'Growing' Microchips from Proteins
03/01/2004
BY PIERRE DEYMIER
In recent years, the exponential growth in semiconductor technology has been sustained by extending the capabilities of top-down manufacturing processes to smaller and smaller dimensions. Unfortunately, the costs of these top-down approaches are becoming prohibitive at sizes and tolerances in the nanometer size range. In response, a new paradigm has arisen for the mass replication of nanoscale electronic circuits based on a bottom-up, molecular engineering approach that promises to be cheaper, more flexible and efficient. In particular, control of device interconnections emerges as one of the major challenges in the development of these bottom-up approaches. Research suggests that proteins and assemblies of proteins offer the control necessary for inexpensive and reliable fabrication of nanoscale interconnects. Proteins and other biomolecules are complex molecular entities with many potentially useful materials properties for non-biological applications. Furthermore, the ability to design and fabricate proteins with structure and functionality not found in nature makes protein-based materials science and engineering a compelling field of study with fabulous long-term potential in a broad range of applications.
 Figure 1. This SEM photo of Ni-plated microtubules shows the metal coating, which is ~10 nm thick, yielding nanowires with an overall diameter of ~40-50 nm. |
Microtubules are self-assembling, dynamic, tubular-shaped biopolymers with nanometer-size diameters and large aspect ratios, that can be used as templates for fabricating nanoscale interconnects, interconnect arrays and networks. Microtubule ends are polarized such that each end exhibits unique and specific biochemical moieties. A microtubule is composed of a fast-polymerizing (+) end, and a slow-growing (-) end. The polar microtubules can be oriented between targets on a microchip through chemical anchors or linkers that specifically recognize either end of the microtubule. In this manner, an excess of microtubules will be grown from a starting site on a microchip to which biological nucleating centers (γ-tubulin protein) are attached. The growth is polarized — as a microtubule grows it exposes its (+) end with its (-) end bound to the nucleating site. Molecular recognition is used to bind the growing (+) end to a target site functionalized with a (+) end-specific linker. Linkers are being designed that also act to stabilize microtubules once attached. Many microtubules will not reach the second target site on the microchip. These microtubules, unattached to a target site through their (+) end, are readily depolymerized by altering the chemistry of the solution, leaving only the attached and oriented interconnections. Using a library of biochemical linkers, interconnections between several nucleating and target sites with different orientations can be assembled simultaneously.
With the stabilized interconnects in place, the microtubules serve as a template for subsequent outer or inner metallization by non-magnetic, low electrical conductivity metals and packaging in a mechanically strong and chemically inert protective layer. Metallization of microtubules with Ni via electroless plating is a well-known (but not completely understood) process resulting in 10-nm-thick Ni-based films on the outer surface of microtubules. However, presently, copper or gold is the metal of choice for interconnects in the microelectronics industry. Here, the challenges of coating microtubules with Cu are associated with the harshness of the environment required for electroless plating with Cu. A basic requirement for an interconnect is that the DC voltage drop caused by the interconnect resistance should not exceed a small fraction (1 percent) of the voltage swing. Calculations for microtubule interconnect structures metallized on a.) both the inside (with estimated metal diameter of 10 nm) and outside (with estimated coating thickness of 10 nm) of the microtubule, or b.) only the outside of the microtubule (with estimated coating thickness of 10 nm) yield resistance per 1 µm of length listed in Table 1.
 Table 1. Resistances of metallized microtubules. |
Consequently, the resistances of interconnects built using microtubules as templates are adequate for use with nano-scale circuits. In some considered applications, we would have interconnects of 1 to 10 µm in length involving devices with a voltage swing of 1 V and a current on the order of µA, causing a 50 to 500 µV voltage drop.
In summary, the microtubule is a polarized, dynamic biopolymer with each end exhibiting unique and specific biochemical moieties. Based on these properties, a fabrication process is being developed in which interconnects between multiple, varied devices (both micro and nano) and interfaces are built, in situ, through directed and controlled polymerization of microtubules. Once in place, the assembled microtubules will be coated and/or metallized for increased electrical conductivity by derivatizing the microtubule surfaces. This work represents a new science of nanomaterials, in which the structure/processing/properties/utilization paradigm is based on the specificity of biomolecular precursors and biopolymer dynamics.
PIERRE DEYMIER, professor of Materials Science and Engineering, may be contacted at the University of Arizona, Department of Materials Science and Engineering, Mines Bldg., Room 125 E, Tucson, AZ 85721; (520) 621-6080; e-mail: [email protected].