New 3-D transistors promising future chips, lighter laptops

Dec 06, 2011 by Emil Venere

(PhysOrg.com) -- Researchers from Purdue and Harvard universities have created a new type of transistor made from a material that could replace silicon and have a 3-D structure instead of conventional flat computer chips.

The approach could enable engineers to build faster, more compact and efficient and lighter laptops that generate less heat than today's. The transistors contain tiny made not of silicon, like conventional transistors, but from a material called indium-gallium-arsenide.

The device was created using a so-called "top-down" method, which is akin to to precisely etch and position components in transistors. Because the approach is compatible with conventional , it is promising for adoption by industry, said Peide "Peter" Ye, a professor of electrical and at Purdue.

A new generation of silicon computer chips, due to debut in 2012, will contain transistors having a vertical structure instead of a conventional flat design. However, because silicon has a limited "" - how fast electrons flow - other materials will likely be needed soon to continue advancing transistors with this 3-D approach, Ye said.

Indium-gallium-arsenide is among several promising semiconductors being studied to replace silicon. Such semiconductors are called III-V materials because they combine elements from the third and fifth groups of the periodic table.

"Industry and academia are racing to develop transistors from the III-V materials," Ye said. "Here, we have made the world's first 3-D gate-all-around transistor on much higher-mobility material than silicon, the indium-gallium-arsenide."

Findings will be detailed in a paper to be presented during the International Electron Devices Meeting on December 5-7 in Washington, DC. The work is led by Purdue doctoral student Jiangjiang Gu; Harvard doctoral student Yiqun Liu; Roy Gordon, Harvard's Thomas D. Cabot Professor of Chemistry; and Ye.

Transistors contain critical components called gates, which enable the devices to switch on and off and to direct the flow of electrical current. In today's chips, the length of these gates is about 45 nanometers, or billionths of a meter. However, in 2012 industry will introduce silicon-based 3-D transistors having a gate length of 22 nanometers.

"Next year if you buy a computer it will have the 22-nanometer gate length and 3-D silicon transistors," Ye said.

The 3-D design is critical because the 22-nanometer gate lengths will not work in a flat design.

"Once you shrink gate lengths down to 22 nanometers on silicon you have to do more complicated structure design," Ye said. "The ideal gate is a necklike, gate-all-around structure so that the gate surrounds the transistor on all sides."

The nanowires are coated with a "dielectric," which acts as a gate. Engineers are working to develop transistors that use even smaller gate lengths, 14 nanometers, by 2015.

However, further size reductions beyond 14 nanometers and additional performance improvements are likely not possible using silicon, meaning new designs and materials will be needed to continue progress, Ye said.

"Nanowires made of III-V alloys will get us to the 10 nanometer range," he said.

The new findings confirmed that the device made using a III-V material has the potential to conduct five times faster than silicon.

Creating smaller transistors also will require finding a new type of insulating layer essential for the devices to switch off. As gate lengths shrink smaller than 14 nanometers, the silicon dioxide insulator used in conventional transistors fails to perform properly and is said to "leak" electrical charge.

One potential solution to this leaking problem is to replace silicon dioxide with materials that have a higher insulating value, or "dielectric constant," such as hafnium dioxide or aluminum oxide.

In the new work, the researchers applied a dielectric coating made of aluminum oxide using a method called atomic layer deposition. Because atomic layer deposition is commonly used in industry, the new design may represent a practical solution to the coming limits of conventional silicon transistors.

Using atomic layer deposition might enable engineers to design having thinner oxide and metal layers for the gates, possibly consuming far less electricity than devices.

"A thinner dielectric layer means speed goes up and voltage requirements go down," Ye said.

The work is funded by the National Science Foundation and the Semiconductor Research Corp. and is based at the Birck Nanotechnology Center in Purdue's Discovery Park. The latest research is similar to, but fundamentally different from, research reported by Ye's group in 2009. That work involved a design called a finFET, for fin field-effect transistor, which uses a finlike structure instead of the conventional flat design. The new design uses nanowires instead of the fin design.

Explore further: X-ray detector on plastic delivers medical imaging performance

More information: First Experimental Demonstration of Gate-all-around III-V MOSFETs by Top-down Approach, www.his.com/~iedm/

ABSTRACT
The first inversion-mode gate-all-around (GAA) III-V MOSFETs are experimentally demonstrated with a high mobility In0.53Ga0.47As channel and atomic-layer-deposited (ALD) Al2O3/WN gate stacks by a top-down approach. A well-controlled InGaAs nanowire release process and a novel ALD high-k/metal gate process has been developed to enable the fabrication of III-V GAA MOSFETs. Well-behaved on-state and off-state performance has been achieved with channel length (Lch) down to 50nm. A detailed scaling metrics study (S.S., DIBL, VT) with Lch of 50nm - 110nm and fin width of 30nm - 50nm are carried out, showing the immunity to short channel effects with the advanced 3-D structure. The GAA structure has provided a viable path towards ultimate scaling of III-V MOSFETs.

Related Stories

It's a wrap! Nanowire opens gate to new devices

Apr 07, 2011

(PhysOrg.com) -- In an interesting feat of nanoscale engineering, researchers at Lund University in Sweden and the University of New South Wales have made the first nanowire transistor featuring a concentric ...

Nanowires key to future transistors, electronics

Nov 26, 2009

(PhysOrg.com) -- A new generation of ultrasmall transistors and more powerful computer chips using tiny structures called semiconducting nanowires are closer to reality after a key discovery by researchers ...

Recommended for you

IBM posts lower 1Q earnings amid hardware slump

40 minutes ago

IBM's first-quarter earnings fell and revenue came in below Wall Street's expectations amid an ongoing decline in its hardware business, one that was exasperated by weaker demand in China and emerging markets.

Microsoft CEO is driving data-culture mindset

1 hour ago

(Phys.org) —Microsoft's future strategy: is all about leveraging data, from different sources, coming together using one cohesive Microsoft architecture. Microsoft CEO Satya Nadella on Tuesday, both in ...

User comments : 11

Adjust slider to filter visible comments by rank

Display comments: newest first

Osiris1
1 / 5 (1) Dec 06, 2011
When do we get giga-quad memory chips for our new Star Trek computer? No longer science fiction pretty soon! Especially couple that with advances in nano tubes and nano tech. Could the sons of Col Corso and Bob Lazar with offices off the 'Extratesticle Highway' be reverse engineering this stuff, or are we just being given it by Uncle Baldr and his government as long as we are good little children of the universe.
Vendicar_Decarian
0.1 / 5 (36) Dec 06, 2011
"No longer science fiction pretty soon!" - Osiris1

You clearly didn't read or weren't capable of comprehending the article.

You won't see such devices. Ever. The ultimate limits of practical construction are within sight.
Nanobanano
4 / 5 (4) Dec 06, 2011
VD:

You're talking about existing techniques and positional assembly.

Eventually, with some sort of bottom up approach, it should be possible to put around a quadrillion transistors in a chip of about 1 cubic centimeter size.

You're also forgetting they are talking about traditional electronic transistors.

We all know photonic and spintronic technology is being developed which is going to require a tiny fraction of the energy, and make almost no waste heat. Addtionally, the operating frequencies of these spintronic and photonic devices will be so much higher, it will be a totally new paradigm in computing power when they are fully developed.

This shit we got now won't even be a toy. It'll be a piece of garbage.
Twin
not rated yet Dec 06, 2011
@VD
Everything that can be invented has been invented. (Charles H. Duell, Commissioner of US patent office in 1899)
Vendicar_Decarian
0.1 / 5 (37) Dec 06, 2011
"Eventually, with some sort of bottom up approach, it should be possible to put around a quadrillion transistors in a chip of about 1 cubic centimeter size." - Nano

And watch it burn up due to it's inability to radiate it's own internal heat from computation.

Such a chip would require laying down 45 million layers of transistors and will therefore cost about 45 million times what a current square centimeter of silicon costs.

At current wafer production prices that brings a cubic centimeter chip in at a cost of about $3.2 billion.

Good luck with that.
Vendicar_Decarian
0.1 / 5 (36) Dec 06, 2011
"Everything that can be invented has been invented." - Twin

"Not everything has been invented hence nothing is impossible." - Twin by implication.
Vendicar_Decarian
0.1 / 5 (36) Dec 06, 2011
The world of fantasy is always spectacular.
fmfbrestel
not rated yet Dec 07, 2011
cost per transistor has fallen at the same time transistor density has increased. No reason to see otherwise moving forward at least for next 10-20 years. Heat dissipation will also resolve itself due to less electricity being required, and more efficient gates. If we can get to spintronics or fully optical chips, heat dissipation becomes almost a non-issue.

Personally, i wish a hard wall would quickly approach, as our society could use more time to adapt to the changing technologies. I just don't see any major hurdles to continued advancement for a good long time.
Vendicar_Decarian
0.1 / 5 (35) Dec 07, 2011
"I just don't see any major hurdles to continued advancement for a good long time." - FmF

In the land of magic and make believe all things are possible.

Buyck
not rated yet Dec 11, 2011
Intel hase already produced 14nm in a testlab! 14nm chips are feasible today see link: http://www.xbitla...ort.html
fmfbrestel
not rated yet Dec 11, 2011
"In the land of magic and make believe all things are possible."

What? I am I supposed to cower under the inescapable truth of your logic and walk away with my tail between my legs? In the land of D-bags and idiots you are king.

More news stories

IBM posts lower 1Q earnings amid hardware slump

IBM's first-quarter earnings fell and revenue came in below Wall Street's expectations amid an ongoing decline in its hardware business, one that was exasperated by weaker demand in China and emerging markets.

Microsoft CEO is driving data-culture mindset

(Phys.org) —Microsoft's future strategy: is all about leveraging data, from different sources, coming together using one cohesive Microsoft architecture. Microsoft CEO Satya Nadella on Tuesday, both in ...

Researchers see hospitalization records as additional tool

Comparing hospitalization records with data reported to local boards of health presents a more accurate way to monitor how well communities track disease outbreaks, according to a paper published April 16 in the journal PLOS ON ...

Ebola virus in Africa outbreak is a new strain

Scientists say that the Ebola (ee-BOH'-lah) virus that has killed scores of people this year in Guinea (GIH'-nee) is a new strain. That means it did not spread there from outbreaks in some other African nations.