Nanodrawing of Aligned Single Carbon Nanotubes with a Nanopen

Nanodrawing of Aligned Single Carbon Nanotubes
with a Nanopen

Talia Yeshua,1,2 Christian Lehmann,3 Uwe Hübner,4 Suzanna Azoubel2,5 , Shlomo Magdassi,2,5 Eleanor E. B. Campbell,6,7 Stephanie Reich,3 and Aaron Lewis*1,2

1. Department of Applied Physics, Selim and Rachel Benin School of Engineering and Computer Science, The Hebrew University, Givat Ram, Jerusalem 9190401, Israel

2. The Center for Nanoscience and Nanotechnology, The Hebrew Givat Ram, Jerusalem 9190401, Israel

3. Department of Physics, Freie Universität Berlin, Arnimallee 14, Berlin 14195, Germany

4. Leibniz Institute of Photonic Technology, A. Einstein 9, Jena 07745, Germany

5. Casali Center for Applied Chemistry, Institute of Chemistry, The Hebrew University, Givat Ram, Jerusalem 9190401, Israel

6. EaStCHEM, School of Chemistry, Edinburgh University, Edinburgh EH9 3FJ, Scotland

7. Department of Physics, Konkuk University, Seoul 143701, Korea.

Abstract

Single-walled carbon nanotubes (SWCNTs) are considered pivotal components for molecular electronics. Techniques for SWCNT lithography today lack simplicity, flexibility and speed of direct, oriented deposition at specific target locations. In this paper SWCNTs are directly drawn and placed with chemical identification and demonstrated orientation using fountain pen nanolithography (FPN) under ambient conditions. Placement across specific electrical contacts with such alignment is demonstrated and characterized. The fundamental basis of the drawing process with alignment has potential applications for other related systems such as inorganic nanotubes, polymers, and biological molecules.

SWCNTs are nanoelements that are both fundamentally and technologically of considerable significance.1 In terms of nanoelectronics they are considered the molecular electrical building blocks that can be used for interconnects,2 transistors,3-5 and sensors.6-9 This is especially significant as microelectronics is moving toward length scales that require new methods and material systems to achieve the desire for smaller, faster, and higher efficiency circuits.10,11

SWCNTs have excellent electrical and physical properties such as high conductivity, low energies of activation and mechanical robustness.12 Specifically, SWCNTs pose a way of achieving device interconnects with an electrical performance far better than the present state of the art.12,13 This is highlighted in the recent fabrication of a functional SWCNT computer14.15 using gross deposition techniques that are hard to control at a single connection level.

In our paper we focus on this control of placement at single target locations. This is done with simplicity, flexibility, and speed of placement.

As has been noted,1 there are few if any methods that focus on the targeted placement of SWCNTs. Nonetheless, there are important approaches that have to be mentioned that deal with the problem of how one could assemble generally over a targeted area aligned carbon nanotubes. These methods involve a variety of innovative methods. One such method is chemical vapor deposition for directly growing aligned SWCNTs on the device substrate.16-19 This procedure can also involve the transfer of SWCNTs from a “growth” substrate to a device substrate.14,20 Another approach is to use contact printing and the procedure of contact printing itself aligns the SWCNTs on the patterned device.21,22 A third method involves deposition of SWCNTs from solution onto large areas of the substrate by either chemical assembly,23-26 dielectrophoretic control,27,28 or using a blown bubble film.29 An intermediate approach is to use a contact printing with a solution in between the contact printing surfaces. However this work focused on nanotubes of sizes that were considerably larger than SWCNTs.30

All of the above methods apply very innovative methodologies to address this important problem, but they are either incapable of simply choosing a specific location or certainly do not have the one step efficiency of our method.

In this paper we describe the drawing and Raman characterization procedure developed for SWCNT placement, proof of SWCNT alignment, optimization of the drawing parameters, and the subsequent placement in predefined lithographic structures for the demonstration of electrical conductivity.

The FPN method uses a simple glass cantilever as a nanofountain pen to draw a nanotube solution directly onto a substrate with atomic force feedback.31 The drawing speed is microns per seconds. The SWCNTs are deposited and aligned by capillary forces in the drawing direction. After the drawing operation, the substrate is washed with water, so only SWCNTs which are strongly held by electrostatic or other forces are left on the surface. Therefore, it is possible to draw and place oriented SWCNTs at desired targets rapidly.

FPN has overtones of the dip pen nanolithography (DPN)32 method but has some powerful advantages. FPN does not have DPNs preference for hydrophilic substrates like gold and does not need to deposit linker molecules for deposition of SWCNTs24 as is the case in DPN. In addition, the nanopens used in FPN are optically transparent and have their drawing tip highly exposed to the optical axis for accurate placement within a device structure.33 Thus, an upright optical microscope can readily be used on opaque device substrates to effectively position the drawing, at the highest resolutions of optical microscopy (a submicron level). Furthermore, the demonstrated ability to achieve atomic force microscopic (AFM) imaging with nanopens allows AFM resolution for the placement on a nanometric scale.

Initial experiments34 indicated that FPN could deposit aligned SWCNT structures. In the present report these initial studies have been advanced in many directions in order to make the writing of SWCNTs a general tool in nanotechnology. For taking this next step, the writing parameters, previously not investigated for any material, have been optimized. This was a critical step for repeatably writing aligned SWCNTs. These parameters included concentration of the SWCNTs in the dispersion, speed of writing, orifice diameter, settling time before washing etc. This has given the ability to accurately draw complex patterns repeatably, without chemical intervention, i.e. surface modifications, and to place aligned SWCNT for effective conducting contacts from one patterned electrode to another with intermediate regions of silicon and/or silicon oxide. This proves the ability with FPN to draw on a variety of materials of considerable importance for prototyping. With this ability of effective control and alignment we have been able to deposit even single conducting and semiconducting SWCNTs between gold electrodes on SiO2 and to choose specific electrodes that are to be connected. The result is the ability to go from deposition to direct electrical characterization without any additional processing steps.

Figure 1 shows images of the drawn structures taken by AFM and chemical Raman mapping of the G+ mode to demonstrate that in fact SWCNTs have been deposited by the drawing process. We show the letters “HU” written by a nanopen with 150 nm orifice on a silicon substrate covered with 300 nm of SiO2. The drawing speed was 5 µm/s and Dispersion I (see Methods) was used. In order to write a new line the tip was retracted and brought into contact again. As such it was used like a pen (see Supporting Information, Movie 1 and Movie 2, for movies of the drawing of such structures). The drawn lines before and after washing are shown in Fig 1d-g. After washing a few well aligned SWCNTs remained on the surface.

Figure 1 | Controlled direct drawing of SWCNTs on SiO2. (a) A diagrammatic representation of the drawing process. The SWCNT alignment occurs by capillary forces in the direction of drawing. (b) Raman mapping of the written structure of the G+ band together with (c) a Raman spectrum at a point in this image. (d) AFM of the written structure before washing with (e) a line scan of a particular location. The height of the drawing is 200-250nm and the line width is 600-800 nm. (f) AFM of a portion of the written structure that has been washed after 4 days of drying, with (g) a line scan at a particular location. After washing only the SWCNTs remained on the substrate with the associated surfactant and solution being washed away. The height of the drawn structure is between 0.8 and 2.9 nm

In Figure 2 an AFM of straight lines of SWCNTs is presented before washing along with a Raman image after washing that clearly shows the presence of SWCNTs even after washing in most of the sample. Dispersion I was drawn on a SiO2 substrate using a nanopen with an orifice of 150 nm and a drawing speed of 8 m/s. The AFM after washing (Fig 2b,c) shows good alignment in the direction of the drawing at a specific location in a written line. This is supported by Raman polarization studies35 which show alignment globally in the sample (see Figure 2e and Supporting Information section b).

Figure 2 | Drawing of individually aligned SWCNTs. (a) AFM image before washing. (b) After a drying period of 3 h and subsequent washing, the boxed area in panel a imaged by AFM with (c) an associated line scan (c). Note that the height is ~0.7 nm which clearly indicates single SWCNTs. (d) Raman map of the G+ mode after washing. (e) Histogram of the S parameter of the AFM line scan in panel a. S=1 implies that the tubes are totally aligned in the vertical direction. S=-0.5 implies total horizontal alignment. As can be seen, nearly total alignment has been achieved.

A variety of drawing parameters were investigated. These include the concentration of the SWCNT dispersion, the size of the nanopen orifice, the speed of drawing, the set point of the AFM and the direction of drawing vis a vis the nanopen tip. As can be seen on Figure 3a,b in spite of the orifice of the pen being 100nm the drawing speed affects the dimensions of the drawn lines while all other parameters are held constant. In contrast in Figure 3c the 3D AFM image of the lines shows no influence of the set-point of the AFM signal on the line dimension. The set-point indicates the alteration in the cantilever height that is used to monitor the feedback during the drawing operation. Set-point values from 0.3V to 0.9V resulted in uniform lines of 120 nm width and heights of 40 nm. All of the above pertain to the resulting drawing before washing. For a delineation of the other parameters mentioned above see the Supporting Information, section c.

As clearly shown in SEM images (Figure 3 d,e) the washing process after drawing is critical to remove associated surfactants and other chemicals present in the SWCNT dispersion. Specifically, as is seen in Figure 3d, the SEM cannot detect SWCNT after the drawing due to the presence of these substances. After a washing operation (as described in the Supporting Information, section d) the remaining SWCNTs are clearly seen but due to the fact that a SEM image was taken before the washing process, residue around the line is seen. This residue is caused by deposition of these chemicals due to interaction with the SEM beam. If the washing was done before the SEM imaging then the residue would not have been seen. In Figure 3e the position of the line of SWCNTs relative to the residue, as seen in this SEM image, is generally what is seen after such a washing operation. In this case a SWCNT line of ~50 nm is left. In Figure 3f a slow speed (0.5 m/sec) produces a line that has a width of 200 nm, while the same probe with a speed of 8 m/sec produces a line of < 10 nm (see Figure 3g). With faster speeds as shown in Figure 3a the line dimensions were too small for continuous deposition of SWCNTs under the above conditions of concentration and orifice diameter. In fact depending on the conditions, drawing speeds of as much as 400 m/s can be achieved. No effect has been seen, in terms of drawing speed, as to whether there was or was not a linker molecule on the surface. The principal parameter for effective deposition is the time given for the SWCNTs to settle after the drawing to interact via surface/molecular forces of interaction or if a linker molecule is present to react chemically with the surface.

Figure 3 | Controlling the effective drawing parameters.(a) 3D AFM image of drawn lines with different speeds (0.5-32 m/s) and (b) an associated graph that summarizes the line dimensions. The faster the drawing of the line, the smaller the dimensions of both height and width. (c) Changing the set-point of the error signal from 0.3V to 0.9V shows no influence of such set point changes on the line dimensions for the ~120nm width and ~40 nm height lines shown (nanopen aperture 100 nm). (d) SEM image of a line of drawn SWCNTs before washing. (e) SEM image of the same drawn line in d after washing. The SEM imaging before the washing step caused in e the deposition of residual material around a thin central line ~50 nm of SWCNTs. (f) SEM image after the washing process of a line drawn at slow speed (0.5 m/s) using a nanopen with an orifice of 150 nm. As can be seen the SWCNTs have less alignment and cover a larger area. (g) SEM image after washing of a line drawn with higher speed (8 m/s) using the same nanopen as in g. Clear alignment can be detected, and the line width of the SWCNTs is < 10 nm.

Based on this demonstration of good alignment of the SWCNTs the technique was applied to write electrical interconnects over a pre-patterned substrate consisting of gold electrodes (see Figure 4). A diagrammatic illustration of the substrate used is shown in Figure 4a. As is seen in this illustration, groups of electrically isolated electrodes were patterned close to the gold pad with gaps of 350 and 850 nm, respectively. SWCNT dispersions were written along the gold electrodes and across the gaps to the gold pad in order to make electrical connections. The SEM image, after the drawing and washing step, is shown in Figure 4b. For the group of electrodes seen in this image on the left (850nm gap), the bright and dark lines are those electrodes that were either connected (bright lines) or not connected (dark lines) to the gold pad. The group of electrodes on the right (350nm gap) was all connected and so all of these electrodes appear bright in the SEM image.

In another series of experiments shown in in Figure 4c a bundle of carbon nanotubes was drawn using a nanopen orifice of 250 nm and a concentration that is three times the amount of Dispersion II (see Methods). The SEM allows for a large field of view at high resolution and these images clearly show the fidelity of the drawing process of these SWCNT. The electrical measurements are seen in Figure 4d. This IV curve shows partial semiconductor character with a resistance of ~20 kΩ. Under the original conditions of concentration noted above (namely Dispersion II) and a nanopen orifice of 150 nm an extremely narrow SWCNT structure was capable of interconnecting a gap of 350 nm (Figure 4e). This is close to a single SWCNT connection, although SEM imaging at this level is not always clear.20 The associated current voltage (IV) curve is shown for this connection (Figure 4f). Additional examples of drawing single SWCNTs between electrodes are shown in the Supporting Information, section d.

Figure 4 | Drawing electrical interconnects with aligned SWCNTs on a pre-patterned chip. (a) Schematic of the pre-patterned chip. Groups of electrodes 50nm thick are patterned and isolated from a large pad for demonstrating electrical contact with SWCNTs. Gaps between the electrodes and the pad were patterned with 850 nm and 350 nm gaps with one group of electrodes having the larger separation and the other group the smaller separation. The drawing direction of the SWCNTs was from the electrodes to the large contacting pad across the gap. (b) SEM image of the chip after the SWCNTs deposition process and washing is shown. The SWCNT interconnects were drawn in this example with a nanopen with an orifice of 150 nm. Dispersion II was used with a drawing speed of 5 µm/s and 3 h before washing. The bright electrodes are those that were contacted with the large pad. (c) Deposition of long bundle of SWCNTs over the electrode with SWCNT concentration of three times Dispersion II with (d) electrical measurements indicate a resistance of ~20 kΩ over the gap.(e) SEM images of single SWCNT over 350 nm gap with (f) an I-V curve measurement.

In summary, this paper presents a method of drawing carbon nanotubes nanometrically that should have general applicability for facilitating the vibrant SWCNT device development that is occurring today. The factors that control the drawing of single SWCNT depend on the optimization of concentration, the orifice diameter of the nanopen and the drawing parameters which control the line width. In addition to this, a critical factor is the settling time to establish the surface/molecular forces of interaction. Optimization of these parameters can result in a wide variety of types of SWCNT deposition from single SWCNTs to bundles. Furthermore, the capillary action of the nanopen drawing process results in SWCNT alignment. This together with the detailed advances that have been made in the drawing parameters developed in this paper could have broad potential for applications to other types of materials such as inorganic nanotubes and molecular chains such as DNA.

Methods. SWCNT dispersion preparation. Two different SWCNT dispersion preparations were used in this paper:

“Dispersion I”: This dispersion contains SCWNTs that were produced by pulsed laser vaporization (PLV). This method uses graphite targets doped with Ni and Co (Toyo Tanso) in an Ar gas flow at a temperature of 1000°C. Prior to further use, the PLV SWCNT soot was purified by washing it three times with dimethylformamide (DMF). A portion of 5 mg of the SCWNTs material was suspended in 30 mL of D2O with 1 wt. % of sodium cholate (Sigma Aldrich) using a tip sonicator (Bandelin, 200 W, 20 kHz) applied for 1 hour at 10% power while cooling the sample in an ice bath. After that the dispersion was centrifuged at 20000 rpm for half an hour to remove larger agglomerates.36 This dispersion was used in the results of Figures 1,2 and S1-S4.

The second dispersion, “Dispersion II”, had an initial concentration of 0.2 mg/mL CoMoCat SWCNT's in deionized water. After adding 20 g/L sodium cholate (Sigma Aldrich) as surfactant, the dispersion was sonicated for 60 minutes using a 70 W tip sonicator. Afterward the tube dispersion was centrifuged at 18000 rpm for 180 minutes with a subsequent settling time of 4 days at 8°C before deposition. This dispersion was used in the results of Figures 3,4 and S5-S11.