Let's assume they're right. Even so, they emphasize that the structure of LK99 is so finicky, so sensitive to a misplaced atom here or there, so prone to confusing performance in the real world due to inhomogeneity, that it's practically impossible to "get it right" with current materials synthesis technologies. It's something that literally has to be grown from the bottom-up, atom by atom --- in bulk, that's essentially a UFO-tier tech.
And that's if their models and assumptions are correct. If it works.
The way I see it, the best this material can do for us is advance our theoretical understanding of superconductivity. There's no grand theory of how molecular/crystal structure relates to superconductivity. But maybe we've gotten a little bit closer.
This isn’t that different to semi conductors, we’re at about the range of single atom doping already.
If this does prove as a path to high temperature super conductivity under normal or at least achievable pressures then it would open a whole new world of potential materials for material scientists to explore.
Our understanding of superconductivity is already incomplete we do not have a working theory for current high temperature (~100k) superconductors either.
UFO-tier or semiconductor fab tier[0]. But even then there are very few physical and economical situations where you can say "let's deposit X number of atoms per square nanometer of substrate" and not get laughed at.
> That's silly. 15 years ago, China celebrated finally being able to domestically make a ball point pen. Why?
> Because the steel ball needs to be a perfect sphere, and no one in China could produce such finely honed materials prior to then.
This rumor has been refuted many times. It is not hard core science to produce such "finely honed material", no one bothered to make it because the total market volume is way too small for typical modern steelmakers. Only 100 tons of such "finely honed material" is consumed in China each year and it is declining for obvious reasons.
There were heated debate in China for the particular matter - should China enter all such tiny specific segment to compete with every country. You don't need a smart brain to come to the right conclusion here.
> If something has value, and would this ever have value!!, trillions will be spent, over 20 years, making it cheap to do.
Obviously this is not the case.
Americans are not building high speed rails, not because it doesn't bring in the so called value, it is because such investment would threat the interests of those airliners. How much the US invested in high speed rail in the last 20 years compared to its total $ spent on wars?
It is also lagging on productions of EVs, maybe EVs don't bring in values? LOL
> And those improvements in manufacturing capabilities will spread through all otger industries, the entire ecosystem improving itself over and over.
until you get yourself de-industrialized and then it is over.
> Americans are not building high speed rails, not because it doesn't bring in the so called value, it is because such investment would threat the interests of those airliners.
Aquiring the rights of way is an enormous undertaking due to rule of law. Getting construction permits is an enormous undertaking due to environmental law. Determining the path and stops is a giant political mess because nobody wants to miss out on being on the rail, but if you let everyone on, it won't be high speed anymore.
The whole thing won't have a lot of return until decades after it begins and only if it offers good comparative utility and value when it's delivered, which depends on the state of roads, air travel, virtual presence, and other things that affect demamd for travel.
"Americans are not building high speed rails, not because it doesn't bring in the so called value, it is because such investment would threat the interests of those airliners. How much the US invested in high speed rail in the last 20 years compared to its total $ spent on wars?"
America is pretty bad at building other infrastructure, too.
NY subway does not threaten interests of airlines, but its expansion is notoriously slow and expensive.
>not because it doesn't bring in the so called value
It doesn't bring in so called value versus the investment assuming actual final costs of building such a system. Except the north east however the cost of buying land to build it there and revamping existing infra like stations is a massive. People keep wishing it did but a system that is slower (door to door) and less convenient (lack of local public transit) than airplanes isn't particularly good.
I could not believe this, so I looked it up. What I found was that in 2015 some Chinese minister complained that Chinese ball point pens were not on par with respect to writing quality when compared with non-Chinese ones. In 2017 they had this worked out and that made headlines in the West. Maybe there is more to this which I could not find right away, but as of now I could not find any evidence that China could not produce ball point pens 15 years ago, only that the quality was not top notch until 5 years ago.
Ya, it’s a quality issue. Sort of like with jet turbines: China has no problem producing fast ones, but is not yet able to produce fast ones that don’t need to be overhauled every flight. Or semiconductors: it can produce modern sized chips, but the yields are not good enough to be economically competitive yet. But these are mostly a matter of going through the same 50 years of experimentation of materials and process that other countries did.
For the balls in pens, I think there was one factory in Japan that did that very well, and they had no qualms about exporting them to China at reasonable prices, so China could make pens fine enough, just the ball used in n high quality ones (which, were the ones that would actually sell) used a ball imported from Japan. The minister was angry the pens weren’t 100% Chinese, rather than just 99% Chinese.
One can easily imagine that there are many such components produced in China that other countries aren’t currently tooled to produce economically, so it’s not weird or very unprecedented that China wasn’t previously set up to make perfect balls for pens.
Yes, basically the incentive wasn't there for China to make the pen ball, not that it was impossible. The same dynamic is in play with bringing back manufacturing to the US. China is the manufacturing center of the world because it's more cost effective to do so there than in other parts of the world, including setting up supply chains. If that incentive is diminished or gone then other places will emerge. It took 20 years for China to get there basically from scratch. It will take much less to rebuild the capabilities in previous manufacturing hubs.
What is silly is the uncritical belief that all problems are equally solvable. This seems to be a particular characteristic of software engineers, even though software engineering itself is replete with fundamentally difficult problems.
Having worked in research for a long time, once you've invested a lot of time an effort into some line of work, you're always looking for some way to continue, spin, or create the least amount of iterative difference in previous works to find funding or push an acceptable paper out. That's just the nature of the environment. Even if most the world moves on from something, the effort to change tracks is often higher than continuing down less promising routes.
Now, there is some value in continued work that ignores consensus or popularity. Consensus doesn't mean absolute certainty and historically big leaps have been made in areas people consider lost causes or wastes of time pursing. So, its good that we don't abandon searching the knowledge spaces entirely, but I will say that's often not the driving motive. It has more to do with careers and putting food on the table, such is the ecosystem that is research. LK99 won't disappear for quite awhile and people will continue novel research in the area.
> It's something that literally has to be grown from the bottom-up, atom by atom --- in bulk, that's essentially a UFO-tier tech.
The UFO-tears were also at our level at some point. How far out in the future are we talking before humans can do this? Is there anything fundamental that would prevents us from doing it?
Also, it's the replicator from Star Trek, isn't it?
> It's something that literally has to be grown from the bottom-up, atom by atom --- in bulk, that's essentially a UFO-tier tech.
isn't that how your CPU is made?
No, your CPU is made from the top-down, via lithographic techniques.
The silicon wafers are grown from the bottom-up, but silicon is as chemically simple as it gets -- a pure element!
LK99 would be tougher to grow (correctly, and assuming that there is a specific superconducting configuration from among several semiconducting configurations,) by at least several orders of magnitude. Growing pristine crystals of such a complex and entropic material is something that has never been attempted.
They did show a thin film sample in their paper, where they show the superconductivity measurement setup. They even mention it is a thin film sample of the material. looks transparent on the probing station. That is what they measured for superconductivity, not the rock/ingot form. That seems to be only a sort of magnetic demo for the material, nothing more. But the superconductive applications material seems to be thin film tech, as they state in their original paper.
not sure why everyone completely overlooks the fact that the superconductivity measurement from their paper is made on a thin film sample which nobody replicated as there's no manufacturing details for that version of the material.
whoever (from the original team) said that the ingot itself is to be tested for superconductivity?
Every new innovation is impossible at first, then possible but difficult, and finally trivial. If this research somehow crossed the barrier from impossible to merely difficult, that's a significant result.
This article contains various unfounded statements.
The manuscript that it refers to does not contain any calculations concerning superconducivity of the material, so it cannot say anything about it aside from vague guesses. There are no manuscripts I know on arXiv that contain calculations that suggest the material "should" be superconducting. This paper and several other discuss flat bands, but as noted in https://arxiv.org/abs/2308.05143 and elsewhere flat bands are no reason to expect superconductivity, especially the kinds that this material supports.
So there is no theoretical support for this material to be in any way good for superconductivity.
There is also no good experimental support. The crystal structure suggested in the original papers has been reproduced, but is found to be an insulator. Reproduction attempts of superconducting features have universally failed, and alternative explanations for the original measurements (which are fairly sloppy) exist. So there's not much reason to believe there ever was anything interesting there.
Was this not already pointed out in one of the theoretical papers during the initial replication attempt including the judgment that the interesting crystal structure is unlikely to be producable because it is not energetically favored?
What everyone following the stuff online seems to forget is that the replication attempts and samples so far are based on a leaked, apparently incomplete version of the LK99 paper.
Supposedly, a University in Taiwan is working to validate the full version of the paper and the material, and the authors of the paper say that once confirmation is complete (and is successful) they will release the full version of the paper... sometime in 2024.
Even just the bare minimal size for testing this concept conclusively?
Ok then, perhaps using ULV lithography (aka microprocessor design) could do it.
Pity those machines take billions of $$$ to work with, so it's unlikely someone would try. Hmmm, unless there is a research institute doing something like those lines.
With today's tech, yes. With some of the nanotechnology we think is very possible, this is one of the things we could potentially see massive gains in.
Theoretically, if you’re adding one atom at a time making a meaningfully thick object requires trillions of layers. Add 1 atom a second across a full layer ~= several years per centimeter of thickness.
Lithography is about as close as we can do, but that takes multiple stages to do each individual layer otherwise you can’t get patterns just undifferentiated layers.
If you don’t need atomic precision for arbitrary patterns there are many options. Crystals can grow much faster, but you have minimal control over how they grow etc.
That was already assuming an unachievable level of parallel behavior and operations per second. Something like 10^15th print heads and feed lines each slapping down 100’s of thousands of atoms per second, is still orders of magnitude to slow to be practical for human scale objects.
Further, that’s the tip of the iceberg in terms of difficulty here. It’s ignoring chemical interactions, thermal issues, sensors, control mechanisms etc etc.
I don't see why this is a theoretical limit. Clearly the existence of crystal growth and similar things show that there is of a much faster speed. Especially with repeating patterns those can be built in parallel and combined. Crystals can grow really fast. If it took so long for atoms to combine and make things then we wouldn't have anything actually made. You can definitely call molecular binding a "trick" when building materials from atomic levels. This is a parallelization method. And we can certainly do trillions of operations a second with atomic level combinations. Just take pure sodium and pure chloride and put them together. Or oxygen and pure iron. These reactions happen down at the "atomic level" and we can do these things at quite a macro level. So I'm just not understanding your argument here.
If the goal was simply to have whatever then you don’t need to do anything, a random rock is perfectly fine. However, if you have specific needs then that creates new limitations. As I pointed out you can grow specific crystal structures faster, but crystals don’t allow for arbitrary structures.
Trying to slap two arbitrary atomic structures together doesn’t automatically work. Sure, you could use some binding agent like glue etc, but now you’re limited to structures which include that binding material. Thus if the goal is to be able to produce arbitrary structures only limited by their stability you can’t necessarily join together separate pieces.
Pons and Fleischmann also announced they were still searching for the real killer after their initial claims couldn't be replicated. As far as I know, they still are.
No, because you're misunderstanding what electricity is and how it propagates.
Electricity is disturbances in the electromagnetic field. A difference between the electromagnetic field on the surface of the conductor, super or not, and the empty space around it. What conductors do is to propagate that field, perfectly in the case of superconductors, across their surface. It is not necessary for surfaces of pieces of the superconductor to actually touch to propagate that field. There can even be stuff in between, if it doesn't conduct it won't heat up and usually won't disturb the superconductivity.
(what it does is it slightly lowers the point, in current, where superconductivity breaks down)
So you can just stack them close to each other in our 3 traditional dimensions, because the superconductive region doesn't have to be continuous, just close enough. That will work fine, that's how it works in current superconductors.
I bet a friend a nice bottle of wine that LK-99 was bullshit. Looks like I’m never getting that bottle but neither is he. It is, obviously, bullshit and always has been.
It's a theoretical paper. We don't have a good theory about high temperature superconductivity, so they are calculating related properties like "flat bands". [It's not my are, so I'm not sure how good is the relation.]
My guess is that they started this during peak LK-99, and it got published just now, too late for the news wave.
I was very optimistic during the first week, but the evidence so far is not convincing. Unless you work in superconductors, you can probably just ignore this until some experimental evidence appears.
And that's if their models and assumptions are correct. If it works.
The way I see it, the best this material can do for us is advance our theoretical understanding of superconductivity. There's no grand theory of how molecular/crystal structure relates to superconductivity. But maybe we've gotten a little bit closer.