Patent Application
20070224804
The present invention relates to carbon nanotube-based electronic devices. In particular, this invention relates to carbon nanotube-based electronic devices made by electrolytic deposition.
The semiconductor industry is facing increasingly difficult technological challenges, as it moves into the production of features at sizes below 100 nanometers. Particular challenges are to achieve affordable scaling and achieve affordable lithography with dimensions below 100 nanometers, utilize new materials and structures, and achieve gigahertz frequency operations and very high device densities on chips. There is a lack of consensus in the industry about how to solve the fabrication challenges that lie beyond the 100 nanometer barrier. The problem confronting the industry is that the dominant technology used to make chips, optical lithography, uses light to form patterns on silicon. Below 100 nanometers, the wavelength of light that is, typically, employed in chip production (193 nanometers and 157 nanometers) is too large to be useful. Several candidate technologies are currently vying for selection as successors to optical lithography. These include extreme ultraviolet lithography (EUV), an electron beam method called scalpel, and x-ray lithography. None has yet emerged as the preferred choice.
It is widely recognized that the development of molecular electronics based on carbon nanotubes would enable logic devices to be built that have billions of transistors. Such computers would be orders of magnitude more powerful than today's machines. In order for this to become a reality, a method must be found to mass produce the molecular electronic devices. Scanning probe methods have proven feasible for fabricating single devices one nanotube at a time, but no way has been found yet to speed up the process sufficiently to make billions of transistors practical. Chemical based self-assembly processes have also been suggested, but so far, only the simplest structures have been built by use of this method. The problem of combining different materials and assembling molecular electronic devices with specific features remains a significant challenge. Therefore, it would be desirable to demonstrate the feasibility of cost-effectively fabricating carbon nanotube molecular electronic devices that have a nanosize diameter (e.g., 0.7-50 nanometers), micron-to-submicron-sized length (e.g., 100-1000 nanometers), and a gate structure that is a few nanometers long (e.g., 0.1-5 nanometers).
A nanotube or nanotube bundle/rope is typically much longer that 1 nanometer. Therefore, many inputs or junctions are needed along the length of each nanotube or nanotube rope to achieve desired nanoscale density. Nanotube junctions, or active nanotube junctions, are locations or points were nanotubes are in close proximity to each other and can be modified electrically.
Theoretical work by Chico et al., “Pure carbon nanoscale devices: nanotube heterojunctions,” Physical Review Letters, 1996, has suggested that introducing pentagon-heptagon pair defects into otherwise hexagonal nanotube structure may create junctions between two topologically or electrically different nanotubes, as bases for nanoscale nanotube devices. S. Saito, “Carbon nanotubes for next generation electronic devices,” Science, 1997, describes possible theoretical designs of a carbon nanotube that may function as a molecular electronic device. Those and other similar theoretical works outline the possibility to use carbon nanotubes as molecular devices, but fail to propose a design of such device and a method of its fabrication.
Collins et al., “Nanoscale electronic devices on carbon nanotubes,” Fifth Foresight Conference on Molecular Nanotechnology, 1997, have demonstrated experimentally the rectification properties of single-wall carbon nanotubes. This work also fails to propose a design for carbon nanotube molecular electronic devices and a method of fabrication.
Therefore, in order to overcome current fabrication approaches that are expensive and impractical (e.g., placing individual nanotubes on a substrate with an atomic force microscope), a method is needed to mass produce carbon nanotube-based electronic devices in a manner that is efficient, cost-effective, and scalable.
The present invention is carbon nanotube-based electronic devices made by electrolytic deposition and applications thereof. The present invention includes nanotube-based electronic devices that are made by electrolytic deposition, such as, but not limited to, chemical or biological sensors, carbon nanotube field-effect transistors (CNFETs), tunnel junctions, Schottky junctions, and a two-dimensional array of nanotube junctions that is suitable for use as a building block in a signal processing application that requires high circuit density.
The present invention includes a novel method of fabricating single-wall nanotube devices that includes the combination of an electrolytic deposition process, followed by an operation to selectively “burn out” the percolated metallic nanotubes and, thereby, form a semiconducting nanotube-based electronic device.
Furthermore, the fabrication method of the present invention provides an efficient, cost-effective process for mass producing nanotube-based electronic devices that is scalable.
In a preferred embodiment, the present invention provides a device comprising at least one array of active carbon nanotube junctions deposited on at least one microelectronic substrate. In another preferred embodiment, the present invention provides a device comprising a substrate, at least one pair of electrodes disposed on the substrate, wherein one or more pairs of electrodes are connected to a power source, and a bundle of carbon nanotubes disposed between at least one pair of electrodes wherein said bundle consists essentially of semiconductive carbon nanotubes. In another embodiment, the bundle of carbon nanotubes consists of semiconductive carbon nanotubes and isolated metallic nanotubes. In another preferred embodiment, a semiconducting device formed by electrodeposition of carbon nanotubes between two electrodes is provided.
The invention also provides preferred methods of forming a semiconductive device by ramping a bias voltage across a single-wall carbon nanotube rope. The single-wall carbon nanotube rope preferably comprises a plurality of semiconducting single-wall carbon nanotubes and a plurality of metallic single-wall carbon nanotubes. The plurality of metallic single-wall carbon nanotubes are removed (e.g., by application of a bias voltage) in an amount sufficient to form the semiconducting device.
It is an object of the invention to provide electrical devices that are formed by carbon nanotube technology.
It is another object of this invention to provide an economic fabrication process for mass producing carbon nanotube electrical devices.
It is yet another object of this invention to provide increased circuit density, by use of carbon nanotube devices.
Substrate 110 can be formed from any electrically non-conductive material that is commonly used in semiconductor manufacturing, such as silicon nitride (SiN), silicon dioxide (SiO2), and silicon (Si). Electrode 112a and electrode 112b can be formed from an electrically conductive material, such as gold (Au) and serve as the electrical contacts for the SWNT device 100. Electrodes 112 have a thickness of, for example, between about 0.1 and 50 microns and a width of, for example, between about 0.5 and 75 microns; however, the width of electrode 112 where SWNT rope 114 is in contact therewith can be ˜0.3 microns. The spacing between electrodes 112a and 112b is, for example but not limited to, between about 0.5 and 75 microns.
SWNT rope 114 can be a bundle of individual semiconducting SWNTs arranged in a rope-like structure spanning electrode 112a and electrode 112b. The semiconducting SWNTs that form SWNT rope 114 are arranged in parallel and in contact with another. The outer diameter of the SWNTs that form SWNT rope 114 ranges from about 0.7 to 3 nanometers, and the SWNTs can be up to a few microns in length. The overall diameter of SWNT rope 114 is, for example, between about 0.7 and 500 nanometers. SWNT rope 114 can formed by an electro-deposition process, which is described in more detail in reference to
With continuing reference to
Method 200 is a preferred method of producing semiconducting SWNT devices combining an SWNT fabrication process and a selective “burn out” operation. A preferred fabrication process by P. Jaynes, T. Tiano, M. Roylance, C. Carey and K. McElrath, “Alignment and Deposition of Single Wall Carbon Nanotubes under the Influence of an Electric Field,” in “Nano- and Microelectromechanical Systems (NEMS and MEMS) and Molecular Machines,” Eds. D. A. LaVan, A. A. Ayon, T. E. Buchheit, and M. J. Madou, MRS Proceedings vol. 741, (2003) pp. J8.5.1-J8.5.6 (incorporated herein by reference in its entirety) is summarized with reference to steps 210 through 218 of method 200. The step of selectively burning out the metallic SWNTs is described with reference to step 220 of method 200 and
With reference to method 200, the variables that contribute to the ability to align SWNTs in an electric field include, for example, SWNT suspension concentration, deposition time, electric field intensity (voltage, V), electric field frequency (MHz), and electrode design (shape, line widths, and spaces).
At step 210, method 200 first includes obtaining an electrode assembly. The electrode assembly, such as a planar electrode assembly, includes a substrate (e.g., substrate 110 of
In one embodiment, the degree of dispersion of the nanotubes (e.g., the length of the nanotube rope, ranging between a single nanotube to ropes consisting of hundreds of nanotubes) can be varied by changing the time of insonification, by varying the amplitude of the acoustic excitation, or by the choice of the fluid in which the nanotubes are dispersed. Fluids which wet the nanotubes increase the degree of dispersion by reducing the tendency of the nanotubes to recombine. The degree of dispersion depends on the surface energy of the solvent and that of the nanotube as measured, for example, by inverse chromatography. Dispersion can also be controlled by functionalization of the nanotube or nanotube rope (e.g., chemically attaching molecules to the nanotube aggregates).
The concentration of nanotubes can be important in controlling the dispersion of nanotubes. The recombination of nanotubes in nanotube linear bundles into larger agglomerates is directly proportional to the concentration of linear nanotube bundles squared. In one embodiment, concentrations of nanotubes can range from 10−4 gm linear nanotube ropes per cc solvent for chemically modified nanotubes to 10−8 gm/cc to obtain single nanotubes with useful suspension times.
Dispersion can be used to control the tendency of nanotubes form three dimensional aggregates. Nanotubes have a tendency to form three dimensional aggregates because they are smooth on the nanometer scale and therefore have very large van Der Waals forces that make them stick together. Strategies to break the three-dimensional aggregates can be used to suspend linear aggregates of nanotubes in solution. In one embodiment, electrophoresis of linear aggregates starts from a solution of linear aggregates of nanotubes.
Ultrasonic dispersion is the most general method for making a solution of linear aggregates of nanotubes. Such a solution is an inherently non-equilibrium (unstable) state. The rate of break-up of the nanotube agglomerates is proportional to the acoustic energy applied to the solution. The rate of recombination is proportional to the nth power of the concentration where n=2 or somewhat greater. Therefore, in one embodiment, continuous insonification is applied. Alternatively, dispersion can be achieved without continuous insonification in a solution sufficiently dilute to avoid recombination of nanotube aggregates during the deposition time period.
At step 212, SWNTs are dispersed in a solution of organic solvent, such as ethanol, to form the electrodeposition solution. In one example, the SWNT suspension concentration is 5.059×10−7 gm/cc. Dispersion is performed by use of high amplitude ultrasonic energy, in order to achieve highly dispersed SWNT suspensions in polar organic solvents. In this example, the dispersion apparatus is a Branson titanium wedge tip ultrasonic welding horn and a Branson 940B power supply. An example of the dispersion apparatus settings is frequency=40 kHz, amplitude gain=2:1, power=700 W continuous, amplitude=45% maximum. The electrodeposition solution can be prepared by the addition of the SWNTs to the solvent and sonicating them for a period of, for example, 30 minutes. The electrodeposition solution includes a mixture of metallic, semi-metallic, and semi-conductive SWNTs.
At step 214, the electrode pairs of the electrode assembly are electrically connected to the power supply and the planar electrode assembly is submersed in the electrodeposition solution, by suspending them therein.
At step 216, the power supply that is connected to electrodes 112 is activated and, thus, an electrodeposition process is performed to form an SWNT rope therebetween, such as SWNT rope 114 of
At step 218, the power supply is deactivated, and the electrode assembly that has a plurality of deposited SWNT devices, such as SWNT device 100, formed thereon is removed from the electrodeposition solution.
At step 220, the metallic SWNTs within the SWNT rope, such as SWNT rope 114, are selectively burned out along the percolation paths therein, in order to produce a semiconducting device. The selective burn out operation is performed by ramping a bias voltage from, for example, about −1.0 to +1.0 volts across the two electrodes, such as electrodes 112a and 112b, and the resulting current flow performs a bulk burn out of the conductive metallic SWNTs, so as to produce a semiconducting device. The voltage can be ramped over a period of, for example, about 0.1 to 5 seconds. The process and result are illustrated, for example, in reference to
In this embodiment of the invention, the carbon nanotubes of the resulting semiconducting device preferably consist essentially of semiconductive carbon nanotubes. The term “consisting essentially of” includes materials that those that do not materially affect the basic and novel characteristics of the semiconducting device. For example, the presence of an electrically isolated metallic carbon nanotube would not materially affect the semiconducting properties of the device.
With reference to
Dielectric layer 410 can be formed of an insulating material, such as silicon dioxide (SiO2) or SiN, as is commonly used in semiconductor manufacture, that has a thickness up to, for example, about 150 nanometers. Gate layer 412 is formed of doped silicon and has a standard wafer. Doping in this case is the deliberate introduction of specific impurity atoms into semiconductor crystal lattice, in order to change its electrical properties. In this example, electrodes 112a and 112b form the source and drain of CNFET 400, while an electrical connection (not shown) to gate layer 412 forms the gate. In the example of
In another embodiment of the invention, multiple nansocale junctions (e.g., 0.7 to 100 nm feature size, densely spaced (10-1000 nm on center separation)) can be formed along the length of linear nanotube aggregates and embedded in a conventional micro electronic circuit or a chip using electrophoresis. For example, circuits can be made in this manner having nanometer feature size using standard lithography. In this example, conventional lithography is used to make a standard micro electronic circuits and the electrode arrays which guide the electrophoretic deposition of the nanotubes. Conventional electronics made in this example can be sealed, the chip can be placed in a electrophoresis bath, and nanotubes can be deposited in accordance with the invention to implement nanoscale circuits.
In another embodiment, nanotube devices can be formed on the deposited linear nanotube aggregates by self assembly of electrically active molecules. For example, the biological sensors described herein can be formed in this manner. Alternatively, a two dimensional or multi-dimensional array of electrodes can be formed from deposited linear nanotube aggregations which contact at each other at steep angles (e.g., crossing or contacting at right angles). As illustrated in
In another embodiment, a two-dimensional or multi-dimensional array of electrodes can be excited one pair at a time. In this manner, a linear nanotube aggregate can be deposited to connect the excited pair of electrodes. In this example, the deposited linear nanotube aggregates can be metallic (having continuous metallic paths connecting the electrodes), semiconducting (having no continuous metallic paths connecting the electrodes), or may be single metallic or semiconducting nanotubes.
In another embodiment, the selection of a pair of electrodes to bridged and the application of burnout voltages may be applied automatically (e.g., under computer control). Automating this exemplary process permits large numbers of nanotubes to be deposited. As the number of nanotubes bridging the space between the electrodes increases, the bridging of the next pair of electrodes creates a larger and larger number of junctions. The computer driven electrophoretic deposition therefore provides an efficient process for manufacturing nano-scale device circuits. Once an array of nanotube bridges are deposited, a planar film of some appropriate microelectronic material can be deposited on the nanotube and the electrodes can be used to deposit and contact the nanotubes. Another layer of nanotube can then be deposited on top of the first film. Thus, in this embodiment of the invention, three-dimensional arrays of nanotube devices can be fabricated.
NT biological sensor 500 of
The exemplary operation of a biosensor, such as NT biological sensor 500, is described as follows. A biosensor of a given functional group has a certain electronic conductivity between its electrodes. When the toxin for which the FET is functionalized binds to the SWNT rope (e.g., BoNT binding to its receptor), the electronegativity of the toxin withdraws electrons from the active region of the device and thereby changing the electrical response of the biosensor. The change in electrical response can be detected by measurement of its current or voltage, via electrodes 112. For example, if the biosensor is a depletion device, when electrons are drawn out of the toxin (e.g., BoNT), an increase in the resistance of SWNT rope 114 is detected. Example methods of attaching the BoNT sensing element to the SWNT for forming a BoNT biosensor, such as NT biological sensor 500, are provided below.
A method of attaching the BoNT sensing element, e.g., BoNT antibodies 510, to the SWNT, (e.g., SWNT rope 114), can be through non-covalent binding of pyrene-modified BoNT receptors, e.g., pyrene receptors 512, to its surface. The specific details for chemically attaching pyrene to the BoNT antibody and GT1b, the natural receptor for BoNT, are described in more detail in reference to
Binding of the modified pyrene to the carbon nanotube surface can be performed by exposing the SWNTs to a solution containing the modified pyrene for 24 hours at room temperature. Unreacted material is removed by rinsing with water. Attachment to the SWNTs can be confirmed by use of optical and microscopic techniques, such as Raman, fluorescence, near-field scanning microscopy, atomic force microscopy (AFM), and transmission electron microscopy (TEM). In addition, nuclear magnetic resonance (NMR) spectroscopy can be used to determine the conformation of the BoNT receptor attached to carbon nanotubes.
First, protection of the hydroxyl group of GT1b is accomplished by the addition of a 1:2 mixture of acetic anhydride and pyridine dried GT1b. Oxidization of the sphingosine double bond of process step 1 of
Attaching pyrene to GT1b begins with 1-pyrenebutanoic acid, succinimidyl ester (Pyr-NHS), which is converted to a terminal amine functionalized compound, by its reacting with an aqueous 5% ethylene diamine solution. Coupling of the amino functionalized pyrene, as shown in process step 3 of
Attachment process 700 of attaching GT1b to pyrene, as shown in the exemplary method depicted in
An alternative method of attaching the BoNT sensing element to the SWNT is based on a covalent tether. The method involves functionalizing the sidewalls of the SWNTs with amine linkages. The amine-functionalized SWNT is the starting material for bonding both the antibody and the GT1b receptor. For the GT1b, the receptor is reacted with the amine-functionalized SWNT, by use of the same procedure described for reacting it with pyreneamine, as described in reference to
In yet another example of a carbon nanotube electrical device,
Electrodes 812, electrodes 814, electrode 816, and electrode 818 are, for example, 0.5 μm metal lines that are formed of an electrically conductive material, such as gold, and that are deposited atop substrate 810 by any conventional process. The spacing between opposite electrodes, such as electrode 812a and 814a or electrode 816 and 818, can be, for example, between about 0.5 and 75 microns. Substrate 810 is formed of any electrically non-conductive material that is commonly used in semiconductor manufacturing, such as SiN, SiO, and Si. Alternatively, substrate 810 is a flexible substrate that is formed of, for example, plastic, liquid crystal polymer films, or polyimide.
Single nanotubes or SWNT ropes 820 are as described in reference to SWNT rope 114 of
2D nanotube array 800, which can be formed by use of method 200, is representative of a method of creating nano-scale arrays of electronic devices, such as memory arrays, from nanotubes, in a manner that is much more scalable, practical, and efficient than current approaches that involve placing individual nanotubes on a substrate with an atomic force microscope. 2D nanotube array 800, formed by use of method 200, takes advantage of the high aspect ratio of nanotubes, as compared to other nano-particles. The average length of SWNTs is 1 μm, and they form single nanotubes or SWNT ropes that can be, for example, several nanometers wide and 5 micrometers long. In another embodiment, multidimensional arrays can be formed having multiple layers of nanotubes arranged, for example, in nanotube bundle or rope structures.
When burnt-out ropes that contain both metallic and semiconducting nanotubes are used, some crossover nanotube junctions 822 will form Schottky junctions (metal-to-semiconductor contacts), and some will form resistive junctions (metal-to-metal contacts). The semiconducting nanotubes in a rope that has percolating metallic nanotube paths will look like metallic media, because the metal nanotube are in tangential contact with semiconducting nanotubes, as shown in
In yet another exemplary carbon nanotube electrical device,
2D nanotube array 900 can be formed by (1) depositing a set of single nanotubes or SWNT ropes that form first nanotube layer 910, via method 200 of
With continuing reference to
As described with reference to
Very dense arrays of junctions may be made by use of crossover nanotube junctions, such as crossover nanotube junctions 822 of 2D nanotube array 800 or crossover nanotube junctions 916 of 2D nanotube array 900. These arrays use the ˜1 micron length nanotubes, in order to interface with control, i.e., input and output lines that are formed by conventional lithography. A crossbar geometry is used to define junctions on a submicron (50 nm) scale.
This application is a divisional of application Ser. No. 11/090,193, filed Mar. 28, 2005, which claims priority to U.S. Provisional Patent Application No. 60/557,118 filed on Mar. 26, 2004 which is hereby incorporated by reference in its entirety.
The United States Government may have certain rights in this invention pursuant to Grant Numbers F19628-03-C-0075 and N41756-02-M-1043.
| Number | Date | Country | |
|---|---|---|---|
| 60557118 | Mar 2004 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 11090193 | Mar 2005 | US |
| Child | 11638571 | Dec 2006 | US |
