The ability that some molecules have to arrange themselves into patterns on surfaces is crucial for many technological applications. However, it has been virtually impossible to predict or control the result of such self-organization. Now, a group of researchers led by Bianca Hermann of the Center for Nanoscience at Ludwig Maximilian University of Munich, reports it has been able to formulate a simple model that can predict the patterns observed. “With the help of the model, we can generate a wide variety of patterns that reproduce surprisingly well the arrangements observed experimentally,” Hermann said. “We want to extend this approach to other surface symmetries. Already now, the areas of molecular electronics, sensor applications, surface catalysis and organic photovoltaics can profit from our model. Its ability to predict structures formed by self-organization allows optimization of molecular building blocks prior to synthesis.” Scanning tunneling microscopy (STM) image displays a pattern reminiscent of tire tracks. The high-resolution measurement on the left is faded into the corresponding molecular mechanics (MM) energy minimizations on the right. These MM simulations provide the interaction strength and main geometrical considerations as input for an interaction site model. Monte Carlo simulations of this coarse-grained model predict four patterns with striking resemblance to the measurements in long range and local ordering. (Image: Copyright B.A. Hermann, WMI, Garching) Left to themselves, some molecules can self-organize into complex structures – a first step in the formation of membranes, cells and other structures. The principle of self-organization, which allows very economical use of resources, also is exploited in the production of functionalized surfaces required in molecular electronics, sensors, catalysis and photovoltaic components. The idea of the manufacturing process is that molecular components are brought into contact with a substrate material, and then “magically” find their preferred positions in the desired molecular network. The starting components are selected to display specific structural and chemical features intended for the envisaged application. However, the optimization of the molecular adlayers depends largely on a trial-and-error approach, and is therefore complicated and time-consuming. To develop a new molecular-interaction site model, Herrmann’s group collaborated with Thomas Franosch and Erwin Frey within the Cluster of Excellence “Nanosystems Initiative Munich” (NIM). The problem was tackled using an approach from statistical physics known as the Monte Carlo method, which allows one to conduct a detailed computer simulation on the statistics of molecular interactions. The structural motifs so generated were compared with experimental high-resolution images of molecular patterns obtained by scanning tunneling microscopy (STM). Marta Balbás Gambra, a doctoral student, began each simulation with a mathematical representation of a collection of hundreds of randomly oriented particles of defined conformation. These schematic molecules were then perturbed by – computationally – adding energy, causing the population to adopt a new configuration. This schematic illustrates the data simulations created by the Munich-based researchers. (Illustration: Copyright B.A. Hermann, WMI, Garching) Using this simulation strategy, one can generate a greater variety of patterns than are found naturally, and many of these corresponded closely to the real molecular patterns revealed by STM. “In one case we actually predicted a pattern that was only later verified with STM,” said doctoral student Carsten Rohr. According to the laws of thermodynamics, physical systems tend to adopt the state with the most favorable (i.e., lowest) energy. Experimental tests showed that different molecular configurations interconvert until an arrangement predominates that is reminiscent of tire tracks. And, indeed, the Monte Carlo approach predicted that this arrangement corresponds to the state with the lowest energy. “In the end, we were able to show that the molecular geometry and a few salient features encode the structural motifs observed,” Franosch said. According to Hermann, the group plans to extend the approach to other types of surface symmetries. “The model already provides an important theoretical tool, because it helps us to forecast the type of surface pattern that a given functional molecule will form,” he said. “This means that the design of molecules can be optimized during the synthetic phase, so as to obtain surfaces with the desired characteristics.” For more information, visit: www.wmi.badw-muenchen.de/spm