APPLIED PHYSICS LETTERS
VOLUME 77, NUMBER 21
20 NOVEMBER 2000
Improved fabrication approach for carbon nanotube probe devices R. Stevens, C. Nguyen, A. Cassell, L. Delzeit, M. Meyyappan, and Jie Hana) NASA Ames Research Center, Moffett Field, California 94035
共Received 27 June 2000; accepted for publication 26 September 2000兲 An improved process is developed for simple and efficient fabrication of carbon nanotube probe devices. This process requires only two steps to make nanotube probes. First a nanotube cartridge is created using chemical vapor deposition, then the nanotubes are transferred from the cartridge to a device using an electric field. Multiwall nanotube probes are made into different device geometries in this approach. Their applications are illustrated by atomic force microscopy imaging of the surface of a terrestrial rock granule selected to simulate the morphology and consistency of a grain of Mars dust and nanolithography on a silicon substrate. 关S0003-6951共00兲04347-3兴
Carbon nanotubes can be viewed as hollow cylinders formed by rolling up a graphite sheet. Multiwall and single wall nanotubes 共MWNT and SWNT兲 have small diameters of ⬃10 and ⬃1 nm, respectively, a high aspect ratio of 102 – 104 , a high Young’s modulus of ⬃1 TPa, and are electrically conducting. These characteristics make them ideal for probe microscopy tips. Nanotubes have previously been made into atomic force microscopy 共AFM兲 tips and have proven to have great advantages in imaging and manipulation over conventional silicon and silicon nitride tips.1–12 Other applications of our interest include using carbon nanotube probes for in situ deep space nanoscale imaging. In this application, nanotube probes are required to be attached perpendicular to a planar surface. However, problems have plagued the fabrication of nanotube probe microscopy tips. The earliest approach was to pick and stick a MWNT bundle to the tips with an acrylic adhesive under an optical microscope.1–5 This process is difficult and time consuming. Therefore, chemical vapor deposition 共CVD兲 has been used to grow MWNT or SWNT on a catalyst deposited tip surface.6–9 Nanotubes grow from the deposited catalyst but there is no control over whether the nanotube grows in the desired location and protrudes from the tip. The third approach is the electric field induced MWNT probe attachment.10–12 It involves several steps such as nanotube growth, purification, transfer to a cartridge using electrophoresis and adhesion, and transfer from nanotube probe cartridge to the tips under an electric field. This approach provides advantages in control of the orientation of nanotube probes because of the role of electric field in alignment and electrostatic attraction of nanotube probes. However, the whole process is time consuming, and two transfer stages need to be monitored by a scanning electron microscopy 共SEM兲. We have developed a simple process for efficient and reliable fabrication of nanotube probe devices. This approach only needs two steps: CVD growth of a carbon nanotube probe cartridge and an electric field transfer. It integrates the three steps in the previous electric field approach into one step without need of SEM. It allows the fabrication of nanotube probe microscopy tips with control over the probe oria兲
Electronic mail:
[email protected]
entation which is difficult in the CVD direct growth approach. Because of its simplicity and flexibility, we have been able to make nanotube probes not only protruding from pyramidal tip geometries, but also other geometries such as perpendicular to a planar surface. The first step is to make a nanotube cartridge based on a well-developed metal-catalyzed CVD method.13 A simple procedure for the synthesis of individual MWNTs directly on the nanotube mounting source 共cartridge兲 was developed in this work. The key steps are the preparation of a liquid phase catalyst precursor material, catalyst transfer to the mounting source, and CVD of ethylene to obtain the source for individual nanotube tips. The liquid phase catalyst precursor was made by dissolving 0.50 g of a triblock copolymer 共P-123, BASF Corp., Mount Olive, NJ兲 structure-directing agent into 15 mL of an alcohol solution 共10 mL ethanol, 5 mL methanol兲. Next, 0.85 mL of SiCl4 (7.0 mmol) was slowly added via syringe while stirring the solution with a magnetic stir bar. Then, Fe共NO3兲3•9H2O 共1.0 g, 2.3 mmol兲 was dissolved into the solution to complete the preparation of the catalyst precursor. Next, a Pt/Ir wire 共0.05 in. diameter兲 was snipped to 1 cm length, and then submerged halfway into the liquid phase catalyst precursor solution for 30 s. The catalystcoated wire was then placed into a 1 in. diameter tube furnace, and heated in air for 4 h at 700 °C to decompose the catalyst precursor into active catalyst. Argon replaced air at a flow rate of 0.5 L/min as the furnace was heated to 750 °C, and then CVD was initiated by flowing ethylene at a flow rate of 0.5 L/min for 30 min. This technique provides a nanotube mounting source containing an abundance of individual MWNTs. Figure 1 shows an example of nanotube probe cartridges grown by CVD. The second step is to transfer a nanotube from the cartridge onto the cantilever to create a usable probe. This is done using an optical microscope 共Zeiss兲. The nanotube cartridge and the tips were held by translation stages under the optical microscope. Using the microtranslators, the cantilever and the cartridge were brought into close proximity to each other under direct view of the optical microscope. When nanotube and cantilever are in close proximity, an electric field is applied between the cartridge 共positive elec-
3453 Downloaded 14 Apr 2003 to 143.232.211.49. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp
3454
Appl. Phys. Lett., Vol. 77, No. 21, 20 November 2000
Stevens et al.
FIG. 3. MWNT probe AFM image of a silicon surface patterned using a MWNT probe to oxidize 20 nm wide lines with a bias voltage of 17 V applied between the MWNT probe shown in Fig. 2共a兲 and the silicon surface.
FIG. 1. SEM image of low and high density of carbon nanotube probe cartridge grown by CVD. The tube diameter is from 10 to 30 nm.
trode兲 and the cantilever 共negative electrode兲. The cantilever is coated with 15 nm of nickel in order to improve the conductive path. With a low voltage applied 共3–10 V兲 the nanotube is attracted to the cantilever tip and favorably aligned with the apex of the tip. The attraction is due to the induced electrostatic dipole moment in the nanotubes and the alignment is due to the geometry of the field created by the cantilever tip shape. When the nanotube is suitably aligned the voltage is increased 共10–40 V兲 until the nanotube is energetically disassociated from the cartridge. The nanotube source is multiwalled nanotubes grown in low temperature CVD and has small defects at which point the resistance locally heats the nanotube until it oxidizes and divides. This event is observable in the optical microscope as an arc discharge. The point at which the nanotube splits can be somewhat determined by using the attraction between the tip and nanotube to mechanically bend the nanotube to induce strain defects at a desired point. After an arc occurs the nanotube can be seen attached to the cantilever free from the cartridge. The low temperature CVD allows the formation of spatially separated MWNTs that contain some defects which are ideal
for this attachment method. Curly or coiled MWNTs make up a portion of the MWNT sample population and can also be attached to cantilevers but their curvature increases their effective tip radius and so, were not used as probes. The nanotube is affixed to the cantilever through chemical force between the nanotube and the metallic tip. A high electric current was exerted to the tip electrode which serves as a resistor to locally generate a high temperature environment, possibly enabling the formation of a carbide at the contact site as well as causing the nanotube to partially disassociate at this interface and deposit amorphous carbon. The nanotube is also held in place by the Van der Waals force between the nanotube, metallic tip, and amorphous carbon. Once attached the nanotube can be shortened to the desired length by following the procedure detailed by Refs. 6–9. Figure 2共a兲 shows a single MWNT probe microscopy tip in which the inset is a bundled MWNT probe. The simplicity and flexibility of this approach allows attachment of nanotube probes to other geometries. As an example, Fig. 2共b兲 shows a SEM image in which a multiwall nanotube probe is attached perpendicular to a solid surface. This structure will be used for profilometer applications in our space exploration programs. It is very difficult for other approaches such as mechanical attachment1–5 and direct
FIG. 4. 共a兲 Conventional silicon probe AFM image of a vertical sided line space array that shows how the shape of the imaging probe can introduce FIG. 2. SEM images of individual and bundle 共the inset兲 MWNT probes imaging artifacts. The sloping sides of the features in the image reflect the attached to 共a兲 a conventional cantilever tip and 共b兲 a planar metallic surface sloping sides of the imaging probe and not the vertical sides of the sample using the approach of this work. The MWNT diameter is about 20 nm 共a兲, features. 共b兲 MWNT image of line space array showing that the MWNT tip and the bundle width is ⬃60–10 nm from the contact to the freely standing is able to track these vertical sided features. nanotube edge 2共b兲 and 2共a兲. Downloaded 14 Apr 2003 to 143.232.211.49. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp
Stevens et al.
Appl. Phys. Lett., Vol. 77, No. 21, 20 November 2000
3455
work. Figure 3 shows an AFM image in which words were written and read by an individual MWNT probe by following a reported procedure using AFM for nanolithography.4 Figures 4共a兲 and 4共b兲 are images of a line-space array of patterned resist on silicon. The topography of the sample in Fig. 4 is an example of a surface that is characteristically very difficult to resolve in AFM. Figure 4共a兲 was imaged using a conventional silicon probe and Fig. 4共b兲 was imaged using a MWNT probe. This sample has 300 nm tall vertical sided features and the conventional pyramidal tip shape is unable to accurately track the surface. The MWNT tip is able to image the surface at the base of the tall feature but the Si probe’s sloping sides create imaging artifacts by contacting the tall features before the probe’s tip can reach the surface. The cross section for the conventional Si probe shows sloping sided features but the true topography is more accurately reflected by the cross section of the MWNT tip image in Fig. 4共b兲. Figures 5共a兲 and 5共b兲 are AFM images of the same surface of a grain of terrestrial rock selected to simulate the morphology and consistency of a grain of Mars dust. These images show the effect that artifacts such as those demonstrated in Fig. 4共a兲 can have on SPM data from samples with large vertical variation. Figures 5共a兲 and 5共b兲 are exactly the same area on the grain but Fig. 5共a兲 was imaged with a conventional probe and shows such a large number of imaging artifacts that when compared with Fig. 5共b兲, the surface is almost unrecognizable. Figures 4共a兲 and 4共b兲 imply that the MWNT image in Fig. 5共b兲 more closely resembles the true surface topography. In conclusion, we have developed a process for fabrication of multiwall carbon nanotube probe devices including microscopy tips and profilometry. Our technique relies on CVD growth of aligned, spatially separated MWNTs that allows a high degree of selection and ease of attachment. This process has only two steps and is simple and reliable. 1
FIG. 5. 共a兲 Conventional silicon probe AFM image of simulated Mars dust. This image is dominated by tip artifacts resulting from the tip–sample interaction demonstrated in Fig. 4共a兲. These images show features with smoothly sloping sides that are a result of the side of the conventional pyramidal tip making contact with a tall feature before the apex of the tip makes contact with the surface, and 共b兲 a MWNT probe AFM image of the same surface as in Fig. 5共a兲. This image shows that the MWNT probe is able to more accurately track the sample surface thereby resolving fine features and does not generate the smooth sloped tip artifacts that dominate Fig. 5共a兲.
CVD growth6–9 to obtain such nanotube probe configurations. Finally we give examples of the wide range of applications for nanotube probe microscopy tips obtained from this
H. Dai, J. H. Hafner, A. G. Rinzler, D. T. Colbert, and R. E. Smalley, Nature 共London兲 384, 147 共1996兲. 2 S. S. Wong, E. Joselevich, A. T. Wolley, C. L. Cheung, and C. M. Lieber, Nature 共London兲 394, 52 共1998兲. 3 S. S. Wong, A. T. Wolley, T. W. Odom, J.-L. Huang, P. Kim, D. V. Vezenov, and C. M. Lieber, Appl. Phys. Lett. 73, 3465 共1998兲. 4 H. Dai, N. Franklin, and J. Han, Appl. Phys. Lett. 73, 1508 共1998兲. 5 J. Li, A. Cassell, and H. Dai, Surf. Interface Anal. 28, 8 共1999兲. 6 J. H. Hafner, C. L. Cheung, and C. M. Lieber, Nature 共London兲 398, 761 共1999兲. 7 J. H. Hafner, C. L. Cheung, and C. M. Lieber, J. Am. Chem. Soc. 121, 9750 共1999兲. 8 C. L. Cheung, J. H. Hafner, and C. M. Lieber, Proc. Natl. Acad. Sci. USA 97, 3809 共2000兲. 9 C. L. Cheung, J. H. Hafner, T. W. Odom, K. Kim, and C. M. Lieber, Appl. Phys. Lett. 76, 3136 共2000兲. 10 H. Nishijima, S. Kamo, S. Akita, Y. Nakayama, K. I. Hohmura, S. H. Yoshimura, and K. Takeyasu, Appl. Phys. Lett. 74, 4061 共1999兲. 11 Y. Nakayama, H. Nishijima, S. Akita, K. I. Hohmura, S. H. Yoshimura, and K. Yakeyasu, J. Vac. Sci. Technol. B 18, 661 共2000兲. 12 R. M. D. Stevens, N. A. Frederick, B. L. Smith, D. E. Morse, G. D. Stucky, and P. K. Hansma, Nanotechnology 11, 1 共2000兲. 13 A. Cassell, N. R. Franklin, T. W. Tombler, E. M. Chan, J. Han, and H. Dai, J. Am. Chem. Soc. 121, 7975 共1999兲.
Downloaded 14 Apr 2003 to 143.232.211.49. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp
