A. Open-cell configuration using ionic liquid-based electrolyte

Over the last few years, substantial progress has been made toward developing methodologies for in situ direct observation of structural and chemical evolution of electrodes used for Li-ion batteries^{Reference Brazier, Dupont, Dantras-Laffront, Kuwata, Kawamura and Tarascon49,Reference Wang, Xu, Liu, Choi, Arey, Saraf, Zhang, Yang, Thevuthasan, Baer and Salmon53,Reference Wang, Xu, Liu, Zhang, Saraf, Arey, Choi, Yang, Xiao, Thevuthasan and Baer54,Reference Huang, Zhong, Wang, Sullivan, Xu, Zhang, Mao, Hudak, Liu, Subramanian, Fan, Qi, Kushima and Li60–Reference Mai, Dong, Xu and Han69} —most notably, the development of an in situ TEM cell based on an open-cell configuration. The fundamental concept of the open-cell configuration for in situ TEM studies of the Li-ion battery was pioneered by Wang et al.^{Reference Wang, Xu, Liu, Choi, Arey, Saraf, Zhang, Yang, Thevuthasan, Baer and Salmon53} and Huang et al.^{Reference Huang, Zhong, Wang, Sullivan, Xu, Zhang, Mao, Hudak, Liu, Subramanian, Fan, Qi, Kushima and Li60} The basic operating principle of the cell is schematically illustrated in Fig. 1, where the ionic liquid was used as an electrolyte.^{Reference McDowell and Cui40,Reference Wang, Xu, Liu, Choi, Arey, Saraf, Zhang, Yang, Thevuthasan, Baer and Salmon53,Reference Huang, Zhong, Wang, Sullivan, Xu, Zhang, Mao, Hudak, Liu, Subramanian, Fan, Qi, Kushima and Li60,Reference Liu, Zhang, Zhong, Liu, Zheng, Wang, Cho, Dayeh, Picraux, Sullivan, Mao, Ye and Huang63} The essential components of the cell are a single nanowire as the observable electrode, vacuum-compatible ionic liquid as the electrolyte, and LiCoO₂ as the counter electrode. A typical ionic liquid electrolyte is lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) dissolved in a hydrophobic ionic liquid 1-butyl-1-methylpyrrolidium (P₁₄) TFSI (P₁₄TFSI). The overall composition of the electrolyte is 10% LiTFSI in P₁₄TFSI. The open-cell configuration offers the possibility of atomic-level spatial resolution and analytical capability to study Li-ion insertion mechanisms into electrode materials during the charge/discharge cycles.^{Reference Ghassemi, Au, Chen, Heiden and Yassar65,Reference Ghassemi, Au, Chen, Heiden and Yassar70–Reference Yang, Huang, Huang, Fan, Liang, Liu, Chen, Huang, Li, Zhu and Zhang73} Since its invention, this technique has helped to reveal many details with respect to the lithiation mechanisms and structural evolution behavior of a range of materials, especially anode materials including Si,^{Reference McDowell, Ryu, Lee, Wang, Nix and Cui61–Reference Ghassemi, Au, Chen, Heiden and Yassar65,Reference Ghassemi, Au, Chen, Heiden and Yassar70,Reference Yang, Huang, Huang, Fan, Liang, Liu, Chen, Huang, Li, Zhu and Zhang73,Reference Liu, Zhong, Huang, Mao, Zhu and Huang74} Ge,^{Reference Liu, Huang, Picraux, Li, Zhu and Huang75} Al₂O₃,^{Reference Liu, Hudak, Huber, Limmer, Sullivan and Huang76} SnO₂,^{Reference Wang, Xu, Liu, Zhang, Saraf, Arey, Choi, Yang, Xiao, Thevuthasan and Baer54,Reference Huang, Zhong, Wang, Sullivan, Xu, Zhang, Mao, Hudak, Liu, Subramanian, Fan, Qi, Kushima and Li60,Reference Nie, Gan, Cheng, Asayesh-Ardakani, Li, Dong, Tao, Mashayek, Wang, Schwingenschlögl, Klie and Yassar72,Reference Zhang, Liu, Perng, Cho, Chang, Mao, Ye and Huang77,Reference Zhong, Liu, Wang, Mao and Huang78} ZnO,^{Reference Kushima, Liu, Zhu, Wang, Huang and Li79} graphene,^{Reference Liu, Wang, Liu, Zheng, Kushima, Huang, Zhu, Mao, Li, Zhang, Lu, Tour and Huang80} Sn,^{Reference Li, Wang, Feng, Mao, Liu, Mao and Wang81} and carbon nanotubes.^{Reference Liu, Zheng, Liu, Huang, Zhu, Wang, Kushima, Hudak, Huang, Zhang, Mao, Qian, Li and Huang82} It has been noticed that the ionic liquid will spread along the nanowire surface to form a thin layer of coating. Therefore, this configuration can, in some degree, mimic the real battery configuration such that the liquid electrolyte forms a conformal coating around the active component in the electrode. The drawback of this method is the polymerization of the ionic liquid electrolyte under the imaging electron beam. As such, the cell only can be cycled several times, which is far short for revealing the structural evolution of the electrode materials relevant to a real battery.

FIG. 1. (a) Schematic drawing showing the experimental setup of the open-cell approach using ionic liquid as the electrolyte, (b) TEM image showing a working nanobattery in TEM column, where the single nanowire anode can be imaged during charge and discharge of this nanobattery.

B. Open-cell configuration based on Li metal and metal oxide as the electrolyte

A variation of the open-cell configuration based on an ionic electrolyte (previously described) is the open-cell configuration using lithium oxide as the electrolyte. In this configuration, a Li metal is used as the anode, and a single nanowire is used as the cathode (shown schematically in Fig. 2). During insertion of the TEM holder into the TEM column, the Li metal surface is instantaneously oxidized to form a thin layer of Li₂O, covering the surface of the Li metal. It is this layer of Li₂O that serves as a solid electrolyte.^{Reference Wang, Yu, Chen, Wu, Pereira, Thornton, Van der Ven, Zhu, Amatucci and Graetz67,Reference Islam and Bredow83,Reference Liu, Zheng, Zhong, Huang, Karki, Zhang, Liu, Kushima, Liang, Wang, Cho, Epstein, Dayeh, Picraux, Zhu, Li, Sullivan, Cumings, Wang, Mao, Ye, Zhang and Huang84} In a typical example, a Si nanowire is used as one electrode, Li₂O on Li metal is the solid electrolyte, and a bulk Li metal is the counter electrode. In principle, the battery assembled in this way is a charged battery, and connecting the circuit will lead to the discharging process. However, this will not be the case due to the low conductivity of the Li ion in Li₂O at room temperature.^{Reference Islam and Bredow83} A negative potential of typically 2–4 V normally is applied between Si and Li to drive Li ions into Si. The propagation of the lithiation front can be clearly visualized by TEM imaging. The open-cell configuration offers the advantage of high spatial resolution imaging and chemical composition analysis by spectroscopic method. In a typical example, Fig. 3 shows the measured lithiation length of Si as a function of lithiation time, illustrating near-parabolic lithiation behavior. Simultaneously, the chemical composition evolution can be mapped using electron energy loss spectroscopy, or EELS.^{Reference Liu, Zhang, Zhong, Liu, Zheng, Wang, Cho, Dayeh, Picraux, Sullivan, Mao, Ye and Huang63,Reference Wang, Li, Wang, Xu, Liu, Gao, Kovarik, Zhang, Howe, Burton, Liu, Xiao, Thevuthasan and Baer64} The silicon map is obtained by an integration of Si L edges (99–170 eV), and the Li map is acquired by an integration of Li K edges (55–85 eV) after background subtractions. Notably, this type of experiment is not a real rechargeable battery. However, the structural response of Si with the Li-ion insertion adequately simulates what happens in a real battery. Therefore, this open-cell configuration provides a quick and convenient way for probing the intrinsic response of material to Li-ion insertion or extraction. A similar principle can be used with other metallic ion systems for studying ionic insertion and extraction behavior, such as in Na-, Mg-, and Ca-ion batteries.^{Reference Liu, Zheng, Zhong, Huang, Karki, Zhang, Liu, Kushima, Liang, Wang, Cho, Epstein, Dayeh, Picraux, Zhu, Li, Sullivan, Cumings, Wang, Mao, Ye, Zhang and Huang84} Another variation of the open-cell configuration has been developed by others using a standard TEM grid as described by Wang et al.^{Reference Miller, Proff, Wen, Abraham and Bareño66} The advantage of using a TEM grid is that it can increase the system's stability, affording high spatial resolution imaging.

FIG. 2. (a) Schematic drawing showing the experimental setup of the open-cell approach using Li metal as the lithium source and Li₂O as the solid electrolyte, (b) TEM image of a nanobattery with a single nanowire as cathode, which allows the direct observation of the structural and chemical evolution during charge and discharge.

FIG. 3. The lithiation process observed using an open-cell configuration for Si. (a) Progression of lithiation of Si in a core–shell fashion. (b) Measured lithiation length as a function of time. The average lithiation speed is ∼25.5 nm/s. In (a), the Li⁺ diffusion direction is labeled by the red arrows, and the lithiation reaction fronts are marked by green arrows. The electron dose is 1.55 A/m². (c) STEM-high-angle annular dark-field (HAADF) image and EELS mapping of Si, Li, and overlaid Si and Li composite, revealing a core–shell lithiation.

C. The closed electrochemical liquid-cell for direct correlation of structure and electrochemical properties

In situ TEM work conducted on open-cell configurations has provided insightful information on the structural and chemical evolution of electrodes upon lithiation/delithiation. However, three typical deficiencies are associated with the open-cell configuration. First, for the open-cell, the electrolyte is only in point contact with the electrode, which may inadvertently modify the diffusion pattern of the Li ion in the electrode. Therefore, what has been observed is not necessarily a representative case for the electrode being fully immersed in the liquid electrolyte in a real battery. Second, in using Li₂O as the electrolyte, a large overpotential is normally applied to drive the Li ions into the electrode, which may change the kinetics and phase behaviors of solid-state electrode lithiation. Third, using the ionic liquid or Li₂O electrolyte excludes some of the fundamental processes that occur only in real electrolytes and battery-operating conditions, such as the interaction between the electrolyte and electrode and the SEI layer formation.

To address the shortcomings of the open-cell (already described), recent work has focused heavily on developing a liquid-cell for in situ or more precisely operando TEM studies of Li-ion batteries using a battery-relevant liquid electrolyte and a lithium metal counter electrode. The in situ TEM study of electrochemical processes was initiated following the work depositing copper nanoparticles from a CuSO₄ electrolyte under galvanostatic conditions using a liquid-cell into the TEM column.^{Reference Williamson, Tromp, Vereecken, Hull and Ross25} Since then, the microfabricated liquid-cell concept has inspired rapid development of in situ TEM imaging under liquid or gas environments for which most of the studies have focused on nucleation and growth of nanoparticles, a process that is essentially stimulated by the imaging electron beam.^{Reference Zheng, Smith, Jun, Kisielowski, Dahmen and Alivisatos22,Reference Liao, Cui, Whitelam and Zheng23,Reference Meng Gu, Kushima, Shao, Zhang, Liu, Browning, Li and Wang85–Reference de Jonge and Ross90} The progression of liquid-cell microscopy, in turn, has helped push the development of an electrochemical cell process.^{Reference Woehl, Park, Evans, Arslan, Ristenpart and Browning91,Reference Chen, Noh, Wen and Dillon92,Reference Evans, Jungjohann, Wong, Chiu, Dutrow, Arslan and Browning93} This approach already has allowed direct observation of beam-sensitive systems, such as macromolecular complexes^{Reference Evans, Jungjohann, Wong, Chiu, Dutrow, Arslan and Browning93,Reference Mirsaidov, Zheng, Casana and Matsudaira94} and soft materials,^{Reference Huang, Liu, Chuang, Hsieh, Tsai, Wu, Tsai, Mirsaidov, Matsudaira, Chang, Tseng and Chen95,Reference Proetto, Rush, Chien, Abellan Baeza, Patterson, Thompson, Olson, Moore, Rheingold, Andolina, Millstone, Howell, Browning, Evans and Gianneschi96} and of processes that span from electrochemical deposition of metals^{Reference Williamson, Tromp, Vereecken, Hull and Ross25,Reference White, Singer, Augustyn, Hubbard, Mecklenburg, Dunn and Regan97} to growth of different nanostructures.^{Reference Liao, Cui, Whitelam and Zheng23,Reference Evans, Jungjohann, Browning and Arslan86,Reference Jungjohann, Bliznakov, Sutter, Stach and Sutter88,Reference Woehl, Evans, Arslan, Ristenpart and Browning89,Reference Zheng, Claridge, Minor, Alivisatos and Dahmen98,Reference Parent, Robinson, Cappillino, Hartnett, Abellán, Evans, Browning and Arslan99}

Gu et al.^{Reference Gu, Parent, Mehdi, Unocic, McDowell, Sacci, Xu, Connell, Xu, Abellan, Chen, Zhang, Perea, Evans, Lauhon, Zhang, Liu, Browning, Cui, Arslan and Wang100} successfully demonstrated the first working closed liquid-cell for a rechargeable battery (schematically illustrated in Fig. 4). Subsequently, similar devices have been demonstrated by other groups as well.^{Reference Gu, Parent, Mehdi, Unocic, McDowell, Sacci, Xu, Connell, Xu, Abellan, Chen, Zhang, Perea, Evans, Lauhon, Zhang, Liu, Browning, Cui, Arslan and Wang100–Reference Unocic, Sacci, Brown, Veith, Dudney, More, Walden, Gardiner, Damiano and Nackashi104} In a typical example, the working electrode is a single Si nanowire, while the counter electrode is a Li metal. This electrode geometry was implemented using a SiN_x membrane deposited on Si chips (illustrated in Fig. 4). An ∼50 nm-thick SiN_x membrane is used to seal the liquid while still allowing transmission of the high-energy electrons for imaging. The biasing chip has six Pt electrodes. The Pt electrodes extend from the SiN_x window to the edge of the chip, connecting the electrode to the outside circuit. A single or multiple Si nanowires can be mounted on one of the Pt electrodes using focused ion beam (FIB) manipulation and Pt deposition welding. The welded Si NWs extend to the electron-transparent SiN_x membrane region to enable imaging of the nanowire in electron transmission mode. A droplet of 1.0 M of lithium perchlorate, LiClO₄—containing mixed ethylene carbonate (EC) and dimethyl carbonate (DMC) electrolytes (3:7, v/v)—was applied to the top surface of the SiN_x membrane (fully immersing all Pt electrodes, Li metal, and Si NWs), and a blank chip with a SiN_x membrane facing down was placed over the biasing chip to seal the liquid electrolyte. The sealing is completed based on a three-O-ring technique, and the whole device is implanted on a biasing in situ TEM liquid holder (Hummingbird Scientific, Lacey, WA). The assembly process was completed in an argon-filled glove box to avoid atmospheric degradation of the electrolyte/electrodes. The viewing window dimensions are 50 μm × 50 μm. The normal thickness of liquid layer is 500–1000 nm. However, after loading the cell into the TEM column, the membrane will bulge outward due to the pressure differences.

FIG. 4. (a) Schematic drawing showing the setup of the liquid-cell battery. (b) SEM image of the inner side of the biasing chip, (c) magnified view of the region labeled by the orange rectangle, and (d) SEM image showing the welded Si NW electrode onto the Pt contact. Note that the Li location is labeled by the light blue color object in panel (based on the TEM holder system developed by Hummingbird Scientific).

FIG. 5. In situ liquid-cell TEM observation of the lithiation of Cu-coated Si (Cu–Si) NW. (a) TEM image showing the pristine state of the Cu–Si NW at 0 s, and (b) core–shell formation of the Cu–Si NW during lithiation at 2462 s. (c) Current versus voltage plot during the delithiation process. (d–f) STEM Z-contrast image and bright-field images of the nanowire at different states of delithiation (the left side of each panel in (d–f) shows the HAADF Z-contrast image, and the right side shows the corresponding bright-field STEM image acquired simultaneously). Note that the white arrows in (d) indicate the deposited Pt markers for Ref. Reference Gu, Parent, Mehdi, Unocic, McDowell, Sacci, Xu, Connell, Xu, Abellan, Chen, Zhang, Perea, Evans, Lauhon, Zhang, Liu, Browning, Cui, Arslan and Wang100.

D. Direct in situ TEM observation of lithiation/delithiation of Si, Li ion transport in LiFePO₄, and SEI layer formation on an Au electrode

The lithiation of Cu-coated crystalline Si NW in the liquid-cell is performed by holding the voltage of the Cu–Si anode to ∼0.03 V range. The structural evolution of the nanowire upon lithiation is illustrated by the captured video frames shown in Figs. 5(a) and 5(b). The lithiation of the Si nanowire immersed in the liquid electrolyte progresses in the core–shell fashion. The total diameter of the wire changes from 100 nm to 391 nm at 2462 s. Based on the projected radial dimension increase, this indicates that the radial direction of this wire is along the <110> direction (maximum volume expansion direction of Si upon lithiation).^{Reference Lee, McDowell, Choi and Cui41,Reference Yang, Huang, Huang, Fan, Liang, Liu, Chen, Huang, Li, Zhu and Zhang73} The increase in the diameter is quicker at the beginning of lithiation and slows down with progression of the lithiation process. This phenomenon is related to the interface stress generated by the volume expansion, which limits the diffusion of Li ions further into the core. The slowdown of lithiation after partial lithiation also has been consistently reported by earlier researchers.^{Reference McDowell, Ryu, Lee, Wang, Nix and Cui61,Reference Gu, Li, Li, Hu, Zhang, Xu, Thevuthasan, Baer, Zhang, Liu and Wang71,Reference Liu, Fan, Yang, Zhang, Huang and Zhu105}

The delithiation process of a pure Si nanowire is shown in Figs. 5(d)–5(f). The delithiation process is performed by scanning the voltage from 0 to 0.65 V at an increment of 0.3 mV/s, and the current versus voltage curve is plotted in Fig. 5(c). The image shown in Figs. 5(d)–5(f) represents a pair of dark-field (left) and bright-field (right) STEM images. In Fig. 5(d), the white arrows in the dark-field image indicate the Pt markers, which were intentionally deposited on the nanowire using electron beam deposition in the FIB SEM. The Pt markers show a higher Z-contrast in the dark-field images, highlighting the wire's positions in the liquid-cell. The diameter of the lithiated Si nanowire is 195 nm [shown in Fig. 5(d)]. The diameter of the nanowire shrank when the voltage scanned to 0.45 V, as shown in Fig. 5(e). The diameter kept decreasing as lithium ions were extracted. In the end, the diameter shrank to ∼92 nm in the final delithiated state, as shown in Fig. 5(f) at 0.65 V. The large volume change, as measured based on the reduction in diameter from 195 to 92 nm, indicates that most of the Li ions were extracted during this process.

Using a closed liquid-cell configuration, Holtz et al.^{Reference Holtz, Yu, Gunceler, Gao, Sundararaman, Schwarz, Arias, Abruña and Muller101} studied the Li-ion transport kinetics and degradation mechanisms in LiFePO₄ particles during the charge/discharge cycles. Observations of Li transport are very challenging. However, in their experiment, they used energy-filtered TEM imaging by selecting a valence energy loss peak at 5 eV, which is unique for FePO₄ but not for LiFePO₄. Therefore, environmental transmission electron microscopy (ETEM) images captured with the energy window selected at this region can be used to probe the phase transition and Li transport characteristics between LiFePO₄ and FePO₄. This method allowed Holtz et al.^{Reference Holtz, Yu, Gunceler, Gao, Sundararaman, Schwarz, Arias, Abruña and Muller101} to determine the lithiation state of a LiFePO₄ electrode and surrounding aqueous electrolyte in real time with nanoscale resolution during electrochemical charge and discharge, enabling them to track lithium transfer between the electrode and electrolyte and image charging dynamics in the cathode.

Zeng et al.^{Reference Zeng, Liang, Liao, Xin, Chu and Zheng103} reported the formation of the SEI layer on an Au electrode during electrochemical lithiation and delithiation of Au anodes in lithium hexafluorophosphate (LiPF₆)/EC/diethyl carbonate (DEC) electrolyte. Upon the cyclic voltammetry (CV) scan, they noticed the inhomogeneous lithiation of the Au electrode, lithium metal dendritic growth, electrolyte decomposition, and SEI formation. The SEI layer growth appears to be concurrently accompanied by the Li dendrite growth. The SEI layer is observed to uniformly develop on the Au electrode surface. In a similar effort, Sacci et al.^{Reference Sacci, Dudney, More, Parent, Arslan, Browning and Unocic102} observed the SEI layer on an Au electrode in 1.2 M LiPF₆ EC/DMC battery-grade electrolyte. They noticed that SEI formed prior to the deposition of Li. Most dramatically, they found that the SEI layer shows a feature of dendritic morphology rather than a uniform layer: the dendritic SEI forms prior to Li deposition, and the SEI layer remains on the surface after Li electrodissolution. Therefore, the formation of SEI dendrites begins prior to Li deposition, suggesting that the electrolyte composition (and its electrodecomposition) affects Li deposition. It should be noted that for the case of the SEI layer growth work, no Li source is supplied in the liquid-cell. All observed phenomena result from the decomposition and deposition of the ions from the electrolyte. Even if a CV scan is done to offer a general view of the system's electrochemical behavior, more systematic work must be carried out with a proper Li source and a reference electrode.

E. Solid-state battery configuration

The solid-state battery configuration using a solid electrolyte provides another option to probe the structural and chemical evolution of electrode materials. For the solid-state configuration, it may be related to the solid-state system as a battery itself or as a platform for in situ TEM studies of a battery system.^{Reference Wang, Xu, Liu, Choi, Arey, Saraf, Zhang, Yang, Thevuthasan, Baer and Salmon53,Reference Meng, McGilvray, Yang, Gostovic, Wang, Zeng, Zhu and Graetz68} Brazier et al.^{Reference Brazier, Dupont, Dantras-Laffront, Kuwata, Kawamura and Tarascon49} developed the first cross-section for an all-solid-state Li-ion nanobattery for in situ TEM observation (illustrated in Fig. 6). The fundamental concept of this configuration is to use FIB to make a “nanobattery” from an all-solid-state battery prepared by pulsed laser deposition (PLD). Using a similar configuration, Yamamoto et al.^{Reference Yamamoto, Iriyama, Asaka, Hirayama, Fujita, Fisher, Nonaka, Sugita and Ogumi52} observed changes of electric potential in an all-solid-state Li-ion battery in situ with electron holography (EH). They mapped the two-dimensional potential distribution resulting from movement of lithium ions near the positive-electrode/electrolyte interface. Meng et al.^{Reference Meng, McGilvray, Yang, Gostovic, Wang, Zeng, Zhu and Graetz68} probed the dynamic phenomena in an all-solid-state nanobattery based on imaging and EELS. The solid-state configuration offers the possibility of high spatial resolution imaging. However, due to the rigidness of the nanobattery configuration, it is difficult to maneuver the sample to the appropriate orientation for lattice resolution imaging. A similar configuration based on a solid-state electrolyte and a single nanowire electrode also has been attempted.^{Reference Mai, Dong, Xu and Han69,Reference Yang, Xie, Ruffo, Peng, Kim and Cui106} Overall, the solid-state battery configuration provides a platform for coupled imaging, diffraction, and spectroscopy for comprehensive structural and chemical analysis of nanobatteries under battery-operating conditions. In particular, tracking the Li may be possible via developed Li K-edge spectroscopy and mapping.

FIG. 6. Schematic drawing showing an all-solid-state battery machined by the FIB lift-out procedure for in situ TEM study.

F. Quasi-in situ TEM study of battery process

The in situ TEM cell based on the open-cell, closed liquid-cell, and all-solid-state cell is an especially useful way to probe Li-ion transport and directly visualize the structural evolution of electrode materials. However, the nanobattery only can be cycled in situ in the TEM column for a very limited number of cycles, typically only up to 10 cycles. Thus, this type of in situ nanobattery lacks a direct correlation with electrochemical data for a large number of cycles. To address these issues, a quasi-in situ TEM observation of energy materials has been reported by Lin et al.^{Reference Lin, Nordlund, Weng, Zhu, Ban, Richards and Xin107} The fundamental concept is to load the material of interest on a standard TEM grid that is embedded into the button cell with another battery material to perform a standard electrochemical testing. Following the electrochemical testing, the TEM grid is removed and loaded into a TEM for direct observation. This experimental scheme is mostly suited for probing the structural and chemical evolution of the electrode materials, but it is not suitable for probing the SEI layers due to the inevitable exposure of the sample in air following electrochemical testing. In principle, this appears to be a type of ex situ experiment. However, it provides the advantage of observing the same particle before and after electrochemical testing, as long as the particle is appropriately observed and marked.

G. Electron-beam-induced effect in the open-cell

As with all imaging situations, one issue for the open-cell in situ TEM study of lithiation is the electron-beam-induced effect. For electrochemical experiments, in addition to the generally observed beam effect such as knock-on damage, heating, and ionization of the materials, the electron beam also will affect the electrochemical process. It has been observed that the electron beam can either accelerate or retard the lithiation process, depending on the materials and electron dose. With uniform illumination of the whole nanowire by the electron beam, the beam effect cannot be visualized, yielding a uniform core–shell structure with silicon crystalline core and Li_xSi amorphous shell (as shown in Fig. 3). The electron-beam-induced retardation of lithiation is vividly demonstrated by the lithiation of Si nanowires, as illustrated in Fig. 7. To identify the beam effect, the electron beam was focused on part of the nanowire with a dose of 22.3 A/m² (shown in Fig. 7). The region illuminated by the electron beam is marked by red arrows. The silicon core at the region under electron beam irradiation is lithiated at a slower rate than the region without electron beam exposure. Typically, the diameter of the silicon core in the region exposed to the electron beam is 37.6 nm compared to 26.9 nm in the region without the electron beam. Clearly, the electron beam inhibits the lithiation process.

FIG. 7. (a) Schematic drawing showing the retardation of the lithiation of Si nanowire by the imaging electron. (b) TEM image showing the electron beam effect during lithiation. The imaging electron beam with an acceleration voltage of 300 kV was focused on the region indicated by the red circle, where lithiation is retarded as indicated by a crystalline Si core width of 37.6 nm compared to 26.9 nm in the region without electron beam exposure. Using this imaging condition, the electron dose is 22.3 A/m². TEM image shows electron-beam-induced delithiation of lithiated Si NWs. (c) Lithiated Si NW with Si crystalline core and amorphous Li_xSi lithiated region. (d) The same region in (c) is delithiated following a prolonged electron beam exposure. (e) Higher magnification TEM image showing the formation of Li metal on the surface of the delithiated Si NW. The inset shows the Li K edge EELS from the Li metal.

The electron beam inhibition of lithiation appears to be consistent with the effect of electron-beam-induced delithiation of amorphous Li_xSi, as illustrated in Figs. 7(c)–7(e). The lithiated Si NW exhibits a Si crystalline core and amorphous Li_xSi-shell with a total diameter of ∼132 nm. After irradiating the lithiated Si NW for ∼26 min with an electron dose of 0.9 A/m², a delithiation process was noticed. Li metal formed on the surface of Si NW, and the diameter of the Si NW shrank to ∼82 nm, as shown in Figs. 7(c)–7(e). A higher magnification TEM image in Fig. 7(c) clearly revealed that the formed Li metal exhibit crystalline contrast and some even with good surface facets. The EELS composition analysis, shown as the inset of Fig. 7(c), illustrates the Li K edges from the formed Li metal on the surface. It has been noticed that in the case of extremely high electron doses, the delithiation process happens without the formation of Li metal. This is due to the displacement damage of Li metal by the high-flux electron beam. Electron beams with a lower acceleration voltage, such as 80 kV, yield even higher damage to the lithiated Li_xSi layer. Wang et al.^{Reference Wang, Graetz, Moreno, Ma, Wu, Volkov and Zhu108} calculated the displacement cross section of Li and Li-containing compounds and concluded that electrons at lower acceleration voltage (even 60 kV) cause even more damage than at higher acceleration voltage (400 kV). Contrary to the case of Si, Liu et al. observed that electron irradiation of SnO₂ nanowire and Li₂O junction on a TEM grid leads to the lithiation of SnO₂.^{Reference Liu, Liu, Kushima, Zhang, Zhu, Li and Huang109} They proposed that electron beam irradiation of Li₂O leads to the decomposition of Li₂O into elemental Li and volatile gas,^{Reference Vajda and Beuneu110,Reference Krexner, Prem, Beuneu and Vajda111} such that 2Li₂O → 4Li + O₂↑, followed by xLi + SnO₂ → Li_xSnO₂. Therefore, during in situ TEM imaging, the effect of the imaging electron beam on the electrochemical process requires careful evaluation.