Absolute and arbitrary orientation of single-molecule shapes

Introduction: Molecular and particulate nanodevices such as carbon nanotubes and semiconductor nanowires exhibit properties that are difficult to achieve with conventional silicon microfabrication. Unfortunately, most such devices must be synthesized or processed in solution. To combine nanodevices into larger circuits, or simply to connect them with the macroscopic world, scientists use a range of directed self-assembly techniques to deposit them at specific locations on microfabricated chips. Many such methods work well with spherical devices for which orientation is irrelevant. For linear wire-like devices, flow or field alignment works for applications involving a single global orientation. However, a general solution for multiple orientations or less symmetric devices (e.g., diodes or transistors) has remained elusive. Rationale: Single-molecule DNA origami shapes can simultaneously act as templates to create nanodevices and as adaptors to integrate them onto chips. With 200 attachment sites just 5 nm apart, origami can organize any material that can be linked to DNA; for example, carbon nanotube crosses have been templated to yield field-effect transistors. With ~100-nm outlines, origami are large enough that shape-matched binding sites can be written at arbitrary positions on chips using electron-beam lithography. Our prior work used equilateral triangles that stuck to binding sites in six degenerate orientations. Here, we asked whether origami shapes could provide both absolute orientation (to uniquely orient asymmetric devices) and arbitrary orientation (to independently orient each device). Success depended on finding a suitably asymmetric shape. Results: To break up-down symmetry and to ensure that each shape was deposited right-side up, we added adhesion-decreasing single-stranded DNAs to one side of each origami. The binding of asymmetric right triangles to shape-matched sites gave orientation distributions consistent with strong kinetic trapping, as predicted by the volumes of basins of attraction around local minima. This motivated the design of a "small moon" shape whose energy landscape has a single minimum. Fluorescent molecular dipoles fixed to small moons served as model nanodevices and allowed us to measure variability in orientation (±3.2°) by polarization microscopy. Large-scale integration was demonstrated by an array of 3456 small moons in 12 orientations, which we used as a fluorescence polarimeter to indicate excitation polarization. The utility of orientation for optimizing device performance was shown by aligning fluorescent dipoles within microfabricated optical cavities, which showed a factor of 4.5 increase in emission. Conclusion: Control over optical dipole orientation may enable metal nanorod metasurfaces at visible wavelengths, optimized coupling of emitters to nanoantennas, lumped nanocircuits, and coherence effects between small numbers of emitters. Still, these applications and the devices we present do not demonstrate the full power of the small moons: Dipolar devices can rotate 180° and still function. Completely asymmetric nanodevices requiring absolute orientation (e.g., molecular bipolar junction transistors) have yet to be developed; now that orientation can be controlled, there is motivation to invent them. In the meantime, the wiring of existing devices into circuits may be greatly simplified.