The authors present a robust process for fabricating passive silicon photonic components by direct-write electron beam lithography (EBL). Using waveguide transmission loss as a metric, we study the impact of EBL writing parameters on waveguide performance and writing time. As expected, write strategies that reduce sidewall roughness improve waveguide loss and yield. In particular, averaging techniques such as overlap or field shift writing reduce loss, however, the biggest improvement comes from writing using the smaller field-size option of our EBL system. The authors quantify the improvement for each variation and option, along with the tradeoff in writing time. V

High resolution patterning on nonplanar substrates with large height variation using electron beam lithography is reported. Using an automatic, high precision, noncontact laser probe microscope, a three-dimensional map of the nonplanar substrate to be patterned is obtained first. This data are converted to a format for the electron beam lithography system, which performs the write by adjusting the plane of electron beam focus based on the mapping data. As a proof of concept of this patterning scheme, three different kinds of nonplanar substrates were used including a tilted Si wafer for a uniform and unidirectional tilt, a planoconvex lens for a multidirectional tilt, and deep Si trenches with a step height variation. The patterning scheme was tested for resolution, field stitching accuracy, and field placement accuracy. The results of these tests are in very close agreement with typical results that are obtained for flat substrates with similar patterning conditions. For wafers with 10 mm tilt, 50 nm gratings show a line width variation of 18%, average field stitching error of 1.06 nm with a 3σ of 24.62 nm, and field placement error of 20.53 nm with a 3σ variation of 31.92 nm. The simplicity of this method and the fact that it can be applied for various nonplanar substrates with nanometer scale precision and over large areas open up a new avenue for the fabrication of devices based on nonplanar substrates.

Articles you may be interested inHydrogen silsesquioxane bilayer resists-Combining high resolution electron beam lithography and gentle resist removal J. Vac. Sci. Technol. B 31, 06F102 (2013); 10.1116/1.4822136Time-dependent exposure dose of hydrogen silsesquioxane when used as a negative electron-beam resist Hydrogen silsesquioxane for direct electron-beam patterning of step and flash imprint lithography templates Hydrogen silsesquioxane (HSQ) is a high resolution negative tone electron beam (e-beam) resist with a resolution well below 10 nm. However, it is known that the time delay between spinning and e-beam exposure has an effect on the contrast and sensitivity. Significant effort has been placed on finding the best developer conditions (e.g., time, temperature, concentration, etc.) to maximize the performance of the resist. However, to date, little progress has been made to mitigate the problem of temporal dependence of optimal results. The authors report a significantly improved time dependent performance of HSQ by using a spin on top coat developed by Showa Denko. The material contains a conductive, water-soluble polymer (i.e., poly[isothianaphthene sulfonate]) and surfactants. Exposure delays from 0 to 12 h were performed in vacuum and the effect on sensitivity and ultimate resolution was compared to results without a top coat. Various bottom layers and top coats were also used and the change in performance was evaluated. In all cases a significant improvement is observed when a top coat is applied and little change occurs with other parameters. The time lag effect is shown to be the equivalent of a decrease in sensitivity over time with about 30% degradation from the optimal dose over the test period if no top coat is used. The use of the top coat reduces the effect to about 5% while also improving initial sensitivity. The authors believe this represents a widely applicable solution to the time lag issue common to HSQ processes.