What is Toshiba's germanium breakthrough for 16nm chips?

  • Level: Graduate 
  • Thread starter Thread starter MTd2
  • Start date Start date
Join the discussion
Registration is free. Start your own thread to ask a follow-up.
Sign up Log in
7 replies · 3K views
Physics news on Phys.org

Discussion

It seems they've found a good buffer layer.
Useful and interesting to people who work with those materials but not terribly exciting to the rest of us...
 
To the OP: Normally gate terminal (where you control the switch to be ON/OFF) is separated from the channel by an insulator (you don't want it to directly interfere with the ongoing current flow below in the channel --- of course it moves the bands up or down)

But to keep the transistor structure intact (keeping all fields constant) you must scale that thin insulator while you down-size your channel.

Nowadays it's a 6-7 nm layer (physical), but they report it in EOT (equivalent oxide thickness) units where the physical thickness is converted to a equivalent thickness as if the material is SiO2 (which was the common insulator for more than 30 years). They just very recently switched to metal-oxide insulators. In the last few generations the oxide thickness came close to 1.1 nm. And they stopped using SiO2

Because they can't make it smaller than a nanometer (gate leakage becomes unbearable leading eventually to a hard oxide breakdown) they had to make it thicker -- but now you need a higher permittivity for the insulator to get the equivalent field.

Now Toshiba is saying that they improved the process and the leakage will not be huge problem because they inserted something in between the channel and the insulator. Doing all this with a HIGHER MOBILITY (thus faster) material (Ge is a low band-gap semiconductor with lower effective mass near the conduction band -- thus higher mobiliy) is fascinating because then it means you can go on scaling.IT might actually be "terribly" exciting for all of us since it may potentially make it easier to produce 16 nm -nodes.

We are just a few nodes away from the ultimate limit of the size of a MOS-transistor and it's not certain as of now, whether Silicon will survive up to 8 nm.

That might have a direct impact on our lives (considering it's Toshiba and they are pretty damn serious about it) in laptop prices, higher performance, lower power etc...

With all due respect, VLSI business is where science begins to be REAL and really really IMPORTANT considering Moore's Law and the entailing technology has revolutionized our modern society.
 
Thank you very much, sokrates. Please, do you know where to get Toshiba's paper?
 
I don't know whether there's paper yet, or not. I was aware of this following the link you posted. They might not even publish it.

but if it's ever going to be published, that's probably going to be ISSCC
 
I just realized that they don't need to publish, but they DO need a patent.
 
sokrates said:
IT might actually be "terribly" exciting for all of us since it may potentially make it easier to produce 16 nm -nodes.

We are just a few nodes away from the ultimate limit of the size of a MOS-transistor and it's not certain as of now, whether Silicon will survive up to 8 nm.

That might have a direct impact on our lives (considering it's Toshiba and they are pretty damn serious about it) in laptop prices, higher performance, lower power etc...

With all due respect, VLSI business is where science begins to be REAL and really really IMPORTANT considering Moore's Law and the entailing technology has revolutionized our modern society.

I am fully aware of how important buffer layers can be, I did my PhD working with devices made from oxide materials and even though I didn't grow our films I've spent many hours discussing the pros and cons of different buffers (I used to work with someone who was really good at depositing CeO2).

However, a buffer layer is only one small step of the process and even when you've found a good buffer there is still the problem of introducing it in a production line (at a resonable cost). Hence, there is no guarantee that this will ever be used in mass produced electronics.
Most semiconductor companies announce "breakthroughs" every 6 months or so and I've been around long enough to know that most of it never goes anywhere.

Moreover, important is not the same thing as exciting. Like most physicits I am primarily interested in the PHYSICS, not the applications of what I do.
Si technology in general is in my view pretty boring, it has been around for a long time and there are too many people working in the field (and this is from someone who just submited a paper where about half the experimental data comes from SiO2).

Another "problem" is that semiconductor companies are only interested in what they can sell (obviously), we have been able to to make devices smaller than anything that will be in the 16nm node for many years now using direct-write e-beam lithography, scanning probe techniques etc although these techniques are obviusly only suitable for making single devices.
 
f95toli said:
However, a buffer layer is only one small step of the process and even when you've found a good buffer there is still the problem of introducing it in a production line (at a resonable cost). Hence, there is no guarantee that this will ever be used in mass produced electronics.

In our times, every step is a small step and there's no big jump in scientific progress, especially when it comes to engineering. And all the papers that have been published in the last 10 years have no guarantee that they will ever be useful.

So I am not seeing anything different for this breakthrough. How can you know it's different than any other breakthrough? Are you saying anything new here?

f95toli said:
Most semiconductor companies announce "breakthroughs" every 6 months or so and I've been around long enough to know that most of it never goes anywhere.

I cannot agree with this. It is the same semiconductor companies that bestowed this technology a trillion-fold increase in every sense possible. It is the same semiconductor companies that revolutionized the world with computers, made possible the cell phone you are using right now, the network backbone your net connection is on, your car, your digital watch, your tv, i.e things that you cannot even breathe without, right now.

These are all thanks to those "breakthroughs" that went NOWHERE -- but NOT due to a number of physicists who were "excited" by a weird "quantum mechanical" effect that could ONLY be replicated with a 50$ million worth experimental apparatus.

f95toli said:
Moreover, important is not the same thing as exciting. Like most physicits I am primarily interested in the PHYSICS, not the applications of what I do.

I don't know how you come to that conclusion on behalf of "most physicists". You must have an enormous amount of data, based on polls, personal opinions when it comes to physicists' taste and perception on the delicate difference between "exciting and important" in their professional lives.
----------
My point is (and was) this: It is NOT the people who were working on asymptotic freedom theories back in the 70's who made everything you see today possible. No it was not string theory, not some weird "tunneling time" experiment, or not quantum gravity theories.

It was those semiconductor companies, and it was their little discoveries. To announce a breakthrough like that on behalf of a huge organization like Toshiba takes YEARS of research and involves a SWARM of scientists.

It's not like submitting a paper to PRB and pretending it is the MOST EXCITING experimental work ever!

So, please...