Ask HN: Room temperature superconductor: what to expect?
Suppose room temperature superconductor technology is validated and replicated. What to expect in the coming months, years and decades in the consumer space?
If there was a viable low cost superconductor that worked at temperatures up to 50 degrees Celcuis and atmospheric pressure, the implications would include:
- way less heat generated in electronics, so faster and more efficient CPUs, GPUs and memory.
- power transmission without loss. Very long range power transmission becomes viable and low cost (buried cables near roads?). For instance, solar energy transmitted in real time from where the sun shines to where it is dark or cloudy.
- a major obstacle to fusion power is removed as superconductors make it much easier, lighter and smaller to use magnetic fields to confine plasma
- medical imaging and scanning gets much cheaper as, for instance, MRI machines get smaller, lighter and simpler.
- mag lev trains could become the lowest cost way to travel by rail. Maybe even cars and trucks become maglev and roads are replaced by magnetic rail.
It's important to note that a superconductor is a superconductor not just as a function of temperature, but also the amount of current carried, the critical current density[1], and magnetic field strength[2].
If the superconductor has a really poor critical current density, and the indicated numbers seemed to suggest so, it might not be viable for a lot of interesting applications in electronics and similar.
[1]: https://en.wikipedia.org/wiki/Superconducting_wire#Important...
[2]: https://en.wikipedia.org/wiki/Superconducting_magnet#Magnet_...
For practical purposes this is only true if the cost of the material and support systems for the superconducting transmission is substantially less than the cost of the power lost in the existing system. Existing power transmission systems are already quite efficient, and we don't bother making them more efficient with HVDC because the infrastructure costs don't justify the power savings. It's almost certainly the case that superconductor cable would be significantly more expensive than that.
What we really need is more capacity, which this material as presented doesn't seem to help with.
I thought one of the benefits of superconducting wires was that they could carry a lot more. Is that generally not true of superconductors?
I doubt that there's ever going to be superconductors used commonly in long range power transmission. The economics just won't make sense, the conductive properties of the material are not the only ones that matter: (lifespan, toughness, flexibility, etc. matter a LOT.)
Fusion reactors will be a lot easier to construct when you don't need such ridiculous cooling needs, but I still don't get why we don't just build more fission reactors. They work great, and we could have been carbon neutral decades ago. You think fusion reactors will solve the construction cost problem?
Imaging and other sensors I think will be huge.
Man, mag lev trains aren't ubiquitous in the states because of the cost of running them. They are just almost impossible to build in the present political climate where the extreme cost of right-of-way acquisition through eminent domain needs and environmental regulation and litigation come into play.
It's already way too expensive to just build roads out of rock. No way we replace roads with magnetic rails.
I don't think DT reactors will solve this problem. If anything, they will make it worse. The volumetric power density of a DT fusion reactor is going to be at least an order of magnitude worse than a fission reactor, for fundamental reasons involving limits on heat transfer across surfaces.
This is why people are excited about Helion, which sidesteps the problem by extracting energy from the plasma electromagnetically rather than as heat (heat is still dissipated, but it does not have to be run through a heat engine.)
BTW, for transmission, superconductors would be more interesting for buried power lines, where voltage is limited and heat dissipation is problematic. Sueprconductors are also interesting in aerospace, particularly for hybrid electric aircraft.
Even without superconductors, advances in permanent magnets make some innovation practical. Here's an article about a portable MRI that sells for $250,000[1]
[1] https://www.science.org/content/article/mri-all-cheap-portab...
Edit: "65-millitesla" or 0.065 tesla.
- electronics need miniturisation, and wires - can it be made into wires? Can it do 10^11 features/chip? Can it form transistors? that's were most of the loss is currently.
- power transmission gets a little more efficient, and for long distance, which is already feasible with HV DC, it would be more efficient still. But in both cases, only if it's cheap enough.
- confinement is far from the only technical obstacle to fusion. The main obstacle is actually economics. It will simply never be cost effective compared to alternatives.
- Yes, smaller/simpler MRI etc is a real benefit.
- levitation itself is only a small piece of the puzzle, and again, cost.
The real benefits are in things like sensors, and current research where a strong magnetic field, or the ability to rejecting one, is needed.
But to touche your point, if ambient superconductors, then long distance transmission from remote solar panels, apparently.
There's just no way fusion is ever going to compete economically as mainstream energy, with ever cheaper solar (and wind, tide, etc), geothermal, and improved fission, at least with the current high-temperature plasma-confinement, and especially D-T, 'chemistry'
What if the RTSC is too brittle to make wires out of it or something like that?
In a few weeks there will be serious attempts at replication and probably some consensus in a few months.
They are reporting a material which does not have high performance in terms of maximum current, those samples aren’t going to outperform traditional conductors. Somebody might figure out how to make higher quality crystals, there may be enough details that you’d need to license several patents to make something useful.
Liquid nitrogen superconductors were discovered in 1986 but were very slow to find applications. Liquid nitrogen is easy to handle (we could demonstrate magnetic levitation in our high school physics lab) and nitrogen is an abundant element (most of the atmosphere) compared to helium which is common in the universe but rare in Earth. People thought it would be a revolution but it wasn’t. There are practical applications of them but when they build new power lines they almost always use ordinary metals.
Now it could be the other way around, the effect is real, people are able to make a high performing material at an affordable price, people solve all the practical problems, etc. Or it could be this round of materials is “close but no cigar” but in 10 years there is a real breakthrough. We don’t know at this point.
https://www.esd.whs.mil/Portals/54/Documents/FOID/Reading%20...
Rebranding is often a bad smell, such as “Facebook” to “Meta” or “Twitter” to “X”.
For anyone else reading this, there's an autocorrect mishaps here: it's supposed to be “cuprate superconductors”.
> people are able to make a high performing material at an affordable price, people solve all the practical problems,
What are the implications of that? What are the effects? How does that change my life? How does it change society?
Better particle accelerators, MRI machines, fusion reactors. (If we can greatly strengthen the magnetic field the reactor gets smaller are more economical.)
Better magnetic field sensors
https://en.wikipedia.org/wiki/SQUID
Possibly better logic gates post-Silicon
https://en.wikipedia.org/wiki/Josephson_effect
https://en.wikipedia.org/wiki/Rapid_single_flux_quantum
Better magnetic levitation rail
https://en.wikipedia.org/wiki/Maglev
When I think of all those things I think the RSFQ logic gates are most likely to be revolutionary.
* You can make charger cables out of it but it will make no practical difference, actually marking things worse due to subpar mechanical aspects.
* In-device interconnects, motherboard-level - maybe. Tiny bump in efficiency.
* CPUs etc - highly unlikely, we are talking dozens-atoms-level stuff, going to be really hard to replicate superconductor properties AND the usual semiconductor stuff
* Networking - this one could be great. No signal repeaters, faster (optical cables are slow due to weird curved path light takes).
Out if ideas. Anything else? Hi-power applications are more straightforward - better electric engines, etc. World-changing power transmission stuff...
P.S. Cold Fusion exists, just not practical atm https://en.m.wikipedia.org/wiki/Muon-catalyzed_fusion
CPUs etc - highly unlikely, we are talking dozens-atoms-level
stuff, going to be really hard to replicate superconductor
properties AND the usual semiconductor stuff
Maybe a RTSC based chip doesnt need to compete with modern chips in terms of process size or transistor count?
What if you could make an RTSC based chip with complexity and process size comparable to e.g. a 1993 era RISC chip but it runs at 500ghz or something
(Of course you'd need memory, or at least a heck of a cache, that could keep up)
Note silicon (probably any semiconductor) is about to hit the wall in terms of feature size. An obvious answer is to stack logic elements in top of each other but it is already difficult to extract heat from 2-d silicon chips. If superconducting gates really use 1000 times less power that stacking might be possible.
Personally, I want to own a SQUID[1] or two for some experiments, which is otherwise FAR outside my budget. They've been made with HTS, I see no reason why they couldn't be made with this.
The chemistry and processes can be made to be compatible.
So, in theory you could assemble a device with multiple SQUIDs on it, sampling the A field[1] in 3 dimensions, instead of the B field (the Curl of the A field) that is usually measured, just for one instance.
However there is one application that would likely be able to overcome any basic engineering or cost/performance challenges and get built relatively quickly (~3-5 years timeframe), in my opinion: The development of very small, room temp SQUIDs [0] for use in small, portable fMRI-type machines. Think a helmet you put on and have the ability to monitor spatial brain activity at very high resolution and without the need for massive liquid helium cooled machines. The machine in link [1] would be able to potentially be shrunken down into something the size of a football helmet and potentially be much, much cheaper, allowing for a cascade of new research into uses of such a machine in ways possibly outside the typically niche medical research.
At best so far from my understanding, this is a new starting point and milestone not a magic bullet.
And as you mention, they need to be reproduced and commercialized (i.e. mass produced or at least affordably produced).
People are allowed to dream and hope, but I'd rather not give into marketing and hype. To each their own.
Yeesh. What an embarrassingly confused, short sighted, and self-serving perspective of science.
I follow science news, but it is a little exhausting how much of it is bullshit.
Outliers and measurement errors disproportionately create interesting stories, and so things that don't pan out end up being overrepresented in the scientific press (like that supposed tachyon detection a few years ago).
There's a lot of stories about seemingly promising technologies that just don't really seem to go anywhere.
When these things get a lot of attention, they're followed by a breathless hype train and the attendant grifters.
A lot of it feels like a wasted effort. I suspect it gives some the impression that science itself is a giant grift where nothing useful is produced, and contributes to antiscience attitudes.
For some reason the poster you're responding to felt the need to add "immediately" to their interpretation of what I said to make my words seem ridiculous.
English is spoken very differently in the world.
The same way that we follow mirroring rules in the Southern Hemisphere.
For instance, in Guyana (South America) it's ⱯI, and in Australia, IⱯ
That's a joke, right?
In the next months and years the focus will be on "is this useful" and "how can we make this useful". To make use of a sc you need to bring it into shapes we can use. Wires of all sizes and shapes are the big one. Can this material be made into wires like current sc (bonding to a base material to fix the brittleness) or even better can it be drawn into wires? If not, that's a new research area.
Also, we still have no real idea how high temperature sc work (there are two competing theories, but afaik neither can fully describe what we see). That's probably even more important for this new ones. A flurry of research will be into related materials, to see if we can find ones which combine the sc with better properties, but having a theoretical understanding would make this far easier than just randomly poking at materials.
So, assuming we get to the state of usable room-temp sc everything else depends on the price. The lower the price the more they can replace until in the last instance all cables, coils and everything could use them (this will probably not happen for a very long time if ever). In the nearer future they could replace all of their predecessors in places where they are used today: In MRTs, in the LHC and comparable machines, fusion reactors, various cables, some small energy storage systems using them already. Such things.
They also could solve the biggest remaining obstacle for renewable grids based on wind and solar (storage) by allowing far better SMES: https://en.wikipedia.org/wiki/Superconducting_magnetic_energ...
Who knows where we'll go from there, but I think that's enough food for thought for the moment.
The reason is because while yes, the PRC produces and uses a lot of hydrophones, their range is a function of how noisy what they're detecting is, and US subs are very, very quiet. Networking and advanced signal processing can help, but only so far (and generally more with accuracy than sensitivity). SQUIDS are different, and exactly how far away they can detect a large metal body underwater is classified, but RTAPS would undeniably significantly improve their sensitivity and thus their range, meaning that whichever side can deploy and maintain a network of them (signal processing works for these guys, too) should have very, very good visibility on where any large bodies of metal are near their network, underwater or not. Both the US and the PRC would be able to capitalise on RTAPS for building SQUIDS, and both could base them in westpac (off the coasts of Japan, PH, Taiwan, ROK for US, eastern seaboard and their actually controlled islands for the PRC). So they both benefit from this technology and it's a wash, right? Wrong, both sides having this technology dramatically harms the usability of submarines in general, and the US is significantly ahead in submarine technology and submarines play a pivotal role in US naval doctrine, so this technology would significantly reduce the relative advantage of the US military compared to the PLA, which does not have submarines that are as militarily effective and whose offensive force structure is based around massing fires largely from the mainland, rather than using submarines to screen carriers which provide fires through sorties.
Basically, RTAPS improves everyones ASW suite, which sucks for the US because they're the kings of submarine warfare and submarine warfare is a large part of how they protect their assets near the PRC and how they project force into the PRC's expanding sphere of influence.
A room temperature superconductor will have different properties, maybe it can handle a higher magnetic field or not, we cant yet say. But what we do know is that the current limit to building HTS magnets is no longer the superconductor, but the strength of the supporting material which is steel. Find a stronger material and you can build a stronger magnet. A superconductor, better than current HTS, won't change that.
I don't expect anything until the results are replicated, multiple times, and confirmed in the peer-review process.