Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
China Hardware

Did Peking U. Just Make the World's Fastest Transistor - Without Using Silicon? (tomshardware.com) 83

"It is the fastest, most efficient transistor ever," proclaims an announcment from Peking University. "And most important of all, there's no trace of silicon involved," adds ZME Science. From the South China Morning Post: A team of researchers at Peking University claims to have shattered chip performance limits and proven that China can use new materials to "change lanes" in the semiconductor race by circumventing silicon-based roadblocks entirely.

The researchers, led by physical chemistry professor Peng Hailin, said their self-engineered 2D transistor could operate 40 per cent faster than Intel and TSMC's cutting-edge 3-nanometre silicon chips, while consuming 10 per cent less energy.... "While this path is born out of necessity due to current sanctions, it also forces researchers to find solutions from fresh perspectives," [Hailin] added.

"Peking's major innovation comes from the two-dimensional nature of their transistors, facilitated by using an element other than silicon," writes Tom's Hardware: BiâOâSe, or bismuth oxyselenide, is a semiconductor material studied for its use in sub-1nm process nodes for years, largely thanks to its ability to be a 2D semiconductor. Two-dimensional semiconductors, like 2D BiâOâSe, are more flexible and sturdy at a small scale than silicon, which runs into reduced carrier mobility at even the 10nm node. Such breakthroughs into stacked 2D transistors and the move from silicon to bismuth are exciting for the future of semiconductors and are necessary for the Chinese industry to compete on the leading edge of semiconductors.
ZME Science adds this note of skepticism. "Turning laboratory breakthroughs into commercial chips typically takes years — sometimes decades..."

Thanks to Slashdot reader schwit1 for sharing the article.
This discussion has been archived. No new comments can be posted.

Did Peking U. Just Make the World's Fastest Transistor - Without Using Silicon?

Comments Filter:
  • Hey EditorDavid, remember your site doesn't support unicode.

    > sent from my PDP-11

    • On your iPhone - Settings -> General -> Keyboard -> Smart Punctuation = off. Congratulations, your iPhone is now /. "compatible."
      • by Teun ( 17872 )
        True, but it would be even better that the editors would avoid this crap or... fix the back-end!
        Because it is Bi2O2Se with two times a subscript 2, nor BiâOâSe.
        • It's 2025 and you still can't edit on /. There is zero chance and back-end update will ever be done to this site.
          • * a not and. Fuck you /. and your lack of editing.
          • Re: lol (Score:5, Informative)

            by Dragonslicer ( 991472 ) on Sunday May 04, 2025 @07:44AM (#65350933)
            The inability to edit posts is an intentional design decision, not due to lack of caring or skill. The problems with being able to edit a post after someone has moderated it or replied to it should be obvious to everyone here.
            • The inability to edit posts is an intentional design decision, not due to lack of caring or skill. The problems with being able to edit a post after someone has moderated it or replied to it should be obvious to everyone here.

              I'll meet you half-way, how about the ability to delete your post.

              • Why should you have the benefit of scrubbing stupid things from the record if you don't like being called out? If you don't like how discussions unfold or how other people may view your opinion you have the option of not saying something.

                Forget what year it is, or what state technology is in. The real question is are you old enough to act like an adult and stand behind the things you have said? Even if that means admitting you were wrong in a follow up post?

            • The inability to edit posts is an intentional design decision, not due to lack of caring or skill. The problems with being able to edit a post after someone has moderated it or replied to it should be obvious to everyone here.

              A brief window of time for editing an unmoderated and reply-free post might be a reasonable compromise.

            • While that is true, it would still be nice to have the ability to edit for a certain time period or until it is moderated or replied to. I can't imagine how many stupid mistakes I have made and didn't see and wish I could have edited it a few minutes after posting.

              Yes, I should have used "Preview" and then reviewed it first. My bad. Let's be honest- how many of us actually do that.

          • by Teun ( 17872 )
            Yes you can edit, that's what the preview is for.
          • I'm sure b i zx dot info will fix it real any day now as slashdot is the crown jewel of their portfolio.

            • roflmao! slashdot won't even let you post the name of their owner without mangling it; it triggers the "lameness filter". try it for yourself!

              that is truly shameful!

      • The unicode was on the Toms Hardware article. Not an iPhone artifact (this time).

        The unicode characters in use were subscripted 2's. Bismuth_sub_2, Oxygen_sub_2, Selenium

        I tried to post the line corrected for its unicode, but apparently Slashdot can't handle a <sub> tag.

        2D Bi2O2Se, are more flexible and sturdy at a small scale than silicon, which runs into reduced carrier mobility at even the 10nm node.

  • Not even close (Score:3, Insightful)

    by locater16 ( 2326718 ) on Sunday May 04, 2025 @03:42AM (#65350745)
    This isn't even close to the world's fastest transistor, nor the only 2d one. By the way Tomshardware is a lying clickbait shithole of a site and should be de facto banned from ever appearing on /. or anywhere else.
    • Re:Not even close (Score:5, Interesting)

      by AmiMoJo ( 196126 ) on Sunday May 04, 2025 @05:18AM (#65350809) Homepage Journal

      On what do you base that claim? TFA doesn't even give any figures for speed, so how have you compared it to other transistors?

      I'm reminded of the scepticism that other Chinese technological advances were met with. LLMs and automotive battery technology being two recent ones. The derision soon turns to claims it was stolen (with that damn time machine of theirs), and finally import bans because somehow that will fix our lack of investment in R&D.

      • Re:Not even close (Score:5, Insightful)

        by Mr. Dollar Ton ( 5495648 ) on Sunday May 04, 2025 @05:28AM (#65350823)

        TFA doesn't even give any figures for speed

        TFA (the original one) doesn't give anything more than newspaper propaganda. Maybe it is the fastest, maybe it isn't, maybe it exist, maybe it doesn't.

        Not enough information for a meaningful discussion, so let's dump all known cliches on the topic of "China".

      • by znrt ( 2424692 )

        MB: Americans are frightened by China not because of what China does, or even its more efficient SUVs, or its infrastructure. United States is frightened of China because it exists. It's the purest example of existential anxiety, the word for which is dread. In other words, it's China by its very success and existence which challenges this sort of foundational assumption of american collective identity that we're superior, exceptional. We're born to be number one and always will be, and the notion of being

      • by CEC-P ( 10248912 )
        Wow, someone promoting China who has far left bullshit in their sig. Never seen that before. I bet you went to a REALLY prestigious university too and learned your socialist bullshit. Anyway, what's really behind this is: China leads global bismuth production, contributing approximately 16,000 metric tons annually—84% of worldwide output. Silicon, on the other hand, is sand. Everyone has sand. So this propaganda article about an unrealistic claim from the Chinese, who invent nothing and can only steal
        • I would not go far as to call it propaganda as much as unsubstantiated. It is possible that these researchers did exactly what they claim they did. However that does not mean that this technology is ready to start churning out chips tomorrow. It may take years to decades to get make chips at scale and if that scale is not cheaper than silicon, few would use it. There are many technologies that show promise at the research level but do not make it into large scale production.
    • by Hodr ( 219920 )

      Also kind of weird how they are calling out "silicon" as being roadblocked. Silicon is literally a component of sand, it's available everywhere on earth in huge quantities. The roadblocks are with EUV lithography equipment, which makes me wonder how are they getting sub-1nm features on this material if they were unable to do so with silicon? Since it's "2d" are they counting the height as the smallest feature?

      There's been tons of different semiconductors that are faster and more efficient than silicon in

      • Silicon has grades of quality, and while many beaches are made of silicon, barely any of it can be used for high quality purposes.

        • But do you know how they make electronic grade silicon? The take lesser grade (metallurgical grade) and further refine it. The process is not exactly a secret. The OP's point is that silicon even electronic grade silicon is not the limitation that China has when producing chips. Obtaining EUV equipment has been a problem. Maybe this new transistor does not require EUV but the issue is that researchers have to massively upscale this process to make nearly 20 billion of transistors on a chip to catch up to to
      • by Khyber ( 864651 )

        "The roadblocks are with EUV lithography equipment"

        Nah, it's getting tougher and tougher to get sourced silicon pure enough to be worth spending the extra energy to get those few extra Ns to ensure you get a sufficient defect-free yield.

        Get back to me when you can take raw beach sand and make a working processor at even ~100um node.

        • Yes but I would guarantee you that silicon is far more common to source than bismuth or selenium. And those elements still have to be ultra pure to be used to make transistors. The main benefit for this technology is if it does not require EUV to scale it to make billions of transistors per chip . But scaling this up will take years to decades to do.
    • I must have missed something, when did Tom's Hardware go to hell?

  • If headline asks a question then the answer is most likely no, or so Betteridge's law of headlines tells us. https://en.wikipedia.org/wiki/... [wikipedia.org]

    In this case the answer could be yes because we've known of all kinds of different semiconductors for ages, we just stuck to silicon for so long because it's been proven the cheapest. Part of what keeps it so cheap is that silicon has taken such a large chunk of the semiconductor market that it takes some powerful motivations to do different, such as a need for extre

    • How many more years am I going to have to keep reading comments on this site from people attempting to appear smart by constantly wielding this law of news headlines?
      • by MacMann ( 7518492 ) on Sunday May 04, 2025 @04:50AM (#65350789)

        How many more years am I going to have to keep reading comments on this site from people attempting to appear smart by constantly wielding this law of news headlines?

        You will likely see comments about Betteridge's law until news outlets learn to have some journalistic integrity.

        I don't know if this is a common lesson in American high schools but I can recall a lesson on how to write a news article. The headline is to summarize as best as possible who, what, when, and where, then the news article is to expand on those details. What we have now are headlines that fail to give the important details because if they gave all the important details in the headline then people might not click on the link to the article to learn more, and most news outlets are paid by advertisers.

        I suspect such lessons on writing news articles in high school was also about learning how to separate news from rumor and gossip. If we know what happened, when, and where, but not who was involved then that's not news, that's a rumor. This is played with a lot when there's claims of some discovery of historical events, they leave out the when. It's not like we need to know the exact date for some event, just putting it within 1000 years might be enough to know if it is bullshit or not.

        The point of Betteridge's law is to point out how headlines that ask questions are most often click bait and the article under that headline is often a steaming pile of bullshit that is a waste of time to read. I don't know how to end the click bait headlines but it might help if Slashdot refused to highlight news articles with click bait headlines.

      • Forever. No old joke has stopped being funny on slashdot, no age discrimination here.

      • I think the editors do this purposely to create engagement. Or submissions with easily correctable mistakes, or missing information. If they made the perfect post, all information perfectly clear, there wouldn't be comments.

        • I've always believed the headline writers were thinking they couldn't present some outlandish claim as fact, but rephrasing it as a question let them off that hook. So not "Peking U makes fastest transistor", which as several commenters have pointed out would require some knowledge and an actual explanation, but rather, "Did Peking U make fastest transistor?" From there, Betteridge's Law is just a corollary.
      • How many more years am I going to have to keep reading comments on this site from people attempting to appear smart by constantly wielding this law of news headlines?

        No.

      • How many more years am I going to have to keep reading comments on this site from people attempting to appear smart by constantly wielding this law of news headlines?

        3 years, 2 months and 12 days 6 hours, 10 minutes, +/- 30 seconds

    • The first question is how many times did they have to try to successfully make a second transistor? The second question is, what is the density of the transistors they can make? The third question is, how easy is it to automate the process they're using? Depending on the answers to these questions, then even if we ignore cost of materials their transistor could be 100 times faster and ten times more efficient and still be completely useless.

    • There's more to the speed of electronic devices than their transistors. There's wires that cause delays, and while that might seem trivial to those unfamiliar with microelectronics this is an issue that needs to be kept in mind during the design process.

      Oh hell yeah. At present day computer speeds, the speed of light is an issue, those signals take some time to get from one place to another.And the faster the transistor, the more the wires and other devices come to play. One of the reasons for shrinking chips is to avoid some of this added time. A rule of thumb is about a foot per nanosecond - but that's just for concept.

      And to complicate matters, there is Velocity Factor, or Vf. And it's different for different materials. For our work, it will be aroun

  • by msauve ( 701917 ) on Sunday May 04, 2025 @04:44AM (#65350781)
    >BiâOâSe

    I think Elon Musk has a kid with that name.
  • by weirdow ( 9298 ) on Sunday May 04, 2025 @05:40AM (#65350839) Homepage
    Recently, two weeks ago, did a video about this bismuth compound

    https://www.youtube.com/watch?... [youtube.com]

    Microchip Breakthrough: World’s First Silicon-Free Processor [youtube.com]
  • SCMP? (Score:3, Insightful)

    by hmilz ( 3035377 ) on Sunday May 04, 2025 @05:45AM (#65350841)
    I'd be extremely careful with reports from this outlet. Let's talk again when this has been peer reviewed and confirmed by independent research groups.
  • Silicon is at the end of what it can do. What this story really shows is that China is catching up, fast. Even if the results here are overstated, in 10-20 years or so, they will be able to compete at the top. And that is bad news. Now, if they were dependent on ASML machines, but they will not be. Nice example of how trade restrictions backfire.

    • Read the ZME and Tom's articles. The headline is pretty misleading. Sounds like they have made a few devices and tested them with very good results. As you say, probably out a ways time wise. To me the real headline is that China is increasing investment in research, while the US keeps cutting it. That is a problem for the future. It is the short term thinking of capitalism for next quarter's numbers that is going to be the end of the west.
    • by Slayer ( 6656 )

      Now, if they were dependent on ASML machines, but they will not be. Nice example of how trade restrictions backfire.

      They may have a fancy new semiconductor compound to work with, but all these wafers have to be patterned to tiny structures if they are going to build competitive chips. This is where ASML machines come in, and the article suggests nothing that these were recreated or substituted or made irrelevant somehow.

    • Yeah, keep thinking like that, 'in 10-20 they'll catch up'. You do know all these US bans are put into place because China was way ahead of USING the advanced chips compared to the US, so the US needed a way to slow them down by barring them from easily getting their hands on the advanced chips, but now China has a necessity to create advanced chips themselves as they cannot easily get them anymore, so they invest heavily into technology to create new advanced chips themselves, you know the saying give them
      • by gweihir ( 88907 )

        Catching up in this in 2-3 years is completely impossible. That is not even enough time to build and test the machinery when you already know how to do it. 10-20 years is _fast_.

      • China is advancing so f-ing hard these days due to the bans, it isn't 10-20 before they catch up, it'll be 2-3 years or even sooner.

        Only if you ignore the reality of the situation. Even if the claims of this research are true; they have made a few transistors. A few. The latest AMD 9900x has 17.2 billion transistors. That is a long way to go from scaling research to production especially in this case where it has taken 70 years from first silicon transistor to producing thousands of 9900xs.

  • Opening up a can of posts: Is this technology patented?
  • Seriously, can it just be a /. thing, can we ban headlines with question marks in them?

    It's such clickbait nonsense, every time. Just make it a policy.

  • by groobly ( 6155920 ) on Sunday May 04, 2025 @12:07PM (#65351287)

    So what? Germanium transistors have been around for a long time. Also, other compounds. The main advantage of silicon is that manufacturing is much easier.

    • Easier and cheaper. Silicon being abundant in the Earth's crust helps with both factors. The main problem with transistors that are not based on silicon is the availability of raw materials. Bismuth (0.0085 ppm) and selenium (0.05 ppm) are not exactly common being less common than silver but more common than gold. Germanium is far more common than either at 1.5 ppm.
  • The mention of sanctions in a technical news release, saying it was the driver for this purported advance, gives the release a BS smell.

    I have no idea of the accuracy or significance of the release. Or even whether sanctions were in fact a driver. But the way it's phrased reduces its credibility.

    • The only way this research was done to counter sanctions is China is in short supply of silicon and must use bismuth and selenium. Considering that silicon is basically sand, I do not think that developing an entire process not using silicon is more practical than China refining sand to make electronic grade silicon. The other thing which is an unknown is how they scale up making billions of transistors per chip from making a few transistors in research. Maybe this process will not involve EUV which is poss
  • It doesn't do any good if they make a CPU that will only last a few months.
  • by Meneth ( 872868 ) on Monday May 05, 2025 @04:30AM (#65352869)
    Is this the paper the article refers to? Publication date is rougly correct... https://pubs.acs.org/doi/full/... [acs.org]

Any sufficiently advanced technology is indistinguishable from magic. -- Arthur C. Clarke

Working...