High-contrast alignment marks are promising optical elements for the high-precision alignment of X-ray gratings in phase-contrast X-ray imaging systems. These marks contain micrometer-scale bands of optically absorptive and reflective areas that determine their relative degree of optical contrast. Nanostructured black silicon (n-BSi) enabling the broadband, quasi-omnidirectional absorption of light, is a promising material to define these absorptive areas. In this work, we have developed an optimized deep reactive-ion etching (DRIE) process to fabricate n-BSi for these alignment marks. The DRIE process was optimized by systematically changing the polymer deposition time, T, while keeping all other process conditions constant. As T varied from 0.6 to 2.5 s, three distinct etching regimes were observed: (1) smooth etching, (2) n-BSi, and (3) etching stop. During the n-BSi regime (T = 1.6 to 2.1 s), varying T altered the n-BSi morphology, resulting in a broad spectrum of nanostructures: nanopillars to nanopores. Our investigation of the process–structure–property relationships among the process etching, morphology evolution, and reflectance of n-BSi revealed that morphologies with an increased height, aspect ratio, and degree of tapering and a lower base spacing were most effective at suppressing reflection for λ = 500–800 nm. Our results demonstrated the lowest reflectance of the absorptive areas, Rabs = 0.01 (∼1%) and Rrel_abs = 0.03 (relative with respect to Si) at λ = 600 nm for T = 1.9 s (100 cycles), thus giving the highest contrast ratio, ϕ = 0.9. To increase the device throughput, we varied the cycle number from 5 to 100 times at T = 1.9 s to study its effect on growth, morphology, and reflectance of n-BSi. Our results demonstrate that Rabs, Rrel_abs, and ϕ plateaued at 50 cycles, cutting in half our total fabrication time, t. In addition, our alignment marks demonstrated rotational and translational precision alignment capabilities. Lastly, we present two design principles for T and the cycle number to fabricate the optimal n-BSi morphology for the targeted application using any DRIE tool.
Read full abstract