Monday, June 5th 2017

IBM Research Alliance Builds New Transistor for 5 nm Technology

IBM, its Research Alliance partners GLOBALFOUNDRIES and Samsung, and equipment suppliers have developed an industry-first process to build silicon nanosheet transistors that will enable 5 nanometer (nm) chips. The details of the process will be presented at the 2017 Symposia on VLSI Technology and Circuits conference in Kyoto, Japan. In less than two years since developing a 7 nm test node chip with 20 billion transistors, scientists have paved the way for 30 billion switches on a fingernail-sized chip.

The resulting increase in performance will help accelerate cognitive computing, the Internet of Things (IoT), and other data-intensive applications delivered in the cloud. The power savings could also mean that the batteries in smartphones and other mobile products could last two to three times longer than today's devices, before needing to be charged.
Scientists working as part of the IBM-led Research Alliance at the SUNY Polytechnic Institute Colleges of Nanoscale Science and Engineering's NanoTech Complex in Albany, NY achieved the breakthrough by using stacks of silicon nanosheets as the device structure of the transistor, instead of the standard FinFET architecture, which is the blueprint for the semiconductor industry up through 7 nm node technology.

"For business and society to meet the demands of cognitive and cloud computing in the coming years, advancement in semiconductor technology is essential," said Arvind Krishna, senior vice president, Hybrid Cloud, and director, IBM Research. "That's why IBM aggressively pursues new and different architectures and materials that push the limits of this industry, and brings them to market in technologies like mainframes and our cognitive systems."
The silicon nanosheet transistor demonstration, as detailed in the Research Alliance paper Stacked Nanosheet Gate-All-Around Transistor to Enable Scaling Beyond FinFET, and published by VLSI, proves that 5nm chips are possible, more powerful, and not too far off in the future.

Compared to the leading edge 10nm technology available in the market, a nanosheet-based 5nm technology can deliver 40 percent performance enhancement at fixed power, or 75 percent power savings at matched performance. This improvement enables a significant boost to meeting the future demands of artificial intelligence (AI) systems, virtual reality and mobile devices.

Building a New Switch
"This announcement is the latest example of the world-class research that continues to emerge from our groundbreaking public-private partnership in New York," said Gary Patton, CTO and Head of Worldwide R&D at GLOBALFOUNDRIES. "As we make progress toward commercializing 7nm in 2018 at our Fab 8 manufacturing facility, we are actively pursuing next-generation technologies at 5nm and beyond to maintain technology leadership and enable our customers to produce a smaller, faster, and more cost efficient generation of semiconductors."

IBM Research has explored nanosheet semiconductor technology for more than 10 years. This work is the first in the industry to demonstrate the feasibility to design and fabricate stacked nanosheet devices with electrical properties superior to FinFET architecture.

This same Extreme Ultraviolet (EUV) lithography approach used to produce the 7nm test node and its 20 billion transistors was applied to the nanosheet transistor architecture. Using EUV lithography, the width of the nanosheets can be adjusted continuously, all within a single manufacturing process or chip design. This adjustability permits the fine-tuning of performance and power for specific circuits - something not possible with today's FinFET transistor architecture production, which is limited by its current-carrying fin height. Therefore, while FinFET chips can scale to 5nm, simply reducing the amount of space between fins does not provide increased current flow for additional performance.

"Today's announcement continues the public-private model collaboration with IBM that is energizing SUNY-Polytechnic's, Albany's, and New York State's leadership and innovation in developing next generation technologies," said Dr. Bahgat Sammakia, Interim President, SUNY Polytechnic Institute. "We believe that enabling the first 5nm transistor is a significant milestone for the entire semiconductor industry as we continue to push beyond the limitations of our current capabilities. SUNY Poly's partnership with IBM and Empire State Development is a perfect example of how Industry, Government and Academia can successfully collaborate and have a broad and positive impact on society."

Part of IBM's $3 billion, five-year investment in chip R&D (announced in 2014), the proof of nanosheet architecture scaling to a 5nm node continues IBM's legacy of historic contributions to silicon and semiconductor innovation. They include the invention or first implementation of the single cell DRAM, the Dennard Scaling Laws, chemically amplified photoresists, copper interconnect wiring, Silicon on Insulator, strained engineering, multi core microprocessors, immersion lithography, high speed SiGe, High-k gate dielectrics, embedded DRAM, 3D chip stacking and Air gap insulators.
Add your own comment

12 Comments on IBM Research Alliance Builds New Transistor for 5 nm Technology

#1
Chaitanya
Nice to see Samsung, IBM and GF collaborating on silicon fab(doubt we will see anything go smaller than 5nm).
Posted on Reply
#2
Captain_Tom
This is what's nailing Intel longterm - everyone else has allied to beat Intel's foundries.



AMD could be to 7nm before Intel gets to 10nm. (At the very least they will both compete at these nodes at the same time).

I know that node size is mostly marketing at this point, but I believe GF's 7nm = Intel's 10nm (roughly), or even beats it by a little bit in the density department.
Posted on Reply
#3
Captain_Tom
Chaitanya said:
Nice to see Samsung, IBM and GF collaborating on silicon fab(doubt we will see anything go smaller than 5nm).
I wouldn't be surprised if we saw one of the foundries market their final silicon node as 4nm. It wouldn't be "Real" 4nm though....
Posted on Reply
#4
Aenra
Chaitanya said:
doubt we will see anything go smaller than 5nm
Really? :)
Few months ago, we all doubted we'd reach a 5nm state acceptable for mass production standards and yet here we are apparently, quantum tunneling notwithstanding.

Also, be advised that even as we speak, research is being conducted for transistors that actually take advantage of tunneling, rather than trying to mitigate its effects ^^

Never say never when money's involved.
Posted on Reply
#5
Vayra86
Honestly, all this news about the new smaller nodes sounds a lot to me like all the news about the next best battery technology.

In the end, none of it makes it to market.

Keep in mind that all of this still needs to be economically viable. The one big positive I see in this news is that GlobalFoundries is racing hard to overtake TSMC, which is good for various players in the market (hi AMD) that can use a process advantage very, very well.
Posted on Reply
#6
Prince Valiant
Vayra86 said:
Honestly, all this news about the new smaller nodes sounds a lot to me like all the news about the next best battery technology.

In the end, none of it makes it to market.

Keep in mind that all of this still needs to be economically viable. The one big positive I see in this news is that GlobalFoundries is racing hard to overtake TSMC, which is good for various players in the market (hi AMD) that can use a process advantage very, very well.
Foundries make themselves more attractive to customers by having and offering smaller/better/more effective nodes. The server and supercomputer markets both want better technology at a fairly constant clip and the consumer market is a decent chunk of change too.

Battery manufacturers would surely have worse long term profit by offering a much better battery. It would be nice for customers and probably produce a lot less waste but what do they care about such silly things :P.
Posted on Reply
#7
R-T-B
Aenra said:
Really? :)
Few months ago, we all doubted we'd reach a 5nm state acceptable for mass production standards and yet here we are apparently, quantum tunneling notwithstanding.
He simply means on the silicon process. Of course tech will march on, but silicon is reaching it's limits.
Posted on Reply
#8
Aenra
R-T-B said:
He simply means on the silicon process. Of course tech will march on, but silicon is reaching it's limits.
And again, we've heard that before :)
Also, he didn't specify. That's you superimposing ^^
Posted on Reply
#9
NdMk2o1o
btarunr said:
The power savings could also mean that the batteries in smartphones and other mobile products could last two to three times longer than today's devices, before needing to be charged.
Have seen this mentioned time and time again year after year with nothing ever coming to market that delivers, taken with a bucket load of salt
Posted on Reply
#10
HopelesslyFaithful
BTW its possible to go to 1nm or even sub 1nm IIRC. But your dealing with just a few atoms I think.

I am sure thats 20 years away.
Posted on Reply
#11
theoneandonlymrk
Captain_Tom said:
This is what's nailing Intel longterm - everyone else has allied to beat Intel's foundries.



AMD could be to 7nm before Intel gets to 10nm. (At the very least they will both compete at these nodes at the same time).

I know that node size is mostly marketing at this point, but I believe GF's 7nm = Intel's 10nm (roughly), or even beats it by a little bit in the density department.
This is also in Amds timeline , i remember thinking it strange seeing gate around transistors being mentioned for a future Amd cpu yet I can't remember where now either because it wasn't front page ,it was suttle like in a roadmap or something.
Post zen.
Posted on Reply
#12
Steevo
NdMk2o1o said:
Have seen this mentioned time and time again year after year with nothing ever coming to market that delivers, taken with a bucket load of salt
I blame bloated code for this as well as hardware choices. Do I really need two quad core processors in a phone? Nope. Would more efficient code written specifically for the hardware in each instead of generic code allow as fast of operation with lower speeds? Damn straight.

It's strange that Windows 95 came on a set of floppies and is smaller than the Facebook and Messenger apps and we find this acceptable.
Posted on Reply
Add your own comment