Room Temperature Local Synthesis of Carbon Nanotubes

Dane Christensen, Ongi Englander, Jongbaeg Kim, and Liwei Lin

Berkeley Sensor and Actuator Center University of California at Berkeley

Berkeley, CA 94720

Abstract – We report the synthesis of carbon nanotubes (CNT) by localized resistive heating of a MEMS structure in a room temperature chamber. This is the first known vapor-deposition CNT growth method that does not require globally elevated temperatures. The localized, selective, and scalable process is compatible with on-chip microelectronics and removes necessity of post-synthesis assembly of nanostructures to form integrated devices. Synthesized nanotube dimensions are 5-50 nm in diameter and up to 7 µµµµm in length. Growth rates of up to 0.25 µµµµm/min were observed. This accomplishment makes possible the direct integration of CNT devices with on-chip transduction, readout, processing, and communications circuitry, facilitating integration of nanotechnology with larger-scale systems.

I. INTRODUCTION

Carbon Nanotubes are attractive for many applications due to their unique electrical, chemical, and mechanical properties [1-4]. Current synthesis processes limit their use due to the high furnace temperatures required for growth and the difficulty in handling and maneuvering the nanostructures to integrate with microelectronics to form more complex devices [3-6]. These obstacles have led to creative methods for maneuvering and electrically connecting nanotubes [6-7]. but the cost and time required for these methods restricts their use.

We present an approach that allows for the synthesis, in a room-temperature chamber, of carbon nanotubes at a pre-specified location while eliminating the requirement of later assembly processes. The method is localized, selective, and capable for the direct integration of nanotubes with larger scale system such as foundry-based microelectronics. Resistive heating causes a suspended microstructure to achieve the thermal requirement for vapor-deposition synthesis, while the remainder of the chip and substrate remains at room temperature. This approach is shown to yield carbon nanotubes, and is extensible to other vapor-deposition-synthesized nanomaterials.

II. MICROFABRICATION

Two types of suspended Microelectromechanical (MEMS) structures were fabricated to serve as localized micro resistive heaters for the synthesis process: polysilicon microstructures fabricated using a standard surface micromachining process [8] and bulk-etched single crystal

silicon (SCS) microstructures based on a silicon on insulator (SOI) wafer [9] as seen in Figure 1(a-c). In both cases the microstructures were heavily doped and suspended 2 µm, as defined by the sacrificial silicon dioxide layer, above a silicon substrate for electrical and thermal isolation. The typical thickness of the bridges was 2 µm for polysilicon microstructures and 40 µm for SCS microstructures. The wet chemical release etching process naturally created recessed regions underneath the device such that a maskless catalyst deposition process could not cause an electrical short-circuit. A thin metal layer was then evaporated over the entire chip to serve as a catalyst for the growth as in Figure 1(d). Approximately 5 nm of pure nickel, pure iron, or Ni Fe (80% – 20% by weight) were used, with growth rates observed to be greater when iron is present. After wirebonding for electrical connectivity, the microstructures were placed into a room temperature vacuum chamber. The organic vapor phase, acetylene (C2H2), was introduced at

Figure 1: Synthesis Process. (a) initial 3-layer wafer composed of silicon over an oxide sacrificial layer. (b) etched microstructural

layer. (c) timed (wet) oxide etch. (d) directional catalyst evaporation. (e) wirebonding and actuation in acetylene ambient. (f) synthesized

nanostructures. (g) vacuum chamber schematic.

0-7803-7977-2/03/$17.00 (C) IEEE

245 mTorr and the microstructure was heated to initiate nanotube synthesis as shown in Figure 1(e).

III. NANOTUBE SYNTHESIS

In practice, by monitoring the current while applying and steadily incrementing voltage, localized resistive heating of the suspended bridge can be controlled. The temperature on the microstructures is assessed based on the geometry, doping level, and current-voltage characteristics [10]. We have found in practice that growth tends to be fastest in the regions where the microstructure is glowing lightly in the optical spectrum. When the temperature gets too high, as evidenced by a very strong glow, the rate of growth diminishes greatly. An experimental current–voltage relationship is shown in Figure 2(k). This curve is characteristically linear under low input power and becomes non-linear as input power is elevated due to local high temperatures and secondary effects [11]. Thermal modeling shows that the temperature profile [12,13] should be nearly parabolic, as seen in Figure 2(k, inset). The chemical vapor deposition synthesis process is understood to be composed of a series of chemical reactions; all depend strongly on the local temperature [14-16]. The thin catalyst layer breaks down into discrete nanoparticles, as seen in Figure 3; at the catalyst surface the acetylene vapor decomposes into carbon and hydrogen gas,

subsequently forming a liquid catalyst-carbon alloy; during the synthesis process, the alloy continues to absorb carbon until it becomes sufficiently saturated; at the liquid-solid interfaces carbon then precipitates from the alloy and self-organizes in the form of nanotubes. The dependence on temperature of each of these reactions implies that a minimum temperature that must be reached before synthesis is activated. Since the temperature profile on the microstructure spans a room- to high-temperature range (25-1000˚C or higher), it is possible to examine the temperature dependence of the vapor deposition synthesis process using microstructure heating. By rapidly heating the microstructure prior to growth, the catalyst can be caused to conglomerate in certain regions, while nanodot morphology is restricted to a narrow region. Catalyst nanodots are desired for single-walled CNT growth, and the geometry of CNTs is directly linked to the size and shape of the catalyst nanoparticle [5]. Thus, by localizing the nanodot region, synthesis parameters associated with desirable CNTs will be localized accordingly. By choosing the catalyst appropriately, diffusion and alloying with the Silicon will not alter the electrical properties of the microstructure by catalyst overannealing.

IV. RESULTS

Experimentally, nanotube growth occurred at a rate of approximately 0.25 µm/min in the optimal-temperature regions. CNTs have been synthesized up to 7 µm in length, and 5-50 nm in diameter. The diameters of CNTs are very uniform across each growth region. Further investigation is ongoing to analyze the dependence of growth parameters on

Figure 2: Sample microstructure I-V curve. (a-j) Microphotos of a

150 µm-long microstructure during actuation. Photos correspond to points on the graph. (k) I-V curve for an actuated microstructure.

This microstructure is similar to the device shown in Figure 5. Inset: Simulated temperature profile of a 150 µm-long

microstructure during actuation.

Figure 3: Thin-film catalyst behavior under resistive heating of underlying microstructure. Rapid overannealing before growth causes the thin film to conglomerate, undesirable

for CNT growth.

0-7803-7977-2/03/$17.00 (C) IEEE

gas pressure and composition, catalyst composition and preparation, and local electrical field. Figure 4(a) shows CNT growth across the central portion of a “pointed” microstructure. These nanotubes resulted from a 10-minute synthesis period, and are approximately 0.5-2 µm long in the region shown. Note that there are no CNTs on the substrate, though catalyst was deposited on that surface. Closer SEM examination shows that the catalyst under and near the bridge has received insufficient heating to melt it or alter its morphology in any visible way. This indicates that the substrate was able to dissipate the input radiative and conductive thermal energy without experiencing a significant temperature elevation, which is significant if the substrate contains microelectronics from prior processing. Work is ongoing to examine the actual local temperature profile resulting from this actuation method. Figure 4(b) shows CNTs near the center of this actuated region. Since there is no external force to orient or control the growth parameters, they appear to grow imperfectly. These CNTs are estimated to be 15 nm in diameter. Figure 5 shows CNT growth localized mainly to the legs of a “u-shaped” microstructure and Figure 2(a-j) gives a series of optical photographs of a similar microstructure when electrically actuated. There is a distinct temperature range at which carbon nanotubes will grow [5], and this can be used to select the synthesis region, between an area that is too cold (< 700°C for single-walled CNT) and one that is too hot (> 1000°C). This series shows that the growth region (barely glowing) can be isolated to a desired location on the microstructure. Each “hot spot” is stationary and stable for sufficient time to grow nanotubes of at least 5 µm. Figure 2(j) illustrates the microstructure just before failure due to melting (~1400°C). Figure 4(a) displays nanotube growth localized to the center portion of a

“pointed” microstructure and Figure 4(b) is an enlarged view of nanotubes at the center portion of this microstructure. Figure 2(k) gives the electrical characteristics typical to the resistive heating of a suspended microstructure. The marked data points correspond to the pictures in Figure 2(a-j). The growth shown in Figure 5(a) occurred on a microstructure powered between data points f and h in Figure 4(k). The nanotubes shown in these figures are 10-50 nm in diameter and up to 5 µm in length. It is believed that these nanotubes are multi-walled due to their size and the structural integrity visible in SEM microphotos. Carbon nanotube growth response to an applied electric field is well documented [17]; we have also demonstrated the orientation of nanotubes as a function of the applied electric field, shown in Figure 5. The bottom right side of this “u-shaped” microstructure was taken to approximately 7 V, while the bottom left was grounded, yielding an electric field of approximately 0.12 V/µm where the longest growth occurred. It is seen that the growth direction is fairly well aligned to the electric field. The inset oblique view shows the CNT from outside the microstructure curving with the electric field. This electric field is weaker than desirable for oriented CNT growth [17] but gives evidence that the electric field due to the synthesis process alone can be used to orient the tubes. The addition of additional electric field, by increasing the potential difference between nearby microstructures, will enhance this effect. The microstructures’ maximum temperature is observed to be located not at the center, but skewed off-center by as much as 20% of the microstructure’s length. The skew was always in the direction of the cathode for both polysilicon and SOI microstructures and was most pronounced in the “U-shaped” devices. This off-center heating is presumed to be due to the Thomson Effect [18] whereby heat is evolved and absorbed in different regions of a conducting element due to the electrical current. When

Figure 5: Nanotube Synthesis. (a) Synthesis localized to microstruture

“legs.” Growth occurs largely in the direction of local E-field. (b) High-resolution SEM of the right leg section (oblique view).

Note the CNTs curving to follow the local E-field.

Figure 4: (a) CNTs localized to the center of a “pointed” microstructure. These tubes are the result of a 10-minute growth, and are approximately

0.5-2 µm long and 15 nm in diameter. (b) High-Resolution SEM of CNT growth near the center of the microstructure.

0-7803-7977-2/03/$17.00 (C) IEEE

designing microstructures for localized synthesis, this should be taken into account. In addition, if the microstructures are insufficiently doped, the I-V characteristic is altered drastically and temperatures are much less predictable. This can be attributed to secondary effects such as electromigration, grain growth, and localized melting [18]. Highly doped microstructures were significantly easier to drive to a repeatable temperature profile and made it possible to maintain the desired profile for the duration of the synthesis.

V. CONCLUSIONS

The synthesis of nanotubes in direct contact with MEMS structures has been demonstrated in a room temperature environment. The method has yielded localized regions of carbon nanotubes, 5-50 nm in diameter and up to 7 µm in length. Growth rates of as high as 0.25 µm/min were observed. This process allows for the direct integration of nanostructures into larger scale systems and permits the placement of these nanostructures at desired locations along the surface of the larger scale system using localized heating of a highly doped microstructure. In addition to eliminating the need for post-synthesis CNT assembly, this process provides a technique for integration of nanotechnology that is fully microelectronics-compatible.

ACKNOWLEDGMENT

The authors wish to thank Mu Chiao for help with microfabrication, Ron Wilson for SEM assistance, and Pepe Kaksonen, Bill Flounders, Bob Hamilton and Katalin Voros of the UC Berkeley Microlab for their guidance. These devices were fabricated in the UC Berkeley Microlab.

REFERENCES

[1] J.H. Hafner, C-L. Cheung, A.T. Woolley, and C.M. Lieber, “Structural and functional imaging with carbon nanotube AFM probes,” Progress in Biophys. & Mol. Bio., vol. 77, pp. 73-110, 2001.

[2] S.J. Tans, A.R.M. Verschueren, and C. Dekker, “Room-Temperature Transistor Based on a Single Carbon Nanotube,” Nature, vol. 393, pp. 49-51, 1998.

[3] M. Dresselhaus, G. Dresselhaus, and P. Avouris (eds.) Carbon Nanotubes - Synthesis, Structure, Properties & Application, Springer-Verlag, New York (2001).

[4] D. Appell, “Nanotechnology – Wired for Success,” Nature, vol. 419, pp. 553-555, 2002.

[5] Y. Huang, X. Duan, Q. Wei, and C. M. Lieber, “Directed Assembly of One Dimensional Nanostructures into Functional Networks,” Science, vol. 291, pp. 630-633, 2001.

[6] P. Avouris, T. Hertel, R. Martel, T. Schmidt, H.R. Shea, and R.E. Walkup, "Carbon Nanotubes: Nanomechamics, Manipulation, and Electronic Devices," Appl. Surf. Sci., vol. 141, pp. 201-210, 1999.

[7] Y.Y. Wei and G. Eres, “Directed Assembly of Carbon Nanotube Electronic Circuits,” Appl. Phys. Lett., vol. 76, pp. 3759-3761, 2000.

[8] J.M. Bustillo, R.T. Howe, and R.S. Muller, “Surface Micromachining for Microelectromechanical Systems,” Proc. of the IEEE, vol. 86, pp. 1552-1573, 1998.

[9] G.T.A. Kovacs, N.I. Maluf, and K.E. Petersen, “Bulk Micromachining of Silicon,” Proc. of the IEEE, vol. 86, pp. 1536-1551, 1998.

[10] M. Chiao and L. Lin, "Electrothermal Responses of Lineshape Microstructures," Sensors and Actuators, vol. 55, pp. 35-41, 1996.

[11] H.A. Schafft, “Second Breakdown-A Comprehensive Review,” Proc. of the IEEE, vol. 55, pp. 1272-1299, 1991.

[12] M. Chiao and L. Lin, "Self-Buckling of Micromachined Beams Under Resistive Heating," J. of MEMS, vol. 9, pp. 146-151, 2000.

[13] L. Lin, A.P. Pisano, and V.P. Carey, "Thermal Bubble Formations on Polysilicon Micro Resistors," Am. Soc. Mech. Eng., J. of Heat Transfer, vol. 120, pp. 735-742, 1998.

[14] S.B. Sinnott, R. Andrews, D. Quan, A.M. Rao, Z. Mao, E.C. Dickey and F. Derbyshire, “Model of Carbon Nanotube Growth Through Chemical Vapor Deposition,” Chem. Phys. Lett., vol. 315, pp. 25-39, 1999.

[15] M. J. Yacaman, M. M. Yoshida, L. Rendon, J.G. Santiesteban, “Catalytic Growth of Carbon Microtubules with Fullerene Structure,” Appl. Phys. Let., vol. 62(6), pp. 657, 1993.

[16] C.J. Lee, J. Park, Y. Huh, and J.Y. Lee, “Temperature Effect on the Growth of Carbon Nanotubes Using Thermal Chemical Vapor Deposition,” Chem. Phys. Lett., vol. 343. pp. 33-38, 2001.

[17] Y. Zhang, A. Chang, J. Cao, Q. Wang, W. Kim, Y. Li, N. Morris, E. Yenilmez, J. Kong, and H. Dai, “Electric-Field-Directed Growth of Aligned Single-Walled Carbon Nanotubes,” Appl. Phys. Lett., vol. 79, pp. 3155-3157, 2001.

[18] J. Tauc, Photo- and Thermoelectric Effects in Semiconductors. Pergamon Press, New York, (1962).

0-7803-7977-2/03/$17.00 (C) IEEE