Hacker News new | past | comments | ask | show | jobs | submit login
Hydrides come within a whisker of room temperature superconductivity (chemistryworld.com)
82 points by lachlan-sneff 48 days ago | hide | past | web | favorite | 38 comments



Is there a measurement for how much power a superconductor of a given diameter can actually conduct? Of course we're all familiar with the "resistance = 0" concept but I don't think that means you could pump ten billion amps through a superconductor the diameter of a hair, or does it? What other forces limit high-density current flow besides simple resistance?


What limits it is that superconductors can only handle a certain density of magnetic flux; more than that will make it stop superconducting. (This property was used in the "cryotron", an electrical switch used in some early computers before transistors became common for the same function.) Any current-carrying wire produces a magnetic field; so, with a strong enough current, the magnetic flux will cause the superconductor to switch itself off. This can be limited by using a larger superconductor, so the flux density is lower.


1) what was never clear to me was if the resistivity of an infinitesimal volume element of superconductor is strictly a function of temperature, and the norm of the magnetic field, or also of the direction of current with respect to the direction of the magnetic field? near the critical current density, or even before, will the resistance depend on the direction of current?

2) while a general bulk high current superconductor is limited in currentcarrying capacity this way, most of the time one is not interested in sending current one way, but also completing the circuit and getting current back the opposite way. Why can they not make power distribution cables with indefinite currentcarrying capacity the following way: consider a "multilayer" coaxial cable where the conductors are superconductors, and to prevent a magnetic field from arising the layers have alternating current directions, so that in any layer the net magnetic field is below the critical magnetic flux? (this would only be for power distribution, since this can't be used to build stronger superconductive electromagnets). The total current would be the sum of all the even, or of all the odd coaxial layers (the 2 sums are identical since they complete a circuit...) in essence co-locate the 2 problematic magnetic fields such that they cancel... A superconductive "humbucker"


What happens if a conductor (regular wire, not a superconductor), is wrapped in a magnet that "matches" the magnetic field produced by the wire?


You can do this to enhance or change the magnetic field of the magnet. Core rope memory and such use this principle to read, write and refresh the contents of the cores. They send a pulse down one of the lines, read line sends one voltage, write line sends a higher one designed to flip the polarity if necessary, and then sense the reflected pulse on the third wire, called the sense wire.


Thank you for the explanation. I'm still confused. So if a wire is wrapped in a specially crafted magnet, then when sending an electric pulse through it the pulse will "bounce back"?


The term "critical current" is often used for the current that generates the maximum magnetic field a superconductor of a given size/geometry can sustain. In practice the actual maximum current before quench is much lower (orders of magnitudes in some cases, I've heard) due to imperfections in metallic lattices.

This [1] is worth a few listens.

1. https://omegataupodcast.net/285-superconductivity/


Magnetic fields, which are created by the current, destroy the super conducting effect.

The colder you run a superconductor, the more current it can conduct.

T quoted for superconductivity is the phase transition for I = 0 A


The limiting factor is the critical field of the superconductor. Superconductors tend to expel magnetic fields, but if the field is strong enough it will bring the material back to normal conduction. Since a current will generate a magnetic field, this is what sets the maximum current.

e: looks like you got your answer haha


I'm not an expert, but I believe that after a certain point, the magnetic field created by the wire destroys the superconducting state.


This article is from January but things do seem to be suddenly getting exciting around room temperature superconductors again. The preprint mentioned in the paper became an article in Nature in May. Moreover, the fact that there now seem to be way to design superconductive materials rather just stumble into them could be a game changer.


A few months back, a team reported that they'd found room-temperature superconduction in gold/silver nanoparticles. Unfortunately, there has been no more news about it since.


I wonder if the path to room-temperature superconductors involves finding ways of containing hydrides at very high pressures, rather than finding materials that superconduct and ambiant temperatures and pressures.


I have the feeling it will be more economical to just use a low-temperature superconductor and cool it down with a closed-cycle cryocooler, rather than try to get a large enough pressure vessel that can maintain such high pressures.


A few years ago, that was the thinking when a 'high temperature' (much colder than room temperature) superconducting cable was trialed in Germany [1].

[1]: https://www.tdworld.com/overhead-distribution/ampacity-proje...

[2 with follow-up presentation]: https://snf.ieeecsc.org/abstracts/stp475-ampacity-project-%E...


With current technology, definitely.


You want high temperature superconductors because they can stand larger magnetic fields. So, it is certainly easier to keep them cool than pressurized, but the cold ones are much less useful.


Wrap them in carbon nanotubes :)


My immediate thought was to wrap them in some sort of diamondoid shell through some sort of mechanosynthesis process, but maybe a carbon nanotube would be stronger?


couple it with some double walled carbon nanotubes and you can get upto 700 GPa

https://www.nextbigfuture.com/2019/01/fabrication-of-ultrast...


I wonder if one could just stuff the tubes with hydrogen, then lanthanum powder, then just let them expand? They might auto-heat from the pressure, and assemble themselves, if you just unfroze them.


> Attaining superconductivity requires the diamond anvil cell to exerts pressures of up to 200GPa, two million times that of the Earth’s atmosphere.


At high pressure might be to superconductivity research what in mice is to medical research.

That doesn't mean that it is not useful to conduct these experiments. They could give further insights into how superconductivity comes to be - or maybe discover a superconducting material that requires high (but not too high * ) pressures to form but remains stable when they are lowered back to normal pressure similar to how diamonds are stable at normal pressures but are require high pressure to be formed.

* Pressures only reachable by using diamond anvils would be highly impractical to produce any useful amount of material ;)


I think he must meant that the headline is misleading. The holy gain is room temperature superconductivity at room pressure.


Depending on the application, mechanically useful pressures and higher temperatures might be good enough.


"Room temperature" already makes this concession as it means above 0C. In the context of superconductivity I would assume "room pressure" to fairly encompass pressures higher than 1 atm, though I don't know if there's a practical cut-off as there is for temperature.


I wonder if there might be a way to utilize the same effect that creates a Prince Rupert's drop to create that sort of stress in a superconductor. Sure, it would be fragile but it would also be stable without external forces applied. The stresses contained within the surface layer of the bulbous end reach as high as 700mpa which might be increased by using better materials and colder "shock" processes.


I would love a CPU that ran at room temperature.

OK, it's a cart and horse affair as it is the resistance that causes the heat, but I'm mindful how lab work pans out into real-work. But even reducing resisting a bit in a practical, consumer useful way has massive benefits.


here you go: https://www.arm.com/products/silicon-ip-cpu/cortex-a/cortex-...

> OK, it's a cart and horse affair as it is the resistance that causes the heat,

Most of the heat comes from semiconductor conduction, leakage, and switching. Superconductors don't fix that. Not to mention, a major factor in resistance is diffractive losses is the ohmic layer surrounding transistors. The ohmic layer is necessary because direct metal-semiconductor connections create blocking diodes. The width of those wires at the connection points is so small that electrons could barely squeeze through. That was the reason we switched to "high-K" dielectrics, which have higher classical resistance but lower electron scattering. Superconductors have similar problems with entry and exit into superconducting areas.


Technically, yes, "resistance causes the heat" - but why? Well - because in a CPU, voltages are being manipulated.

Those voltages represent binary values. Usually a positive voltage is designated as a binary "1" and a zero voltage as a binary "0". Putting a "1" into a register (or whatever) is easy - just add the appropriate voltage. But how do you change that "1" into a "0"?

You do so by "grounding" the circuit - that is, a circuit is made where the voltage level representing the "1" is connected to the reference voltage level that represents "0". Since that voltage level is "0 volts" (usually - but not always - but it is usually lower than the voltage level representing "1"); that change in voltage happens to occur across a conductor, and that conductor has some resistance, and that ultimately is what generates the heat.

ASIDE: It is possible to reverse all of this - that is, use a negative or zero voltage to represent a "1" and a positive voltage to represent a "0", and so on - I'm not sure, though, that this is done in common practice, but I am certain it has been done somewhere at least once.

So - what does that mean? Well - it means that in a CPU, just by the fact that it is manipulating voltages that represent the binary symbols "1" and "0" - it is going to end up generating some amount of heat. The fasting these changes occur, the more heat that is output. The larger the voltages involved, the more heat is output (note how this explains why, as CPUs have gotten faster, their operating voltages have become lower).

How do you combat this? Well - it ain't easy - but what if you could, instead of grounding a voltage representing a "1" to make it a "0" - you instead used it elsewhere (in some manner) to generate a "1". That is, instead of wasting that voltage and energy as heat due to resistance, you used it instead to perform the opposite function elsewhere in the CPU?

Well - you can - though it ain't easy from what I understand. It's called:

https://en.wikipedia.org/wiki/Reversible_computing


To vastly simplify: Since one signal is zero resistance (zero loss) and one signal is infinite resistance (zero loss), switching between the two requires a finite time spent at a resistance between the two (nonzero loss). Every change requires a change in entropy to represent it. This is also known as Landauer's principle: https://en.wikipedia.org/wiki/Landauer%27s_principle

> How do you combat this? Well - it ain't easy - but what if you could, instead of grounding a voltage representing a "1" to make it a "0" - you instead used it elsewhere (in some manner) to generate a "1". That is, instead of wasting that voltage and energy as heat due to resistance, you used it instead to perform the opposite function elsewhere in the CPU?

That's different and wouldn't really apply to a concept like resistance. It also doesn't need to be the opposite function, just... something.


>>but what if you could, instead of grounding a voltage representing a "1" to make it a "0" - you instead used it elsewhere (in some manner) to generate a "1"

that leans towards reversible computation, which would need a reversible computer (getting close to unitary transformations in QM)


Is it correct to think that because of entropy of heat, it is harder to make atomic structures stable enough to form a perfect path for electrons to run without loss at higher temps?


I hope this is the right description because it would make such intuitive sense.

But superconductivity is also a quantum phenomenon so it probably defies intuition.


I ponder if the sun is a really large super conductor? Since the sun has really high pressure like in the article and it has largely a lot of hydrogen like in the article. https://en.wikipedia.org/wiki/Sun


I don't know if it is superconductor, but plasma is somewhat conductive because it is consists of charged particles.


It's also a lot hotter than room temperature, especially where the high pressure is.


Do these concepts even make sense at the temperature of the sun?




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: