[HN Gopher] A Common Semiconductor Just Became a Superconductor
___________________________________________________________________
A Common Semiconductor Just Became a Superconductor
Author : tsenturk
Score : 73 points
Date : 2025-11-05 05:33 UTC (9 days ago)
(HTM) web link (www.sciencedaily.com)
(TXT) w3m dump (www.sciencedaily.com)
| tsenturk wrote:
| Researchers have for the first time turned germanium--a widely
| used semiconductor--into a superconducting material by embedding
| gallium atoms in its crystal structure. This breakthrough could
| usher in a new era of quantum devices and ultra-efficient
| electronics.
| algesten wrote:
| > ...allows it to carry current with zero resistance at 3.5
| Kelvin (about -453 degrees Fahrenheit)
|
| Seems to me this is a problem.
| fnands wrote:
| It's an interesting result, but yeah, not a room temperature
| superconductor.
| zahlman wrote:
| For that matter, we've had superconductors for decades that
| work at much higher temperatures than this one.
| rwmj wrote:
| It seems the breakthrough is that you could use familiar
| semiconductor manufacturing processes. However the
| temperature is still going to be a major issue. I don't
| want a computer that requires liquid helium cooling.
| zahlman wrote:
| > you could use familiar semiconductor manufacturing
| processes.
|
| Unclear to me why that's helpful. Materials that
| superconduct at a higher temperature than this one aren't
| hard to come by, or obscure:
|
| > In 1913, lead was found to superconduct at 7 K,
| lazide wrote:
| Probably because they don't behave well for normal
| lithography techniques? The high temp superconductors I
| know of are weird meta materials, and good luck getting
| them to exist in chip form at all.
| pwg wrote:
| > I don't want a computer that requires liquid helium
| cooling.
|
| True, but I /can/ see someone, such as Sandia National
| Labs, very much willing to install a liquid helium cooled
| computer if it provides a significant performance
| increase above their existing supercomputer
| installations.
| throwaway173738 wrote:
| Isn't that very close to the practical limit for cooling in a
| lab?
| analog31 wrote:
| Not that hard. A dilution fridge, used for instance for
| cooling quantum computers, can go much lower:
|
| https://en.wikipedia.org/wiki/Dilution_refrigerator
| wizardforhire wrote:
| Thanks!
|
| Was gonna be lazy and say... temp or is doesn't matter.
| wasabi991011 wrote:
| Quantum devices are already cooled to that temperature (at
| least for some technologies), so it's not a problem in that
| use case.
| jeffwass wrote:
| Title is a bit misleading - it's not pure germanium that
| superconducts here, it's germanium doped w/ Gallium atoms.
|
| Superconducting germanium alloys have been known for decades, I
| used a Molybdenum/Germanium superconducting alloy in my PhD
| research 20 years ago, with much higher Tc.
|
| The interesting aspect of this current experiment is the precise
| alignment of the Ga atoms into specific points of the Ge lattice,
| so preserving the crystalline structure order which leads to some
| interesting effects.
| zahlman wrote:
| > which leads to some interesting effects.
|
| Such as?
| cbondurant wrote:
| Leaves me wondering if this will allow for superconducting
| cryogenic transistors? If my hobby level understanding of how
| silicon doping works, this new superconducting germanium would
| be a p-type? I could imagine something like ion implantation
| could be able to establish n-type regions within the germanium
| while allowing bulk regions of the lattice to maintain
| superconducting properties.
|
| Though admittedly, I'm not actually aware what parts of a
| semiconductor circuit are the biggest power dissipation
| sources, so I guess its entirely possible that most of the
| power is dissipated across the p-n junctions themselves.
| sevensor wrote:
| Yes, this would be P type. Boron is usually the P type dopant
| of choice. I'm not sure what role they have in mind for this,
| but probably to replace polysilicon and metals as conductors.
| What you have to watch out for is that this will make diodes
| wherever it bumps up against n-type material. This is a
| problem for metals as well, because you can get accidental
| schottky junctions, and we usually solve it with degenerate
| doping under the contract. I'm not sure what a junction with
| this material would do though.
| pfdietz wrote:
| > Boron is usually the P type dopant of choice.
|
| I want to note that in what has become the largest (by
| mass) application of semiconductors, silicon PV cells,
| boron has been replaced by gallium as the P type dopant of
| choice. Boron suffers from an annoying form of light-
| induced efficiency degradation that gallium avoids.
| sevensor wrote:
| Fair enough, my ion implant experience was DRAM / flash.
| I never worked on PV.
| metalman wrote:
| whole article is suspect in that it mentions cryogenic consumer
| products or maybe this is a slip and next gen refrigerator
| advertising will be run by a self hosting AI
| wasabi991011 wrote:
| I don't know if they updated the article, but I don't see any
| reference to cryogenic consumer products.
|
| They mention cryogenic electronics, which are used for high-
| sensitivity electronics in research labs and in medical tests
| (eg SQUIDs for magnetoencephalography).
| zahlman wrote:
| > For decades, researchers have tried to create semiconductor
| materials that can also act as superconductors -- materials
| capable of carrying electric current without resistance.
| Semiconductors, which form the foundation of modern computer
| chips and solar cells, could operate far faster and more
| efficiently if they also possessed superconducting abilities.
|
| Really? First I've heard of it. And it also doesn't make any
| sense, since defintionally a material can't be superconducting
| and semiconducting at the same time, any more than it could be
| conducting and insulating at the same time. Are they imagining
| some new kind of thermal-switching circuitry?
|
| This reads to me like the researchers came up with an irrelevant
| novelty (which is, to be fair, a valid and important part of
| scientific progress; it still expands our understanding of the
| universe) and Science Daily asked an LLM to rationalize it as
| useful.
| mschuster91 wrote:
| > And it also doesn't make any sense, since defintionally a
| material can't be superconducting and semiconducting at the
| same time
|
| I'd say it gets interesting if one can get at least _part_ of a
| die made out of superconductors. Getting power in into the die
| is a huge damn challenge, we 're talking about hundreds of amps
| for modern CPUs and GPUs - if even a part of that could be
| shrunk that would be a huge gain.
| HelloNurse wrote:
| Not if you spend that energy, or more, to cool the device
| hundreds of degrees below room temperature using liquid
| helium.
| dotnet00 wrote:
| The image on the article talks about making Josephson junctions
| with it, and the abstract talks about epitaxial superconductor-
| semiconductor devices.
|
| It feels like the researchers were mainly interested in
| applicability to Josephson junctions, and the article mixed
| them up with semiconductor junctions.
| yxhuvud wrote:
| .. at a temperature of 3.5K. So perhaps not super practical.
| s1mon wrote:
| It's not practical for your desktop computer, but a tank of
| nitrogen and some refrigeration hardware which fits in a single
| rack and you can run at 3.5K in a data center.
| AnimalMuppet wrote:
| Nitrogen freezes at 63K. That makes it a bad coolant for a
| continuously-running process at 3.5K.
| jandrewrogers wrote:
| 3.5K is well below the point where nitrogen is liquid. The
| only option would be helium.
| marcosdumay wrote:
| Just to point, but it would require actively cooled helium.
| You can't just drop it in liquid helium and expect boiling
| to cool your device.
| Pet_Ant wrote:
| Can you make a closed loop helium cooler? Also, that level
| of coldness seems like it would have negative interactions
| with other components.
| quickthrowman wrote:
| > Can you make a closed loop helium cooler?
|
| An MRI machine is a giant magnet with a closed loop
| helium cooler to keep the superconducting coils cold. A
| chiller is used to reject the heat outside.
___________________________________________________________________
(page generated 2025-11-14 23:01 UTC)