Material-specific high-resolution table-top extreme ultraviolet microscopy

Abstract

Microscopy with extreme ultraviolet (EUV) radiation holds promise for high-resolution imaging with excellent material contrast, due to the short wavelength and numerous element-specific absorption edges available in this spectral range. At the same time, EUV radiation has significantly larger penetration depths than electrons. It thus enables a nano-scale view into complex three-dimensional structures that are important for material science, semiconductor metrology, and next-generation nano-devices. Here, we present high-resolution and material-specific microscopy at 13.5 nm wavelength. We combine a highly stable, high photon-flux, table-top EUV source with an interferometrically stabilized ptychography setup. By utilizing structured EUV illumination, we overcome the limitations of conventional EUV focusing optics and demonstrate high-resolution microscopy at a half-pitch lateral resolution of 16 nm. Moreover, we propose mixed-state orthogonal probe relaxation ptychography, enabling robust phase-contrast imaging over wide fields of view and long acquisition times. In this way, the complex transmission of an integrated circuit is precisely reconstructed, allowing for the classification of the material composition of mesoscopic semiconductor systems.

Fig. 1. EUV ptychography setup.

A few-cycle IR-laser is focused in an Argon gas jet where a broad EUV continuum is generated. From the broadband continuum, a narrow bandwidth of 0.2 nm is selected at a wavelength of 13.5 nm by three multilayer mirrors (ML-Mirrors) and focused on a mask (M). The sample (S) is illuminated by a structured beam and the resulting diffraction pattern is recorded by the detector.

Fig. 2. EUV ptychography using structured beams.

a Reconstructed transmissivity of the Siemens star using an unstructured illumination. The reconstruction exhibits spurious mid-spatial-frequency modulations, which are displayed in inset (b). A magnified view of the center of the Siemens star is shown in (c). The corresponding reconstructed probe is shown in (d). e The probe back-propagated into the mask plane. The small inset in (e) shows an SEM image of the mask, which is a 8 µm diameter pinhole. f Reconstructed transmission of the Siemens star using structured light showing fewer modulations in (g) and a higher resolution in (h) as compared to the unstructured reconstruction (b, c). The green circular line in (c), h corresponds to the smallest radius where the spokes are still resolved. i, j Reconstructed illumination in the sample and mask plane. The small inset in (j) shows the SEM image of the mask for comparison. k Azimuthal lineout of the probe phase along the white, dotted path indicated in (i). The scale bar in (a), f indicates 2 µm and the scale bar of (d), e, i, j corresponds to 5 µm. The brightness and hue of (d), e, i, j encode modulus and phase, respectively.

Fig. 3. High-resolution, wide field of view imaging.

a Reconstructed Siemens star over a field of view of 340 µm². The corresponding reconstructed illumination (probe) is shown in (b) and features a charge-1 OAM beam. The smallest features are present in the innermost part of the Siemens star, which is shown in (c). The Fourier ring correlation (FRC) is shown in (d) and indicates a diffraction-limited resolution of 16 nm (31 µm⁻¹). The scale bar in (a), (b) corresponds to 5 µm and the scale bar in (c) corresponds to 1 µm.

Fig. 4. Quantitative amplitude and phase imaging EUV of an integrated circuit.

a Reconstructed complex transmission of a conventional solid-state disc lamella using the m-s reconstruction model. At the edges of the reconstruction, artifacts are visible, which are particularly evident in the region of interest in (b). c Reconstructed complex transmission using the combined m-s/OPR approach. This results in a reduction of artifacts, as highlighted by the white insets (b), d. Since ptychography reconstructions are invariant under a global phase shift and amplitude scaling, the reconstructed complex transmission must be referenced. The reconstructed amplitude is referenced to the surrounding vacuum region, which is indicated in (a), b by a white box labeled “R”. The scale bar in a corresponds to 5 µm and the scale bar in (b), d corresponds to 1 µm.

Fig. 5. Material-specific EUV imaging.

a Scattering quotient for a selected region of interest (compare red dashed box in Fig. 4b). For each of the numbered regions 1 to 5 a histogram is plotted, as shown in (b) for the m-s reconstruction (Fig. 4a) and in (c) for the m-s/OPR reconstruction (Fig. 4c). The tabulated scattering quotient for the materials Al (aluminum, f_q = −0.1), SiO₂ (silicon dioxide f_q = 2.0), and Si₃N₄ (silicon nitride f_q = 2.9) at a photon energy of 92 eV is indicated by a black dotted line in (b), c. The semi-transparent areas indicate plus/minus one standard deviation from the mean scattering quotient in the corresponding areas. Using only the m-s method (b), the different materials cannot be identified reliably. For the combined m-s/OPR method (c), Al, SiO₂, and Si₃N₄ can be clearly distinguished. Energy-dispersive X-ray spectroscopy (EDX) measurements, for a region indicated by a black dashed box in (a), are shown in (d) for nitrogen and oxygen and in (e) for silicon and aluminum. The scale bar in (a), d, e has a size of 1 µm.