Experimental Tip Shape Imaging Routes to Assist Data Reconstruction in Atom Probe Tomography

Offering 3-dimensional elemental mapping with sub-nanometer spatial resolution, Atom Probe Tomography (APT) finds applications in various research domains [1]. Despite an improved description of the fundamental mechanisms underpinning this technique, major challenges remain when analyzing heterogenous materials, such as the reliable and accurate reconstruction of the 3D point cloud of the analyzed volume. This forms a critical step in the workflow of any tomographic method, as it defines its spatial resolution and analytical performance. At present, reconstruction protocols in APT rely on the Bas protocol [1], a quasi-stereographic projection-based model, with a simplified description of the atom probe tip geometry and ion trajectories. When analyzing heterogeneous systems these assumptions are too simplistic and yield to misplaced atoms in the reconstructed volume, causing faulty dimensions and distorted shapes. With the understanding of the root-causes of these artefacts, focus has shifted towards implementing more realistic ion trajectories, based on electrostatic field calculations solved for the actual tip geometry. Consequently, knowledge of the (evolving) geometry of an atom probe tip could represent an important cornerstone for an accurate 3D reconstruction, which lead us to develop two complementary, contact-less imaging routes for such purposes. In this contribution we will showcase images of atom probe tips obtained by these imaging methods, highlight respective advantages, summarize current state-of-the-art, as well as discuss potential challenges for their implementation and use in data reconstruction.

The first approach leverages the advantages of lens-less imaging with short wavelength photons, (soft x-rays or extreme ultraviolet light) in scanning coherent diffractive imaging (CDI) (i.e. ptychography) [2]. For this imaging method a spatial resolution of 15 nm was reported on regular samples when using a laboratory high-harmonic generation source at ∼70 eV [3], a comparable source as used to stimulate field evaporation in [4]. Here we follow-up on our preliminary proof-of-concept study [5] using a synchrotron light source (800 eV). We report chemically resolved images with a relevant field-of-view along the tip axis and a spatial resolution down to 11 nm. At present the image represents a 2D projection of the atom probe tip, but 3D information could be envisioned in the future using tomographic CDI [3]. The chemical and geometrical information obtained by scanning CDI might for example feed a model-driven reconstruction workflow as developed by Fletcher et al. [6].

Atomic Force Microscopy (AFM) is an alternative, contact-less method we had developed [7] for tip shape imaging motivated by the high spatial resolution, especially in z-direction, and the quantitative 3D topographic information that one can obtain [8]. Compared to scanning CDI (at 800 eV), AFM provides highly spatially resolved topography images of the tip apex, with a smaller field-of-view along the tip z-axis. Relying on atomic interaction forces, AFM is inherently sensitive to the surface, as opposed to transmission CDI which probes the whole volume of the tip. Beyond surface topography, other material properties (mechanical, electrical, chemical, etc.) could be mapped by AFM [8], prospectively adding relevant information about the atom probe tip being analyzed. As it is a contact-less method, AFM imaging should not impede further APT data acquisition, as will be demonstrated by examples, allowing for iterative APT-AFM-APT-AFM… measurement cycles in a hybrid APT-AFM tool. Recently, AFM images of an atom probe tip have successfully informed a trajectory-based reconstruction protocol, demonstrating a reduction in density fluctuations in the reconstructed data in this proof-of-concept [7].

While the presented imaging methods have their inherent limitations, we will showcase their capabilities to provide relevant information about the actual atom probe tip geometry (evolution), which could offer valuable feedback for data reconstruction. We therefore believe that our efforts represent an important stepping-stone towards more accurate tomography of heterogeneous systems such as for example semiconductor devices, which is critically needed in the semiconductor industry.