IBM researchers Alfred Grill, Deborah Neumayer and Dinkar Singh have developed a process in which the diameter of carbon nanotubes grown by chemical vapor deposition (CVD) or plasma enhanced (PECVD) is controlled independent of the catalyst size by controlling the residence time of reactive gases in the reactor.
According to U.S. Patent Application 20090278114, IBM’s process offers a significant advantage in terms of catalyst preparation and the growth. When used in conjunction with a catalyst system with a narrow catalyst particle size, carbon nanotubes with a very narrow diameter distribution can be obtained. Also described in the application is a method for forming a novel structure comprising an array of CNTs with well defined diameters and lithographically defined origins. This structure is suitable for forming the channel region of CNT based FETs.
The gas residence time is a measure of the average time of the gas in the reactor. Thus, if the flow is constant and the pressure increases, the residence time increases, and if the pressure is constant and the flow increases the residence time decreases. The inventors “unexpectedly discovered” that by varying the residence they can influence the diameter of SWNT. If the residence time is too high, only pyrolytic carbon is deposited and if the residence time is too low, nothing is deposited. The residence time is typically about 1 minute to about 20 minutes and more typically about 1 to about 10 minutes. The residence time is typically determined by controlling the pressure, flow or both the pressure and flow in the reactor. By varying the residence time (e.g. growth pressure and/or flow rates) of the CNT precursor gases in the CVD or PECVD reactor, SWNT diameter can be varied from about 0.2 nanometers to several nanometers.
Since accurate size control of the catalyst particles is not required, a variety of catalyst systems deposited by a variety of solution deposition, or physical vapor deposition can be utilized for CNT growth. Suitable catalysts include the group of transition metals including Fe, Mo, Co, Ni, Ti, Cr, Ru, Mn, Re, Rh, Pd, V or alloys of them. The catalyst is then ramped up to the desired growth temperature in a suitable ambient prior to initiating carbon nanotube growth using a carbon containing precursor. The growth temperature is typically about 400 to about 1200.degree. C. and more typically about 500 to about 1000.degree. C.
Suitable carbon containing precursors include aliphatic hydrocarbons, aromatic hydrocarbons, carbonyls, halogenated hydrocarbons, silyated hydrocarbons, alcohols, ethers, aldehydes, ketones, acids, phenols, esters, amines, alkylnitrile, thioethers, cyanates, nitroalkyl, alkylnitrate, and/or mixtures. Other sources include methane, ethane, propane, butane, ethylene, acetylene, carbon monoxide, benzene and methylsilane. Other reactive gases such as hydrogen and ammonia, which play an important role in CNT growth, may also be introduced. Also, carrier gases such as argon, nitrogen and helium are used.
One of the main challenges facing carbon nanotube based electronics is the low drive currents of present-day CNTFETs. The low drive current stems from the extremely small diameter of SWNTs effectively resulting in a transistor with a narrow width. Using arrays of SWNTs for the channel region will increase the drive current, making CNTFET based technologies feasible. However, at the present time no controlled ways exist forming arrays of CNTs with a well defined pitch. Thus the ability to grow arrays of SWNTs with lithographically controlled origins (limited by ebeam resolution) and small diameters (<5 nanometers) is crucial to the success of CNT electronics. IBM’s discovery is a step towards carbon nanotube based electronics.