News & Analysis
Comment
iniewski
Larger bandgap means lower leakage. That is basic physics. But it does not mean ...
sranje
Colin rocks
New material for semis said to beat silicon
R Colin Johnson
2/1/2011 11:43 AM EST
PORTLAND, Ore.—A new semiconductor material called molybdenite (MoS2) is claimed to be 100,000 times lower power than silicon, plus will allow the fabrication of much smaller transistors, according to researchers at Switzerland's Ecole Polytechnique Federale de Lausanne (EPGL).
As a next-generation semiconductor material, molybdenite also beats graphene by possessing a bandgap, according the EPGL.
The EPFL claims that molybdenite is an abundant mineral which is already used in steel alloys and as an additive in lubricants. The material is being developed for the first time as a next-generation semiconductor at its EPGL's Laboratory of Nanoscale Electronics and Structures (LANES).
"[Molybdenite] has real potential in the fabrication of very small transistors, light-emitting diodes and solar cells," said EPFL professor Andras Kis.
According to Kis, one or molybdenite's main attributes is that unlike silicon, which is a three-dimensional crystal, molybdenite is an inherently two-dimensional material, permitting thin films as thin as 6.5 angstrom to be relatively easily fabricated (one nanometers equals 10 angstroms) that have an equal electron mobility to two-nanometer thick layers of silicon.
In addition, unlike graphene—which does not possess a bandgap—molybdenite has a bandgap of 1.8 electron-volts, putting it between gallium arsenide (1.4) and gallium nitride (3.4)—opening the possibility of fabricated chips that can handle both electrical and optical functions.
Molybdenite--MoS2--here is used to create an ultra-lower power field effect transistors (FET) by acting as its channel on a silicon-on-insulator substrate using high-k dielectric (HfO2) gate oxide.
As a next-generation semiconductor material, molybdenite also beats graphene by possessing a bandgap, according the EPGL.
The EPFL claims that molybdenite is an abundant mineral which is already used in steel alloys and as an additive in lubricants. The material is being developed for the first time as a next-generation semiconductor at its EPGL's Laboratory of Nanoscale Electronics and Structures (LANES).
"[Molybdenite] has real potential in the fabrication of very small transistors, light-emitting diodes and solar cells," said EPFL professor Andras Kis.
According to Kis, one or molybdenite's main attributes is that unlike silicon, which is a three-dimensional crystal, molybdenite is an inherently two-dimensional material, permitting thin films as thin as 6.5 angstrom to be relatively easily fabricated (one nanometers equals 10 angstroms) that have an equal electron mobility to two-nanometer thick layers of silicon.
In addition, unlike graphene—which does not possess a bandgap—molybdenite has a bandgap of 1.8 electron-volts, putting it between gallium arsenide (1.4) and gallium nitride (3.4)—opening the possibility of fabricated chips that can handle both electrical and optical functions.
Molybdenite--MoS2--here is used to create an ultra-lower power field effect transistors (FET) by acting as its channel on a silicon-on-insulator substrate using high-k dielectric (HfO2) gate oxide.
|
|
|
|
|
Navigate to related information
prabhakar_deosthali
2/1/2011 12:23 PM EST
This is good discovery which will allow further miniaturization and more densities on the future chips. But the question is whether the Molybdenite is available as abundantly as Silicon?
Sign in to Reply
truworth
2/2/2011 4:22 AM EST
Awesome post. I am really impress with your efforts put in the blog. It's very useful for us. Thanks
http://www.mt2md.com
Sign in to Reply
selinz
2/1/2011 1:30 PM EST
Presumably the comment about "has a bandgap" really means "is a direct bandgap material?"
Sign in to Reply
h46av8r
2/4/2011 11:08 AM EST
@selinz, graphene in its 2D form is a zero-bandgap semiconductor, or "semi-metal". That is, the conduction and valence bands just touch at a so-called Dirac point. However, the article is wrong in that graphene, as envisioned for transistors, would be patterned into ribbons. The lateral quantum confinement then does open a bandgap just like carbon nanotubes only un-rolled. In theory, you could control the bandgap by controlling the width of the ribbon. In fact, turn the ribbon 30 degrees and it goes metallic so in once piece of graphene you could go metal-semiconducting-metal i.e. the source and drain and interconnects are automatic. All this requires, however, controlling the ribbon width with atomic precision...
Sign in to Reply
Robotics Developer
2/1/2011 1:49 PM EST
I would like to know what the leakage current will be with these type of devices. When can we expect the prototype "silicon" devices to come out of the lab for testing and characterization? Is there any increased (or reduced) manufacturing processes involved in the production of these types of devices? Neat discovery! I would love to know more..
Sign in to Reply
Code Monkey
2/1/2011 3:50 PM EST
Maybe molybdenite processes will be ready about the same time EUV is ready for full-scale production, providing the next step in Moore's law.
Sign in to Reply
iniewski
2/1/2011 5:53 PM EST
Colin, saying that a new material is 100,000 more efficient in power is a big stretch to say the least...for starters semiconductor material does not dissipate power per se but semiconductor devices build on on it do...second I highly doubt it will be 100,000 times, seem silly to claim that...plus really you are only talking about the new material for electron channel, everything else as your drawing shows is a silicon technology! dr Kris
Sign in to Reply
Neo1
2/1/2011 9:24 PM EST
I guess their 100,000 number comes from the maximal range i.e. the assumption that one can actually fabricate a device with 6.5 Angstroms gate thickness and make it switch like Si. But the diagram still shows the other junctions to be still Si, so what's the secret here?
Sign in to Reply
iniewski
2/1/2011 9:34 PM EST
thank you @Neo1, but even 6.5A does not produce 100,000 factor in lowering the power, today's MOSFETs have gate oxide 20A thick or so, the difference is not that dramatic...the power "number" likely comes from the bandgap value, 1.8eV vs 1.1eV for silicon but this again highly misleading, there are several wider bandgap materials available, so what, some have bandgaps larger than 3eV, should we claim they offer 100,000,000 gains in lowering the power (if you calculate exp(-EA/kT) factor)...dr Kris
Sign in to Reply
resistion
2/2/2011 12:34 PM EST
The big drawback is molybdenite is a compound, with extra potential for defects due to stoichiometry.
Sign in to Reply
iniewski
2/2/2011 1:40 PM EST
good point @resistion...I work in compound semiconductor industry and defects are big issues (orders of magnitude behind silicon)...Krisd
Sign in to Reply
sranje
2/4/2011 1:31 PM EST
Colin rocks
Sign in to Reply
iniewski
2/9/2011 11:36 AM EST
Larger bandgap means lower leakage. That is basic physics. But it does not mean that you can replace silicon with any wider bandgap material. If that was the case silicon would be finished long time ago ;-)...dr Kris
Sign in to Reply