115
China Just Made the World's Fastest Transistor and It Is Not Made of Silicon
(www.zmescience.com)
this post was submitted on 05 May 2025
115 points (99.1% liked)
technology
23781 readers
159 users here now
On the road to fully automated luxury gay space communism.
Spreading Linux propaganda since 2020
- Ways to run Microsoft/Adobe and more on Linux
- The Ultimate FOSS Guide For Android
- Great libre software on Windows
- Hey you, the lib still using Chrome. Read this post!
Rules:
- 1. Obviously abide by the sitewide code of conduct. Bigotry will be met with an immediate ban
- 2. This community is about technology. Offtopic is permitted as long as it is kept in the comment sections
- 3. Although this is not /c/libre, FOSS related posting is tolerated, and even welcome in the case of effort posts
- 4. We believe technology should be liberating. As such, avoid promoting proprietary and/or bourgeois technology
- 5. Explanatory posts to correct the potential mistakes a comrade made in a post of their own are allowed, as long as they remain respectful
- 6. No crypto (Bitcoin, NFT, etc.) speculation, unless it is purely informative and not too cringe
- 7. Absolutely no tech bro shit. If you have a good opinion of Silicon Valley billionaires please manifest yourself so we can ban you.
founded 4 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Not viable because bismuth is radioactive. We're not moving away from silicon anytime soon, I mean we've heard of the hype stories about gallium nitride and carbon nanotube transistors every few months now for nearly a decade, yet almost everything still uses silicon.
I mean, I work with transistors and gallium nitride transistors are actually a thing. Like you can just buy them yourself. I've looked through a few examples and they seem pretty decent. Certainly there are much fewer of them than the silicon stuff, but I imagine it's a new technology, so that makes sense.
are gallium nitride transistors what make all the new GaN (gallium nitride) USB power delivery devices work? I have one and its awesome. Of course the gallium nitride isn't being used for a microprocessor in that case.
It seems to be the case. The GaN MOSFETs I am seeing on the market seem to have very good power characteristics, much better than the silicon stuff so I am not surprised to see GaN MOSFETs be used for power supplies.
Yes they are obviously a thing, but when we're talking about CMOS logic circuits required for consumer electronics like smartphones, laptops and desktops, they've been around since 2016 in research papers, and haven't really gone anywhere beyond more research papers.
Obviously you won't disrupt those massive sprawling supply chains for the most state of the art chips right away for such a new tech. However, the availability of discretes means that there is some appetite for non silicon tech.
The hyperlink directs to "nl.farnell.com/en-NL"...
Are you Dutch too? What a coincidence, haha.
No, I'm an immigrant.
Wait, you're Dutch? We might actually be going to the same uni if you are studying electrical engineering 😳
A half life of 2×10^19 years still means 100 events per gram per year, 100 new defects in the chip every year.
Modern CPUs have transistors at least in the tens of millions, the most advanced have billions. A gram of bismuth has ~2*10^21 atoms. Pre-existing impurities would probably be a bigger factor by orders of magnitude.
100 per gram of the bismuth part of the transistors per years. But even decades old transistors have volume measured in nanometers cubed which implies a negligible mass.
Assume the transistors are 10 nm, the volumic mass of bismuth is 9750 kg/m^3 = 9.75 * 10^-6 kg/nm, which means, assuming the bismuth part of the transistor is 10 nm^3 it weigh 9.75 * 10^-6 * 10 kg = 9.75 * 10^-5 kg.
100 event per gram per year is 1000 event per kg per year, which means approximately 1000 * 10^-5 = 10^-2 event per 10^-5 kg per year or 1 event per 100 000 years per transistor. assuming there is a billion transistors in a chip that's 10^9 * 10^-5 = 10^4 event per year in a chip.
A bismuth atom has a radius of approximately 160 picometers = 0.16 nm. To simplify, let's assimilate the atoms to cubes of volume 0.16 nm^3. Then, per our previous assumption, a transistor in our chip has a side 10 / 0.16 = 62.5 atoms long which means it contains 62.5^3 = 244 140.625 atoms of bismuth, let's round it down to 244 140 atoms, which means the billion transistors of the chip contain 10^9 * 244 140 =~ 2.441 * 10^14 atoms of bismuth.
Which means that 1000 atoms decay per year out of ~ 2.441 * 10^14 atoms (1000 / ~ 2.441 * 10^14) * 100 =~ 4.096 * 10^-10 % of the bismuth. At this rate it would take well over 5 * 10^9 years, or 5 billion years, for half a percent of the material to have decayed.
TL;DR: AstroStelar [he/him] is right, for all intent and purposes bismuth is stable.
I'm not a material scientist but I'm not sure every single atomic decay will result in defect
Every decay will result in at least one crystal defect. Alpha particles don't travel very far, so it probably cause another crystal defect nearby, and maybe a short chain reaction. I'm sure they can work around it but I don't think it will be easier to deal with than the defects due to ion migration they already have to deal with in Silicon.
Used in stomach medicine.
Bismuth is already used in all kinds of consumer products, and using it in chips would probably be one of the safer uses https://healthyfocus.org/bismuth-potential-dangers/