Objective

Study of SPP interference as a platform to form various desired dynamic patterns with fine resolution.

Results

Since SPPs are essentially two dimensionally confined on a metal-dielectric interface, their interference patterns can be controlled by arranging different 2D geometries of the slits/edges. Figure 1a-d show simulated SPP interference patterns with four geometries. Obviously, not only periodic but also quasi-periodic and even more complicated 2D patterns can be realized. The lithography experimental results (see Figure 1e-h) clearly show the formation of different interference patterns predicted by simulations. A single collimated i-line beam from a mercury lamp (wavelength 365 nm) is used to generate all these patterns with exposure time typically less than 10 s, corresponding to a dose of around 18 mJ/cm2. Compared with the pattern formation by free laser beam interference, one obvious advantage of SPP interference is the higher resolution as we discussed above. In addition, SPP interference requires a much simpler setup. To achieve a complex laser interference pattern, multiple laser beams have to be very precisely directed and controlled by complicated optics. As for SPP interference in our case, all of the complicated optics can simply be replaced by slits/ edges with well-designed shapes and only one excitation beam is required.

For a structure with fixed shape, the interference pattern can still be dynamically adjusted by the polarization and incident angle of the excitation light beam. The SPPs can only be efficiently excited when the incident light has a polarization perpendicular to the slit. Therefore, the strength of the SPPs at different portions of the slit can be arbitrarily tuned by adjusting the polarization of the excitation light beam. For example, in Figure 2 the fluorescent images illustrate how the SPPs propagate along the surface for different geometries. The excitation light beam is normally illuminated to the structure. Nonpolarized light and horizontally polarized light were used in Figure 2, panels a-d and a'-d', respectively. Clearly, all the slits can be treated as SPP sources and get equally excited under nonpolarized light illumination. In the middle of the triangular, square, pentagonal, and hexagonal structure, three, four, five, and six SPP beams form interference patterns, respectively (the interference patters are not distinguishable by the far field fluorescent imaging method). After a polarizer is added in the excitation beam, as indicated by the white arrows, the SPP sources at the horizontal slits are completely turned "off". The interference patterns inside of the triangle, square, pentagon, and hexagon are then formed by two, two, four and four SPP beams, respectively. When the excitation beam is normal to the structure, the SPP propagation direction is always perpendicular to the slit as shown above. If an in-plane wavevector is introduced by an inclined illumination light beam, the SPP propagation direction can be adjusted. This method has been used to tune the focus position of a plasmonic lens and now can also be utilized to further improve the tunability of the SPP interference patterns on the surface. A more complicated dynamic pattern is also possible by rotating the polarization direction and controlling the incident angle at different parts of the slit. Only one beam is used in all of our previous experiments, which has shown the simplicity of our SPP interference method for applications such as nanolithography. One could also notice that each SPP source (slit/edge) can be addressed individually by parallel multiple focusing techniques. Given all the aforementioned flexibilities, the SPP interference can be used as a platform to form various desired dynamic patterns with fine resolution.

.