Scientists Confirm Existence Of Atom-Sized Electronic Devices Within Carbon Nanotubes

   BERKELEY, CA. -- Scientists with the Ernest Orlando Lawrence BerkeleyNational Laboratory (Berkeley Lab) have confirmed the existence of atom-sized electronic devices on nanotubes, hollow cylinders of pure carbonabout 50,000 times more narrow than a human hair in diameter. Nanotubedevices have been predicted by theorists but this is the firstdemonstration that such devices actually exist.

   Alex Zettl, a physicist with Berkeley Lab's Materials SciencesDivision and a professor of physics on the University of California'sBerkeley campus, led a study in which nanotubes of pure carbon wereshown to function as a two-terminal electronic device known as a diode.

   "What we are seeing is the world's smallest room temperaturerectifier, one that is only a handful of atoms in size," says Zettl. "When we grow nanotubes, electronic devices naturally form on them."

   Past attempts to identify nanotube devices employed submicron-sizedelectrode contact pads that could only measure small isolated sectionsof the tube.  Evidently, the experimenters were measuring the wrongsections.  Zettl succeeded by measuring nanotubes along their entirelength.  He accomplished this through the use of the ultrafine tip of ascanning tunneling microscope.

   The research was reported in a recent issue of the magazine Science(10/10/97).  Co-authoring the paper with Zettl were Phil Collins, ofZettl's research group; Hiroshi Bando, from the ElectrotechnicalLaboratory in Japan; and Andreas Thess and Richard Smalley of RiceUniversity.

   Nanotubes are only a few nanometers (billionths of a meter) indiameter. When made exclusively from carbon molecules, they arechemically inert, about 100 times stronger than steel, and offer a fullrange of electrical and thermal conductivity possibilities. 

   Carbon nanotubes were discovered by the Japanese electronmicroscopist Sumio Iijima.  They are created by heating ordinary carbonuntil it vaporizes, then allowing it to condense in a vacuum or an inertgas.  The carbon condenses in a series of hexagons, like sheets ofgraphite, that curl and connect into hollow tubes.

   Depending upon its diameter, a pure carbon nanotube can conduct anelectrical current as if it were a metal, or it can act as asemiconductor, meaning it will only conduct a current beyond a criticalvoltage.  According to a theory proposed by Berkeley Lab physicistsMarvin Cohen and Steven Louie, both also with UC Berkeley, an electronicdevice could be created at the interface between two dissimilarnanotubes, one that acts as a metal and one that acts as asemiconductor.  This would create a "Schottky barrier," which means thecurrent will only flow in one direction -- from the semiconductor to themetal.  Under the scheme envisioned by Cohen and Louie, the twodissimilar tubes would be connected by the introduction of pentagon-heptagon pair defects (rings of five and seven carbon atoms) into theinterface region.

   Zettl and Collins have been able to confirm that Schottky barriers doexist along carbon nanotubes.  The key to their success was the scanningtunneling microscope or STM.  An STM features a metallic tip that is theworld's smallest pyramid: a few layers of atoms descending in numberdown to a single atom at the point.  The Berkeley researchers wouldbring the tip of an STM into contact with a tangle of nanotubes on asubstrate then slowly withdraw it.  Van der Waals forces would induce asingle nanotube to stick to the tip of the STM and the researchers wouldcarefully stretch it out from the other nanotubes on the substrate, muchlike unravelling a single fibre from a nest of thread.  Once a singlenanotube was extracted, the researchers would then slide the STM tipacross its entire surface to measure variations in an electrical currentpassing through.

   "We measured distinct changes in the conductivity as the activelength of the nanotube was increased, suggesting that different segmentsof the nanotube exhibit different electronic properties," says Zettl. "The changes occurred over very short lengths and were suggestive of on-tube nanodevices."

   Zettl does not expect nanotubes to replace silicon overnight in theelectronics industry but can see this as a possibility down the road. Silicon must be doped with other atoms to make an electronic device.  Asthe size of a device shrinks, the dopant atoms eventually begin to moveabout, degrading the device's performance.  Heat also becomes a problemdespite silicon's good thermal conductivity.  The use of diamond film,with its exceptionally high thermal conductivity, has been proposed toprotect silicon-based devices but this adds further complications to themanufacturing process.  Size and heat are no issue for nanotubes becausethey are covalently bonded (which means their atoms are locked firmlyinto place) and are predicted to be even better thermal conductors thaneither silicon or diamond at room temperature.

   "Silicon is eventually going to hit a brick wall where devices can'tbe made any smaller," Zettl says.  "Nanotubes are already smaller anddon't have a problem with heat.  You could not ask for anything betterin a material."

   Rather than wiring individual devices in nanotubes for specificpurposes, as is done with silicon chips, Zettl suggests a betterapproach might be to make a "tube cube," a block of nanotubes that wouldbe densely packed with billions upon billions of devices.  The tubecould then be wired to form a random network of "nanocomputers."  Thisrandom network would be able to train itself to perform tasks,reconfiguring its input/output architecture to improve its performanceas it learns and develops.  In other words, this random computer wouldnot just get older, it would get better.

   "The idea is not as far off as you might think," says Zettl, whosegroup has already constructed and wired up a tube cube of sorts.  Thecube cannot yet perform any useful function, but Zettl says it doesyield some "interesting" responses to input signals.

   "Nanotube technology might be exploited in a conventional manner orwe might have to go off in a completely different direction," saysZettl.  "The technology simply has too much potential to not figure outhow to use it."

   The Berkeley Lab is a U.S. Department of Energy national laboratorylocated in Berkeley, California.  It conducts unclassified scientificresearch and is managed by the University of California.