Faster and Faster - Moore's Law Continues to Amaze
PCWorld
IBM Details World's Fastest Graphene Transistor
Joab Jackson, IDG News Service
Feb 5, 2010 12:20 pm
Thanks to a change in recipe, IBM has created a graphene-based processor that can execute 100 billion cycles per second (100GHz), almost four times the speed of previous experimental graphene chips.
With this research, IBM has also shown that graphene-based transistors can be produced by the wafer, which could pave the way for commercial-scale production of graphene chips, said Yu–Ming Lin, the IBM researcher who led the work.
If commercialized, such graphene processors could be the basis of superior signal processing componentry, improving the fidelity of audio and video recording, radar processing and medical imaging.
IBM conducted the work on behalf of the U.S. Defense Department's DARPA (Defense Advanced Research Projects Agency), under a program to develop high-performance RF (radio frequency) transistors. A write-up of the results has been published in the Feb. 5 issue of Science.
Graphene, a single-atom-thick honeycomb lattice of carbon atoms, can transport electrons more quickly than other semiconductors, a quality called electron mobility. "That makes graphene a promising material for high-speed or high-frequency electronic components," Lin said.
This prototype processor was created on a 2-inch wafer, though in principle it could be done on even larger wafers, which should bring the production costs down, Lin said. Graphene is produced by heating a silicon carbide wafer, allowing the silicon to evaporate.
Until now, the downside of graphene has been that it is very sensitive to the environment. During the fabrication process, an oxide layer is deposited over the graphene to form the gate insulator. Typically, this deposition degrades the graphene's electron mobility, due to defects in the oxide that scatter electrons in the graphene. The IBM researchers minimized the damage by separating the graphene from the oxide with a very thin polymer layer.
This new approach has been instrumental in allowing the researchers to almost quadruple the frequency of graphene chips. Last year, research teams from IBM and the Massachusetts Institute of Technology both demonstrated graphene processors capable of frequencies around 26GHz. By comparison, silicon-based transistors of the same gate length (240 nanometers) have only been able to scale up to a clock rate of 40Ghz or so.
It also sets the stage for commercial production. The research shows that "high-quality graphene can be produced on a wafer scale, and graphene transistors can be fabricated with those processes used in the semiconductor industries," Lin said.
Lin cautioned against thinking of graphene as a substitute for the silicon-based microprocessors used in today's computers, at least at anytime in the near future. One major roadblock is that graphene does not work easily with discrete electronic signals, he explained. Because graphene is a zero bandgap semiconductor, meaning there is no energy difference between its conductive and nonconductive states, transistors made of the semiconductor cannot be turned on and off. In contrast, silicon has a bandgap of one electron volt, making it good for processing discrete digital signals, Lin said.
Instead, graphene is better suited for making analog transistors, such as signal processors and amplifiers. Today, such circuitry is largely made from GaAs (gallium arsenide), though GaAs offers nowhere near the same electron mobility, Lin said.
February 5, 2010
IBM hits graphene transistor breakthrough
by Larry Dignan
IBM graphene transistor
IBM Research on Friday announced that it has demonstrated a radio-frequency graphene transistor with the highest frequency so far: 100GHz.
Graphene is a special form of graphite, consisting of a layer of carbon atoms packed in honeycomb lattice. In a nutshell, graphene is like "atomic scale chick wire." Graphene's properties could lead to faster transistors.
One-step graphene doping could enable complementary metal oxide graphene transistors
By Suzanne Deffree, Managing Editor, News -- Electronic News, 2/17/2010
Researchers at the Georgia Institute of Technology have claimed a one-step process that produces both n-type and p-type doping of large-area graphene surfaces and that could facilitate use of the material for future electronic devices.
The doping technique -- produced by applying a commercially available SOG (spin-on-glass) material to graphene and then exposing it to electron-beam radiation -- can also be used to increase conductivity in graphene nanoribbons used for interconnects, according to the university.
Both types of doping were created by varying the exposure time to the to e-beam radiation, the university said, explaining that higher levels of e-beam energy produced p-type areas, while lower levels produced n-type areas.
The technique was used to fabricate high-resolution p-n junctions. When properly passivated, the doping created by the SOG is expected to remain indefinitely in the graphene sheets studied by the researchers, Georgia Tech said.
"This is an enabling step toward making possible complementary metal oxide graphene transistors," said Raghunath Murali, a senior research engineer in Georgia Tech's Nanotechnology Research Center, in a statement.
In the doping process, Murali and graduate student Kevin Brenner began by removing flakes of graphene one to four layers thick from a block of graphite. Next, they placed the material onto a surface of oxidized silicon, then fabricated a four-point contact device. They then spun on films of HSQ (hydrogen silsesquoxane) and cured certain portions of the resulting thin film using e-beam radiation. According to Georgia Tech, the technique provides precise control over the amount of radiation and where it is applied to the graphene, with higher levels of energy corresponding to more cross-linking of the HSQ.
"We gave varying doses of electron-beam radiation and then studied how it influenced the properties of carriers in the graphene lattice," Murali said. "The e-beam gave us a fine range of control that could be valuable for fabricating nanoscale devices. We can use an electron beam with a diameter of four or five nanometers that allows very precise doping patterns."
Electronic measurements showed that a graphene p-n junction created by the new technique had large energy separations, indicating strong doping effects, he added.
Researchers elsewhere have demonstrated graphene doping using a variety of processes including soaking the material in various solutions and exposing it to a variety of gases. Georgia Tech said it believes its process is the first to provide both electron (n-type) and hole (p-type) doping from a single dopant material.
In the process, the doping is believed to introduce atoms of hydrogen and oxygen in the vicinity of the carbon lattice. The oxygen and hydrogen do not replace carbon atoms, but instead occupy locations atop the lattice structure, the university said.
In volume manufacturing, the e-beam radiation would likely be replaced by a conventional lithography process, Murali said. Varying the reflectance or transmission of the mask set would control the amount of radiation reaching the SOG, and that would determine whether n-type or p-type areas are created.
"Making everything in a single step would avoid some of the expensive lithography steps," he said. "Gray-scale lithography would allow fine control of doping across the entire surface of the wafer."
For doping bulk areas such as interconnects that do not require patterning, the researchers coat the area with HSQ and expose it to a plasma source. The technique can make the nanoribbons up to 10 times more conductive than untreated grapheme, Georgia Tech claimed.
However, the researchers noted that a better understanding of how the process works and whether other polymers might provide better results is needed.
"We need to have a better understanding of how to control this process because variability is one of the issues that must be controlled to make manufacturing feasible," Murali said. "We are trying to identify other polymers that may provide better control or stronger doping levels."
A paper describing the technique appeared February 10, in the journal Applied Physics Letters. The research was supported by the Semiconductor Research Corp and the Defense Advanced Research Projects Agency through the Interconnect Focus Center.
IBM Details World's Fastest Graphene Transistor
Joab Jackson, IDG News Service
Feb 5, 2010 12:20 pm
Thanks to a change in recipe, IBM has created a graphene-based processor that can execute 100 billion cycles per second (100GHz), almost four times the speed of previous experimental graphene chips.
With this research, IBM has also shown that graphene-based transistors can be produced by the wafer, which could pave the way for commercial-scale production of graphene chips, said Yu–Ming Lin, the IBM researcher who led the work.
If commercialized, such graphene processors could be the basis of superior signal processing componentry, improving the fidelity of audio and video recording, radar processing and medical imaging.
IBM conducted the work on behalf of the U.S. Defense Department's DARPA (Defense Advanced Research Projects Agency), under a program to develop high-performance RF (radio frequency) transistors. A write-up of the results has been published in the Feb. 5 issue of Science.
Graphene, a single-atom-thick honeycomb lattice of carbon atoms, can transport electrons more quickly than other semiconductors, a quality called electron mobility. "That makes graphene a promising material for high-speed or high-frequency electronic components," Lin said.
This prototype processor was created on a 2-inch wafer, though in principle it could be done on even larger wafers, which should bring the production costs down, Lin said. Graphene is produced by heating a silicon carbide wafer, allowing the silicon to evaporate.
Until now, the downside of graphene has been that it is very sensitive to the environment. During the fabrication process, an oxide layer is deposited over the graphene to form the gate insulator. Typically, this deposition degrades the graphene's electron mobility, due to defects in the oxide that scatter electrons in the graphene. The IBM researchers minimized the damage by separating the graphene from the oxide with a very thin polymer layer.
This new approach has been instrumental in allowing the researchers to almost quadruple the frequency of graphene chips. Last year, research teams from IBM and the Massachusetts Institute of Technology both demonstrated graphene processors capable of frequencies around 26GHz. By comparison, silicon-based transistors of the same gate length (240 nanometers) have only been able to scale up to a clock rate of 40Ghz or so.
It also sets the stage for commercial production. The research shows that "high-quality graphene can be produced on a wafer scale, and graphene transistors can be fabricated with those processes used in the semiconductor industries," Lin said.
Lin cautioned against thinking of graphene as a substitute for the silicon-based microprocessors used in today's computers, at least at anytime in the near future. One major roadblock is that graphene does not work easily with discrete electronic signals, he explained. Because graphene is a zero bandgap semiconductor, meaning there is no energy difference between its conductive and nonconductive states, transistors made of the semiconductor cannot be turned on and off. In contrast, silicon has a bandgap of one electron volt, making it good for processing discrete digital signals, Lin said.
Instead, graphene is better suited for making analog transistors, such as signal processors and amplifiers. Today, such circuitry is largely made from GaAs (gallium arsenide), though GaAs offers nowhere near the same electron mobility, Lin said.
February 5, 2010
IBM hits graphene transistor breakthrough
by Larry Dignan
IBM graphene transistor
IBM Research on Friday announced that it has demonstrated a radio-frequency graphene transistor with the highest frequency so far: 100GHz.
Graphene is a special form of graphite, consisting of a layer of carbon atoms packed in honeycomb lattice. In a nutshell, graphene is like "atomic scale chick wire." Graphene's properties could lead to faster transistors.
One-step graphene doping could enable complementary metal oxide graphene transistors
By Suzanne Deffree, Managing Editor, News -- Electronic News, 2/17/2010
Researchers at the Georgia Institute of Technology have claimed a one-step process that produces both n-type and p-type doping of large-area graphene surfaces and that could facilitate use of the material for future electronic devices.
The doping technique -- produced by applying a commercially available SOG (spin-on-glass) material to graphene and then exposing it to electron-beam radiation -- can also be used to increase conductivity in graphene nanoribbons used for interconnects, according to the university.
Both types of doping were created by varying the exposure time to the to e-beam radiation, the university said, explaining that higher levels of e-beam energy produced p-type areas, while lower levels produced n-type areas.
The technique was used to fabricate high-resolution p-n junctions. When properly passivated, the doping created by the SOG is expected to remain indefinitely in the graphene sheets studied by the researchers, Georgia Tech said.
"This is an enabling step toward making possible complementary metal oxide graphene transistors," said Raghunath Murali, a senior research engineer in Georgia Tech's Nanotechnology Research Center, in a statement.
In the doping process, Murali and graduate student Kevin Brenner began by removing flakes of graphene one to four layers thick from a block of graphite. Next, they placed the material onto a surface of oxidized silicon, then fabricated a four-point contact device. They then spun on films of HSQ (hydrogen silsesquoxane) and cured certain portions of the resulting thin film using e-beam radiation. According to Georgia Tech, the technique provides precise control over the amount of radiation and where it is applied to the graphene, with higher levels of energy corresponding to more cross-linking of the HSQ.
"We gave varying doses of electron-beam radiation and then studied how it influenced the properties of carriers in the graphene lattice," Murali said. "The e-beam gave us a fine range of control that could be valuable for fabricating nanoscale devices. We can use an electron beam with a diameter of four or five nanometers that allows very precise doping patterns."
Electronic measurements showed that a graphene p-n junction created by the new technique had large energy separations, indicating strong doping effects, he added.
Researchers elsewhere have demonstrated graphene doping using a variety of processes including soaking the material in various solutions and exposing it to a variety of gases. Georgia Tech said it believes its process is the first to provide both electron (n-type) and hole (p-type) doping from a single dopant material.
In the process, the doping is believed to introduce atoms of hydrogen and oxygen in the vicinity of the carbon lattice. The oxygen and hydrogen do not replace carbon atoms, but instead occupy locations atop the lattice structure, the university said.
In volume manufacturing, the e-beam radiation would likely be replaced by a conventional lithography process, Murali said. Varying the reflectance or transmission of the mask set would control the amount of radiation reaching the SOG, and that would determine whether n-type or p-type areas are created.
"Making everything in a single step would avoid some of the expensive lithography steps," he said. "Gray-scale lithography would allow fine control of doping across the entire surface of the wafer."
For doping bulk areas such as interconnects that do not require patterning, the researchers coat the area with HSQ and expose it to a plasma source. The technique can make the nanoribbons up to 10 times more conductive than untreated grapheme, Georgia Tech claimed.
However, the researchers noted that a better understanding of how the process works and whether other polymers might provide better results is needed.
"We need to have a better understanding of how to control this process because variability is one of the issues that must be controlled to make manufacturing feasible," Murali said. "We are trying to identify other polymers that may provide better control or stronger doping levels."
A paper describing the technique appeared February 10, in the journal Applied Physics Letters. The research was supported by the Semiconductor Research Corp and the Defense Advanced Research Projects Agency through the Interconnect Focus Center.
posted by Bob at 1:55 PM
0 comments
