Amorphous Nanomaterials. Lin Guo

steps involve multiphase reactions, resulting in a low Coulombic efficiency (CE), a large overpotential, and a fast capacity fading upon cycling [71]. In addition, even for materials with the same metal cations, there is still a difference in conversion kinetics. Taking Fe as an example, an intermediate phase can be confirmed as LixFe₃O₄ before the conversion of nanosized Fe₃O₄ to Fe [72], while no similar phenomena can be observed with nanosized Fe₂O₃ [73]. In addition, for iron sulfide as a sodium battery electrode, the coexistence of Fe₃S₄, FeS₂, and FeS can be observed by HRTEM, and those quantum-sized FeSx ensured a synergistic and highly reversible conversion reaction, leading to a superior cyclability and rate capability [74–76]. Recently, we also investigated the charge storage mechanism of bismuth as a promising anode material for the state of the art rechargeable batteries [77]. In our work, a 2D structure of few-layer bismuthene was designed (as shown in Figure 2.5), which undergoes a two-step mechanism of ion intercalation, followed by a reversible crystalline phase evolution. This structure can alleviate the stress accumulated along the critical z-axis and allow sodium ions to rapidly diffuse due to a shorter diffusion distance, which is very helpful to develop high-performance batteries. Meanwhile, ex-situ TEM can also be a powerful tool to study the structural evolution of an amorphous electrode transition. For example, we prepared a Prussian blue analog (PBA) Co₃[Co(CN)₆]₂ as nonaqueous potassium-ion anode material [78]. The HRTEM image revealed that tiny crystallites of metallic Co of 5 nm in size are dispersed in the amorphous matrix. This observation is quite similar to the lithiation behavior of metal oxides, in which metal nanoparticles are found in the Li₂O matrix [71]. The metallic Co formed during the lithiation process may enhance the electronic conductivity. The amorphous matrix might contribute to the good cycling performance because the isotropic nature of the amorphous materials can tolerate homogeneous volume changes and accommodate volume strain. Besides, we have also carried out research with amorphous FeVO₄, which is a bimetallic element oxide for K-ion battery [79]. Local structural information of amorphous FeVO₄ after potassiation can be obtained based on the HRTEM images, which displayed tiny crystallites of VO₂, V₂O₃, and FeO with sizes below 5 nm. These tiny crystallites are surrounded by amorphous materials with solid electrode interface (SEI) or other potassiated products. The conversion was reversible because those crystalline phases partially recover to amorphous FeVO₄ after subsequent depotassiation. The particle size for the crystalline counterpart is much larger, showing lower potassiation/depotassiation capacities. This observation suggested that particle size plays a role in determining the electrochemical performance of amorphous materials. Another important issue is the observation of SEI and Li (de)plating on the electrode/electrolyte interface with high spatial resolution. By using in-situ TEM, Li dendrite growth and SEI formation or decomposition in a LiPF₆/ethylene carbonate (EC)/diethyl carbonate (DEC) electrolyte can be recorded at nanoscale resolution [80]. The SEI formation is not uniform but in the shape of dendrites. The growth kinetic of SEI can further be a valuable reference for understanding battery failure. Moreover, the beam irradiation with hundreds of keVs can cause side reactions in targeted materials and affect the imaging process, including atomic displacement, e-beam sputtering due to the elastic scattering, and heating or contamination damage due to inelastic scattering [81]. For a solid-state open cell, the main concern is the stability of Li₂O under the electron beam and the subsequent effect on the battery electrochemical performance. An effective solution to reduce the electron dosage to a safe value (approximately 1 A cm⁻²) is to suppress the chemical lithiation [82]. Meanwhile, in situ liquid cell TEM is subject to more side reactions such as the electrolyte breakdown [83] and the nanoparticles’ precipitation/dissolution [84]. The LiPF₆-based electrolyte has proven to be stable as the formation of fewer and smaller nanoparticles, and the SEI nucleation and growth can be captured on the Li deposit [85]. In addition, similar cyclic voltammetry (CV) curves were observed under and without electron beam irradiation, which demonstrates the suitability of applying liquid TEM in a real battery system [86].

Figure 2.5 In-situ TEM experiments. (a) Schematics of structural evolution of bismuth to Na₃Bi during electrochemical sodiation. Purple, Bi; Yellow, Na. (b) High-resolution TEM image of a pristine bismuth flake. (c)–(e) Time-lapse TEM images of sodiation in bismuth. Source: Reproduced with permission from Zhou et al. [77]. Copyright 2019, WILEY-VCH Verlag GMbH& Co. KGaA, Weinheim.

Apart from the TEM analysis on batteries, gas-phase reactions have also attracted a lot of attention. The degradation study during catalysis – Ostwald ripening – is a good example. Helveg et al. studied the shrinking behavior of Pt nanoparticles with amorphous Al₂O₃ as a support under O₂ and N₂ atmosphere [87]. As shown in Figure 2.6, by calculating the size change of a large amount of individual nanoparticles, they found that the larger particles grew larger while the smaller ones dissolved and finally disappeared, which undergoes the Ostwald ripening process. A similar phenomenon has also been observed in other catalysis systems, such as the system of Fe nanoparticles that worked for carbon nanotube growth [88]. Besides, photocatalysts and optical materials were also investigated under a continuous but less extreme irradiation condition. Crozier et al. observed the surface changes of TiO₂ particles in ETEM [89]. The initial surface of the particles showed a good crystallinity. After treatment under H₂O environment for one hour, a disordered layer appeared, which became more obvious after seven hours of light illumination. Based on the X-ray photoelectron spectroscopy (XPS) results, Ti³⁺ species can be found in the amorphous surface layer, indicating that the water-splitting was associated with the reduction of TiO₂ during photocatalysis. Another example using the gas-phase ETEM is the CO oxidation with the metal/metal oxides as catalysts. Two typical phenomena occurred on the surface of these catalysts. One is the formation of a sublayer of oxide underneath the outer most layer of the metal nanoparticles because of the diffusion of oxygen into the metal [90]. The other is the thermodynamicallly driven surface reconstruction of nanoparticles because of the facet-preferential adsorption of CO molecules during catalysis [91]. Taking the Au/CeO₂ system as an example [92], the absorbed CO molecules were bound to the on-top sites of Au atoms in a close-packed hexagonal orientation, leading to the surface restructuration. Because of this tensile bonding configuration between the surface layer and the sublayer, the surface reconstructed Au nanoparticles would absorb more CO molecules than the original nanoparticles. However, the challenge for gas-phase introduced TEM is the image resolution. Especially when the electron beams pass through the sample area and interact with gas, a large percentage of electrons would be scattered, resulting in the reduction of image resolution [93, 94]. The other challenges in observing the gas-phase chemical reaction in ETEM is that the ionized gases might generate the reactive gas species, giving rise to the unwanted chemical reaction at the material surface which therefore disturbs the reaction behavior.

Meanwhile, the dose rate in TEM is much higher than that of other external radiation sources [95, 96], making it easier to transfer energy to the specimens under the irradiation of 200–300 keV. Under such excitation, many ionic species would be generated to help initialize many side chemical reactions and produce various other species. When a solution of metal salt or precursor is under irradiation in TEM, the intermediated species will be produced during this process, e.g. hydrated electrons can act as a reducing agent to reduce metal cations into metal nuclei [97]. The metal nuclei are controllable within the TEM, as the dose rate can determine the speed to produce hydrated electrons and influence the reduction rate of metal cations, which then generates various morphologies either through single-atom deposition or through cluster-based oriented attachment. The observation under TEM can then provide