Editor: Editorial Department Silicon, is it the end of all electronics?
This record was broken by graphene!
Researchers at Tianjin University and Georgia Tech have created the world's first functional semiconductor made of graphene.
The team's breakthrough opened the door to new electronic products. Research has been published in nature.
*Address: This research has successfully overcome the key technical problems that have long hindered the development of graphene electronics, opened the graphene band gap, and achieved a breakthrough from 0 to 1.
Some netizens exclaimed: This is simply a revolution in electronics, and the breakthrough of epitaxial graphene can make Moore's Law continue to live for decades!
It turns out that silicon is just the beginning.
This discovery could change computing and electronics forever.
Graphene research has been broken through decades of obstacles.
Semiconductors are materials that conduct electricity under certain conditions and are the basic components of electronic devices.
The team's discovery comes at a time when silicon's performance is reaching its limits.
In the past, silicon was the raw material for almost all modern electronics, but faster calculations and smaller and smaller electronic devices have made this route more and more stretched.
Nvidia CEO Lao Huang often says that Moore's Law is dead.
At this time, graphene came into play!
Graphene is a monolithic carbon atom held together by the strongest known bonds.
You must know that natural graphene is not a semiconductor, nor a metal, but a semi-metal.
However, a team led by Walter de Heer, a physics professor at Georgia Tech, has created a graphene semiconductor that is compatible with traditional microelectronic processing methods.
As a result, this semiconductor can be an alternative to silicon.
Why didn't anyone think of replacing silicon with graphene?
That's because there has been one of the biggest hurdles plaguing graphene research for decades, so much so that many people are convinced that graphene can't be used as a semiconductor.
The obstacle is that graphene has no band gap.
At this point, the excited electrons can jump from one energy band to another. This effectively turns the current on and off, thus controlling the conductive switch, while creating a binary system of 0s and 1s used in digital computers.
Band gap plots showing different sizes of bandgaps for conductors, semiconductors, and insulators.
This obstacle was overcome by Professor de Heer and his team.
Professor De Heer said, "Today we have a very strong graphene semiconductor with a mobility rate of 10 times that of silicon, and it has unique properties that silicon does not have. 」
But what we've been racking our brains for every day for the last decade is – can we make graphene materials better, can we become semiconductors? 」
He knew about the potential of graphene 20 years ago.
Graphene's fame is related to the story of two British scientists who won the Nobel Prize in 10 years with hand-torn scotch tape.
Until then, though, there were many people who believed in the potential of graphene in electronics.
When stacked in sheets, graphene can form crystalline transparent structures with unique properties, known as miracle materials.
It is one of the thinnest and lightest materials known, and it is estimated that graphene is harder than diamond and about 100 to 300 times stronger than structural steel.
One square meter of graphene weighs only 00077 grams, but can support up to 4 kg. It can also bend up to 20% of its own length without breaking.
The honeycomb arrangement of carbon atoms in graphene promotes the free movement of electrons, and the ultra-high load of electrons can make electrons run very fast and achieve many cool sci-fi material properties, such as touch screens, airplanes, and so on.
Early in his career, Professor de Heer began exploring the capabilities of carbon-based materials as potential semiconductors, and in 2001 he turned his attention to two-dimensional graphene.
The team hopes to introduce three properties of graphene into electronics:1Sturdy; 2.Handles very large currents; 3.Works without heating and separation.
During the experiment, the team came up with the idea of using a special furnace to grow graphene on silicon carbide wafers.
They made a breakthrough and produced a single layer of epitaxial graphene that grows on the crystal surface of silicon carbide.
They found that if made correctly, epitaxial graphene would chemically bind to silicon carbide and begin to exhibit semiconductor properties.
For the next decade, the Georgia Tech team worked on the material, collaborating with the Tianjin International Research Center for Nanoparticles and Nanosystems at Tianjin University.
TCNN Director Ma Lei.
A key breakthrough: donating electrons to the system with a 10-fold higher mobility than silicon.
Naturally, graphene is neither a semiconductor nor a metal, but a semi-metal.
A bandgap is a material that can be turned on and off when an electric field is applied, and all transistors and silicon electronic devices rely on this working principle.
The main problem in graphene electronics research is how to open and close the band gap so that graphene works like silicon.
However, if you want to make a functional transistor, you have to make most of the semiconductor material controllable, which can compromise the performance of graphene.
To prove that graphene can function as a semiconductor, the team needed to measure its electronic properties without damaging it.
Researchers place atoms on graphene and donate electrons to the system – a technique known as doping to see if a material is a good conductor. In this way, there is no need to damage the material or properties of graphene.
The researchers used heated silicon carbide wafers that forced the silicon to evaporate before the carbon, effectively leaving a layer of graphene on the surface.
The results show that the mobility of graphene semiconductors is 10 times higher than that of silicon.
Electrons can move with extremely low resistance, which translates into faster calculations in electronics.
It's like driving on gravel roads instead of highways. The former is more efficient, doesn't heat up excessively, and is fast enough to allow electrons to move quickly. Professor de Heer explained.
This graphene product is currently the only two-dimensional semiconductor with the necessary characteristics of nanoelectronics, and its electronic performance is far superior to other two-dimensional semiconductors currently under development.
Ma Lei, director of the Tianjin International Research Center for Nanoparticles and Nanosystems, and co-author, said
A long-standing problem with graphene electronics is that graphene doesn't have the right bandgap to open and close in the right proportions. Our technology enables the bandgap, which is the most critical step in achieving graphene-based electronics.Wright Brothers moments.
This epitaxial graphene is likely to cause a paradigm shift in the field of electronics and give rise to many new technologies.
It allows the use of the quantum mechanical wave properties of electrons, which is required by quantum computing.
According to Professor de Heer, the next generation of electronics can be expected. Before silicon, there were vacuum tubes, and before that, there were wires and telegraphs.
In the history of electronics, silicon is only one of the forms for a while, and the next step is probably graphene.
Professor de Heer said that for him, it was like a Wright Brothers moment.
The Wright brothers built an airplane that could fly 300 feet in the air. Skeptics ask: why do we need airplanes when there are already trains and ships in the world? But they persevered, and after that, airplanes could take people across the ocean.
Ultra-high mobility semiconductors.
There is a lack of intrinsic band gaps in graphene. Over the past two decades, attempts to alter the band gap through quantum confinement or chemical functionalization have been unsuccessful.
In this work, the researchers demonstrated that the semiconductor surface stone ene (SEG) on a single crystal silicon carbide substrate has 06 EV with a bandgap and reached more than 5000
The room temperature mobility is 10 times greater than that of silicon and 20 times greater than that of other 2D semiconductors.
In other words, viable semiconductor-graphene was born.
When silicon evaporates from the surface of a silicon carbide crystal, the carbon-rich surface crystallizes to produce a graphene multilayer film. The first graphite layer formed on the silicon end face of the SiC is an insulating epidermene layer, which is partially covalently bonded to the surface of the SiC.
Spectroscopic measurements of the buffer layer show semiconductor characteristics, but the mobility of this layer is limited due to disorder.
In this paper, the researchers demonstrate a quasi-equilibrium annealing method that generates SEG (i.e., an ordered buffer layer) on macroscopic atomically flat terraces with the SEG lattice aligned with the SIC substrate.
Seg is chemically, mechanically, and thermally robust, can be patterned using traditional semiconductor fabrication techniques, and seamlessly attaches to semi-metallic epigraphene. These fundamental properties make SEG suitable for nanoelectronics.
As shown in Figure (a)(b) below, the conventional epigraphene and buffer layer are grown in a closed controlled sublimation (CCS) furnace, where 35mm × 4.The 5 mm semi-insulated SiC chips are annealed in 1 bar AR in a cylindrical graphite crucible with a temperature range of 1300 °C to 1600 °C (shown in figure (C) below).
The crucible is heated by an eddy current induced in the RF source** circle, there is a small leak in the crucible, and the rate at which the silicon escapes from the crucible determines the rate at which graphene forms on the surface. As a result, the growth temperature and graphene formation rate are controlled.
Stack the two chips on top of each other with the C side of the bottom chip (Source) facing the Si side of the top chip (Seed).
At high temperatures, a small temperature difference between the chips results in a net mass flow from the bottom chip to the top chip, resulting in the gradual growth of large terraces on the seed chips and the growth of a uniform SEG film on them.
SEG grows in three stages:
In the first stage, the chip is heated to 900 °C in a vacuum for about 25 minutes to clean the surface;
In the second stage, the sample is heated to 1,300 °C for approximately 25 min in 1 bar AR, resulting in a regular array of double-layer SiC steps and approximately 02 m wide staircase.
In the third stage, the SEG-coated terraces are grown in AR at 1600°C, 1 Bar, where stepped bunching and stepped flow produce large atomically flat terraces, and the buffer layer grows under quasi-equilibrium conditions established between the C and Si planes.
The most important parameters in the process are the temperature t, the temperature difference between the chips δt and the annealing time t, when t = 1600 1700 °C, the annealing time is usually 1-2 hours. The temperature difference δt depends on the crucible design and is estimated to be around 10 °C to provide sufficient vapor pressure difference required for mass transfer between the two chips.
SEG characterization.
Figure (a) below illustrates 35 mm×4.Compound electron microscope (SEM) image of a 5 mm wafer.
SEM has been adjusted to provide a sharp contrast between the SIC (white area) and SEG (gray area). Approximately 80% of the surface is covered by SEG. Graphene will appear as dark patches (the black dots seen here are dust particles). The largest step-free area is about 05mm×0.3mm。
Figure (B) is a low-temperature atomic-resolution image of SEG using scanning tunneling microscopy (STM).
The STM image shows a graphene honeycomb lattice (green) that is spatially covered.
The superperiodic structure (red rhombic and purple hexagon) modulated, corresponding to the SEG height modulation of about 100 pm, is covalently bonded to the substrate in part.
Low-energy electron diffraction (LEED) is used to identify SEG and verify its atomic registration with the SIC substrate.
Figure (c) above shows the characteristic 6 3 6 6 3 3 R30° diffraction pattern (LEED) of the SEG lattice, showing the graphene crystal structure of the SEG and the crystal arrangement of the SEG relative to the atoms on the SiC substrate. There are no abundant traces of graphene in conventionally produced buffer layer samples.
Figure (d) is a Raman plot of the 50 m 50 m region with a resolution of 1 m, the Raman spectrum (1 100 m) is very sensitive to graphene and SEG, and the traces of graphene are easily identified by its strong characteristic 2D peaks, which show that there is no graphene on the surface.
Figure (E) shows a low-temperature STS image of SEG, mapping the density of states (DOS) of SEG as a function of Fermi energy. The image shows 0Clear bandgap of 6 EV.
Figure (a) shows that the conductivity of the sample increases monotonically with increasing temperature. The room temperature conductivity ranges from 1e-3 s to 8e-3 s, corresponding to a resistivity of 125 to 330. The low temperature value can be reduced by up to 1000 times.
Figure (B) represents the charge density, STS measurements show that SEG is charge-neutral in nature, so the charge is caused by ambient gases (including trace amounts of volatile organic compounds) and residual resistance from photolithography.
Figure (d) shows that the mobility of the material increases with increasing temperature, tending to saturate at higher temperatures. The maximum mobility measured was 5500
Room temperature SEG conductivity, charge density, and mobility are all in the typical range for epigraphene. However, the temperature dependence is similar to that of doped semiconductors with a deep body state.
By measuring the semiconductors and DOS, we can** the response of the field-effect transistor:
Figure (a) shows the calculated resistivity of the DOS**SEG channel, assuming an ideal dielectric with an SEG mobility of 4000
*Breaking ratio of more than 1e6 at room temperature.
Figure (b) shows the charge density as a function of fermi energy. At t = 300K, the on-voltages of the n and p branches are expected to be +0., respectively34v and 023v。
References:
List of high-quality authors