News & Analysis
Comment
Steven Huo
Thanks so many guys for putting forward their opinions. As a third party of TI, ...
goafrit
Can they even post it online. Thank you.
TI growing large-area graphene
R Colin Johnson
3/3/2011 7:55 PM EST
PORTLAND, Ore.—Preparing the way for carbon-based semiconductors, Texas Instruments Inc. is perfecting the growth of graphene sheets. By carefully characterizing a method of growing graphene monolayers, TI hopes to pave the way for faster, smaller and lower power electronics based on carbon instead of silicon.
"This research is teaching us a lot on fundamental mechanisms of graphene growth," said TI Fellow Luigi Colombo. "We believe that the results of this work will lead scientists and engineers to further increase the size of the single crystal graphene and also improve the electrical characteristics of the material."
TI recently demonstrated growing large-grain graphene crystals—up to half a millimeter in diameter—in collaboration with IBM's T.J. Watson Research Center, the Nanoelectronic Research Initiative (NRI), and the University of Texas. Using low-pressure chemical vapor deposition inside copper-foil enclosures, with methane as a precursor, the resulting graphene films were then transferred to doped single-crystal silicon and silicon-on-insulator substrates where their electrical characteristics were tested. The team is also pursuing ways of growing graphene directly on dielectric substrates.
"Growth on other substrates, preferably dielectrics or metals more stable than copper, are also being investigated," said Colombo.
Large-grain crystals of graphene start as perfect hexagons but grow more quickly at their tips (rather than the sides of the hexagon) resulting in snowflake-like fractal islands.
For several years now, TI has been quietly pursuing graphene growth methods, previously reporting surface nucleation followed by a two-dimensional growth process on copper substrates with CVD. However, the domains were only 10-to-20 microns, 30 times smaller than the half-millimeter (500 micron) domains on which TI is now reporting.
Using low-energy electron microscopy the research team was able to confirm that the films were relatively uniform single-crystal graphene monolayers. Using Raman spectroscopy, the researchers were able to confirm an electron mobility of 4000 square centimeters per volt second--compared to 1,400 cm2/Vs for silicon and 8500 cm2/Vs for gallium arsenide. Theoretically, graphene can achieve electron mobilities of 10,000-to-100,000 cm2/Vs.
"4000 [cm2/Vs] is reasonably high, but not as high as the highest value possible for exfoliated films, so we still have some improvements to make in our process," said Colombo. "But we have high hopes for these large-domain films."
The large-domain graphene growth was promoted inside the copper-foil cage at a relatively high temperature—over 1,035 degrees Celsius (1,895 Fahrenheit). After processing the films, field-effect transistors were fabricated by transferring the films to highly doped silicon substrates, which served as the back-gate contact, with source and drain electrodes made from nickel. Electron mobility was inferred from measuring resistance was as a function of back-gate voltage.
Funding was provided by the National Science Foundation, the Office of Naval Research and Semiconductor Research Corp.'s NRI at the SouthWest Academy of Nanoelectronics.
"This research is teaching us a lot on fundamental mechanisms of graphene growth," said TI Fellow Luigi Colombo. "We believe that the results of this work will lead scientists and engineers to further increase the size of the single crystal graphene and also improve the electrical characteristics of the material."
TI recently demonstrated growing large-grain graphene crystals—up to half a millimeter in diameter—in collaboration with IBM's T.J. Watson Research Center, the Nanoelectronic Research Initiative (NRI), and the University of Texas. Using low-pressure chemical vapor deposition inside copper-foil enclosures, with methane as a precursor, the resulting graphene films were then transferred to doped single-crystal silicon and silicon-on-insulator substrates where their electrical characteristics were tested. The team is also pursuing ways of growing graphene directly on dielectric substrates.
"Growth on other substrates, preferably dielectrics or metals more stable than copper, are also being investigated," said Colombo.
Large-grain crystals of graphene start as perfect hexagons but grow more quickly at their tips (rather than the sides of the hexagon) resulting in snowflake-like fractal islands.
For several years now, TI has been quietly pursuing graphene growth methods, previously reporting surface nucleation followed by a two-dimensional growth process on copper substrates with CVD. However, the domains were only 10-to-20 microns, 30 times smaller than the half-millimeter (500 micron) domains on which TI is now reporting.
Using low-energy electron microscopy the research team was able to confirm that the films were relatively uniform single-crystal graphene monolayers. Using Raman spectroscopy, the researchers were able to confirm an electron mobility of 4000 square centimeters per volt second--compared to 1,400 cm2/Vs for silicon and 8500 cm2/Vs for gallium arsenide. Theoretically, graphene can achieve electron mobilities of 10,000-to-100,000 cm2/Vs.
"4000 [cm2/Vs] is reasonably high, but not as high as the highest value possible for exfoliated films, so we still have some improvements to make in our process," said Colombo. "But we have high hopes for these large-domain films."
The large-domain graphene growth was promoted inside the copper-foil cage at a relatively high temperature—over 1,035 degrees Celsius (1,895 Fahrenheit). After processing the films, field-effect transistors were fabricated by transferring the films to highly doped silicon substrates, which served as the back-gate contact, with source and drain electrodes made from nickel. Electron mobility was inferred from measuring resistance was as a function of back-gate voltage.
Funding was provided by the National Science Foundation, the Office of Naval Research and Semiconductor Research Corp.'s NRI at the SouthWest Academy of Nanoelectronics.
|
|
|
|
|
Navigate to related information
yalanand
3/3/2011 10:57 PM EST
"Theoretically, graphene can achieve electron mobilities of 10,000-to-100,000 cm2/Vs." does it mean the graphene transistor will be 4 times faster than the normal CMOS transistors ?
Sign in to Reply
GroovyGeek
3/4/2011 2:27 AM EST
Theoretically is the key word... Mobility in nanoribbons is more like 100-200. Furthermore graphene has no bandgap so it makes an incredibly lousy switch. Many have given up on that particular application.
Graphene is the new nanotube - the next great white hope that will go nowhere, at least as far as high performance CMOS replacement is concerned.
Sign in to Reply
Mark.Hewitt_#1
3/4/2011 12:15 PM EST
Perhaps you are not familiar with doping techniques that open the band gap in Graphene - in eetimes December the feature about IBM opening the band gap for Graphene's.
We are clearly in new territory here with Graphene's however I see the functional use to far exceed the expatiations of nano tubes.
Sign in to Reply
iniewski
3/4/2011 11:26 AM EST
I don't agree GroovyGeek, I think this is the most exciting material that will eventuallu replace silicon in certain applications. Material properties are very unique. We just don't know yet all manufacturing issues...Kris
Sign in to Reply
R_Colin_Johnson
3/4/2011 12:06 PM EST
Texas Instruments made clear to me that their goal is to eventually reach the wafer scale in growing graphene monolayers--the same way silicon is grown today. However, it will be many years before anybody progress from growing millimeter-scale patches to achieving the kind of control necessary for wafer scale growth.
Sign in to Reply
yalanand
3/5/2011 7:14 AM EST
@R_Colin_Jhonson Could you please comment on the speed of Graphene transistor compared to the CMOS trasistor ?
Sign in to Reply
Rchandta1
3/4/2011 3:07 PM EST
What are the theoretical limit of electron's mobility in Si and GaAs and GaN?
Sign in to Reply
iniewski
3/4/2011 4:19 PM EST
I agree Colin, sounds they will be able to grow graphene monolayers across the whole wafer one day but material science always takes much longer than people believe...I would give 10 years until graphene is commercially available...and it will help achieve Kurzweil's singularity ;-)...Kris
Sign in to Reply
Colombo
3/4/2011 4:50 PM EST
We just want to clarify that the graphene growth referred to in this article was all performed in Prof. Rod Ruoff’s laboratory at UT Austin, in which we have been working with his group over the past 4 years in developing the process.
Furthermore, the precise nature of the collaboration with IBM on this work was on the verification that the graphene grown at UT Austin by this method was indeed single crystal as described in our recently accepted joint article currently online on the Journal of the American Chemical Society website.
Sign in to Reply
iniewski
3/4/2011 4:58 PM EST
Colombo: would it be possible to get a copy of that article that you are describing? kris.iniewski@gmail.com
Sign in to Reply
goafrit
3/6/2011 11:08 AM EST
Can they even post it online. Thank you.
Sign in to Reply
resistion
3/5/2011 7:51 AM EST
The temperature is way too high.
Sign in to Reply
david.may
3/5/2011 11:21 PM EST
we probably need a layman's guide on all this and what it mean's.
but it seems there was lot's of talk about Superconductivity at 'room temperature' with graphene and using that in CPU design etc...
http://torcuil.wordpress.com/2008/03/24/superconductivity-at-room-temperature-with-graphene/
stating room temperature CPU's always struck me as an odd goal as silicon CPU's have always run hotter inside, even if they only get warm by the time that generated heat reaches the outside packaging/lid
however there are high-temperature uperconductors and talk of making Graphene: A Superconductor
http://physics.aps.org/synopsis-for/10.1103/PhysRevLett.104.136803
http://techgenie.com/latest/the-creation-of-graphene-a-superconductor/
Sign in to Reply
david.may
3/5/2011 11:21 PM EST
by far the most interesting though is apparently this latest practical tool to finally measure and even effect Andreev bound states (ABS)?
http://www.sciencedaily.com/releases/2011/02/110214142348.htm
"If you have two superconductors with a normal metal between, you can actually transport the superconductivity across the normal material via these bound states, even though the normal material doesn't have the electron pairing that the superconductors do," Mason said.
ABS are extremely difficult to measure or to observe directly. Researchers can measure conduction and overall magnitude of a current, but have not been able to study individual ABS to understand the fundamental physics contributing to these unique states.
Mason's group developed a method of isolating individual ABS by connecting superconducting probes to tiny, nanoscale flakes of graphene called quantum dots. ..."
so assuming their superconducting probes where not this superconducting graphene but standard graphene then cant they also take even more advantage of the higher temperature inside your average silicon and make use of that heat to induce some super conductivity property in a hybrid core today...
or did i misunderstand the concepts due the scientific ether/or thinking in these links? why cant we have both states and make use of them interacting.
Sign in to Reply
Steven Huo
3/7/2011 6:30 AM EST
Thanks so many guys for putting forward their opinions. As a third party of TI, Embest info&tech Co.,Ltd http://www.embedinfo.com/english/index.asp are able to expect a more bright future in business.
Sign in to Reply