Atomic-resolution 3D imaging of metallic nanoparticles under realistic conditions

Abstract text (incl. figure legends and references)

Determining the three-dimensional (3D) atomic structure of nanoparticles (NPs) is critical to understand their functional properties for many applications. It is hereby essential to perform such analyses under conditions relevant to the envisioned application. In this talk, a refined methodology will be presented for the 3D atomically resolved structural characterization of metallic NPs on an oxide support at elevated temperatures, relevant for catalytic applications.

In situ scanning transmission electron microscopy (STEM) experiments are valuable to analyze NPs under application-relevant conditions. Nonetheless, these investigations are usually inadequate because they only provide a projected image of a 3D structure. Electron tomography is a powerful method for the 3D reconstructions of NPs, but the long acquisition time required to collect a tilt series limits this technique when one wants to observe dynamic changes with atomic resolution. As an alternative, the 3D atomic structure of NPs can be extracted from annular dark-field (ADF) STEM images acquired along a single zone axis orientation during in situ experiments. Due to the thickness sensitivity of ADF STEM images, the number of atoms can be reliably estimated by the so-called atom counting method. The outcome serves as an input for energy minimization by Monte Carlo (MC) or molecular dynamics (MD) simulations. [1,2] However, this approach cannot be directly applied to supported metal NPs at elevated temperatures. The reason is that increased thermal displacements of atoms and particle-support interactions should also be taken into account as they influence the morphology of the particle. Moreover, it is expected that the structure of NPs at elevated temperatures may significantly deviate from their ground state configuration, which is difficult to determine using purely energy minimization approaches.

To overcome these limitations, we proposed an iterative local minima search algorithm based on atom-counting results by taking both the temperature effect and the particle-support interaction into account. The proposed method explores different local minima in the energy landscape while quantitatively validating the reconstructed structure with the observations from ADF STEM images (Fig 1). It has been demonstrated that our approach outperformed the previously applied methodologies (Fig 2) and enabled 3D atomic-scale investigations of both stable and metastable states of the NPs that may appear at high temperatures. [3]

References

[1] A. De Backer et al., Nanoscale 9 (2017), 8791.

[2] T. Altantzis, I. Lobato et al., Nano Lett. 19 (2019), 1, 477–481.

[3] E. Arslan Irmak et al., Small Methods 5 (2021), 2101150.

[4] This work was supported by the European Research Council (770887 PICOMETRICS to SVA and 815128 REALNANO to SB, 823717 ESTEEM3), and Research Foundation Flanders (G.0267.18N, G.0502.18N, and G.0346.21N).

Fig 1. 3D reconstruction from an experimental ADF STEM image of a supported NP (a) by combining atom-counting (b) and the iterative local minima search algorithm (c) resulting in final 3D model (d).

Fig 2. Comparison of different approaches. a) Original test structure. The final structures obtained from the combination of atom counting and b) MC, c) MD and d) the proposed approach. e) Comparison of the surface structure of the reconstructed NPs and the original NP based on the coordination number (CN) of the atoms.