science, and have heard stories of them just... not doing paperwork because it's annoying.
To enrich uranium legally in the US you would have to get a license by justifying why you were doing it and there are not many legitimate reasons to do so outside slightly enriched uranium for power plants or highly enriched for a few medical isotopes.
But if you just decide to do it, yes it’s illegal and the NRC will probably find you.
Another possibity is building a breeder reactor first to breed Plutonium.
Also, where to get Uranium ores in the first place? You can be sure that governments keep close tabs on who takes funny stones out of certain mines.
I would STRONGLY disagree that the Germans were ever more advanced than the Manhattan project. The British were a bit (~9 months) ahead of the US before S-1 started (December 1941), but I've never seen any evidence that the Germans were ahead of either country at any point. Certainly the fact that they never managed to replicate CP-1, which went critical in December 1942 (and was still years of hard work away from an actual nuclear bomb) suggests that they were never very close at all.
The closest I can come up with is that the discovery of nuclear fission in 1938. In that work, there were five people who collaborated closely enough that today they would have all had their names on the one paper (because of Nazis the work was published in two papers, Aryans on one, Jews on the other): Otto Hahn, Fritz Strassman, Lise Meitner, Otto Frisch, and Wilhelm Traube. Hahn and Strassman were considered Aryan enough to stay in Germany. Meitner and Frisch escaped- Frisch ended up working on the MAUD report and then at Los Alamos, Meitner stayed in neutral Sweden. Traube did not manage to escape and died in Gestapo custody in 1942. So 20% of this team ended up at Los Alamos, and only 40% managed to stay in Germany, which does somewhat point to some of the underlying problems any German bomb project would have had.
Yes, but you don't know if those plans are viable and correct. This whole thing about the gun firing a cylinder is a case in point. People *thought" they knew how it was built sans some detailed measurements and such, but it turns out they were completely wrong on a rather critical aspect of the design.
Please, please don't say that. :(
Especially take note of Carey Sublette's comment on the design history in the blog comments: Mr. Sublette is definitely a name to be reckoned with, as far as unclassified nuclear analysis goes. His thought, that the design was basically taken direct from the Thin Man bomb design, is an especially interesting one.
As for why the unclassified world thought what it thought for so long, I always presumed it was because men in the 1940's naturally assumed that the rod moves into the long tube, not the tube moves to surround the rod. (Cut to shot of train going into tunnel.)
I know that technically it's a "scheme", but with the history of this technology they should probably use different language
Scheme itself means the same thing but it's fallen out of use and the only times I've encountered it is when the speaker wants to distance themself from it: get rich quick scheme, hare-brained scheme, nefarious scheme, malicious scheme, etc.
I note we here use schema, which is partly why I interpret a scheme neutrally.
I don't know how the word carries so substantially different moral leanings, but that's English. I'm sure there are many other words with evil twins.
There's a book published by some Brits in Oxford that very usefully researches the history of English words. They have found examples of all the meanings discussed here from the mid-to-late 17th and early 18th centuries, FWIW.
Edit: Fusion makes ~ 10^15 neutrons for every watt-hour of energy released (for the easiest kind of fusion)
Something like ... the difference between folding a paper airplane and designing and building an airliner.
- "With a huge research budget, we found a nifty new way to reliably set a few tiny lumps of coal on fire in our lab."
and
- "We can reliably build useful and practical locomotives, ships, and electrical generating plants which are powered by burning coal...and are long-term economically viable in a world which has several other ways of powering locomotives, ships, and electrical generating plants."
Except that with coal, making a far bigger fire is incredibly easy. With fusion, all the $Billions in the world don't seem capable of making even a modestly bigger...
But neither has fusion power shown anything even remotely resembling the real-world promise of fission power - which went from the first major attempt at a proof-of-concept reactor (Chicago Pile-1, Dec. 1942, ~1/2 watt thermal power output) to powering a large, high-performance warship (USS Nautilus, Jan. 1955, ~10MW on the propeller shafts) in just 12 years.
https://www.mpg.de/18250857/jet-fusion-facility-new-world-en...
The plasma was stable and they could have gone longer except instead of superconductors, JET uses copper coils that would melt if they went longer.
Their input energy was about three times their total output. But fusion output scales with the square of reactor volume and the fourth power of magnetic field strength, and modern REBCO superconductors can support much stronger fields than JET was using.
Meanwhile, a set of 5 30-year-old diesel-electric railroad locomotives can reliably put out ~10MW of usable electrical power (vs. thermal production). Vastly cheaper, with a proven track record and 100% duty cycle. (Generously figuring 3 running, 1 standby, 1 down for maintenance.)
( Wikipedia reference on JET: https://en.wikipedia.org/wiki/Fusion_power#1990s )
After 1997, the only way to scale up was reactor size, and that started with ITER, the 20-story-tall reactor in France. That soaked up a lot of fusion money, has been slow to build, and it's still not running. But more recently REBCO hit the market, and the same scaling laws say a reactor smaller than JET using those should get substantial energy gain. Two projects are building such reactors, and at least one will be ready around 2025.
(In any case, I wouldn't say five diesel locomotives are comparable to "burning a few tiny lumps of coal.")
The "bullet" is fast enough to compress the gas inside the cube, creating fusion.
It works. But in the scenario it does work, a machine is manually opened, loaded, prepared, and then they do the shot. Whole process takes days to prepare.
For it to be viable they need to do this every five seconds.
That is a hard problem to solve. First lights business model is not to solve that problem, but rather producing the "fuel", the tiny cubes with gas inside.
They say there is many details in how they are built which increases efficiency.
But someone still has to figure out how to build a machine that can continuously reload both the fuel and the bullet.
Conceptually, maybe nothing, but we aren't really talking about your average MG-42 here.
Wait'll they hear about what those folks at CERN are doing.
Finally, fusion in only ten to twenty years. I've been waiting ten to twenty years for this!
(be sure to click for its very relevant alt-text)
* has been "10 years away" for many years: could be tomorrow
An example would be "How long before a computer beats a grandmaster at Go?" The answer was "10 years away" for decades, right up to 2015, and then one day in 2016, that day was "today".
I've heard this a few times. In 2014 I was doing an MSc in Intelligent Systems ("AI" after the Winter) and Go was discussed in class in the context of Russel and Norvig. I don't remember the tutor saying that beating a grandmaster (? do they have grandmasters in Go?) was "10 years away". I remember him saying that Go was the last of the classic board games where humans still dominated machines because it requires intuition.
So, can you say where the "10 years away" quote comes from? Is it an actual quote? Do you know someone who actually said beating [a top human player] in Go is "10 years away" at some time before 2015?
>In early 2014, Coulom's Go-playing program, Crazystone, challenged grandmaster Norimoto Yoda at a tournament in Japan. And it won. But the win came with caveat: the machine had a four-move head start, a significant advantage. At the time, Coulom predicted that it would be another 10 years before machines beat the best players without a head start.
I'm sure there's more to find, but of course google now biases towards the articles about AlphaGo actually winning.
More to the point, when you commented above that "The answer was "10 years away" for decades, right up to 2015", did you have the Wired article in mind? I mean to say, did you read that article in 2014 and think that machines dominating Go is still 10 years away or is it more something you found with a search yesterday? Did you think in 2014 that machines dominating Go was 10 years away?
What I really want to know is what this "prediction" means. Was there really some kind of consensus on "10 years"? How seriously was this taken? It's all so vague and anyone could have said anything and meant anything.
But, ok, seriously, yes I am an armchair futurist, and also used to play go with my uncle (I am terrible at it), so since chess fell I've been waiting for go to fall too.
I am a believer in exceptional people and teams entering a space and turning established thought on its head. Another example would be spaceX. If you'd have said, ten years ago, that we'd have reusable rockets, that land on their tails like a 1950s sci-fi movie, you'd have been laughed at. On a boat! Ha ha! Crazy! Boeing still don't believe it, given what they've just managed to roll out. I've personally built my career on doing what others have said is impossible.
So, seriously then, sure, if enough people have been saying that something is still 10 years away, then other people have been listening, noticing the opportunity, working on it quietly, and a solution could happen later today.
My perspective on this is that Go (I don't know anything about rockets), was a technological advance that should not have been surprising for anyone who knew the state of the art. One reason I don't know how to interpret Coulom's quote is that he must have been one of the few people who understood computer go as well.
You'd say, yeah, that's the point, but if you look at the AlphaGo paper it really has nothing new in it. There is no sudden breakthrough in scientific understanding, in algorithm design, in efficiency or accuracy of a search, or anything of the sort. All there is, is Monte Carlo Tree Search and neural nets. That's as old as bread and about as exciting.
The only reason why it was AlphaGo that dominated Go, and not some other system, is that it was created by a company with the resources of DeepMind, meaning Google. It had to be such a big company exactly because there was no real advance in the scientific understanding of the game of Go, and so progress could only be made by turning up the compute to 11.
The same thing happenned back in the 1980's with DeepBlue and chess, also: it had to be IBM because few others could make a minimax system that searched so far and deep, because few others had IBM's computing resources.
And you know, perhaps that's what Coulom meant when he said "10 years". That it would take 10 years for the scientific understanding of Go to advance to the point where a machine can beat a human. It just so happenned that turning on the firehose and spamming the dollars at the problem made it go away, so there's no need to understand anything anymore.
Same thing happenned with chess, also. Alan Turing, Claude Shannon, John McCarthy, all those AI greats, they thought that chess was a measure of human intelligence and they wanted to understand how humans play chess using our human intelligence. Well, it turned out that you don't need human intelligence to play chess, as you don't need it to play Go, and a dumb machine with a blind search can look deeper into a game tree and find better moves than a human. But as far as I can tell, predictions of "10 years" for chess or for go or for anything else, were trying to predict when the scientific understanding of those games and of human and artificial intelligence would have advanced far enough that we could make an intelligent machine capable of playing go like a human, but better. That's what McCarthy would have wanted- he once made a bet with a Grandmaster that "in 10 years" a machine would beat the grandmaster in chess; and lost the bet. Because McCarthy was thiking about understaning intelligence, not winning games. And he underestimated the difficult of the former and overestimated that of the latter. Coulom, in the Wired quote, probably made the same mistake, thinking along the same lines.
If so, then he wasn't wrong, because while we have machines that can play better Go than humans, we still don't understand how humans played better Go than machines for so long.