50-nm-resolution full-field X-ray microscope without chromatic aberration using total-reflection imaging mirrors

Abstract

X-ray spectromicroscopy with a full-field imaging technique is a powerful method for chemical analysis of heterogeneous complex materials with a nano-scale spatial resolution. For imaging optics, an X-ray reflective optical system has excellent capabilities with highly efficient, achromatic, and long-working-distance properties. An advanced Kirkpatrick-Baez geometry that combines four independent mirrors with elliptic and hyperbolic shapes in both horizontal and vertical directions was developed for this purpose, although the complexity of the system has a limited applicable range. Here, we present an optical system consisting of two monolithic imaging mirrors. Elliptic and hyperbolic shapes were formed on a single substrate to achieve both high resolution and sufficient stability. The mirrors were finished with a ~1-nm shape accuracy using elastic emission machining. The performance was tested at SPring-8 with a photon energy of approximately 10 keV. We could clearly resolve 50-nm features in a Siemens star without chromatic aberration and with high stability over 20 h. We applied this system to X-ray absorption fine structure spectromicroscopy and identified elements and chemical states in specimens of zinc and tungsten micron-size particles.

Figure 1. Advanced Kirkpatrick–Baez (KB) mirror optics based on two monolithic mirrors.

(a) Mirror arrangement. (b,c) Whole mirror shapes. (d) Shapes and residual shape errors on the each section. ‘E-’ and ‘H-’ respectively represent the ellipse and hyperbola. ‘Vert.’ and ‘Hori.’ represent vertical and horizontal directions, respectively.

Figure 2. Simulated tolerances of the shape errors and actually measured errors.

‘V’ and ‘H’ denote the mirrors in vertical and horizontal directions, respectively.

Figure 3. Experimental setup of the microscope system.

Figure 4

Bright-field X-ray image of (a) whole image and (b) magnified image. Results of analysis with (c) contrast analysis and (d) PSA. Exposure = 500 s. X-ray energy = 9.881 keV. Bar = 2 μm.

Figure 5. X-ray energy dependence between 8 and 12 keV.

(a) Bright-field X-ray images. (b) Results of PSA. Exposure is shown below each image. Bar = 2 μm.

Figure 6. Long-term stability.

(a) Time-lapse bright-field X-ray images. (b) Results of PSA. Exposure = 60 s. X-ray energy = 9.881 keV. Bar = 2 μm.

Figure 7

(a) SEM image. (b) X-ray image averaged over XAFS images between 10,159 and 10,655 eV, showing the existence of Zn and W particles. (c,d) Distributions of standard deviation (σ) of a series of XAFS images, showing the drastically changing area for image contrast during the XAFS measurement, i.e. Zn and W distributions, respectively. (e) Peak-shift map to identify W and WC. (f) XAFS spectra averaged over a 100 × 100 nm² square area. Energy scan: (c) 9640–9690 eV every 2 eV, and (d) 10195–10225 eV every 1 eV. (e) Red and blue regions represent W and WC, respectively. (f) Solid lines represent the obtained spectra on the different particles. Dash lines represent the reference spectra, which were obtained from the XAFS database (Institute for Catalyst, Hokkaido University) (data information: sample = Zn foil; correspondence = Kiyotake Asakura; date = 2006.12.13) for Zn, and the article (Fig. 2) published by Uo et al. for W. All images were obtained with an exposure of 60 s. Bar = 2 μm.

Figure 8

(a) Bright-field X-ray images before and after correction using deconvolution processing. (b) Result of PSA. The image was deconvoluted with a Gaussian with FWHM of 73 nm (36 nm) in the vertical (horizontal) direction. Bar = 2 μm.