Patent Application
20090045061
1. Technical Field
The present invention generally relates to a nanotube device and method of forming the device, and more particularly, to a method for controllably depositing one or more nanotubes in a defined region. The present invention further generally relates to a method of forming a nanotube vertical field effect transistor.
2. Background
There are many applications where a nanotube, e.g., a carbon nanotube (CNT), or an array of nanotubes, can be employed as a sensing or active device element in an electrical probe or electronic device. In these applications, electrical contact must be made with the nanotube, which requires accurate positioning of the nanotube with respect to various conductive links (i.e. interconnects) and other circuitry.
Aside from the need for precise alignment, properties of the nanotube also need to be controlled in order to provide device performance according to desired specifications. For example, many transistor applications for CNTs are best achieved with single wall carbon nanotubes (SWNT) rather than multi-wall carbon nanotubes (MWNT). Furthermore, as an active element of a transistor, a semiconducting SWNT, rather than a metallic SWNT, is required. For other applications such as interconnects and nanoprobes, however, a metallic CNT is preferred.
Existing fabrication methods for CNT devices do not fully address both needs for alignment and property control. In addition, in CNT electrical device fabrication, at least one interconnect level may be processed before CNT deposition. The most common metallization schemes, e.g., with aluminum and copper interconnects, often impose thermal budget constraints for subsequent processing steps. Chemical vapor deposition (CVD) methods, which are typically used for depositing CNTs, are not compatible with aluminum or copper interconnects because of the relatively high temperatures involved.
In the fabrication of CNT devices, there is often a need to provide a vertically oriented CNT inside an aperture. In transistor fabrication processes, depending on the specific stage or levels, the aperture is also referred to as a via.
Embodiments of the present invention provide a method of depositing nanotubes in a region defined by an aperture, with control over the number of nanotubes to be deposited, as well as the pattern and spacing of nanotubes. Specifically, electrophoretic deposition, along with proper configuration of the aperture, allows at least one nanotube to be deposited in a target region with nanometer scale precision. Pre-sorting of nanotubes, e.g., according to their geometries or other properties, may be used in conjunction with embodiments of the invention to facilitate fabrication of devices with specific performance requirements.
Embodiments of the present invention also provide a method of controlling the number of nanotubes to be deposited and their spacings in a given region. Such a method is useful for many applications where it is desirable to deposit more than one nanotube in a defined region. For example, certain vertical field effect transistor (VFET) designs may benefit from having more than one nanotube forming a channel to allow more current to flow through the device. Thus, by controlling the number of nanotubes to be deposited, one can ensure that the VFET output can be designed with sufficient current to meet the parameters of a logic circuit input.
Additional advantageous features, functions and implementations of the present disclosure will be apparent from the description which follows, particularly when read in conjunction with the accompanying figures.
The teachings of the present invention can be readily understood by considering the following detailed description in conjunction with the accompanying drawings, in which:
a is an SEM image of oriented single wall carbon nanotube “forests” electrochemically functionalized by enzyme as discussed in text above (alignment is perpendicular to the plane of the image displayed) and
To facilitate understanding, identical reference numerals have been used, where possible, to designate identical elements that are common to the figures.
The present disclosure is described herein with reference to the accompanying figures, wherein exemplary embodiments are schematically depicted or otherwise described.
Embodiments of the present invention allow the CNT 110 to be deposited inside the aperture 106, to the exclusion of other CNTs. The aperture 106, which has to be sufficiently large to accommodate the CNT 110, may be patterned using different lithographic processes. Thus, in one embodiment, the aperture 106 may have a diameter (D) ranging from about the lower limit (e.g., resolution) of the lithography process to about 100 nm. For example, existing lithography at 193 nm readily provides a resolution limit of about 90 nm. In one embodiment, the substrate 102 has a lateral dimension (e.g., extending across the aperture) sufficiently large to meet level-to-level overlay constraints with respect to the aperture 106. As will be shown below, the CNT 110 can be deposited proximate the center of the aperture 106, e.g., with a lateral alignment precision of a few nanometers. Furthermore, the CNT 110 may be pre-selected to have a preferred physical property including multiwall CNT versus single wall CNT and or conducting CNT versus semiconducting CNT.
Successful EPD requires preparation of a stable dispersion. In general, an electrostatically stabilized dispersion can be obtained with particles of high ζ-potential, while keeping the ionic conductivity of the suspensions low. SWNTs have shown high ζ-potential values at low pH values. It is also known that the presence of charging salts can play an important role in improving adhesion of the nanotubes to substrates and increasing the deposition rates.
In one embodiment, 10 mg of purified SWNTs are suspended in 30 ml of distilled water, and 10−4 moles of magnesium nitrate hexahydrate [Mg(NO3)26H20] is added to the suspension and sonicated for about 2-3 hrs. In general, it is preferable that the nanotubes in the liquid bath 220 be pre-sorted for the type of nanotubes according to application needs. For example, while semiconducting SWNTs are used as active elements in transistors, either semiconducting or metallic nanotubes may be used for probes or other devices. A few drops of non-ionic Triton-X surfactant are added to improve the suspension with a final pH of solution at about 4.
Aside from hydrogen ions (H+), shown as circles in
In
In
Since H+ ions have higher mobility than other positively charged species, including the CNTs, H+ ions will arrive at the substrate structure 200 faster than other charged species, and thus, preferentially accumulate on the surface of the insulating layer 204, as shown in
Positively-charged CNTs arriving near the substrate structure 200 are directed by the electric field towards the center of each aperture 206, as shown in
Since different devices often require different properties of the nanotubes for proper operation and/or optimum performance, it may be advantageous to provide a pre-sorting of the nanotubes prior to electrophoretic deposition. For example, nanotubes may be sorted according to their properties such as semiconducting versus metallic, single-walled versus multi-walled, or they may be sorted according to geometries or dimensions such as lengths, diameters, and so on.
Since different types of nanotubes have different mobilities, e.g., longer or multiwalled nanotubes will generally have lower mobility compared to shorter or single-walled nanotubes, electrophoresis may also be used for sorting purposes. Such sorting can be done prior to the electrophoretic deposition so that the nanotubes in the bath have a relatively uniform distribution in terms of properties and/or geometries. Alternatively, if the nanotubes in the electrophoresis bath have a relatively wide distribution in terms of geometries or other properties, a certain degree of sorting may also be achieved “in situ” during deposition by virtue of the different mobilities of the nanotubes.
The degree of focusing that directs the nanotubes towards the aperture is affected by the magnitude and shape of the electric field distribution, along with the configuration of the aperture. To provide control over the number of deposited nanotubes as well as their positioning, a finite element model is used to investigate the electric field distribution as a function of various input parameters. Parameters or factors that are relevant for controlling nanotube deposition include the aperture configuration, nanotube properties, characteristics of the insulating layer and substrate, bias potential, dielectric properties of the solution, among others. The aperture configuration may generally include the shape, dimensions (e.g., width, length, depth, ratios of dimensions), sidewall profile, and so on. The nanotube properties may generally include the dimensions (e.g., length, diameter), single-walled or multi-walled, semiconducting or metallic.
The electric field around the aperture results from a combination of the potential applied to the metal layer on the substrate structure and charges that accumulate on the surface of the insulating layer. The positive charge accumulation on the dielectric layer covering the cathode creates an electric field that opposes the field arising from the bias applied between the anode and cathode. Once the two electric fields become equal and opposite, positive charges will no longer be attracted to the surface of the insulating layer. This “saturation charge density”, σ, which determines the strength of the nanoscopic lens from the resulting electric field distribution, can be calculated from:
σ=∈0∈rE Eq. (1)
where E is the magnitude of the electric field between the anode and cathode, ∈0 is the permittivity of free space, and ∈r is the relative permittivity of the liquid.
As an example, for E=103 V/m, ∈0=8.85×10−12 Farad/meter and the liquid is water ∈r=80, the surface charge density σ is equal to 7.1×10−7 Coulomb/meter2.
Once the specific aperture geometry is selected and the surface charge density is calculated, the electric field in the region near the apertures and the motion of positively charged particles can be calculated using finite element analysis techniques that are well known. Thus, with proper configuration and design, one can obtain an electric field distribution to produce a desired focusing or lens effect to direct the nanotube deposition.
After sufficient charges have accumulated to reach the charge saturation point, the electrostatic lens effect will direct all charged particles towards the center of the aperture 306. The equipotential lines for this geometry favor the focusing of mobile charged nanotubes towards the center of the aperture 306. In this case, the diameter of the aperture 306 is 100 nm and the depth is 50 nm. In this example, since the electric field distribution around the aperture 306 is substantially symmetric with respect to a central longitudinal axis of the aperture, the CNT 310 is also substantially centered inside the aperture 306. Thus, one end of the CNT 310 is attached to a region of the conductive layer 302 defined by the aperture 306 (i.e., the exposed region at the bottom of the aperture), e.g., within a few nanometers of the center of the defined region.
In general, for a fixed potential difference between the reference electrode and the metal contact at the bottom of the aperture, the strength of the focusing effect is inversely proportional to the diameter of the aperture for a fixed aperture depth.
In this case, the electric field distribution will not provide a preferential direction to guide the nanotubes towards the center region of the aperture 406. The final location of the nanotube will depend on the initial position of the nanotube before the bias is applied. For a large aperture, e.g., diameter or lateral dimension of greater than about 100 nm, the unattached end of the first deposited nanotube may still be the focal point for further nanotube deposition. However, when the lateral dimension of the aperture is sufficiently large, the electric field will also direct other nanotubes to other locations on the exposed surface of the conductive layer 402.
Although results suggest that an aperture diameter of about 100 nm provide a transition or reference point below which deposition is restricted to a single nanotube, while apertures larger than about 100 nm tend to favor deposition of more than one nanotubes, it is understood that this reference point may vary with specific combinations of nanotubes and/or structural configurations.
Aside from the aperture diameter (or lateral dimension), other parameters, e.g., shape, aspect ratio (defined as depth or height of aperture divided by lateral dimension), among others, may also be used for the purpose of controlling deposition of nanotubes, for example, by providing different configurations according to the nanotube properties and/or geometries.
Results of another finite element analysis also show that, for nanotubes with a 10 nm diameter and a length of 100 nm, and an aperture formed in silicon nitride with a diameter of 100 nm and a depth (or height) of larger than 18 nm, only one nanotube will be deposited inside the aperture. This suggests that an aperture with an aspect ratio of at least 0.18 or greater may be used to restrict the number of deposited nanotubes to only one. For a nanotube with a smaller diameter, a larger aspect ratio may be required in order to restrict the deposition to only one nanotube. Similar analysis can be used to simulate probable locations of deposited nanotubes for other aperture configurations and nanotube properties. While a two dimensional analysis is suitable for situations in which a plane of symmetry is available, a three dimensional analysis can generally be used for other situations. Thus, finite element analysis can be used for nanoscopic lens design as a guide to providing nanotube deposition with additional levels of control.
Many different nanotube-based devices may be fabricated using the method of the present invention. While the method can generally be applied to the deposition of nanotubes within apertures of different dimensions, it is particularly well-suited for situations in which it is desirable to control the number of nanotubes to be deposited or the lateral positioning or alignment of the nanotube. Examples of nanotube-based devices that can benefit from this method include vertical CNT transistors, chemical sensors or biosensors, among others.
A device similar to that shown in
Depending on the specific sensor applications, different functional molecules 514 are provided to the other end of the nanotube 510. In general, SWNTs are preferred for sensor applications, although there may be situations in which MWNTs may also be used.
The conductive portion 602, e.g., an interconnect metal, can be deposited and patterned using photolithography and resist liftoff techniques known to one skilled in the art. The interconnect metal needs to be suitable for maintaining electrical contact and adhesion with CNTs, and may include, but not limited to, cobalt (Co), nickel (Ni), or iron (Fe).
To facilitate adhesion between the metal interconnect to the quartz substrate, an adhesion layer (not shown) may also be formed prior to the formation of the conductive portion 602. In one embodiment, cobalt (Co) is used as a metal interconnect, and a 20 nm thick chromium (Cr) adhesion layer is used to promote adhesion of cobalt to the quartz surface. The Cr layer may be evaporated at a rate of about 2 Ångstroms per second (Å/s), and a 120 nm thick Co layer can then be evaporated at a rate of about 1 Å/s. A thickness of about 20 nm and 120 nm may be used for the Cr and Co layers, respectively. The cobalt metal interconnect can also serve as the cathode during electrophoretic deposition of the nanotube
In the embodiment where CNTs are deposited using electrophoresis, the conductive portion 602 is configured to be electrically connected to contact pads (not shown) provided at the edge or periphery of the substrate 602. In one embodiment, each conductive portion 602 upon which a CNT will be deposited is provided as part of a continuous conductive layer formed over the substrate 600, in order to simplify the electrical connection paths. This facilitates electrical grounding of the metal during electron-beam lithography (if e-beam is used during fabrication) and provides a single connection point for electrophoretic deposition of CNTs. In one example, electrical connections between different devices are made in the kerf (area between each device), which allows the connections to be broken by a dicing saw when the substrate is cut to facilitate assembling the devices. Alternative configurations may also be used for providing the electrical connections needed during fabrication, and a variety of conventional lithographic and etching processes may be adapted for this purpose.
For each individual sensor device, a metal contact is needed to provide electrical connections to external circuits, for example, by soldering or wire bonding. In one embodiment, gold (Au) is used as the metal contact material. The metal contact can be formed using photolithography and resist liftoff techniques.
Using a suitable lithography process, aperture 610 having a lateral dimension of about 100 nm or less can be formed in the insulating layer 608. The aperture 610 is sufficiently large to accommodate a nanotube to be deposited onto substrate 602. In one embodiment, the aperture 610 has a diameter ranging from about the lower limit (e.g., resolution) of the lithography process to about 100 nm. In one embodiment, optical lithography using 193 nm source illumination may be used to pattern the aperture in photoresist, providing a resolution of about 90 nm. Alternatively, these apertures may also be fabricated using electron-beam lithography or a focused ion beam milling technique. Apertures with dimensions of less than about 100 nm are suitable for providing the electrostatic lens effect during electrophoretic deposition of nanotubes. The lithography technique of the interconnect metal and vias will limit the separation between the nanotube devices.
The structure of
After the nanotube 650 is deposited in the aperture 610, its vertical orientation may be affected by the rinse process or by charging effects. The nanotubes can be re-aligned to the vertical direction by applying an electrical potential between the metal level on which it was deposited and a metal plate above the wafer substrate. This may be done, for example, in a reactor prior to the subsequent deposition process in the fabrication sequence. Plasma processing systems typically have a metal plate above the wafer that is part of the electrical circuit for generating the plasma. By establishing a DC or AC electric field between this metal plate (or another electrode) and the metal level that the nanotube is deposited on, the nanotube can be re-aligned to a desired orientation prior to subsequent processing.
After nanotube deposition, a conformal film 612 of an insulating material having a thickness range of about 2-5 nm may be formed to encapsulate and passivate (or insulate) the nanotube 650. Suitable materials for this encapsulation film include SiNx or suitable polymers, e.g., polytetrafluoroethylene.
Since the sorting of nanotubes before or during electrophoretic deposition usually does not provide sufficiently precise control of the lengths of the nanotubes, additional trimming may be needed in order to provide a nanotube with a certain length specification. This can be achieved by the process steps illustrated in
A portion of the encapsulation film 612 around the tip of the deposited nanotube 650 can then be removed by a brief reactive ion etching (RIE) or chemical etch to uncover the tip of the nanotube 650. The length of nanotube 650 that is uncovered will depend on the etch rate and the time duration of the etch. RIE etches for SiNx are standard processes for fabrication of complementary metal oxide semiconductor (CMOS) integrated circuits, and etch rates are well known and incorporated into commercial SiNx etching apparatus.
The poly-Si layer 613 will then be removed to leave free standing encapsulated nanotube 650, as shown in
The above embodiments and discussions illustrate the capability to controllably deposit a single nanotube with nanoscale lateral precision near a center of a region defined by an aperture. The method is particularly attractive from an implementation or processing viewpoint, because the ability to achieve such controlled deposition within a relatively large region significantly relaxes the requirement for lithographic techniques. As such, the fabrication can readily be performed using optical lithography, without resorting to more complicated lithographic tools (such as e-beam or focused ion beam) to form sufficiently small apertures to define the target deposition region.
Embodiments of the present invention also provide a method of controlling the number of nanotubes to be deposited and their spacings in a given region. Such a method is useful for many applications where it is desirable to deposit more than one nanotube in a defined region. For example, certain vertical field effect transistor (VFET) designs may benefit from having more than one nanotube forming a channel to allow more current to flow through the device. Thus, by controlling the number of nanotubes to be deposited, one can ensure that the VFET output can be designed with sufficient current to meet the parameters of a logic circuit input.
One constraint in the design of the VFET is that the lateral size of the device should be as small as possible to maximize the number of VFETs per unit area. One possibility is to fabricate closely spaced vias and connect each source 802, drain 804, and gate 806 in parallel, as shown in
Embodiments of the present invention will allow a device concept such as that shown in
Furthermore, the number of nanotubes deposited within the slot can be controlled by the length of the slot. Once a first nanotube is deposited in the slot the electric field distribution around the slot will be modified. The new field distribution can be calculated using finite element analysis. The closest separation between adjacent nanotubes can also be calculated by using finite element analysis to predict the trajectory of randomly approaching charged particles that are successively deposited in the slot.
Using this analysis for nanotubes having a length of 100 nm, it has been estimated that the closest separation between nanotubes with 1 nm diameter is about 15 nm. For nanotubes with a diameter of 10 nm and a length of 100 nm, the closest separation for adjacent nanotubes is about 20 nm. The same method can be used to calculate the closest separation of nanotubes with any geometry. An alternative method can be used to calculate the electric field in the vicinity of two closely spaced nanotubes and reduce the spacing until the calculated electric field has a distribution that would exclude deposition of a third nanotube in between the two that are already deposited.
Once the closest separation (s) between nanotubes is known, the number of nanotubes, N, deposited in the slot is given by: N=MOD(L/s). The function MOD( ) truncates the resulting number L/s to an integer. The shape at the ends of the slot may also modify this result, depending on the degree of rounding. The calculation is most accurate if there is no rounding. With the presence of rounding, an additional degree of focusing may reduce the number of deposited nanotubes, and this can be determined using three dimensional finite element analysis for the exact geometry.
As shown above, embodiments of the present invention provide a method for controllable depositing nanotubes using electrophoresis in a defined region. The deposition region may be defined by an aperture, which can be configured to control the number of nanotubes that can be deposited in the region, as well as the spacings of deposited nanotubes. By properly configuring the aperture, e.g., providing a sufficiently small aperture size such as less than about 100 nm, one can also control the deposition such that only a single nanotube is deposited in the region, with lateral alignment precision of a few nanometers.
Embodiments of the invention provide a room temperature process that is readily scalable and compatible with conventional fabrication processes and materials, and allow improved control over the properties of nanotubes being used in device fabrication.
Although some examples have been discussed in the context of the deposition of carbon nanotubes, it is understood that the method can generally be adapted for deposition of other nanotubes. Furthermore, embodiments of the invention can generally be applied to depositing single-walled, multi-walled, semiconducting or metallic nanotubes for fabrication of different devices.
The above embodiments and discussions illustrate the capability to controllably deposit a single nanotube with nanoscale lateral precision near a center of a region defined by an aperture. The method is particularly attractive from an implementation or processing viewpoint, because the ability to achieve such controlled deposition within a relatively large region significantly relaxes the requirement for lithographic techniques. As such, the fabrication can readily be performed using optical lithography, without resorting to more complicated lithographic tools (such as e-beam or focused ion beam) to form sufficiently small apertures to define the target deposition region.
Embodiments of the present invention also provide a method of controlling the number of nanotubes to be deposited and their spacings in a given region. Such a method is useful for many applications where it is desirable to deposit more than one nanotube in a defined region. For example, certain vertical field effect transistor (VFET) designs may benefit from having more than one nanotube forming a channel to allow more current to flow through the device. Thus, by controlling the number of nanotubes to be deposited, one can ensure that the VFET output can be designed with sufficient current to meet the parameters of a logic circuit input.
One constraint in the design of the VFET is that the lateral size of the device should be as small as possible to maximize the number of VFETs per unit area. One possibility is to fabricate closely spaced vias and connect each source 502′, drain 504′, and gate 506′ in parallel, as shown in
Embodiments of the present invention will allow a device concept such as that shown in
Furthermore, the number of nanotubes deposited within the slot can be controlled by the length of the slot. Once a first nanotube is deposited in the slot, the electric field distribution around the slot will be modified. The new field distribution can be calculated using finite element analysis. The closest separation between adjacent nanotubes can also be calculated by using finite element analysis to predict the trajectory of randomly approaching charged particles that are successively deposited in the slot.
Using this analysis for nanotubes having a length of 100 nm, it has been estimated that the closest separation between nanotubes with 1 nm diameter is about 15 nm. For nanotubes with a diameter of 10 nm and a length of 100 nm, the closest separation for adjacent nanotubes is about 20 nm. The same method can be used to calculate the closest separation of nanotubes with any geometry. An alternative method can be used to calculate the electric field in the vicinity of two closely spaced nanotubes and reduce the spacing until the calculated electric field has a distribution that would exclude deposition of a third nanotube in between the two that are already deposited.
Once the closest separation (s) between nanotubes is known, the number of nanotubes, N, deposited in the slot is given by: N=MOD(L/s). The function MOD( ) truncates the resulting number L/s to an integer. The shape at the ends of the slot may also modify this result, depending on the degree of rounding. The calculation is most accurate if there is no rounding. With the presence of rounding, an additional degree of focusing may reduce the number of deposited nanotubes, and this can be determined using three dimensional finite element analysis for the exact geometry.
Embodiments described above can be used for fabricating different carbon nanotube (CNT) devices, e.g., a CNT field effect transistor (CNT-FET). Since the materials and processes for forming a CNT-FET are compatible with those typically used in complementary metal oxide semiconductors (CMOS), such a transistor can readily be integrated with CMOS processing to provide a three dimensional semiconductor structure.
For example, in hybrid structures, the vertical CNT-FET (VFET) process steps are inserted into the process flow of the metal levels (i.e. interconnect levels) of the CMOS device. That is, the patterning for the VFET is done at the same time as one or more of the metal levels of the CMOS (at least three metal levels are needed for a VFET) and the VFET level device logic is incorporated into the interconnect levels along with the “horizontal transistor” interconnects. The VFETs becomes a part of the entire logic diagram along with the “horizontal transistor” logic.
A conductive layer 704′ is formed by depositing a suitable material, e.g., Al, Cu, TiN, or Co, over the insulating layer 702′ and patterned to form a source (or drain) of the vertical CNT-FET. The material for conductive layer 704′ should have sufficient adhesion to the CNT (to be formed as the channel and discussed in a later deposition step, e.g.,
Furthermore, if the CNT-FET is to be integrated with the first metal level of a CMOS device, then the substrate 700′ may correspond to the wafer and all of the process levels preceding the dielectric 1 level, while insulating layer 702′ and conductive layer 704′ may correspond to the dielectric 1 (D1) and metal 1 (M1) levels of the CMOS device, respectively. For example, the patterning of the conductive layer 704′ to form the source/drain of the CNT-FET is performed at the same mask level as M1 of the CMOS device.
After the conductive layer 704′ is patterned, an insulating material is deposited and the resulting structure is polished to form a planarized dielectric layer 706′. The thickness of the dielectric layer 706′, which corresponds to a separation between the source 740′ and the gate (to be formed) of the CNT-FET, is determined by the device specification. Dielectric layer 706′ may correspond to dielectric 2 (D2) level of the CMOS device in an integrated structure.
A conductive material is provided over the dielectric 706′ and patterned to form a conductive layer 708′, which, after further processing to be described below, will form a gate of the CNT-FET. In one embodiment, the material is aluminum (Al) having a thickness in a range from about 10 nm to about 100 nm depending on the requirements of the circuit design. Another suitable material includes copper, Cu. In the case of an integrated CNT-FET and CMOS process, in which the gate metal of the CNT-FET also acts as M2 in the CMOS metal stack, the material for the gate will also have to satisfy all of the current carrying specifications required by M2. Alternatively, the gate metal for the vertical CNT-FET may be provided as a separate metal level between two of the normal metal levels in the CMOS metal stack (i.e., requiring an additional metal level for each level of vertical CNT-FET).
A dielectric layer 710′ is formed by depositing an insulating material and patterning, e.g., with optical lithography, to define an aperture 720′, as shown in the structure of
According to embodiments of the present invention, the MWNT 725′ is deposited by electrophoresis, as previously discussed. For example, a suspension of MWNT nanotubes, e.g., pre-sorted, may be used. The parameters for electrophoresis and the configuration of aperture 720′ are selected such that only one MWNT 725′ is deposited inside aperture 720′. As previously mentioned, aperture 720 is sufficiently large to accommodate the diameter of the MWNT to be deposited within the aperture. Furthermore, the aperture 720′ is configured so that it will allow only a single MWNT to be deposited in the aperture. In general, the maximum diameter of 720′ depends on the diameter of the nanotube, the length of the nanotube, and the depth of the aperture. It can be estimated using finite element analysis for the specific geometry that is required by the circuit design. Furthermore, MWNT 725′ is substantially centered within aperture 720′ such that it is substantially vertically oriented, and its end is proximate a center 708C′ of the defined region on conductive layer 708′, e.g., within a few nanometers from the center 708C′. After deposition of the MWNT 725′, the wafer is rinsed in distilled and deionized water.
In a subsequent step, a dielectric material 712′ is deposited over the structure of
Aperture 730′ is extended through the insulating layer 706′ to stop at the conductive layer 704′. Preferably, the conductive layer 708′ acts as an etch mask for the etch of insulating layer 706′ using a reactive ion etch process. It is also preferable that the etches for 712′ and 706′ do not significantly etch insulating layer 710′. A region of the conductive layer 704′ is thus defined by the aperture 730′, with a lateral dimension d corresponding to the diameter of the MWNT 725′, e.g., in the range of approximately 10 nm to 40 nm. Aperture 720′, on the other hand, has a diameter D larger than aperture 730′, with aperture 730′ being substantially centered with respect to aperture 720′.
A suitable dielectric material (e.g., silicon nitride of ˜2 nm to 30 nm) is then deposited to form a conformal dielectric layer 714′ over the conductive layer 708′ covering the sidewall 708W′ and the SWNT 735′, as shown in
Referring back to
After the deposition of SWNT 735′ and gate dielectric formation, the next step involves positioning or orienting SET 735′ so that it can contact the dielectric layer 714′ at the sidewall 708W′ of the conductive layer 708′. This can be done by applying a voltage across conductive layers 704′ and 708′, as shown in
Referring back to
In the next step (i.e., after SWNT deposition and gate dielectric formation), a sufficiently thick layer of dielectric material 716′ is deposited inside apertures 730′ and 720′ to surround the SWNT 735′ (see
A conductive material (e.g., Al, Cu, TiN, or Co) is deposited over the planarized structure and patterned to form a drain (or source) 718′ of the CNT-FET. The material for conductive layer 718′ should have minimal contact resistance with the CNT 735′ after the processing is completed. The sheet resistance of the material should also be sufficiently low to be compatible with low current operation, as are Al and Cu in standard CMOS integrated circuits. Alternatively, the contact between the CNT 735′ and the VFET drain may be provided as a composite material, which may include a metal known to have good adhesion and low contact resistance combined with another metal with a low sheet resistance. Combinations such as Co/Al, Co/Cu, Fe/Al, Fe/Cu may be suitable with Co or Fe in direct contact with the CNT 735′. Most metals that are used as seed metals in CVD of CNTs may also be suitable to be in direct contact with EPD deposited CNTs.
After completion of the vertical CNT-FET, electrical connections to conductive layers 704′, 708′ that are used for fabrication purpose, e.g., for electrophoresis, are severed using techniques known to one skilled in the art. Another level of vertical FET may be fabricated above the conductive layer 718′.
As shown above, embodiments of the present invention provide a method for controllable depositing nanotubes using electrophoresis in a defined region. The deposition region may be defined by an aperture, which can be configured to control the number of nanotubes that can be deposited in the region, as well as the spacings of deposited nanotubes. By properly configuring the aperture, e.g., providing a sufficiently small aperture size such as less than about 100 nm, one can also control the deposition such that only a single nanotube is deposited in the region, with lateral alignment precision of a few nanometers.
Embodiments of the invention also provide a room temperature process that is readily scalable and compatible with conventional fabrication processes and materials, and allow improved control over the properties of nanotubes being used in device fabrication. Furthermore, the process allows integration of nanotube field effect transistors with CMOS devices.
Although some examples have been discussed in the context of the deposition of carbon nanotubes, it is understood that the method can generally be adapted for deposition of other nanotubes. Furthermore, embodiments of the invention can generally be applied to depositing single-walled, multi-walled, semiconducting or metallic nanotubes for fabrication of different devices.
Turning to the field of nanotube sensors, the patch-clamp technique is a widely used method for investigating electrical activity in cells. In the patch clamp method, a micropipette is fabricated from a glass material and has an incorporated electrical contact. The micropipette is filled with a fluid that whose electrochemistry can be adjusted to mimic different extracellular conditions. The micropipette is attached to the cell membrane such that the local electrical activity of a small portion of membrane can be studied. The circuit is completed with an electrode that is in the liquid bath that surrounds the cell. The patch-clamp technique is used to measure ionic current through intramembrane proteins and to monitor the membrane potential as a function of external and internal stimuli.
It would be desirable to have an electrical probe that can measure electrical activity from within a live cell. Such a probe could be combined with nanomanipulation techniques to spatially resolve variations in cell electrochemistry that are involved in key cell processes. Since, in the majority of cases, there is a mechanical component in the chain of signals that carry out cell processes, it would also be desirable if the electrical probe can be made small enough to not distort the cell during insertion and measurement. This may also allow normal cell motility processes to proceed unencumbered by the probe. A carbon nanotube is a good candidate for such an intracellular probe because the diameter can be made as smaller than the width of the cell membrane (i.e. less than 3 nanometers). A difficulty of such a probe is that the entire shaft of the CNT is conducting, such that the portion of the CNT that is external to the cell will leave a conducting path to the bath electrode (if the patch-clamp circuit arrangement is used with the CNT probe). Embodiments of the present work relate to a method for fabricating a CNT probe that can be electrically connected to an external circuit and whose shaft in insulated up to the tip which is inserted into the cell.
Thus,
In some cases, it will be advantageous to integrate the nanoprobe array into a circuit on the wafer (e.g. silicon) on which the array is deposited. In particular for the purpose of minimizing resistive and capacitive losses in the signal and to reduce noise, it would be advantageous to fabricate amplifier and multiplexing circuits on the same wafer as the nanoprobe array and to position each nanoprobe on an interconnect level that provides direct access to the input of the multiplexing circuit or amplifier input.
An insulating layer such as silicon nitride, SiNx, will then be deposited (
An electrophoresis technique will be used to deposit vertically aligned SWNTs selectively in the small windows (
The tips of the CNT probes can then be uncovered by applying a brief RIE or chemical etch. This etch is to uncover the tips of the SWNTs. The length of SWNT that is left uncovered will depend on the etch rate and the time duration of the etch. RIE etches for SiNx are standard rates are well known and incorporated into commercial SiNx etching apparatus. The poly-Si layer will then be removed to leave free standing encapsulated SWTs (
The SWNT probe array may be used to monitor the near-membrane intracellular signaling events (voltage, [Ca2+], enzymatic activity). It is a very promising device with a multitude of applications in physiology/biophysics, cell signaling, pharmacology, neuroscience and other fields that study various forms of the interface between extracellular and intracellular signaling.
Many cellular processes that are important for normal physiology, as well as therapeutically, occur in the vicinity of the plasma membrane. A large body of research is aimed at the coupling of electrical trans-membrane signals to intracellular events. These include such high-impact research areas as: excitation-contraction coupling, excitation-secretion coupling, and neuronal (post-synaptic) integration. The near-membrane domain is also critical in electrically quiescent tissues that are active communicators: lymphocytes, polar epithelia, endothelia, etc. Measurement of voltage between the tip of a single SWNT electrode and the extracellular space could be very useful, especially if it could be done with the nanometer spatial resolution. Previously, distribution of trans-membrane potential was simplistically inferred in theoretical applications assuming continuum electrostatics (constant field assumptioni,ii and mean-field approximation of Poisson-Nernst-Planckiii) or calculated from molecular dynamics of membranes and channelsiv,v,vi,vii. The trans-membrane distribution of the potential has been widely viewed to be that indicated in FIG. 4.viii The difference in potential between the aqueous solutions on both sides of the cell membrane give rise to ΔΨ(˜100 mV) which is a function of the difference in ionic charge concentrations inside and outside of the cell. The dipole potential, ΨD, is thought to arise from dipoles just below the water/membrane interface and may be on the order of several hundred mV. ΨD affects the permeability of the membrane to hydrophobic ions and the binding of such ions to the membrane. The surface potential ΨS, arises from negatively charged lipids and may be several tens of mV. The distribution of potential (both trans-membrane and tangential) affects movements of voltage dependent and electrogenic transporters (channels, pumps, receptors, etc.). It also depends on their function. Simple recordings of voltage near functional channels have not been done before with sub-optical wavelength resolution and potentially could be of great interest to many. Transporters aside, the nanoprobe will permit direct measurements of charges on membrane surfaces. These have been shown to lower trans-membrane potential from the extracellular side by about 40 mV (averaged to all membrane surface). Extracellular Ca2+ and Mg2+ ions screen them at physiologically relevant concentration.ix However, very little is known about surface charges on the intracellular side and about their tangential distribution. Our preliminary data on how intracellular Ca2+ affects movements of the voltage sensor in Ca2+ channels strongly indicate that negative charges on the intracellular side of plasma membrane exist and contribute to the interaction between Ca2+ ions and Ca2+ channels.x A SWNT nano-electrode can be used to investigate how trans-membrane voltage changes at a spatial scale comparable to sizes of transporters and distances between them.
Functionalized SWNT probes specifically responsive to changes in chemical environment of the tip will be even more useful. Although the near-membrane intracellular space is a hot-spot where many intracellular signaling events originate and/or end, direct measurements there are limited to the micrometer scale of optical techniques. Even the most elegant single-molecule microscopy results involve significant statistical inferences due to the fundamental sub-micrometer optical resolution. A large spectrum of molecules (cyclic nucleotides CAMP and cGMP, IP3, diacylglycerol, etc.) have been extensively studied as second messengers of intracellular events initiated by activation of many types of hormone and neurotransmitter receptors. Various phospholipases, protein kinases and phosphotases (key enzymes whose activity initiates many signaling cascades) are targeted specifically to the plasma membrane. The importance of direct measurements of their activity and/or concentration of their substrates/products in the membrane domain are indicated in several research results.xi
A unifying feature of most pathways interfacing extra- and intracellular events is the engagement of Ca2+ signaling. This is because most of Ca2+ ions are bound/sequestered inside cells, so that concentration of free Ca2+ is very low (˜100 nM) in the bulk of cytoplasm. On a “nano-scale” corresponding to protein sizes, this means that the probability to find one free Ca2+ in a (10 nm)3 volume is about 1%. In the 10-100 nm vicinity of open Ca2+ permeating channel (sources) or active pump (sinks), [Ca2+] rapidly changes in the range from 100 nM to 100 μM. Spatial and temporal distributions of near-membrane [Ca2+] are strongly influenced by gating/kinetic properties of Ca2+ transporters and by various buffers of different affinity and diffusional mobility. Although fluorescent Ca2+ probes with a wide assortment of properties are available to study intracellular Ca2+ signaling, understanding of the near-membrane Ca2+ dynamics, where the bulk of interactions between Ca2+ ions and other signaling events occur, is limited by the lack of direct measurements in the 100 nm domain under the plasma membrane. Ca2+ measurements in the space (˜100 nm) between plasma membrane and sarcoplasmic reticulum (cell organelle originating from endoplasmic reticulum and filled with Ca2+) of various types of cells would be just one, but good, example of the potential use of Ca2+-sensitive nano-electrode. Ca2+ dynamics in these junctions are key determinants of heart beat, blood pressure, muscle contraction, synaptic transmission, opto-electrical transmission in the eye and many other physiologically, as well as medicinally, important mechanisms.xii
A CNT probe array may also be used as force sensors. It is known that single wall carbon nanotubes (SWNT) can undergo a two order of magnitude reduction in conductance under a 3% strain that is induced by bending.xiii The measurements were done on an SWNT suspended over a trench and using an atomic force microscope (AFM) tip the bend the SWNT. A plot of the conductance versus strain is shown in
Using this conductance vs. bending behavior in a force sensor using SWNTs in the geometry shown in
Equation 2 is a slight over estimate of the force since it assumes that the SWNT is solid. Using E=1.2 TPa, L=100 nm, and D=1.2 nm, F as a function of the deflection angle θ is plotted in
The cell's cytoskeleton consists of a complex network of interconnected struts or tracks that function as structural elements (actin filaments, microtubules, and intermediate filaments) providing shape and form and as highways (actin filaments and microtubules) ridden by molecular motors to transport cargo and generate contractile forces. Unlike conventional architectural struts and highways, the cytoskeleton is a highly dynamic network whose organization and function is almost instantaneously changeable during normal physiological activity. Classic cellular examples of these dynamics are muscle contraction, flagellar motility, cell division, and amoeboid crawling. At the protein level these activities are driven by rapid assembly/disassembly of polymer and reorganization of polymer structure under the control of spatially and temporally binding partners that themselves are regulated by intracellular signaling pathwaysxv,xvi,xvii,xviii,xix,xx. The assembled polymers are then used by the myosin protein family (for actin filaments) and the kinesin/dynein protein families (for microtubules) for force transduction via mechanochemical activity coupled to hydrolysis of ATP. Carbon nanotube-based methodologies can be used to explore dynamic mechanical properties of protein-protein interactions, which empower the cytoskeleton with its recognized diversity of structure and function. As the initial step in this development, carbon nanotube sensors will be used for probing the dynamic properties of actin filaments because this cytoskeletal system is easily manipulated for in vitro experimentation that will then be transferable to an in vivo setting.
Actin filaments are composed of actin monomers that freely assemble into polymer under physiological salt conditions with the only constraint being the monomer concentration must exceed the minimal critical concentration needed to nucleate assemblyxviii,xx,xxi. Once assembled the polymer can be modeled as a two-start helical filament consisting of two protofilament strands with right-handed pitch and a repeat spacing of approximately 14 monomersxxii,xxiii. The repeat spacing exhibits some dispersion because of “angular disorder” in the spacing of monomers along the polymerxxiii, which results from internal rotation/translational movement of monomer along the length of the filamentxxiv. The importance of such rotational dynamics lies not in the differences of helical pitch but in the resultant exposure of monomer interfaces created by changes in monomer orientation within the filament. The importance of such translational movements of monomer in the physiological activity properties of actin filaments was first noted in studies with the capping agent cytochalasin Bxxv and has subsequently been recapitulated for binding interactions with members of different actin binding protein familiesxxvi,xxvii,xxviii,xxix.
The atomic structure identifies a globular protein with 4 domains separated by defined clefts that serve as natural bending points for translational movement within a polymer.xxx Given the atomic structure of an actin monomer, any protein or agent that binds to an actin filament has the potential to either limit or enhance translational/vibrational movement of monomer which in turn changes the mechanical properties of the filament. Taking this one step further, changes in mechanical properties of a filament will ultimately dictate it physiological activity in vivo. Thus, any physiological or pharmacological agent that can alter the vibrational modes of an individual monomer within a filament would have the potential to alter the physiological properties of the filament. In one example, it has been shown that the actin binding protein cofilin reduces angular disorder along a filament, which in turn prevents binding of the pharmacological agent phalloidin.xxxi Consequently, determining the dynamic properties of an individual actin monomer within an individual filament will provide fundamental information on how and possibly why particular combinations of actin binding proteins are used to build such diverse structures as stereocilia of hair cells in the cochlea, microvilli of the brush border in intestinal epithelial cells and the thin filaments of sarcomeres in striated muscle. While the aforementioned examples represent extremes of form and function, it is certainly reasonable to hypothesize that an individual cell must have domains of actin exhibiting different physical properties since different regions of the cell are known to exhibit cytoskeletal structures and functions.
At present, there is no real-time method to study the described dynamics with molecular resolution. The resolution of optical techniques using fluorescent analogs or fluorescent actin binding molecules is limited to a fraction of a micron even with sophisticated enhancement techniques and consequently these observations essentially provide bulk measurements from populations of monomers. Optical tweezers have been useful in measuring binding forces, but cannot track the dynamics of the actin monomers.
The present disclosure relates to SWNT-based methods to measure the activity of individual actin monomers and record the dynamics of the attachment of SWNTs to the cytoskeleton as a function of the physiological activity within living cells. The disclosed approach to methods development will build upon analysis of three different actin filament populations—individual filaments, magnesium-induced actin filament paracrystals, and bundles of actin filaments isolated from Limulus spermxxxii,xxxiii. Upon this base, it is possible to expand the application to examining effects of nucleotide, tension, and interactions with binding proteins/myosin on individual monomer dynamics. Importantly, the examination of effects on monomer dynamics that may be elicited by activities occurring microns away from the measurement site may be undertaken. The underlying physical basis of the measurement is that, as a monomer undergoes translational motions, there is an associated production of force, which will cause the SWNT to bend resulting in changes in SWNT conductance. An analysis of the magnitude and frequency of SWNT conductance changes can then be directly related to the forces and dynamics associated with individual monomer movements. Consequently, by using complementary techniques to induce changes in filament properties transformations, it may be possible to establish fingerprints for these dynamics, which in the future would be used to characterize the spatial and temporal dynamics of filament populations within cells.
Functionalization of the SWNTs is a key enabler for the disclosed method. Functionalized carbon nanotubes offer enormous potential as components of nanoscale electronics and sensors. The prospect of these applications has led to successful functionalization of single wall (SWNT) and multiple wall carbon nanotubes (MWNT). These functionalizations may be separated into two categories: a non-covalent wrapping or adsorption and covalent tethering. In the first category, O'Connell et al. showed evidence for the formation of water-soluble SWNTs by wrapping with various polymers.xxxiv Similarly, Zheng et al. have reported DNA wrapping onto SWNTs through relatively weak π-stacking.xxxv On the other hand, several attempts have been made to achieve bonded functionalization of SWNTs in the second category. For example, fluorination of SWNTsxxxvi,xxxvii, 1,3-dipolar additionxxxviii, derivatization of small diameter SWNTsxxxix,xl, glucosamine attachmentxli, and sidewall carboxylic acid functionalization of SWNTsxlii have been shown to occur. As reported by these authors, a common outcome has been the increased solubility of SWNTs either in an organic solvent or water. Previous attempts at achieving enzyme linkage to SWNTsxliii were done via non-covalent adsorption or via diimide activated amidation of SWNTsxliv,xlv. Recently members of this research team reported the first study on functionalization of SWNTs with enzymes, achieved chemically by first acylation of SWNTs followed by amidation with the desired protein.xlvi The two-step chemical method employs mild conditions and results in tethering of the organic or bioorganic functionality through a covalent bond. It is a simple, practical and highly effective protocol. This linkage of chiral molecules and enzymes to SWNTs has opened the possibilities for applications of carbon nanotubes in medicinal and biological fields, and in biosensor or chemically modulated nanoelectronic devices.
The nanotube sidewalls are rather inert and in some cases chemical bonding may become a complex task. To overcome such hurdle, a relatively facile and scaleable electrochemical strategy for the direct functionalization of SWNTs with nitro-groups has been developed.xlvii This raises the possibility of wider use of electrochemical techniques for the controlled and in-situ functionalization of carbon nanotubes in device structures. The electrochemical functionalization of oriented SWNTs with 50 mg of enzyme β-NAD Synthetase was carried out in a three-electrode cell, with 0.1 M of KCl as the electrolyte and 80 ml of pH 7 phosphate buffer (40 mM).xlviii The Si wafer with oriented SWNTs was directly used as working electrode, with a platinum wire as counter electrode and a saturated calomel electrode (SCE) as reference electrode. The enzyme attachment was verified with cyclic voltammetry. After reaction, the SWNT electrode was rinsed with reaction buffer and distilled water, and then dried in air. The SEM image after functionalization is shown in
Measurements of Ca2+ will require an SWNT probe functionalized by adding to its tip compounds that strongly interact with Ca2+, e.g., bisulfite, or by adding other negatively charged molecules that will interact preferentially with strongly charged Ca2+ and/or Mg2+—the main intracellular divalent cations, rather than with K+ and Na+. Charged molecule attachment is best done electrochemically.
To investigate the dynamics of actin filaments, a “vehicle” must be attached to the tip of the SWNT that can subsequently bind to the actin filament when the SWNT is brought into close proximity. There are several possibilities for the type of vehicle and they include a chemical crosslinker, a pharmacological actin-binding factor, or an actin binding protein. Each have their advantages and disadvantages based on the type of question being asked, the required resolution of movement, the experimental set-up, etc. Initial experiments will make use of pharmacological agents such as phalloidin (a mushroom toxin), jasplakinolide (a marine sponge cyclic peptide), and dolastatins (isolated from sea hares). These are commercially available, bind actin filaments with high affinity, and have been chemically modified (fluorescently-labeled) to study actin. It is proposed that phalloidin/jasplakinolide will be attached to the SWNT probes to enable binding to the actin filaments for studying their dynamics. For phalloidin, two possible techniques may be employed for attachment in this application. One would use a fluorescent analog of phalloidin (eg. fluorescein-phalloidin) as an independent optical signal for actin binding by fluorescence energy transfer when using a rhodamine-labeled actin analog.xlix Alternatively, there has been success with attaching Au to phalloidin which could then be used as a spatial marker for TEM investigations of actin.l It is proposed that a similar technique be employed where Au is electrodeposited to the SWNT probe followed by attachment of phalloidin.
The present application is related to and claims priority benefit to a co-pending, commonly assigned provisional patent application entitled “NANOTUBE DEVICES,” which was filed on Jun. 20, 2007 and assigned Ser. No. 60/945,170. The entire contents of the foregoing provisional patent application are incorporated herein by reference.
| Number | Date | Country | |
|---|---|---|---|
| 60945170 | Jun 2007 | US |
