Recently in material science news from China we hear that [Hailin Peng] and his team at Peking University just made the world’s fastest transistor and it’s not made of silicon. Before we tell you about this transistor made from bismuth here’s a whirlwind tour of the history of the transistor.
The Bipolar Junction Transistor (BJT, such as NPN and PNP) was developed soon after the point-contact transistor which was developed at Bell Labs in 1947. Then after Resistor-Transistor Logic (RTL) came Transistor-Transistor Logic (TTL) made with BJTs. The problem with TTL was too much power consumption.
Enter the energy-efficient Field-Effect Transistor (FET). The FET is better suited to processing information as it is voltage-controlled, unlike the BJT which is current-controlled. Advantages of FETs include high input impedance, low power consumption, fast switching speed, being well suited to Very-Large-Scale Integration (VLSI), etc.
The cornerstone of Complementary Metal-Oxide-Semiconductor (CMOS) technology which came to replace TTL was a type of FET known as the Metal-Oxide-Semiconductor Field-Effect Transistor (MOSFET). The type of MOSFET most commonly used in CMOS integrated circuits is the Enhancement-mode MOSFET which is normally off and needs gate voltage to conduct.
A transistor’s technology generation is given with the “process node”, in nanometers (nm). This used to mean the size of the smallest feature that could be fabricated, but these days it’s just a marketing term (smaller is “better”). Planar CMOS MOSFETs were initially dominant (through ~28nm), then came SOI MOSFETs (28nm to 16nm), then FinFETs (16nm to 5nm), and now finally Gate-All-Around FETs (GAAFETs, 3nm and beyond).
All of that in order to say that this new transistor from [Hailin Peng] and his team is a GAAFET. It’s made from bismuth oxyselenide (Bi₂O₂Se) for the channel, and bismuth selenite oxide (Bi₂SeO₅) as the gate material. See the article for further details.
Keep in mind that at this point in time we only have a prototype from a lab and the gory details about how to mass-produce these things, assuming that’s even possible, haven’t yet been worked out. We have previously discussed the difficulty of manufacturing state-of-the-art transistors. If you’re interested in bismuth be sure to check out how to use bismuth for desoldering.
Press F for “gonna believe it when someone independently replicates it”, just like that magic graphene 400 picosecond flash replacement from 3 weeks ago.
wonder what research China stole this time.
Gun Powder, a real killer app from China.
I was just going to say that you can’t mention this kind of thing on YT since the primitives commenting there cling to their ‘China only copies’ propaganda lifebuoy.
But I forgot you guys are here too. Doh.
Doesn’t mean reasonable skepticism doesn’t apply, regardless who made a claim. You well know research like this often takes a decade to make it to production, and that’s if it’s actually useful outside of a lab. Plenty of US technology is the same.
Or the hubbub with the ambient-temperature-superconductor a couple of years ago.
Gunnerkrigg Court Technology!!!
Good lord I just grew a meter long white beard remembering that
Hmm, now you’ve made me think about how long it’s been.
Behind a paywall. Any “alternative” links? Any reviews from someone who has read the paper?
Bismuth is not used in any commercial transistors. If there were some in production there might be a route to advancing an existing fabrication facility to make these new 2D transistors. Even if it works well in the lab (I have my doubts) there is no path to making chips you can buy on AliExpress.
I’m putting this headline in my “Bigfoot sightings” folder.
Gift link:
https://archive.is/ZmCzA
“Bismuth is not used in any commercial transistors.” That’s a bold assertion because how do you know? If you said “Before 1947 blah blah” we could have a conversation.
“I’m putting this headline in my “Bigfoot sightings” folder.”
Oh my, however will we go on?
Making chips you can buy on AliExpress isn’t really useful when the goal is HPC modules for datacenters.
AliExpress isn’t the goal, but it’s populated by both resellers and literal factories, so no that’s just incorrect. Wholesale purchases are usually on Alibaba for overseas sales anyway,
“The Bipolar Junction Transistor (BJT, such as NPN and PNP) was invented by Bell Labs in 1947.”
Not exactly. The transistor demonstrated by Bell Labs in ’47 was a point contact transistor. BJTs came a bit later.
“Enter the energy-efficient Field-Effect Transistor (FET)….”
Interestingly, Julius(?) Lilienfeld was issued a patent for (what amounts to) a FET two decades earlier, though I’ve never seen images of a practical device.
Ah, you are quite right, thanks for picking me up on that one, it was the point-contact transistor developed at Bell Labs in 1947. I will edit the article to clarify. I have recently read about these topics in Principles of Electronic Circuits 2ed by Burns and Bond and I misremembered the detail about the type of transistor developed at Bell Labs (in my head my logic was “it wasn’t the FET therefore it was the BJT, but of course my “logic” was flawed). Chapter 1 of PoEC starts with a brief history of the transistor. They mention Lilienfield (1925), Heil (1935), and Pohl (1938) as all providing precursors to the FET.
Is it just me or did this article bring out the worst of HaD comments? I thought we were all on a roll there being cordial and stuff.
Anyway, I liked the article. I learned some things, maybe I’m a fool but I believe the research, and I thought this was pretty well written…
Probably a silly question, but what is the substrate ? Is it silicon as is used by almost all fabs worldwide. Or is the substrate material bismuth oxyselenide (Bi₂O₂Se) or bismuth selenite oxide (Bi₂SeO₅), because if it is either I can imagine really heavy devices, and new fabs needing to be created. And for 40% faster and 10% less power, the economics are probably not worth it.
This article does not answer your question, but it is informative and mentions a substrate of sapphire for molybdenum disulphide 2D wafers.
https://www.intelligentliving.co/china-2d-chip-atom-thin-processor-memory/
That’s interesting. Looks like they didn’t submit any of the supporting data or test code for this at all though.
and it will keep your computer from having an upset tummy
Ooo, pink CPUs!
Bell Labs think they’re slick. They did not develop the transistor. They received pieces of Germanium in 1947 from the Roswell crash. Airforce personnel at Wright Patterson AFB sent the samples to Bell Labs to be analyzed, since they were the only lab at the time that was capable of determining the material’s composition.
Actually, Bell Labs was the big player in developing advanced systems for the special access programs. But Los Alamos National Labs was also qualified at that time to analyze crash debris samples.
Lols aside, work with germanium started at Purdue in the 1930’s due to need for crystal rectifiers for early radar development.
Commercial researchers at Bell, GE and Westinghouse all followed this work. While ultimately Bell Labs ultimately got the credit, it was only with support from Horovitz’ direction of the Purdue lab’s semiconductor research on germanium that provided the basis forprototype designs by Bardern and Brattain at Bell.
Simultaneously Westinghouse, who notably won the contract for germanium based crystal radar detectors, were concentrating on fulfilling that contract, then developed their own “transitron” independently afterwards.