Formation of crystals filmed in atomic detail

August 28, 2020 - A new study of scientists from the Germany, UK and Japan shed light on the fundamental phenomenon of nucleation and crystal growth. The direct observation of the entire process of nucleation allowed to distinguish the point at which the particles were large enough to become crystalline.

A huge variety of materials are made of crystals, from ice, table salt and sugar, to more specialized formulations used in medicines, transistors and those found in gemstones. Up to now, exactly how crystals initially form has been an unresolved and contentious issue as it happens at incredibly small scales. The nucleation is at the center of the crystallization process and determines the structure and size distribution of crystals and thus the functional properties of all materials.[1]

Till now, two theories including ‘classical nucleation theory’ and ‘two-step nucleation mechanism’ have been developed to describe the nucleation of crystals. This classical nucleation theory (CNT), which was developed to explain the nucleation of crystals, describes the formation of crystals in a single step from monomers (atom, ion or molecule) by adding individual monomers to an ordered structure, overcoming a single barrier for free energy (Figure 1).[2] This can be extended to streamline the formation and nucleation of crystals by thermodynamically metastable precursors (ordered or disordered) in a single step.[2] However, there are a number of experimental observations that cannot be satisfactorily explained by this theory, leading to the postulation of the more complicated and controversial two-step nucleation mechanism (TSNM).[3,4] This assembly route involves the initial formation of an amorphous "precursor" phase which is subsequently replaced by a more stable crystalline phase[5] or the formation of stable species (ordered or disordered) that, even under saturated conditions, can never dissociate and crystallize when more monomer is added becomes. As a result of the two energy barriers in a TSNM, the initially formed stable or metastable precursor phase can either crystallize into a crystalline phase by overcoming an additional energy barrier for crystallization or dissociate again into monomers by overcoming the energy barrier for monomer detachment.[2]

Because the in-situ and atomic observation of nucleation process of metals at the atomic level has been impossible, the mechanism of the metal crystallization remained unknown.

Nano test tubes and electron beams
The scientists now applied in situ low voltage aberration corrected high resolution TEM (AC-HRTEM) to investigate the nucleation of a metal crystal core at the single atom level in an electron transparent test tube, a single walled carbon nanotube (SWNT), which has a well defined atomically smooth surface and has excellent thermal, mechanical and chemical stability under various conditions including an 80 keV electron beam.[6-8] Led by Professor Ute Kaiser, head of the Electron Microscopy of Materials Science in the University of Ulm, and Professor Andrei Khlobystov, head of the University of Nottingham School of Chemistry’s used carbon nanotubes as miniature test tubes for atoms and molecules to record moving images at the atomic scale using transmission electron microscopy (TEM).

In previous work the team had shown that SWNTs can act as effective hosts for extremely small (30–60 atoms) metal clusters.[6-8] In this study, the team at the University of Ulm and Nottingham and the National Institute of Advanced Industrial Science and Technology (AIST) extended this approach to experiments on the atomic scale, which enable the observation of the first steps of the nucleation processes of γ-Fe, Au and Re crystal nuclei on the atomic level and in real time using Low-Voltage AC HRTEM imaging. The SWNT offers an ideal cavity for atom transport and a substrate for the heterogeneous nucleation of metal, which, due to its conductive structure, prevents ionization by the electron beam. Here the electron beam of the TEM is not only an imaging probe, but also a stimulus for the nucleation processes by transferring kinetic energy from the incident electrons to the atoms. The kinetic energy transferred has a maximum value and a calculable distribution and can control the chemical reactions of molecules and the dynamics of metal clusters.[6] In metal clusters, the kinetic energy transferred increases their total free energy, similar to a heating process. Here, the increase in total free energy can be controlled by adjusting the accelerating voltage and the dose rate of the electron beam, thereby promoting the nucleation processes of metal crystallites in a manner similar to thermal activation. In order to study nucleation in the initial stages with atomic resolution, it is important to obtain a solid and observable nucleation nucleus with only a few atoms. As shown in Figure 2a, a cluster containing three Fe atoms located on the wall of a SWNT can serve as a nucleation nucleus in the present experiments. During nucleation in large quantities, the extra atoms or molecules are released to the nucleus through random collisions, which is a thermally driven stochastic process with complex mechanisms. In a SWNT, the situation is considerably simplified: while the wall of the nanotube represents a substrate for nucleation, mobile clusters of amorphous carbon released from ferrocene serve as a vehicle for the release of Fe atoms to the growing core, atom for Atom, and effectively act as an "atom injector".

The group managed to film the crystal nucleation of three different metals with atomic resolution, in particles composed of only one or two dozen metal atoms. Crucially, in all three cases they were able to record before, during, and after the key moment of nucleation.

Amorphous carbon forms a complex with the Fe atoms (Figure 2b), Fe atoms are marked by the red arrow on the far right), which appears to be very mobile and slides along the nanotube cavity on the time scale of a few seconds, probably driven by thermal energy or electron beam excitation due to the extremely low friction of the atomically smooth SWNT.[9-11] The ring-shaped dark field scanning TEM (HAADF-STEM) with a high angle and the corresponding image of the electron energy loss spectroscopy (EELS) (Figure 2c, d) confirm the presence of Fe in the highly mobile atomic injector. The time series AC-HRTEM images in (Figure 2e) show a typical example of the active atom injector. Due to the fact that the Fe cluster moves quickly and stops at two positions in the SWNT during the exposure time of 1 s per image, two images with lower contrast (which indicates partial occupancy) appear in images at 0 s and 52 s. After 60 s the carbon cluster binds to the Fe cluster and then translates to the right side of the SWNT over the next 25 s.

Professor Ute Kaiser said: “The direct observation of the entire process of nucleation allowed us to distinguish the point at which the particles were large enough to become crystalline and showed us how many atoms were required for nucleation in our conditions. This was between 10 and 20 atoms for both iron and rhenium, a remarkably small number. The entire nucleation process of γ-Fe is successfully observed and recorded by AC-HRTEM using 80-keV electrons as a movie.

Three metals, three processes, one mechanism
They have determined that the nucleation of metal crystal proceeds via a two-step mechanism in the case of three different metals: γ-Fe, Au and Re. The main difference between the metals is that the heterogeneous nucleation can be triggered by different causes: the gathering of individual atoms (Fe), ordering of atoms in an amorphous nanocluster (Au), or coalescence of two separate amorphous sub nanometre clusters (Re) (Figure 3). The researchers have directly observed the existence of a liquid-like metastable precursor – a fundamental component of the two-step nucleation mechanism, and demonstrated an additional energy barrier, which the amorphous precursor has to be overcome in order to crystallize. In addition, atomic resolution of AC-TEM allowed us to estimate the number of atoms in the nucleating metal clusters. The team found in the case of all three metals the critical size is below 2 nm, and the number of atoms necessary for successful crystallization of γ-Fe and Re lies between 10 and 20.

Intriguingly, each metal displayed a different example of how the initially amorphous atomic clusters could overcome the size requirement and energy barrier for nucleation:

Professor Andrei Khlobystov said: “The further we develop nanotubes as nanoscale laboratories, the deeper we understand atoms and molecules in action. A combination of nanotubes and electron microscopy imaging reported in this study gives us the most direct experience of atoms and their reactions.”

The existence of the initially amorphous clusters is a key to understanding the underlying mechanism of the nucleation processes, as this and the energy barriers to nucleation are requirements for theoretically proposed processes. Two-step nucleation mechanisms are much more complex and contentious processes compared to classical nucleation theory, with a large degree of scientific debate over both their specific details and whether or not they are required to explain nucleation in a variety of different materials. The research team hopes that by successfully capturing this process in action for the three metals studied, it will shed light on the fundamental phenomenon of nucleation and crystal growth, enabling better-controlled design of materials in the future.

Resource: Cao, K., Biskupek, J., Stoppiello, C.T. et al. Atomic mechanism of metal crystal nucleus formation in a single-walled carbon nanotube (2020) Nature chemistry, doi: 10.1038/s41557-020-0538-9, [PDF], see also the supporting information.

  1. Myerson, A. S., Trout, B. L. Nucleation from solution. Science 341, 855–856 (2013).

  2. Sleutel, M., Lutsko, J., Driessche, A. E. S. V. A. N., Durán-Olivencia, M. A. & Maes, D. Observing classical nucleation theory at work by monitoring phase transitions with molecular precision. Nat. Commun. 5, 5598 (2014).

  3. Gebauer, D., Cölfen, H. Prenucleation clusters and non-classical nucleation. Nano Today 65, 564–584 (2011).

  4. Vekilov, P. G. The two-step mechanism of nucleation of crystals in solution. Nanoscale 2, 2346–2357 (2010).

  5. Nielsen, M. H., Aloni, S., De Yoreo, J. J. In situ TEM imaging of CaCO3 nucleation reveals coexistence of direct and indirect pathways. Science 345, 1158–1162 (2014).

  6. Cao, K. et al. Comparison of atomic scale dynamics for the middle and late transition metal nanocatalysts. Nat. Commun. 9, 3382 (2018).

  7. Khlobystov, A. N. Carbon nanotubes: from nano test tube to nano-reactor. ACS Nano 5, 9306–9312 (2011).

  8. Zoberbier, T. et al. Interactions and reactions of transition metal clusters with the interior of single-walled carbon nanotubes imaged at the atomic scale. J. Am. Chem. Soc. 134, 3073–3079 (2012).

  9. Somada, H., Hirahara, K., Akita, S., Nakayama, Y. A molecular linear motor consisting of carbon nanotubes. Nano Lett. 9, 62–65 (2009).

  10. Warner, J. H. et al. Capturing the motion of molecular nanomaterials encapsulated within carbon nanotubes with ultrahigh temporal resolution. ACS Nano 3, 3037–3044 (2010).

  11. Ran, K., Zuo, J. –M., Chen, Q., Shi, Z. Electron beam stimulated molecular motions. ACS Nano 5, 3367–3372 (2011).