Using principles of mother nature in high-tech industrial processes

An example of a biological self-assembly system is the DNA double helix, and in nature the crystal grids are of the same principle. Organisms consist of cooperating cells forming live tissue. All these structures have in common that they don’t depend on covalent chemical binding, but on weak local forces in a thermodynamic balance. Thus, they can be vulnerable to mechanical or thermal impact, but they can also be self-correcting or self-healing. They are formed through implicit programmatic instructions in every particle or cell, such as surface tension or molecular docking.

The application of self assembly as an industrial tool has a wide scope and markets waiting for it – for all the things that are reaching down to the nano-scale: information processing, material science, environmental control, medical and pharmaceutical techniques. Also very important are semiconductor devices and displays, as well as future circuit assembly techniques. There is intense research activity for self-assembly processes centered at several large universities (most prominently University of Berkeley in the USA), and also in Europe. The EU’s Sixth Framework Program has attempted to fund promising work. Part of it is SANDiE (self-assembled semiconductor nanostructures for new devices in photonics and electronics) and the German DFG Research Focus Area 522 (architecture of nano and micro-dimensional structural elements). A report sponsored by the German Ministry of Education and Research (BMBF) „Kontrollierte Selbstorganisation für zukünftige technische Anwendungen“S evaluates, among other topics, the various international approaches to the goal of mounting electronic and optical systems by self-assembly processes.

As the dimensions of semiconductor and micro-mechanical components are reaching down into the nano-scale, there is concern that conventional assembly and contacting procedures based on mechanical assembly systems and soldering techniques will run into a severe barrier – cost wise and technologically. No wonder that there are some alarmist voices rising – coming, at this point, from the academic research community. Researchers such as D.O. Popa (Micro and Meso Scale Assembly, SME Manufacturing Journal, 2004), are already talking of an impending “packaging gap” or, even more ominous, an “assembly crisis” seen to occur in the decade after 2010 if the evolution according to Moore’s Law continues as prescribed. Already today, he says, the cost of assembly and packaging complex electronic systems is up at 60 to 90% of the total system manufacturing costs.

For this purpose, Popa has developed a classification of current assembly processes and their viability for future developments. Meso-scopic components, in his terminology, meaning components measuring above 1mm per side, are relatively simple to be positioned by current assembly machines. The drawback, becoming more and more visible, is that pick-and-place methodology is a serial procedure, mounting one component at a time. The dominant physical effects are gravitation and friction. If component size shrinks further, there will be part dimensions from millimeters down micrometers to be processed in the foreseeable future. Here, surface effects, electrostatics and Van-de-Waals forces must be used to handle the components.

But serial processing of such small devices yields, in Popa’s view, insufficient throughputs which are no longer commercially viable. Even more pronounced is this paradigm, when real nano-scale components have to be processed in high-volume assembly. Positioning them with mechanical tools, as precise as they might be, is not useful anymore. The dominant effects fashioning position and placement of such components will be the interaction of minuscule molecular forces. Here serial processes based on mechanical procedures will simply go out of business.

That will be the hour of parallel mounting procedures. One approach, promoted by A. Singh et al. (Batch Transfer of Microstructures Using Flip-Chip Solder Bonding, IEEE Journal of Micromechanical Systems, 1999), uses the transfer of a pre-structured wafer image of a complete system onto a substrate, so generating the complete circuit pattern in parallel, by means of “stamping” it on. This is comparable in its result to the idea of ink-jetting or printing the system onto the substrate. In a similar vein is the attempt to place a number of components of meso-scopic scale in parallel processing on a provisory substrate and after interconnecting them take them off this provisory substrate – which can be used over and over as a mounting stage.

Placing the components onto this stage can be achieved by a diffusion process, in a solution or propelled by inkjet printing, which would bring the components in proximity to their final position. An alternative way to force the components to their position would be by exposing them to appropriately directed acoustic-surface waves as is done by Adalytix. Micro-fluidics is another preliminary positioning process taking place at high speed, and achieves high throughput due to the high degree of parallelism. Other principles of propelling the components to their position are electrostatics and magnetism.

The final alignment and positioning close to the desired location, however, is a further complication requiring a sophisticated methodology. In principle, it would be achieved by the prevailing weak short-range forces – by biological recognition or chemical binding forces. This can be fashioned in a hydrophilic or a hydrophobic environment. The well-researched fluidic self-assembly (FSA) method pioneered and patented by Alien Technology Corp./California, is already applied to high-volume, high-speed, low-cost manufacturing of RFID tags. FSA combines preliminary and final positioning in one step. It washes all the necessary ICs suspended in a solution over the substrate and lets them settle into pre-configured troughs of fitting size and direction, as a representation of their lowest possible energy state. This selects (within limits) and positions the ICs (nano blocks) on the substrate. Actual throughput of this technology is 2m per hour. As a measure of success, Alien has so far attracted more than $200m of venture capital.

Another idea for positioning meso-scale components on a substrate is by minimizing the surface energy between devices and substrate-side solder bumps which are heated above the melting point. Selection and correction of false positioning is done by petite vibrating of the assembly, whereby wrongly placed components are shaken lose and then repositioned. This method affords no selectivity in terms of the component positions. H.O. Jacobs et al are developing the procedure at Harvard University. A related concept is pursued by Jacobs with Wei Zheng at Minnesota University, Minneapolis (Fabrication of Multicomponent Microsystems by Directed Three Dimensional Self Assembly).

Electrical fields as directional forces are experimented with within the EU’s SHOT (self assembled hybrid optoelectronic technologies) program and by M. Ozkan et al. at UC San Diego, California. Again, no selectivity being provided, but effective transport of components to their final positions. Magnetic fields are used at the Microsystems Technology Laboratories at Cambridge. Molecular recognition as highly selective force has been investigated at UC San Diego by S.C. Esener at al., and at the Fraunhofer IZM (project BioMount), and at Purdue University by H. McNally et al.

In summary: There is a host of different activities, out of academic curiosity and commercially purposed as well: reaching from GaAs-LEDs on Si-substrates to micro-optics and MEMS, which are demanding in terms of their assembly. That’s a focal point of the EU project SHOTS. Another activity is carried out at IBM: the generation of nano crystals on Si-substrates for ultra-compact flash memories, using self-organizing polymer matrices having grid-like openings of 20nm diameter. At Infineon there is work being done with bundled carbon nanotubes using them as vias for very high current densities of up to 10 (exponent) 6 A/cm² between the substrates of multi-layer devices. The possible combination of peptides and nanotubes is explored within the EU Project DNA-Based Electronics. Other configurations and concepts deal with nano-crystals from 50nm to 500nm in size, coated with nano particles. Equally promising, yet still remote from the impending “packaging gap,” is the generation of quantum dots by self-assembling principles.

Werner Schulz