User:Sqing73/sandbox/Carbon nanotube computer

From Wikipedia, the free encyclopedia

Carbon Nanotube Computer is a computer whose central processor is based entirely on carbon nanotube transistors. In 2013, a group in Stanford University built the first Carbon Nanotube Computer after several years of dealing with some material problems. The first carbon nanotube computer is called Cedric(a rough acronym for "carbon nanotube digital integrated circuit"). Cederic is able to finish some basic tasks just like the first generation silicon computers[1]. Compared with normal silicon transistor computers, carbon nanotube transistor computers operates faster and at a lower supply voltage of the same size[2]. While these advantages give carbon nanotube computers to outperform and replace traditional computers, scientists still are facing a lot of problems to build large-scale computers.

Compositions[edit]

The first carbon nanotube computer is composed of 178 carbon nanotube field-effect transistors(CNFETs) and each CNFET is made of ~10-200 carbon nanotubes(CNTs).[1]

CNT are an allotrope of carbon. They take the form of cylindrical carbon molecules and have novel properties that make potentially useful in a wide variety of applications in nanotechnology, electronics, optics and other fields of materials science.[2] As an efficient electrical material, carbon nanotube are either metallic or semiconducting, which makes it feasible to build carbon nanotube transistors.

CNFETs are quite like metal-oxide-semiconductor field-effect transistors(MOSFET), which are widely used on modern electronic devices. A CNFET is built on a source and drain, where the current flows into and out of the device, a channel between the two where the current flows, and a gate near the channel to control current flow.[4] A complete transistor usually have four major components: a source, a drain, a channel, and a gate. Channels are like bridges that connect all the other components. What really makes a difference is the material of channels and a MOSFET's channel is made totally of silicon. However, a CNFET's channel is fabricated purely of carbon nanotubes.[4]

Capabilities[edit]

Although it runs stored programs and is programmable, just like a silicon-based computer, the first carbon nanotube computer only performs some basic tasks. These tasks include concurrently executing a counting program and an integer-sorting program on a basic operating system, which does demonstrate its multitasking ability. [1]Also, it executes 20 different instructions from the commercial MIPS instruction set. In general, the computer is comparable to Intel 4004[3] and VAX-11(1970)[1], which are also the first silicon transistor computers.

Obstacles[edit]

To build the first carbon nanotube computer, technologists have to face two major obstacles:

Mis-position[1][edit]

Mis-position means that it is difficult to position carbon nanotubes in arrays on a surface. Hence, some of them may extend beyond the confines of their channel and connect to other nanotubes. This chaotic may lead to incorrect logic functionality.

Metallic nanotubes[1][edit]

Metallic carbon nanotubes is caused by carbon synthesis. In the process of growing carbon nanotubes, tubes are inevitably built of different diameters and atomic arrangements. Such variation in the structure of nanotubes will make some of them semiconducting and others metallic. Metallic nanotubes have little or no bandage, just like a wire, resulting in uncontrollable current flows[1]. Such current will further cause power waste and ,more seriously, logic operation errors because it cannot be turned off by the gate.[4]

Solutions[edit]

In 2004, diverse group of students and postdocs at Stanford University, led by professors Subhasish Mitra and H.-S. Philip Wong, set out to tackle these obstacles. After 7 to 8 years study, they finally figure out a way to transform the nano-material from a laboratory curiosity into the stuff of today's chips.[4]

To mis-position[edit]

The approach technologists use to align carbon nanotubes in parallel is to settle them on a crystalline quartz substrate. The CNTs are grown by chemical-vapor deposition with methane at 865 centigrades on an annealed quartz substrate.[1] Then, they are transferred to a traditional chip substrate(made of a layer of silicon dioxide on top of silicon), coated by a thin layer of metal, which acts as an adhesive. This metal is then chemically etched away, leaving the array of highly aligned carbon nanotubes behind. [4]

However, this method still may result in about 0.5 percent of carbon nanotubes not in order. To deal with such minor error, technologists design an algorithm, based on graph theory, determines which areas of the circuit should be etched away. In this way, they don't need to inspect every wafer and carbon nanotube in order to find the misaligned ones. [4]

To metallic nanotubes[edit]

The main idea of the solution adapted by the technologist group is to burn metallic carbon nanotubes with sufficient voltage. After the circuits are made, technologists will apply a voltage to gate(used to control current flows) directly to turn off all of the semiconducting nanotubes. Then they apply a voltage across each source and drain region and current flows only into metallic nanotubes because the rest ones have already been closed. With a sufficient charge, even the relatively low resistance of these bad nanotubes will create brief heat spike that fries the nanotubes.[4]

Future[edit]

The miniaturization of electronic devices has been the principle driving force behind the semiconductor industry.[1] After 50 years of improvements, silicon transistors may have reach their material limits not far in the future. Up till now, carbon nanotube transistors seem to be the most exciting and mature replacement for silicon transistors because of their higher energy efficiency[1] and similar fabrication process as traditional one. Even though they have relatively optimistic future, there are still lots of problems to solve. The next step of development is to build a large-scale computer comparable to modern computers.[5]

Pros[edit]

  • Fabrication and design techniques can be perfectly integrated into today's semiconductor fabrication facility.[4]
  • Transistors are faster and more energy efficient than those made by silicon.[3]
  • Carbon nanotube's heat resistivity can help computers keep low temperature while running fast.[3]

Cons[edit]

  • Only 178 transistors are obviously not enough to complete complex tasks. [3]
  • Messy wire and pin packaging with external devices lowers the computer efficiency.[3]

References[edit]

  1. ^ a b c d e f g h i Shulaker, Max M.; Hills, Gage; Patil, Nishant; Wei, Hai; Chen, Hong-Yu; Wong, H.-S. Philip; Mitra, Subhasish (26 September, 2013). "Carbon nanotube computer". Nature. 501 (7468): 526–530. doi:10.1038/nature12502. ISSN 0028-0836. {{cite journal}}: Check date values in: |date= (help)
  2. ^ "Carbon nanotube". ScienceDaily. Retrieved 2018-11-03.
  3. ^ a b c d e Bourzac, Katherine (September 25, 2013). "The First Carbon Nanotube Computer". MIT Technology Review. Retrieved 2018-10-16.
  4. ^ a b c d e f "How We'll Put a Carbon Nanotube Computer in Your Hand". IEEE Spectrum: Technology, Engineering, and Science News. Retrieved 2018-10-16.
  5. ^ "Scientists: Carbon Nanotubes Would Outperform Silicon Transistors at the Same Scale". IEEE Spectrum: Technology, Engineering, and Science News. Retrieved 2018-11-03.

External links[edit]