X-ray microscopy is a strong analytical tool with a plethora of applications in physics, materials science and life sciences [1]. In many high resolution X-ray microscopes, a focusing optic such as a Fresnel zone plate (FZP) is utilized as a lens to form an image. The FZP is a diffractive optic composed of a series of concentric rings (zones) of radially varying grating period, where the width, Δr, of the outermost zone determines the resolution. Synchrotron radiation facilities are required as bright and coherent X-ray sources as the FZPs are usually quite limited in diffraction efficiency. The diffraction efficiency depends on FZP properties such as geometry, material, structure thickness and on the outermost zone width for small Δr. Conventional FZP profiles are binary and can have up to 10-40 % diffraction efficiency for ideal absorption and phase reversal zone plates, respectively. In practice the efficiencies are lower due to fabrication difficulties and errors. Electron beam lithography (EBL), the usual fabrication method, allows fabrication of binary FZPs with ultra-high resolution via processes of ever increasing complexity. However, due to the scattering of electrons within the resist, it is very difficult to fabricate ultra-high resolution FZPs with high structural thicknesses to achieve high aspect ratios. This, so called, proximity effect limits the aspect ratio, restricting the utilization of high resolution FZPs to softer X-ray energies. One special type of refractive/diffractive X-ray optic is the kinoform lens, with a continuous 3D surface profile. Ideally, it has a theoretical diffraction efficiency of 100 %. While in a real lens absorption would hinder 100 % efficiency, diffraction efficiencies well above 40 % are already demonstrated [2] using these lenses. Nevertheless, in order to fabricate these kinoforms, researchers resorted to step approximations using several consecutive overlay EBL steps, which is complicated and lead to much lower efficiencies than theoretically expected due to the vulnerability of the optics to fabrication errors amplified by error accumulations. To solve the issues that are intrinsic limitations of the EBL-FZPs we introduced novel, precise and mostly direct methods relying on the FIBs. The powerful method potentially allows for higher resolutions than EBL [3] and precise 3D sculpting capability enables the realization of highly efficient X-ray optics with high fidelity [4]. We have used a standard multi-purpose FIB instrument (Nova 600 NanoLab, DualBeam, FEI) to fabricate binary FZPs, kinoform lenses and multilayer FZPs (Figure 1).
Read full abstract