#	Pagina
attuale pagina	/open-h2020/projects/200764/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 2 - CLUSTER (Birth of solids: atomic-scale processes in crystal nucleation)

Teaser

Summary

The main objective of the project CLUSTER is to explore the fundamental processes which trigger the nucleation and growth of solids. Condensed matter is formed by clustering of atoms, ions or molecules. This initial step is key for the onset of crystallization, condensation and precipitate formation. Yet, despite of the scientific and technological significance, on an atomistic level we merely have expectations on how atoms should behave rather than experimental evidence about how the growth of solid matter is initiated. The classical nucleation theory is commonly in agreement with experiments, provided the original and the final stages are inspected qualitatively. However, the classical theory does not define what fundamentally constitutes a pre-nucleation state or how a nucleus is formed at all. CLUSTER aims at investigating the very early stages of crystalline matter formation on an unprecedented length scale. It shall explore the atomic mechanisms which prompt the formation of solids. Complemented by density functional theory calculations and molecular dynamics simulations, in-situ high-resolution electron microscopy shall be used to investigate the formation, dynamics, stability and evolution of tiniest atomic clusters which represent the embryos of solid matter. Firstly, we investigate the 3D structure of clusters deposited on suspended graphene and thin films. Secondly, we focus on cluster formation, the evolution of sub-critical nuclei and the onset of particle growth by thermal activation. Thirdly, using a novel liquid-cell approach in the transmission electron microscope, we control and monitor in-situ cluster formation, precipitation, particle growth and the formation of complex nanoparticles in supersaturated solutions. The results of CLUSTER, which will advance the understanding of the birth of solid matter, are important for the controlled synthesis of (nano-)materials, for cluster science and catalysis and for the development of novel materials.

The importance of this project lies in the fact that we are continuously confronted with solid matter, but we still know very little about what triggers the crystallization of solid matter. Moreover, controlled crystallization reactions are important in various industrial areas, such as pharmaceutical, metallurgic, and, e.g., energy storage sectors. In this respect, shedding light on nucleation reactions might eventually enable engineering of nanomaterials, and potentially creating new materials of uncommon atomic order.

Work performed

Within the project CLUSTER we produced and studied tiniest clusters of Pt atoms (6-50 atoms) on thin films of amorphous carbon. As we implemented a fast atomic-resolution imaging approach, the large sets of data are noisy. We thus developed a deep convolutional neural network to process and analyze the data in order to be able to assess 3D structural data about the clusters. We then explored their 3D structure as a function of temperature and found hitherto undocumented cluster configurations. We find that the clusters down to six atoms in size exhibit at elevated temperature preferentially the stable bulk structure of Pt, namely the face-centered cubic structure. Moreover, while we find ordered, crystalline structures at elevated temperature, at room temperature the clusters are amorphous, seemingly in a molten state. The transition from the ordered at elevated temperature to the disordered state at room temperature is reversible. These findings about the behavior of clusters are of high relevance for understanding the initial states of crystalline matter formation and, considering the disorder-order transition at elevated temperature, reveal the counter-intuitive characteristics of tiniest agglomerates of atoms which can behave strikingly different from nanoparticles and bulk materials.

Within another approach aiming at shedding light on particle nucleation and growth, we investigate the formation of gold and silver particles in a liquid cell in (scanning) transmission electron microscopy (S/TEM). We employ the electron beam to trigger the reaction and are able to control growth mode and particle morphology by adjusting the electron dose. Moreover, exploring more complex reactions, we use silver nanoparticles as precursors for the growth of silver-gold core-shell particles. Aside from galvanic replacement reactions, by carefully tuning electron dose and the local chemistry using surfactants and solvents, we succeeded to nucleate a continuous atomic layer of gold atoms on the surface of the Ag particles and thus extend liquid-cell reactions to form well defined core-shell Ag-Au nanoparticles. The synthesis of core-shell structures is quite common, but not easy, in chemistry. Being able to observe the actual nucleation process of the shell and maintain a controlled growth of the shell in an in-situ experiment in a liquid cell in the transmission electron microscope is new and reveals unprecedented control of such in-situ wet-chemical reactions. Understanding and controlling growth mode and particle morphology of homogeneous and heterogeneous nanoparticles is of high relevance for optimizing syntheses reactions. Moreover, the above mentioned core-shell Ag-Au nanoparticles are supposed to exhibit excellent plasmonic properties for various sectors, including medical applications.

Although atomic-resolution information is difficult to access with a conventional liquid cell in transmission electron microscopy, in another branch of research we explore the possibility of using window-free suspended ionic liquids to study reactions, particle nucleation and growth in liquid phase in S/TEM. Within this area of research we study the structure of tiniest Au and Pt clusters in ionic liquids and the formation and growth of nanoparticles at different temperatures. This part of the project bridges the liquid phase experiments with the work on metallic clusters as the ionic liquids allow for single atom imaging in the liquid phase. All of our experimental approaches are complemented with theoretical efforts, involving density functional theory and molecular dynamics calculations, providing a comprehensive pictures of the complex phenomena observed in the experiments.

Final results

Our findings about tiniest Pt clusters, which were enabled by implementing a novel fast STEM approach combined with machine learning for data analysis, reveal hitherto undocumented cluster configurations, their dynamics and most importantly a disorder-order transition which is counter-intuitive as the ordered structure can be found at elevated temperature. These results advance the understanding about tiniest agglomerates of atoms and underlie the complex behavior of such clusters, which can be strikingly different from corresponding (nano)materials. In the future, we would like to further explore such metallic clusters, of different elements, and to try to generalize our previous findings. This also includes to develop a more advanced theoretical understanding of the experimental observations. Moreover, aside from documenting the dynamics of these pre-nucleation states of matter, we will investigate in more detail their growth mechanism, by e.g. single atom attachment, cluster coalescence or Ostwald ripening. In a parallel effort, we would like to explore the structure, dynamics and growth mode of such clusters in ionic liquids, and thus investigate how the environment can affect the pathway of solid matter formation. In this aspect, our previous results and experience about particle growth in ionic liquids form an excellent platform to further advance these studies, down to single atom dynamics and cluster formation at elevated temperature. In addition, our activities with liquid cells shall be further advanced by developing reproducible ways of generating complex homo- and heterogenous nanoparticles. Moreover, a complementary approach for modelling nucleation shall be pursued, namely by studying the self-assembly and 3D crystallization of agglomerates of monodisperse nanoparticles. While the products of such ordered self-assemblies have been well studied, our preliminary results show that it is possible to control and monitor such self-assembly processes which is of utmost importance for designing new metamaterials.