Researchers at Sandia National Laboratories here have found an important clue to how metal films make a transition from individual droplets to organized film structures that could lead to a lithographically controlled means for creating quantum-dot structures.

Meanwhile, a team of researchers from the University of California at Santa Barbara, Tulane University (New Orleans) and Texas A&M University (College Station, Texas) have found a means of creating defect-free, self-organized structures using the Langmuir-Blodgett technique, one of the oldest means for creating self-organized molecular structures.

And, physicists at the Fritz-Haber-Institut (Berlin) have hit on a new observational technique that can view surface catalytic reactions at both the atomic and the micron scale — two vastly different length scales. Their work has revealed that self-organized systems are much more complex than previously thought.

One way of finding a replacement for semiconductor processes based on photolithography, which are not far from their physical limits, would be to enlist the parallel, self-organizing capabilities of chemical reactions. That is the thrust of the Sandia work. The lab's results suggest that it might be possible to "pre-format" wafers with nanometer-scale patterns that enhance specific patterns, creating an enabling technology for quantum-dot computers.

The small-scale patterning technique could also be applied to various microelectromechanical systems (MEMS) such as on-chip optics using photonic lattices. The newly observed processes could also lead to methods of building circuits with the atomic precision only achievable with the one-by-one atomic probe techniques used currently.

Serendipity

"There was some serendipity in making this discovery, [because] we were actually trying to understand the wetting and spreading properties of lead on copper for soldering and brazing applications," said Sandia physicist Norm Bartelt. "Our original goal was to get a more microscopic view of what is going on during wetting and spreading of solder, but this just jumped out at us." Also on the project were Sandia researchers Richard Plass and Sandia project leader Gary Kellogg.

What the Sandia researchers demonstrated for the first time was that current theories regarding the atomic-scale stages materials go through — whether for soldering or fabricating wafers — are reproducible and highly predictable.

In a nutshell, when a new material is deposited on a substrate to form a new layer, it can be made to undergo three intermediary stages, each of which harbors useful nanostructures, were the process to be interrupted. In particular, the first stage is "droplets," the second stage is "stripes" and the third is inverse droplets or "pits."

"We were able to observe the self-assembly and compare how closely it follows the predicted evolution of patterns that begins with droplets — that is a hexagonal arrangement of circular islands — into a striped pattern, and then finally into an inverse droplet pattern which is again a hexagonal arrangement of circular islands, but the materials are reversed," said Bartelt.

The group's observations, made with a low-energy electron microscope, were of the deposition of lead onto a copper substrate. They recorded real-time images using a low-energy electron microscope, enabling them to see exactly how the nanostructures are self-assembling and how they transform as a function of temperature and the amount of added material. The group is actively profiling the key interatomic force parameters involved in the process in the hopes of extending current theory.

Since 1991, theorists have hypothesized that competing interatomic interactions of opposing attractive and repulsive forces could be harnessed in the service of automatically forming ordered patterns on the nanometer scale — patterns that could be useful as templates for atom-sized devices. (One nanometer is the width of about four silicon atoms). These are the kinds of device feature sizes needed for quantum dots and certain optical devices like photonic lattices. For instance, a field of quantum dots, equally spaced every few nanometers on a wafer, could then have the rest of the chip fabricated on top of them, possibly with 3-D MEMS devices on-chip too.

The experimental verification of this self-assembling process went as predicted to a point, then diverged. First the group saw the predicted field of bumps, each representing a different droplet of lead. As more lead was added, the droplets increased in size until their coverage of the substrate amounted to about 30 percent, all according to theory. At that point the droplets started joining up into a distinctive striped pattern resembling herringbone. As lead was added, eventually the stripes became more and more bloated until they broke into an inverse droplet pattern of pits.

For now the group is looking at all the physical factors that influence the process, such as temperature and pressure, as well as for ways to fine-tune the self-assembly process.

Researchers at the Fritz-Haber Institut are tackling the problem of creating more realistic models of self-organizing reactions. The group used a simple physical system consisting of hydrogen oxidation on a platinum surface. The traditional method of modeling this type of reaction is to simplify the analysis using continuous variables such as the concentration of reactants. Modeling on the atomic scale would be more accurate, but at the same time is computationally intractable.

The platinum substrates were first coated uniformly with oxygen atoms, which attach themselves to the surface. When hydrogen is introduced, a reaction front moves across the wafer as the platinum operates as a catalyst for the formation of water (H₂O) molecules. After passing across the wafer, the reaction front leaves self-organized patterns of water molecules behind it. Conventional reaction equations give only a rough approximation of the resulting patterns.

Imaging reaction

The research group was able to image the reaction at both the nanometer atomic scale and at the mesoscopic (micron-length) scale using a scanning-tunneling microscope. The result was the discovery of an intermediate reaction at the atomic level that is completely left out of the conventional model. The researchers plan to use the new data to create a different approach to catalytically induced, self-organized structures. The new computational model will include Monte-Carlo simulations of atomic behavior along with finite-element modeling on the micron scale.

Back in the United States, the UC Santa Barbara researchers and their colleagues believe they have found a way around the most serious problem posed by the Langmuir-Blodgett technique. While this technique is highly versatile and takes place at room temperature, the resulting self-organized structures typically have large defects, and areas of the pattern sometimes spontaneously reorganize into different structures.

The process begins by forming a single molecular layer on water using various fatty acids, of which there are a great variety. A substrate is passed vertically through the air-water interface where the film floats. If the process is done right, the single-molecular film ends up deposited on the substrate.

The new work suggests that absolutely perfect films can be created by simply increasing the pH of the monolayer. The higher pH has the effect of increasing the concentration of the molecules so that they have less freedom of movement during the transfer to the substrate.

Using atomic-force microscopy, the researchers observed that in the liquid phase, the films had a highly regular herringbone pattern and when transferred to a substrate, they reorganized themselves into a highly regular hexagonal pattern free of defects. However, the hexagonal pattern was still far from any equilibrium state, and further research is needed to determine if the method can be applied to other molecular types used in the Langmuir-Blodgett technique.

An audio recording of reporter R. Colin Johnson's full interview on self-organizing systems can be found online at AmpCast.com/RColinJohnson.

Chappell Brown contributed to this report.