Nanorobotic spot welder

Automatic deposition of copper from nanotubes with attogram precision enables the joining of building blocks for nanoelectronic circuits and nanoelectromechanical systems.
23 May 2007
Lixin Dong and Bradley J. Nelson

As technology progresses toward nanometer-scale systems, one barrier yet to be adequately addressed is the interconnection of miniscule as-synthesized or as-fabricated building blocks into complex structures. Carbon nanotubes (CNTs), nanowires, nanobelts, nanohelixes, and a variety of other nanomaterials and structures present fascinating opportunities. Examples include computing at terahertz frequency, measuring mass in attogram quantities, and sensing forces at femtonewton scales. But to use nanostructures in making electronic circuits and nanoelectromechanical systems (NEMS), reliable ways of soldering them onto electrodes and joining them together—ideally via electrically conducting junctions—must be developed.

Van der Waals forces,1 electron-beam-induced deposition,2 focused-ion-beam chemical vapor deposition,3 and high-intensity electron-beam welding4 are experimentally demonstrated interconnection strategies. All have limitations. Van der Waals forces are generally very weak, and the other methods involve high-energy electron or ion beams. Here we describe a new approach that we call nanorobotic spot welding (see Figure 1) in which copper-filled CNTs5 act as spot welders.1

Figure 1. Schematic illustration of a nanorobotic spot welder. CNT: Carbon nanotube. NW: Nanowire.

Ultra-small mass delivery of encapsulated copper from carbon shells was used to fuse nanoparts together with a nanorobotic manipulator for positioning. In our experiment, the manipulator was installed inside a transmission electron microscope. The CNTs were grown from copper catalysts5 with up to 5μm length and 40–80nm outer diameters. The copper cores were encapsulated in 4–6nm-thick walls.

CNT bundles were attached to a 0.35mm-thick gold wire using silver paint, and the wire was fixed in a specimen holder. The probe was an etched tungsten wire with a tip radius of (∼)100nm. Physical contact was made between the probe and the tip of a nanotube. Applying a voltage between the probe and the sample holder establishes an electrical circuit through a CNT and injects thermal energy into the system via joule heating. By increasing the applied voltage, the local temperature can be increased past the melting point of the copper. In our demonstration, melting occurred at 1.5V. When the bias was increased to 2.5V, the copper core began flowing to the tip of the CNT (see Figure 2).

Figure 2. Transmission electron microscope images from video frames showing the delivery process. The copper core started to flow as the bias reached 2.5V. The shortening rate of the copper core was determined to be 11.6nm/s.

According to the core length change, we calculated the mass flow rate as 120 attograms per second, an extremely slow rate allowing for very precise control of the amount of material delivered from the tube. Due to the polarity dependence, we believe that the copper is moved by electromigration in the presence of the electric field. The application of such controlled transportation of the copper core was then investigated. A copper-filled CNT was attached to a probe and brought into contact with another tube. Melting ensued, and the two were soldered together.

Nanorobotic spot welding has several attractions. A very low current can induce the melting and drive the flow. The welding site can be readily selected using nanorobotic manipulation. Melting occurs rapidly. Time-based control allows the precise delivery of attogram mass. Finally, copper has good compatibility with conventional semiconductor industry processes, and carbon shells protect copper against oxidation. A next step will be to show that deposition can be used for drawing wires and prototyping three-dimensional nanostructures.

Bradley J. Nelson, Lixin Dong
Swiss Federal Institute of Technology (ETH)
Zurich, Switzerland

Bradley J. Nelson is professor of robotics and intelligent systems at ETH Zurich and director of the Institute of Robotics and Intelligent Systems (IRIS). He is currently head of the Department of Mechanical and Process Engineering.

Lixin Dong is a senior research scientist at IRIS, ETH Zurich.

Recent News