1.   Introduction

The ever-growing energy demands of society require the development of more efficient and economical energy storage systems [1–3]. Rechargeable lithium batteries (LBs), given their high theoretical energy density and environmental friendliness, have drawn intense research interest and dominated the market of portable electronic devices. Although LBs have also been adopted for grid-scale energy storage and electric vehicles, their relatively low capacity and high fabrication cost compared with solar, wind, or fossil energy sources pose barriers to their market adoption. The property and performance of LBs are mainly determined by their anode and cathode components. Thus, high-capacity and cost-effective electrode materials must be designed and explored to promote LBs into a wider consumer market.

Various electrode materials, such as Si, Sn, and Li metal for anodes and Ni-rich oxide and S for cathodes, have been explored for LBs. However, despite their higher theoretical capacity than conventional electrode materials, all these emerging materials suffer from grand challenges. For example, although Si possesses more than ten times the theoretical capacity of graphite anodes, its huge volume expansion (～300%) during lithiation causes structural collapse, loss of electrical contact, and accumulated solid electrolyte interface (SEI) formation, which induce a practically low achievable capacity and poor cycling life [4]. Li metal, which possesses the highest theoretical capacity and lowest electrochemical potential, has been considered as the best anode of choice for LBs. However, Li tends to deposit in dendritic form, which causes capacity decay and brings severe safety concerns to batteries [5]. As for S cathodes, the polysulfide dissolution severely harms its cycling life and impedes the commercialization of lithium–sulfur (Li–S) batteries [6–7]. In addition, the recovery and recycling of valuable metal elements in exhausted LBs attract great attention due to resource consumption and cost increase [8–10].

To improve the performance of batteries, scholars should improve their understanding of the chemical and electrochemical reaction processes and the internal degradation mechanisms in the whole battery system from different scales, including the electrode materials, electrodes, and devices. Given such a goal, various material characterization and electrochemical analysis techniques have been developed. Compared with characterization under ex situ conditions, in situ and operando characterization not only can provide the information the same as that from ex situ characterization but can also capture information related to the dynamics of charge/discharge cycles [11].

In the past years, in situ and operando characterization tools have been extensively developed for the monitoring of material structure and surface/interface chemistry evolution under operating conditions. For example, in situ transmission electron microscopy (TEM) is a powerful tool used to visualize morphological and structural evolutions and volume expansion/contraction [12]. In situ X-ray absorption spectroscopy (XAS) shows a unique capability to disclose the atomic and electronic structure of the active host metal in action [13]. In situ atomic force microscopy (AFM) provides direct information on the surface topography; it is especially useful in detecting the SEI layer evolution [14]. In situ Raman spectroscopy is used to analyze new products as the transient states of an electrode surface [15]. Other in situ tools, such as scanning electron microscopy (SEM) [16], X-ray diffraction (XRD) [17], Fourier transform infrared (FTIR) spectroscopy [18], mass spectrometry (MS) [19], nuclear magnetic resonance (NMR) spectroscopy [20], and neutron reflectometry/neutron depth profiling (NDP) [21], have also been developed to study battery failure during battery operation. With the assistance of these techniques, a comprehensive understanding of how the structure and composition of battery components affect their chemical/electrochemical behaviors during battery operation can be obtained. This information can guide people in the optimization and design of suitable electrode and electrolyte materials, thus improving the overall performance of batteries.

In this review, we summarize the recent progress in advanced in situ characterization techniques for LBs, especially focusing on the multi-model coupling of different characterization techniques. We first briefly discuss the advancement of single characterization techniques and the combination of several techniques for the LB system. Then, the characterizations of different battery components, including the cathode, anode, electrode–electrolyte interface, and electrolyte, are discussed. Case studies are used to illustrate how these advanced characterizations are used to study the composition and structural evolution of typical battery systems during cycling and reveal the structure–performance correlation, ranging from the local structure and functional groups at the molecular level to the morphological evolution at the electrode scale and overall electrochemical performance on the device scale. Finally, along with the concluding remarks and remaining challenges, opportunities and possible directions for future development of in situ characterization techniques that will further benefit LBs are discussed.

2.   Current in situ characterization techniques for LBs

In the past years, the great progress in in situ characterization techniques facilitated the comprehensive understanding of the reaction mechanism in LBs. The research has greatly benefited the optimization and design of more efficient battery electrodes and electrolyte materials and contributed to the further improvement of the overall battery performance [22]. Fig. 1 summarizes the important and widely applied in situ techniques with their spatial resolution scales and the corresponding detection objectives in batteries. Depending on the relevant length scale of each unit, these techniques can be selectively used to disclose the structure, morphology, chemistry, and kinetics of batteries. Moreover, with the joint help of different characterization techniques, comprehensive studies have been carried out to understand the complicated issues that arise during cycling with different physical/chemical aspects and at multiple length scales. In section 2.1 and section 2.2, we discuss the progress of in situ characterization using single and joint multi-modal techniques, respectively. In section 2.3, we discuss the design of electrochemical cells for in situ characterization.

Figure  1.  Main in situ characterization techniques with their spatial resolution scales and the corresponding detection objectives in batteries (XANES: X-ray absorption near-edge structure; EXAFS: extended X-ray absorption fine structure; XPS: X-ray photoelectron spectroscopy; TXM: transmission X-ray microscopy; XRT: X-ray tomography; NPD: neutron powder diffraction; NPDF: neutron pair distribution function, NR: neutron reflection; NI: neutron imaging; EELS: electron energy loss spectroscopy; ESM: electrochemical strain microscopy; SICM: scanning ion conductance microscopy; EPR: electron paramagnetic resonance; MRI: magnetic resonance imaging).

下载: 全尺寸图片 幻灯片

2.1   Single in situ characterization technologies

In this part, the progress of typical in situ characterization technologies for LBs based on X-ray, electron, neutron, optical, magnetism, and scanning probe microscopy (SPM) is discussed.

X-ray techniques offer quantitative and qualitative information on the structure and chemical composition of materials [13–17]. In situ XRD is used to investigate the phase constitution, space group symmetry, lattice parameters, and space-averaged atomic structure during cycling. XAS can provide fine insights into the structural evolution at the atomic level, detecting the electronic transition, valence state changes, sites symmetries, bond length, coordination number, degree of disordering, and chemical identity of the nearest atom. TXM and XRT offer direct visualization of the phase transformation, crack creation and growth, and electrode deformation, which are associated with the degradation of battery performance.

Compared with X-rays, neutron-based techniques, typically including NPD, have unique advantages, such as the complementary scattering cross-sections of neutrons, high sample penetration, and sensitivity to low atomic number elements (e.g., H, Li, and O) [23]. NPD, similar to XRD, can precisely determine structural and dynamic information at the atomic level, but it is difficult to use in amorphous materials. NDP has a nanoscale spatial resolution, which is very sensitive to the Li-ion concentration changes with respect to the depth of battery materials. On this basis, other neutron-based techniques, for example, NPDF, NR, and neutron imaging (NI) have been developed and applied to the in situ analysis of electrodes in LBs [21].

In comparison with X-ray and neutron, an electron beam has a shorter de Broglie wavelength and shallower penetration depth. Therefore, electron-based techniques are theoretically capable of obtaining higher resolution [12,16]. TEM can provide a combination of nanoscale down to atomic level spatial resolution. EELS can examine the oxidation state related to the reaction chemistry. Energy dispersive X-ray analysis (EDX) provides information about the elemental distribution and quantitative concentration. The combination of these techniques, in situ TEM turns into a powerful analytical tool to reveal the electrochemical mechanisms. By contrast, the resolution of SEM is about 10 nm, which is usually employed to probe the morphological change.

Optical techniques, which have non-destructive, non-contact, and non-vacuum requirements, are also widely adopted for in situ characterization of LBs. In situ Raman spectroscopy can be applied to monitor the valence and bond variation in the local structure and oxidation state at the atomic scale. With the adoption of the mapping mode, Raman images provide distinct phase-transformed zones accompanying electrochemical reactions [15,24]. FTIR can identify organic chemical species, including ion molecules, solvents, dipolar species, and some chemical breakdown products [18,25]. Optical microscopy (OM), which possesses a low resolution of about 200 nm, is often used to visualize the macrostructural features of batteries [26–27].

Techniques based on magnetism include NMR, EPR, and Mössbauer spectroscopy. NMR helps in monitoring the local electronic environment around the nucleus over the electrode, electrolyte, and their interfaces [20]. MRI can provide spatial information in an opaque cell body [28]. EPR can track the initialization of microstructural growth and the reaction intermediates of electrolyte breakdown and lithium metal without any spectroscopic interference in the signals [29]. MS enables one to study the influence of the electronic environment on nuclear hyperfine energy levels and probe the structure of materials [30].

The main principle of SPM is based on the detection of physical parameters produced by the distance change between the sample and the tip of a fine scanning probe. In situ AFM is available to study the interface of batteries, such as SEI morphology, Li dendrite growth, and mechanical and electrical properties. New analytical tools are developed by modifications of the scanning probe. For example, a periodic high-frequency voltage bias is applied between the cathode and anode, which results in oscillatory surface displacement. With the use of an electrically active AFM probe, ESM can capture oscillatory signals to study the lithium ionic motion at the nanometer scale [31]. Electrolyte-filled nanocapillaries are adopted as the probe for SICM, which can provide an ionic current map with topography [32]. Table 1 summarizes the current main in situ characterization techniques, including their spatial resolution, applicable scenarios and information collected, and the current challenge of each technique.

Table  1.  A summary of current in situ characterization techniques

In situ technique		Spatial resolution	Applicable scenes and obtained information	Facing challenge
X-ray technique	XRD	~μm scale	Quantitatively and qualitatively provide composition, distribution, and conversion of crystallographic structure analyses; study local environment of an element of electrodes.	Amorphous and low atomic number elements in LIBs.
	XAS	1–100 nm
	TXM, XRT	≥15 nm
	XPS	~μm scale
Neutron-based technique	NPD		High penetration and sensitive to low atomic number elements. Precisely determine structural and dynamic information of electrodes.	Amorphous and high atomic number elements in LIBs.
	NR	nm
	NI	~μm scale
	NDP	nm
Electron-based technique	TEM	≥0.1 nm	Provide morphological information of electrodes and SEI; obtain structural information by coupling accessory.	Shallow penetration; high vacuum; high-energy electron damage.
	SEM	10 nm
	EELS
Optical technique	Raman spectroscopy	10 nm–1 μm	Probe the molecule structural features and the chemical composition of electrodes, SEI, and electrolytes. Visualization of macrostructural features.	Specific active component; signal intensity and background interference; bulky or depth analysis.
	FTIR spectroscopy	~μm scale
	OM	~μm scale
Scanning probe technique	AFM	nm	Monitor morphology and changes of surface/interface in diverse environments, SEI, Li dendrite, etc., as well as mechanical properties and electrical properties.	Low resolution
	ESM	10 nm–1μm
	SICM	10 nm–50 nm
Magnetic resonance technique	NMR	mm	Apply to an odd atomic mass or odd atomic number element of LIBs; monitor the local electronic environment around the nucleus through electrodes, electrolytes, and their interfaces.	Specific elements; low resolution;
	EPR
	Mössbauer spectroscopy

下载: 导出CSV 

| 显示表格

Each technique has its advantages and disadvantages, leading to different application scenarios. Hence, different in situ characterization techniques must be combined to conduct comprehensive studies on the understanding of different physical/chemical aspects at multiple length scales. Luo et al. [33] studied the charge compensation in 3d-transition-metal (TM) oxide intercalation cathodes by combining operando MS, soft XAS (SXAS), resonant inelastic X-ray scattering spectroscopy (RIXS), XANES spectroscopy, and Raman spectroscopy (Fig. 2). According to the ¹⁸O-labeled MS results, oxygen was extracted from the lattice of the charging cathodes and not from the electrolyte, which almost entirely generated the capacity to extract charges above 4.5 V. SXAS, RIXS, and XANES spectra were used for the joint investigation of the formation of electron-hole states on oxygen and the dominant mechanism of charge compensation, indicating that localized electron holes formed on the O ions coordinated by Mn⁴⁺ and Li⁺, which promoted the localization and not the formation of true ${\mathrm{O}}_{2}^{2-}$ peroxide and O–O (~0.145 nm). Focusing on O-related questions, various in situ characterization methods provided complementary information from different points of view. This research can guide the design of future materials that can utilize redox reactions on TM cations and oxide anions.

Figure  2.  Combination applications using different in situ techniques: (a) in situ MS resutls of ¹⁸O-labelled Li_xNi_0.13Co_0.13Mn_0.54O₂ cathode and (b) SXAS spectra collected at different states of charge (i–v) in total fluorescence yield mode of Li_1.2Ni_0.13Co_0.13Mn_0.54O₂ cathode; (c) schematic of oxygen anions located in Li_1.2Ni_0.13Co_0.13Mn_0.54O₂: (c1) coordination around O²⁻, (c2) honeycomb-type cation ordering of the TM layer in (c1), (c3) oxygen coordinated by Mn⁴⁺ and Li⁺ in (c2), and (c4) density of states of O⁻/O²⁻ couple . Reproduced by permission from Nature Publishing Group: K. Luo, M.R. Roberts, R. Hao, et al., Nat. Chem., vol. 8, 684-691 (2016) [33]. Copyright 2016.

下载: 全尺寸图片 幻灯片

2.2   Coupling of different in situ characterization technologies

Given that the phenomena pertinent to the functioning of batteries occur over wide spatial and temporal ranges, coupling different in situ characterization technologies with simultaneous measurements at the same spot is favorable to eliminating the information mismatch of time lag. Wang et al. [34] developed an operando TXM–XAS approach to study the dynamic phase transformation evolution and intercalation pathway within multi-particle and individual LiFePO₄ particles during battery operation (Fig. 3(a)–(d)). Therein, TXM images provided direct 2D mapping of the inhomogeneous phase spatial distribution, whereas XANES spectra indicated the composition of each region, corresponding to nearly pure LiFePO₄ and FePO₄ phases, respectively. The single-particle system presented direct evidence of two-phase coexistence for the first time and suggested the two-phase dynamic models of individual LiFePO₄ particles by in operando imaging work. Pérez-Villar et al. [35] exploited an in situ microscopic approach combining Raman spectroscopy and FTIR micro-spectroscopy to investigate the interfacial reactions and spectral characterization of the local structure at the same spot on the surface of electrodes (Fig. 3(e)–(f)). Through micro-FTIR spectroscopy, solvation–desolvation interfacial phenomena were detected, whereas the carbonaceous material was characterized by micro-Raman spectroscopy at the same spot on the glassy carbon surface. This research demonstrated the reliability of the combined microscopy technique and its suitability in obtaining complementary information about both the surface species involved in interfacial reactions. Tsuchiya et al. [36] designed a combined ion-beam analysis of high-energy elastic recoil detection (ERD) and Rutherford backscattering spectrometry (RBS) with 9.0 MeV tetravalent oxygen ion (O⁴⁺) probe beams to study the distribution of Li and its movement around the interface. These typed coupling techniques are based on the same modal or excitation beam, e.g., X-ray, laser, and ion-beam. Given their similar demand and conditions, they are easy to implement by modification of specific devices and cells. Nevertheless, congeneric information alone can be obtained owing to the limitation of the same excitation source.

Figure  3.  Examples of the same modal coupling technique: (a) schematic of operando TXM–XANES approach; (b) TXM image of a multi-particle LiFePO₄ system; (c) XANES of selected regions (squares A and B) at (b); (d) delithiation processes of single-particle LiFePO₄ system revealed by in operando TXM; (e) illustration of the coupling Raman/FTIR microscopy; (f) in situ Raman spectra (f1) and in situ subtractively normalized interfacial Fourier transform infrared spectroscopy (SNIFTIRS) spectra (f2) at various potentials; (g) cross-sectional (g1) and top-view (g2) schematic of in situ ERD and RBS analysis and the beam paths from different directions (SSD represents solid-state detector); (h) RBS (h1) and ERD (h2) spectrum of an all-solid-state battery (LATP represents the solid electrolyte of Li–Al–Ge–Ti–P–O). (a–d) Reproduced by permission from Macmillan Publishers Ltd.: J.J. Wang, Y.C.K. Chen-Wiegart, and J. Wang, Nat. Commun., vol. 5, art. No. 4570 (2014) [34]. Copyright 2014. (e–f) Reproduced from Electrochimica Acta, 106, S. Pérez-Villar, P. Lanz, H. Schneider, and P. Novák, Characterization of a model solid electrolyte interphase/carbon interface by combined in situ Raman/Fourier transform infrared microscopy, 506-515, Copyright 2013, with permission from Elsevier [35]. (g–h) B. Tsuchiya, J. Ohnishi, Y. Sasaki, et al., Adv. Mater. Interfaces, vol. 6, art. No. 1900100 (2019) [36]. Copyright 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.

下载: 全尺寸图片 幻灯片

Techniques should also be coupled based on different modalities [37–39]. Schmidt et al. [38] developed a novel correlative microscopy device by combining Raman microscopy, SEM, and EDX. As shown in Fig. 4(b), the laser beam for the confocal Raman microscope was linked to the SEM chamber through an optical window. SEM provides high-resolution images of surface topography, EDX offers elemental component information, and Raman microscopy delivers molecular bond information, leading to a comprehensive result. Zhang et al. [39] constructed a novel AFM with an environmental TEM (AFM–ETEM) setup to observe in situ growth and measure the stress changes of individual Li whiskers (Fig. 4(c)). In the in situ electro-chemo-mechanical experiment, an arc-discharged carbon nanotube (CNT) was attached to a conducting AFM tip by electron-beam deposition of carbonaceous materials; this assembly was used as the cathode, and the scratched Li metal on the top of a sharp tungsten needle was used as the anode; Li₂CO₃ on the Li surface was used as the solid electrolyte. The morphology and growth process of the whisker can be directly recorded by ETEM, whereas the axial compressive stress can be detected by the AFM tip, corresponding to the critical stress and applied voltage. A sub-micrometer whisker grew at overpotentials at room temperature, generating a growth stress of up to 130 MPa.

Figure  4.  Coupling techniques with different modalities: (a) configuration schematics of the in situ TEM test (a1), high-resolution TEM image of the atomic structure of ω-Li₃V₂O₅ cubic structure (a2), and EELS comparison of the V M_4,5 edges and Li K-edge (a3); (b) illustration of the SEM–Raman microscopy system (b1) and its digital photo (b2); (c) schematic of the AFM–ETEM setup (c1), in situ TEM images of Li wisker growth (c2), stress changes versus applied voltage for eight Li wiskers (c3), and compressive stress–strain curves of six Li wiskers with different growth directions and diameters d₀ (c4). (a) Reproduced with permission from J.K. Zhu, H.H. Shen, X.B. Shi, et al., Revealing the chemical and structural evolution of V₂O₅ nanoribbons in lithium-ion batteries using in situ transmission electron microscopy, Anal. Chem., 91(2019), No. 17, p. 11055 [37]. Copyright 2019 American Chemical Society. (b) Reproduced from R. Schmidt, H. Fitzek, M. Nachtnebel, C. Mayrhofer, H. Schroettner, and A. Zankel, Macromol. Symp., vol. 384, art. No. 1800237 (2019) [38]. Copyright 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission. (c) Reproduced from L.Q. Zhang, T.T. Yang, C.C. Du, et al., Nat. Nanotechnol., vol. 15, 94-98 (2020) [39]. Copyright © 2020 The Author(s), under exclusive licence to Springer Nature Ltd.

下载: 全尺寸图片 幻灯片

To study the lithium diffusion behaviors in LB systems, our group successfully designed and developed a multi-model in situ characterization system coupling NMR technique and Raman spectrum (Fig. 5). The laser can enter into the magnetic field and reach the targeted materials across the optical window through an optical fiber and optical extension pipe, whereas the Raman signals can be collected by the objective lens and reach the spectrograph through the optical fiber. The interference can be avoided by controlling the technical parameters of the instrument and selecting the proper ingredients. Using our system, the Raman spectra of electrodes, NMR signals, or MRI of ¹H or ⁷Li can be obtained simultaneously during the discharge/charge cycling. Raman spectra can reveal structural conversion at the molecular level and deduce electrochemical reaction and kinetics process; NMR data can provide the local environment information of hydrogen and lithium element, whereas MRI can reveal the morphological information about electrodes, electrolytes, and their interfaces. Owing to synchronous data acquisition, the difference in temporal resolutions between NMR and Raman techniques is reduced, which favors the accurate and comprehensive understanding of the battery failure mechanism.

Figure  5.  (a) Schematic and (b) digital photographs of the multi-model coupling design of the NMR system and Raman spectrograph developed by our group.

下载: 全尺寸图片 幻灯片

The multi-model coupling of characterization techniques, which has a deep significance, proceeds beyond the simple addition of different methods. In practice, different characterizations have varying spatial resolutions from nm to the µm scale while also reaching the mm and cm levels and temporal resolutions from several seconds to minutes, hours, or days. To eliminate the influence of acquisition time, scientists artificially set the testing conditions and parameters to retain the target in equilibrium or stable state. Given that the system of batteries is complex, dynamic, and interrelated, the integration of the information collected from each test is a huge challenge. The concept of “in situ” in multi-model coupling needs to be defined and distinguished and requires the consistency of spatial and temporal scales. The characterization sites or zones are spatially unchanged in a complex system, depending on the spatial resolutions of the characterization technique. The applicable objectives are stable during the acquisition period, relying on temporal resolutions. The multi-model coupling has the following features. First, it is favorable to the simultaneous collection of different information, compensating for the applicable objectives limited to a single technology. Then, the results obtained can cover a broad scale range from the atomic to microcosmic, mesoscale, and macroscopic scales, building up the spatial link for whole systems. More importantly, the multi-model coupling can trace the real and complete dynamic change of a process, eliminating the error of temporal mismatching one at a time. The multi-model coupling will be the growing trend of in situ techniques in the future, but it is still in its infancy. For example, temporal resolutions are not only related to the parameters of each instrument but also the specific materials and process of charge/discharge cycling, which is more complex than spatial-temporal. Finding solutions to the temporal resolutions of different characterization techniques is a huge challenge. Hence, great effort needs to be exerted to address this bottleneck of technology.

2.3   Design of electrochemical cells for in situ characterization

To conduct in situ measurements successfully, scholars should carefully optimize the in situ electrochemical cells to properly fit the specific analytical techniques. As shown in Fig. 6, some typical setups and schematics of in situ cells have been reported for different types of in situ experiments [25,40–46]. Some basic principles are followed for their design [47]. First, the design of the in situ cell needs to have an environment similar to that in a commercial or a coin cell. For example, an open-type cell without any covering is adopted for in situ TEM measurement. The excited radial wave can reach the sample region and return to the detector across the packages by various transparent windows. Polyethylene, beryllium, and Kapton film are used as X-ray windows, whereas glass, quartz, sapphire, and CaF₂ are suitable for the in situ Raman test. Meanwhile, metallic parts should be used as sparingly as possible in NMR. Third, the interference from inactive cell components to outgoing detected signals should be at the minimum to truly collect information under in situ operating conditions. The interaction of the particle beam with cell components often generates background signals, which need to be precisely removed afterward from the total signals. Importantly, radiation damages and side effects on the material should be reduced, such that the magnetic field orientation in NMR has a direct influence on the paramagnetic cathode materials, and the heat produced by laser during in situ Raman test may lead to the decomposition of electrolytes. Finally, the in situ setup should be easy to assemble/disassemble and highly reproducible.

Figure  6.  Typical setups and schematic of in situ cells: (a) in situ XRD cell designs (AMPIX means Argonne’s multi-purpose in situ X-ray); (b) in situ neutron diffraction cell designs and photo images; (c) in situ battery for the Raman test; (d) in situ battery for the FTIR test (IR represents infrared reflection); (e) cylindrical in situ NMR cell; (f) in situ TEM electrochemical cell with ionic liquid (IL) and solid Li₂O electrolytes; (g) in situ focused ion beam SEM design; (h) in situ OM cell, where CE, WE, and PE represent counter electrode, working electrode, and polyethylene, respectively. (a) Reproduced from O.J. Borkiewicz, B. Shyam, K.M. Wiaderek, C. Kurtz, P.J. Chupas, and K.W. Chapman, J. Appl. Crystallogr., vol. 45, 1261-1269 (2012) [40]. Copyright International Union of Crystallography. (b) Reproduced from L. Vitoux, M. Reichardt, S. Sallard, P. Novák, D. Sheptyakov, and C. Villevieille, Front. Energy Res., vol. 6, art. No. 76 (2018) [41]. Copyright © 2018 The Author(s). (c) Reproduced from C. Ghanty, B. Markovsky, E.M. Erickson, et al., ChemElectroChem, vol. 2, 1479-1486 (2019) [42]. Copyright 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission. (d) Reproduced from J. Power Sources, 243, K. Hongyou, T. Hattori, Y. Nagai, T. Tanaka, H. Nii, and K. Shoda, Dynamic in situ Fourier transform infrared measurements of chemical bonds of electrolyte solvents during the initial charging process in a Li ion battery, 72-77, Copyright 2013, with permission from Elsevier [25]. (e) Reproduced from Electrochem. Commun., 13, F. Poli, J.S. Kshetrimayum, L. Monconduit, and M. Letellier, New cell design for in situ NMR studies of lithium-ion batteries, 1293-1295, Copyright 2011, with permission from Elsevier [43]. (f) Reproduced with permission from M. Gu, L.R. Parent, B.L. Mehdi, et al., Demonstration of an electrochemical liquid cell for operando transmission electron microscopy observation of the lithiation/delithiation behavior of Si nanowire battery anodes, Nano Lett., 13(2013), No. 12, p. 6106 [44]. Copyright 2013 American Chemical Society. (g) X. Zhou, T. Li, Y. Cui, Y. Fu, Y. Liu, and L. Zhu, In situ focused ion beam scanning electron microscope study of microstructural evolution of single tin particle anode for Li-ion batteries, ACS Appl. Mater. Interfaces, 11(2019), No. 2, p. 1733 [45]. Copyright 2019 American Chemical Society. (h) Reproduced from Electrochim. Acta, 136, J. Steiger, D. Kramer, and R. Mönig, Microscopic observations of the formation, growth and shrinkage of lithium moss during electrodeposition and dissolution, 529-536, Copyright 2014, with permission from Elsevier [46].

下载: 全尺寸图片 幻灯片

3.   In situ characterization of LBs

3.1   In situ characterization of cathodes

Cathode materials are one of the most important factors that determine the performance of LBs. Thus, the geometric structure, phase structure, and composition evolution of cathode materials must be traced during battery operation through in situ characterization.

3.1.1   Lithium-rich cathode materials

Lithium-rich cathode materials, such as lithium-rich manganese-based cathode materials (xLi₂MnO₃·(1−x)LiTMO₂, TM = Ni, Mn, Co, etc.), have extremely high theoretical specific capacities (>350 mAh·g⁻¹) and reversible specific capacities (>250 mAh·g⁻¹) and are considered to be the most promising next-generation Li-ion battery (LIB) cathode materials [48]. However, Li-rich Mn-based cathode materials also face problems, such as the low initial coulombic efficiency, poor multiplicity performance, and severe capacity and voltage decay, which hinder their commercialization. In situ characterization means can help in revealing some internal mechanisms, which can provide new insights and ideas for the future development of Li-rich cathode materials.

Currently, the mechanisms involved in the activation of high-energy LiNi_xCo_yMn_(1−x−y)O₂ (NCM) remain intensely debated. In this context, Lanz et al. [49] compared the changes in the Raman spectra of stoichiometric and high-energy NCM during electrochemical cycling under the same conditions for the first time. Ex situ Raman measurements of NCM with increased overlithiation revealed a shift of the A_1g band toward the A_g band of Li₂MnO₃, whereas no such movement was observed after electrochemical cycling, thus lending support to the Li₂MnO₃ domain model and the irreversibility of Li₂MnO₃ activation, respectively (Fig. 7(a)). Finally, the in situ Raman spectra of stoichiometric and high-energy NCM showed a new reversible band at ~545 cm⁻¹, which was stable over a larger potential range in the latter compound.

Figure  7.  (a) In situ Raman spectra (left) of stoichiometric NCM and corresponding contour plot (right) during the first charge. (b) In situ XRD patterns of LLNMO electrode in the first two cycles (left) and their O 1s XPS spectra at the first charge etching depths (right). (c) Voltage profile (left) and in situ Li NMR spectra (right) of a Li_1.08Mn_1.92O₄ during the first cycle. (a) Reproduced from Electrochim. Acta, 130, P. Lanz, C. Villevieille, and P. Novák, Ex situ and in situ Raman microscopic investigation of the differences between stoichiometric LiMO₂ and high-energy xLi₂MnO₃·(1−x)LiMO₂ (M = Ni, Co, Mn), 206-212, Copyright 2014, with permission from Elsevier [49]. (b) Reproduced from Energy Storage Mater., 38, X. Cao, H.F. Li, Y. Qiao, M. Jia, P. He, J. Cabana, and H.S. Zhou, Achieving stable anionic redox chemistry in Li-excess O2-type layered oxide cathode via chemical ion-exchange strategy, 1-8, Copyright 2021, with permission from Elsevier [50]. (c) Reproduced from L.N. Zhou, M. Leskes, T. Liu, and C.P. Grey, Chem. Int. Ed., vol. 54, 14782-14786 (2015) [51]. Copyright 2015 Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.

下载: 全尺寸图片 幻灯片

Triggering oxygen redox activity is considered a promising strategy to increase the capacity of cathode materials for LBs. However, the irreversible loss of lattice oxygen exacerbates the structural distortion of the spinel phase, leading to severe voltage and capacity decay. Cao et al. [50] evaluated the reversibility of the structural evolution of Li-excess O2-type layered oxide cathode, Li_0.66[Li_0.12Ni_0.15Mn_0.73]O₂ (LLNMO) in the first two Li⁺ (de)intercalation cycles through systematic in situ XRD. The in situ XRD results revealed that the evolution of the peak position at the second charge showed an opposite trend relative to that of the initial discharge. In addition, the layered structure and O2-type oxygen stacking can be maintained after 50 cycles (Fig. 7(b)). The two pseudo-plateaus located at 3.67 and 2.73 V were well preserved, and the voltage decay can be effectively suppressed during long cycling.

Moreover, lithium diffusion is one of the key criteria for achieving high-magnification and high-power batteries. Zhou et al. [51] demonstrated how in situ NMR methods can be adapted to probe the mobility of lithium and the nature of electrode phase transitions in real time during cell cycling (Fig.7(c)). The results showed that the center of mass of the broad static resonance was close to the center of mass of the isotropic resonance observed during rotation. The resonance at about chemical shift (δ) = 250 ppm originated from the lithium metal negative electrode. The spinel resonance shifted to high frequencies and decreased in intensity as lithium was removed from the tetrahedral sites, and Mn³⁺ was oxidized to Mn⁴⁺ during the charging process. The electrochemical trend and shift of the spinel peak position were reversed during discharge.

3.1.2   Nickel-rich cathode materials

Nickel-rich ternary materials possessing a nickel content generally greater than 50wt% are one of the cathode materials used for high-energy-density LIBs. Given the high nickel content, nickel-rich ternary cathode materials have a high reversible discharge capacity. However, numerous problems also occur, and they include unstable surface properties, structural defects and lithium–nickel mixing, intergranular cracks, and microstrain, which lead to poor cycle performance.

To solve the above problems, researchers have carried out a number of studies. Seong et al. [52] suggested a simple in situ method to control the residual lithium chemistry of a high nickel lithium layered oxide, Li(Ni_0.91Co_0.06Mn_0.03)O₂ (NCM9163), with minimal side effects. Further surface-specific analyses were performed using XPS depth analysis and time-of-flight secondary-ion MS to determine the chemical properties of SO₂-treated NCM9163 (Fig. 8(a)). As observed in the XPS curves, the sulfur signal was not detected in the SO₂-treated NCM9163 until 10 min after milling but not in the original NCM9163. They observed that if the preferred reaction is the formation of a low alkalinity lithium compound, then it mitigates the formation of the predominantly basic LiOH or Li₂CO₃.

Figure  8.  (a) Chemical analysis of SO₂-treated NCM9163 particle by in situ XPS spectra (left) and differential electrochemical MS (DEMS) (right) results. (b) Dissolution of TMs from the NMC622 cathode via in situ X-ray diffractograms (left) and their lattice parameter c/a ratio with the change in the delithiation process (right). (c) Schematic of the in situ TEM setup (left), corresponding low-magnification high-angle annular dark-field (HAADF)–scanning TEM (STEM) images (middle), and atomic structure (right) around the phase boundaries of LiNi_0.76Mn_0.14Co_0.10O₂ (NMC76) with the corresponding phase orientations. (a) W.M. Seong, K.H. Cho, J.W. Park, et al., Angew. Chem. Int. Ed., vol. 59, 18662-18669 (2020) [52]. Copyright 2020 Wiley-VCH GmbH. Reproduced with permission. (b) R. Jung, F. Linsenmann, R. Thomas, et al., J. Electrochem. Soc., vol. 166, art. No. A378 (2019) [53]. Copyright 2019 © The Author(s), published by ECS. Reproduced with permission. All rights reserved. (c) S. Li, Z.P. Yao, J.M. Zheng, et al., Angew. Chem. Int. Ed, vol. 59, 22092-22099 (2020) [54]. Copyright 2020 Wiley-VCH GmbH. Reproduced with permission.

下载: 全尺寸图片 幻灯片

Otherwise, the dissolution of TMs from the cathode active material and deposition on the anode will cause significant cell aging. Jung et al. [53] investigated the dissolution of TMs from an Li_1−xNi_0.6Mn_0.2Co_0.2O₂ (NMC622) cathode via in situ and ex situ XRD (Fig. 8(b)). The analysis showed that the aging mechanism for all three metals was the loss of recyclable lithium in graphite SEI. This loss was greater when Mn was present in the electrolyte compared with Ni and Co due to the higher activity of the deposited Mn toward SEI decomposition compared with Ni and Co.

For nickel-rich cathode materials, the role of crystal defects on their structural evolution and subsequent property degradation during electrochemical cycling is unclear. Li et al. [54] studied the structural evolution of Ni-rich LiNi_1−x−yMn_xCo_yO₂ (NMC) cathodes using Li_0.5La_0.5TiO₃ (LLTO) as solid-state electrolyte via in situ TEM (Fig. 8(c)). They identified that the antiphase boundary (APB) and twin boundary (TB) separating layered phases played an essential role during phase change. When lithium was depleted, the APB extended across the laminar structure, and the mixing of Li/TM ions in the lamellar phase induced the formation of rock-salt phases along the coherent TB.

3.1.3   Sulfur-based cathode materials

With the advantages of high specific energy, low cost, and abundant resources, Li–S batteries have received widespread attention from the research community. On the one hand, compared with current commercial cathode materials, monolithic sulfur provides 8–10 times higher theoretical specific capacity. On the other hand, monomass sulfur is abundantly stored and environmentally friendly. However, numerous unresolved issues exist with Li–S batteries. First, the insulating nature of S and Li₂S leads to poor kinetics of the conversion process and low actual capacity utilization. Second, as an intermediate, lithium polysulfides (LiPSs) have a significant “shuttle effect” on most organic liquid electrolytes, where soluble sulfur shuttles between the cathode and anode, deteriorating the reaction interface. In addition, the severe volume change (80%) leads to the degradation of the mechanical properties of the sulfur cathode [55]. The current research on Li–S batteries focuses on suppressing the polysulfide shuttle effect, preventing lithium dendrite growth, and improving the conversion reaction kinetics.

The immobilization of sulfur and LiPSs in various host materials is a common strategy used to overcome some of the problems of the sulfur cathode. However, a phenomenon in which some hosts themselves can be lithiated−delithiated occurs during the cycling process. Liu et al. [56] proposed the concept of dynamic hosts for Li–S batteries for the first time and elucidated the mechanism through which TiS₂ acts in such a fashion via in situ XRD (Fig. 9(a)). TiS₂ was completely converted to LiTiS₂ during the initial discharge process. Upon charging, LiTiS₂ was converted to Li_xTiS₂ (0 < x < 1) and disappeared. During the next charging process, the diffraction peak of Li₂S also disappeared. Similar behavior was observed in the subsequent cycles.

Figure  9.  (a) In situ XRD patterns of a TiS₂ electrode at 50 mA·g⁻¹ during the initial four cycling processes (left) and schematic of the mechanism (right). (b) Galvanostatic curves (left) and the corresponding XRD contour plot for the Li–S battery (right). (c) In situ Raman spectroscopy (left) in 1 M LiTFSI and TEGDME/DIOX (1/1, volume ratio) electrolyte and normalized intensity of S₈ and short-chain polysulfides with the increase in time (right). (d) Schematic of the experimental setup of the solid cell applied with a MEMS heating accessory for in situ TEM observation (left) and TEM images (right) showing the lithiation of S@CNT. (a) Reproduced with permission from X.C. Liu, Y. Yang, J.J. Wu, et al., Dynamic hosts for high-performance Li–S batteries studied by cryogenic transmission electron microscopy and in situ X-ray diffraction, ACS Energy Lett., 3(2018), No. 6, p. 1325 [56]. Copyright 2018 American Chemical Society. (b) Reproduced by permission from Nature Publishing Group: J. Conder, R. Bouchet, S. Trabesinger, C. Marino, L. Gubler, and C. Villevieille, Nat. Energy, vol. 2, art. No. 17069 (2017) [57]. Copyright 2017. (c) Reproduced with permission from H.L. Wu, L.A. Huff, and A.A. Gewirth, In situ Raman spectroscopy of sulfur speciation in lithium-sulfur batteries, ACS Appl. Mater. Interfaces, 7(2015), No. 3, p. 1709 [58]. Copyright 2015 American Chemical Society. (d) Z.F. Wang, Y.F. Tang, L.Q. Zhang, M. Li, Z.W. Shan, and J.Y. Huang, Small, vol. 28, art. No. 2001899 (2020) [59]. Copyright 2020 Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.

下载: 全尺寸图片 幻灯片

Another major problem with Li–S batteries is that the mechanistic understanding of the reaction behind the polysulfide shuttle is still incomplete. Conder et al. [57] identified the characteristics of polysulfides adsorbed on the surface of glass fiber diaphragms and monitored their evolution during cycling via operando XRD (Fig. 9(b)). The results showed that only the long-chain polysulfides were adsorbed on the glass fiber surface, and they remained on it during the repeated lithiation/delithiation processes. These findings not only demonstrated that XRD can be used to study liquid–liquid interactions in Li–S batteries but also showed that fumed silica is a promising electrolyte additive that can significantly improve the performance of Li–S batteries.

Sulfur reduction is a stepwise electrochemical process with a series of highly soluble polysulfides as intermediate species. Wu et al. [58] examined the discharge and charge processes of a Li−S cathode in 1 M lithium bis(trifluoromethane sulfonyl)imide (LiTFSI) and tetraethylene glycol dimethyl ether (TEGDME)/1,3-dioxolane (DIOX) (1/1, volume ratio) via in situ Raman spectroscopy (Fig. 9(c)). Raman spectroscopy showed that ${\mathrm{S}}_{8}^{2-}$ was formed via the S₈ ring opening in the first reduction process at ~2.4 V and short-chain polysulfides, such as ${\mathrm{S}}_{4}^{2-}$, ${\mathrm{S}}_{4}^{-}$, and ${\mathrm{S}}_{3}^{\cdot -}$. With the continued discharge at ~2.3 V in the second reduction process, ${{\mathrm{S}}_{2}\mathrm{O}}_{4}^{2-}$ was observed. Wang et al. [59] investigated the precipitation and decomposition processes of Li₂S at high temperatures using in situ TEM techniques implemented in a microelectromechanical system (MEMS) heating device. The results showed that Li₂S transformed from the amorphous/nanocrystalline state to the polycrystalline state with the electrochemical lithiation process at room temperature, and the precipitation of Li₂S was closer to complete at high temperature than at room temperature (Fig. 9(d)). In addition, with the increase in Li⁺ ion conductivity at high temperatures, the decomposition of Li₂S, which is difficult to achieve at room temperature, occurred easily.

3.2   In situ characterization on anodes

Anode materials are another important factor affecting the energy density and cycle life of batteries. Based on the mechanism of lithium-ion storage, anode materials can be divided into intercalation, conversion, and alloying-type anodes [60]. However, the dynamic evolution of the microstructure, morphology, phase, and chemical composition of anode materials is not very well understood. Therefore, the chemical, material, physical, and mechanical information on LIBs can be obtained by using advanced in situ characterization techniques, which are conducive to further material design and performance improvement [61].

3.2.1   Intercalation-type anodes

Graphite, which is a classical intercalation anode material, is the most commonly used commercial anode material [62]. A series of phase transitions and volume changes is caused by ion intercalation in graphite, and numerous controversial issues about its intercalation mechanism remain. Thus, in situ characterization is beneficial to understanding the working mechanism of graphite [63].

Tao et al. [16] studied the anisotropic deformation and dynamic microstructural evolution of graphite-based commercial electrodes using a combination of in situ SEM digital image correlation. The electrode exhibited an irreversible swelling of 50% in the initial cycle and a reversible anisotropic expansion and contraction deformation process in the subsequent cycle (Fig. 10(a)). The vertical swelling deformation was evident, and the electrode–separator interface was the main area of lithiation and delithiation. In addition, the diffusion path of lithium-ion was obtained based on the evolution of the strain field.

Figure  10.  (a) Microstructural evolution and anisotropic deformation tested by in situ SEM images (a1) and two-dimensional (2D) strain field mappings along z direction (a2) of graphite-based electrodes, where ε_zz represents the maximum strain along the z-axis. (b) Typical lithium intercalation stages and phase transition regimes (b1), evolution of the 002 reflection (b2), and the phase fraction (b3) and total volume change (b4) of graphite vs. lithium content. (c) In situ XRD patterns of ZrNb₁₄O₃₇ NWs in the first cycle (c1), changes in the main reflections peaks (c2 and c4) during cycling, and evolution of lattice parameters in the first cycle (c3); where a, b, and c are cell parameters, and V is cell volume. (d) Relative phase content of the LiFePO₄ (blue), Li₄Ti₅O₁₂ (red), and delithiated FePO₄ (dark blue) phases (d1) and Li 8a (red) and 16c (blue) sites occupation and total Li content (green) in LTO-2 during charge–discharge cycling (d2). (a) Reproduced from Extreme Mech. Lett., 35, R. Tao, J.G. Zhu, Y.F. Zhang, W.L. Song, H.S. Chen, and D.N. Fang, Quantifying the 2D anisotropic displacement and strain fields in graphite-based electrode via in situ scanning electron microscopy and digital image correlation, 100635, Copyright 2020, with permission from Elsevier [16]. (b) Reproduced with permission from S. Schweidler, L.d. Biasi, A. Schiele, P. Hartmann, T. Brezesinski, and J. Janek, Volume changes of graphite anodes revisited: A combined operando X-ray diffraction and in situ pressure analysis study, J. Phys. Chem. C, 122(2018), No. 16, p. 8829 [64]. Copyright 2018 American Chemical Society. (c) Reproduced with permission from Y.H. Li, R.T. Zheng, H.X. Yu, et al., Observation of ZrNb₁₄O₃₇ nanowires as a lithium container via in situ and ex situ techniques for high-performance lithium-ion batteries, ACS Appl. Mater. Interfaces, 11(2019), No. 25, p. 22429 [65]. Copyright 2019 American Chemical Society. (d) Reproduced with permission from W.K. Pang, V.K. Peterson, N. Sharma, J.J. Shiu, and S.H. Wu, Lithium migration in Li₄Ti₅O₁₂ studied using in situ neutron powder diffraction, Chem. Mater., 26(2014), No. 7, p. 2318. [66]. Copyright 2014 American Chemical Society.

下载: 全尺寸图片 幻灯片

Schweidler et al. [64] measured the phase diagram of Li_xC₆ by in situ XRD and gave the single-phase and coexisting phase regions. The study revealed that when C₆ was completely lithiated to LiC₆, the total volume expansion was 13.2%, of which about 5.9% occurred in the early dilution stage. The remaining expansion was due to the transition from phase 2 to phase 1. As a result, lithium was thought to intercalate indiscriminately in graphite, but it intercalated only in one layer in every two layers, leaving all graphene in the same chemical state (Fig. 10(b)). This study provided valuable information indicating that graphite undergoes considerable crystalline strain under partial lithiation.

Li et al. [65] fabricated ZrNb₁₄O₃₇ nanowires (NWs) by electrostatic spinning, and they were used as anodes for LIBs. The basic reaction mechanism of lithium/delihiation was studied by in situ XRD. Fig. 10(c) shows the in situ XRD system used for the study; it consisted of an XRD device and a LAND CT2001A battery tester. The discharge process can be divided into three different regions based on the behavior of the main peak of the in situ XRD pattern. State I was the solid solution reaction region, in which 5.25 Li⁺ insertion occurred in each molecular formula unit of ZrNb₁₄O₃₇ to form the intermediate Li_5.25ZrNb₁₄O₃₇. State II was a two-phase reaction in which an additional 2.87 Li⁺ insertion into the octahedral frame occurred, indicating the transition from Li_5.25ZrNb₁₄O₃₇ to Li_8.12ZrNb₁₄O₃₇. State III was the second solution reaction, in which 12.84 Li⁺ insertion into each formula unit occurred to form the Li_20.96ZrNb₁₄O₃₇ phase. When the battery was fully charged to 3 V, the characteristic peak almost returned to the original position, indicating that its structure had high stability and reversibility.

Pang et al. [66] used in situ NPD to study the transference of Li⁺ in lithium titanate (LTO) anodes with different particle sizes during the battery operation (Fig. 10(d)). Li⁺ migrated from 8a site to 16c site via 32e site during the charging process, accompanied by the rearrangement of oxygen positions in the structure. The increase in the number of Li⁺ migrations was due to the shortening of the path length at the smaller particle size anode rather than a change in the Li⁺ migration path. Li replenishment at position 8a was faster than Li transfer to position 16c. Compared with LTO-1, which had a large particle size, LTO-2 had a smaller particle size and contained a larger proportion of lithium sites.

3.2.2   Conversion reaction-type anodes

Compared with intercalation anode materials, conversion anode materials, mainly including TM oxides, sulfides, nitrides, and phosphides, have a high theoretical specific capacity.

He et al. [67] employed a strain-sensitive, bright-field STEM to study the physical and chemical mechanism of nano-Fe₃O₄ in real time. Explicit visualization of the two-step intercalation process of lithiation in Fe₃O₄ nanocrystals revealed that the initial lithium intercalation formed as two phases of LiFe₃O₄ (Fig. 11(a)). Further lithiation led to a conversion reaction, which resulted in the coexistence of three different phases in a single nanoparticle (NP).

Figure  11.  (a) Phase evolution of Fe₃O₄ nanocrystals tested by in situ TEM images: (a1) pristine Fe₃O₄ (red), Li-inserted Li_xFe₃O₄ (blue), and Fe+Li₂O composite after conversion (green). Scale bar, 20 nm; (a2) time-dependent projected areas of the three phases; (a3) propagation speeds of intercalation and conversion. (b) Lattice evolution of the MnO₂ electrode tracked by in situ XRD based on synchrotron X-ray scattering: (b1) voltage profile for the first cycling at 30 mA·g⁻¹; (b2) in situ XRD patterns. (c) SEM image of a nanomechanical device (c1); illustration for in situ tension of a single SnO₂ NW using this device (c2); in situ tensile snapshots before the test and after fracture (c3–c5). (d) ⁷Li in situ NMR spectra and cell voltage of (d1) a-SiO and (d2) d-SiO (1000°C). (a) Reproduced from K. He, S. Zhang, J. Li, et al., Nat. Commun., vol. 7, art. No. 11441 (2016) [67]. © The Author (s). (b) K. He, Y.F. Yuan, W.T. Yao, et al., Angew. Chem. Int. Ed., vol. 61, art. No. e202113420 (2022) [68]. Copyright 2021 Wiley-VCH GmbH. Reproduced with permission. (c) Reproduced from Nano Energy, 53, B. Song, P. Loya, L.L. Shen, et al., Quantitative in situ fracture testing of tin oxide nanowires for lithium ion battery applications, 277-285, Copyright 2018, with permission from Elsevier [69]. (d) Reproduced with permission from K. Kitada, O. Pecher, P.C.M.M. Magusin, M.F. Groh, R.S. Weatherup, and C.P. Grey, Unraveling the reaction mechanisms of SiO anodes for Li-ion batteries by combining in situ 7Li and ex situ 7Li/²⁹Si solid-state NMR spectroscopy, J. Am. Chem. Soc., 141(2019), No. 17, p. 7014 [70]. Copyright 2019 American Chemical Society.

下载: 全尺寸图片 幻灯片

He et al. [68] first synthesized a 1D pure β-MnO₂ phase with a 1×1 tunnel structure and used it as a cathode for a series of in situ measurements to reveal its energy storage mechanism. The bulk phase changes of β-MnO₂ crystal structure during the cyclic process were studied by in situ XRD. The instability of the frame structure was caused by the incomplete reversible transition of β-MnO₂ to the orthogonal LiMnO₂ phase during cyclic lithiation, which led to capacity decay (Fig. 11(b)). In addition, in situ TEM analysis was performed on the lithium storage behavior of β-MnO₂ nanorods to understand their microstructural changes at the microscale. The findings of this study can provide better guidance for the structural regulation of such materials to further optimize lithium storage properties.

At present, the in situ characterization techniques for tensile mechanical properties and fracture mechanisms of electrode materials are very limited. Song et al. [69] quantitatively studied the tensile properties of electrochemically modified SnO₂ NWs by in situ SEM. The results showed that the transformation of crystal to glass during the lithiation process resulted in large plastic deformation (Fig. 11(c)). The lithiation and delithiation processes, on the other hand, reduced the mechanical properties of SnO₂ NWs. Notably, the mechanical properties of delithiation SnO₂ NWs were generally higher than those of lithiation. This study promotes the understanding of the mechanical properties of SnO₂-based nanomaterials in LIBs.

Kitada et al. [70] observed the electrochemical reaction process of silicon-based materials by in situ NMR. By comparing the electrochemical processes of amorphous lithium silicide, amorphous SiO, and disproportionated silicon monoxide obtained by heat treatment at different temperatures, they pointed out that amorphous silicon oxide has better cyclic properties than crystalline silicon (Fig. 11(d)). This reaction did not follow the typical two-phase reaction but was a gradual solid solution process, thus avoiding the electrode deterioration caused by volume mutation.

3.2.3   Alloying-type anodes

Si is a typical representative of alloying-type anode materials, and it has a high theoretical specific capacity (4200 mAh·g⁻¹) and reserve abundance in the earth’s crust; thus, it is considered an ideal choice for the next generation of LIB anode materials [71]. However, the large volume change during the process of lithium intercalation is one of the main reasons leading to poor cycle performance. Given the limited spatial resolution of characterization methods, the microscopic mechanism of improving the performance of Si anode is still unclear.

Zhao et al. [72] developed a simple and controllable methodto embed multiple Si NPs as cores into hollow multishelled structures (HoMS), and in situ TEM technology was used to study the dynamic morphological evolution of Si NPs during lithiation (Fig. 12(a)). The overall structure of the electrode remains intact during Li⁺ repeated insertion/extraction process. This demonstrated that the inner shell can effectively limit the direction of core expansion, whereas the outer shell maintained a stable interface and overall structural integrity. Therefore, the result showed the excellent electrochemical performance of the electrode material in LIBs. In addition, in situ Raman spectroscopy was used to detect the electrochemical changes in Si NP@2S-CoFe₂O₄ HoMS at different potentials. The peak intensity of the Si–Si bond (516.2 cm⁻¹) decreased with the development of lithiation. When the discharge reached 0.01 V, the peak disappeared completely, indicating the formation of Si–Li alloy.

Figure  12.  (a) Schematic of the in situ TEM device (a1); time-dependent images of Si NP@2S-CoFe₂O₄ during lithiation (a2) and delithiation (a3) processes, respectively; (a4) voltage variation during lithiation and digital image of the cell device (insert image); (a5) Raman patterns at different lithiation states. (b) In situ NDP spectra showing Li transport of the Sn electrode (b1); lithiation spectra plotted every 60 min interval from 20 min to 740 min at 0.4 V versus Li/Li⁺ (b2). (c) Morphological evolutions of a single Ge_0.9Se_0.1 particle by in situ TXM images at different lithiated states (c1); XANES spectra from TXM images (c2). (a) J.L. Zhao, J.Y. Wang, R.Y. Bi, et al., Angew. Chem. Int. Ed., vol. 60, 25719-25722 (2021) [72]. Copyright Wiley-VCH GmbH. Reproduced with permission. (b) D.X. Liu, J.H. Wang, K. Pan, et al., Angew. Chem., vol. 126, 9652-9656 (2014) [73]. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission. (c) T.Y. Li, X.W. Zhou, Y. Cui, et al., ChemSusChem, vol. 14, 1370-1376 (2021) [74]. Copyright Wiley-VCH Verlag GmbH. Reproduced with permission.

下载: 全尺寸图片 幻灯片

Liu et al. [73] investigated the formation of lithium ions in Sn electrodes by in situ NDP and monitored the transfer of lithium ions in the electrode (Fig. 12(b)). In situ NDP enables researchers to understand the transport rate and distribution of Li⁺, thus guiding the development of effective storage mechanisms for materials. It complements other material characterization methods and helps us realize the complex interrelationships among electrochemistry, reaction dynamics, intercalation, and migration. This technique is not limited to materials formed between metals but can also be used to study a wide range of anode and cathode materials.

Li et al. [74] studied the dynamic morphology and phase change of selenium-doped germanium at the single-particle level by synchrotron-radiation in situ TXM. The experimental results showed that the micron-sized Ge_0.9Se_0.1 particles had good reversibility at a 1 C rate, which promotes the learning of their better cycling performance and admirable rate capacity (Fig. 12(c)). The in situ operation technology based on single-particle batteries not only avoids the influence of other components in the coin battery on the active material and the damage induced by X-ray but also provides a reference for understanding the dynamic characteristics of battery materials during the particle-level charge and discharge process.

3.2.4   Lithium metal anodes

Lithium metal is considered to be an ideal material for high-energy-density secondary batteries owing to its low potential platform (−3.04 V vs. standard hydrogen electrode) and high theoretical capacity (3860 mAh·g⁻¹) [75]. However, lithium dendrite growth and consequent serious safety problems seriously limit its commercial applications.

Li et al. [76] designed and formed a microscopic-ordered 3D electrode structure on the surface of a garnet electrolyte by electron-beam evaporation and ion-beam etching method. In situ NDP technique with high spatial resolution and detection sensitivity was used to quantify lithium deposition. Its detection principle is shown in Fig. 13(a). The particles emitted by different depth paths had different energy losses and carried various energies. Based on this condition, the depth distribution information of lithium in the Z direction can be determined to distinguish the deposition distribution of lithium under the hole and Ti film. The voltage and current information of the battery were collected via in situ charge and discharge, and the energy information of outgoing particles was collected synchronously. The discharge capacity and total integral intensity were consistent, indicating the sensitivity and reliability of collecting information on lithium atoms by this method.

Figure  13.  (a) In situ NDP signal count (dark cyan) curves (a1), contour map of in situ NDP spectra collected at 3 min intervals during cycling (a2), and integrated curves of NDP data collected every 1 h (a3). (b) Schematic of in situ X-ray imaging setup with a V-slot Li-electrode holder (b1) and evolution of Li plating at 0.5 (left) and 10.0 mA·cm⁻² (right) (b2); inset of the present voltage profiles of Li plating at different current rates. (c) Cryo–EM image Li metal dendrites (c1) and TEM images of Li metal dendrites exposed to the air (c2) and their electron dose (c3); timed images of Li dendrite from the green-outlined area in (c1) after continuous electron-beam irradiation (c4–c6); (c7–c9) faceting behavior of Li metal dendrites and their corresponding selected-area electron diffraction (SAED) patterns (insets) growing along <111> (c7), <110> (c8), and <211> (c9). (d) ⁷Li NMR spectra obtained during the Li plating process (d1); stack plot of the Li chemical shift spectra (d2) during plating in the first and fourth cycles; (d3) comparison of the fitted spectra at the end of charge in the first and fourth cycles. (a) Reproduced from Nano Energy, Vol 63, Q. Li, T.C. Yi, X.L. Wang, et al., In-situ visualization of lithium plating in all-solid-state lithium-metal battery, art. No. 103895, Copyright 2019, with permission from Elsevier [76]. (b) Reproduced with permission from S.H. Yu, X. Huang, J.D. Brock, and H.D. Abruña, Regulating key variables and visualizing lithium dendrite growth: An operando X-ray study, J. Am. Chem. Soc., 141(2019), No. 21, p. 8441 [77]. Copyright 2019 American Chemical Society. (c) From Y.Z. Li, Y.B. Li, A. Pei, et al., Science, vol. 358, 506-510 (2017) [78]. Reprinted with permission from AAAS. (d) Reproduced with permission from A.B. Gunnarsdóttir, C.V. Amanchukwu, S. Menkin, and C.P. Grey, Noninvasive in situ NMR study of “dead lithium” formation and lithium corrosion in full-cell lithium metal batteries, J. Am. Chem. Soc., 142(2020), No. 49, p. 20814 [79]. Copyright 2020 American Chemical Society.

下载: 全尺寸图片 幻灯片

Yu et al. [77] used in situ X-ray imaging to observe the deposition and stripping behavior of Li metal during battery operation and quantitatively characterized the evolution of Li porosity/consolidation during the electroplating process. A single lithium branch growing at a current density of 0.5 mA·cm⁻² was larger than that at a higher current density of 10.0 mA·cm⁻², whereas the density of lithium branches was smaller (Fig. 13(b)). The effects of lithium-salt concentration, current density, ion strength, different electrolytes, and additives on the morphological evolution of lithium plating/stripping were studied systematically. By adjusting these parameters, the inhibition of Li dendrite formation and formation of uniform lithium deposition were achieved in repeated cycles. This work contributes to understanding the complex lithium plating/stripping mechanism more comprehensively and improving the safety, utilization, and cycle life of lithium metal batteries.

Li et al. [78] captured the first image of atomic lithium dendrites by cryo–electron microscopy (EM). Li dendrites can preserve the original structure and chemical information at low temperatures. In carbon-based electrolytes, lithium dendrites were observed to grow along the <111> (preferred), <110>, or <211> directions, forming single-crystal NWs with a hexahedral structure (Fig. 13(c)). The dendrite growth directions may change at the kink, but no crystal defects were observed. This work provides a simple way to preserve and image the original state of beam-sensitive battery materials at the atomic scale, revealing their detailed nanostructures. The relevant data observed from these experiments can provide a complete understanding of the battery failure mechanism.

Gunnarsdóttir et al. [79] studied the plating, stripping, and corrosion behaviors of lithium ions on a copper collector by using in situ ⁷Li NMR technology. The method allows the tracking of inactive or dead lithium formation during the plating/stripping of lithium in lithium metal batteries. The formation of dead lithium and SEI can be quantified using NMR, and their relative formation rates in carbonate and ether electrolytes can be compared. Fig. 13(d) shows the in situ ⁷Li NMR spectra obtained after plating and stripping. The ⁷Li metal peak in the Cu–LiFePO₄ (LFP) cells appeared at about 275 ppm and shifted to a lower direction in the subsequent cycle. Using in situ NMR, the authors demonstrated that polymer coating and surface chemical modification of the Cu current collector contributed to the stabilization of the lithium metal surface.

3.3   In situ characterization of the electrode–electrolyte interface

In general, a stable and uniform SEI layer prevents the continuous parasitic reaction between electrodes and the electrolyte, which is critical for achieving a high coulombic efficiency. The stable cathode–electrolyte interface (CEI) layer can prevent adverse phase transformation by surface reconstruction of the cathode material [80–81]. Therefore, the development of advanced characterization techniques to systematically study the formation and properties of SEI/CEI layers is critical to assist in the development of optimized anode, cathode, and electrolyte combinations.

Liu et al. [82] characterized the SEI formation process between graphite anode and electrolyte using a collaborative strategy, combining the quantitative, in situ, and on-site characteristics of electrochemical quartz crystal microbalance (EQCM) and the atomic accuracy of AFM. The EQCM can accurately measure the accumulation or loss of substances on the graphite electrode during the lithiation process, and the balance can identify lithium fluoride and lithium alkyl carbonate as the dominant chemical compositions at different potentials (Fig. 14(a)). Furthermore, SEI morphology during the first lithiation process was measured by in situ AFM. These quantitative observations can provide useful guidelines for designing and customizing better interphases for new battery chemistry.

Figure  14.  (a) Cyclic voltammetry curves (black) and deposition responses on the graphite electrode (blue) recorded by EQCM (i); in situ AFM image (ii) and height distribution (black curve) of the SEI on the stage labelled by the blue line in (ii) (iii). (b) Time sequential high-magnification HAADF–STEM (i) and time sequential ABF–STEM (ii) images of the SEI film; changes in the thickness of SEI film as a function of time during lithiation (iii). (c) schematic of the in situ electrochemical cell for optical cross-section imaging (i); in situ OM images at OCP (ii) and tenth cycle (iii); in situ AFM topography images of CE in the first cycling (iv). (d) In situ Raman spectral evolution of the LNMC electrode (i); representative spectra acquired at ~3 V (blue) and ~4.5 V (red) in comparison with that of a 1 M EC/DMC LiPF₆ electrolyte (black) (ii); changes in the integrated intensities for LNMC A_1g band in v_Lo (400–680 cm⁻¹), ν_Mid (800–1900 cm⁻¹), and ν_Hi (2700–3200 cm⁻¹) regions, stored charge (Q), and the cell voltage profile with the acquisition number (iii). (a) Reproduced by permission from Nat. Nanotechnol, vol. 14, In situ quantification of interphasial chemistry in Li-ion battery, T.C. Liu, L.P. Lin, X.X. Bi, et al., Pages 50-56, Copyright © 2018, The Author(s), under exclusive licence to Springer Nature Limited [82]. (b) C. Hou, J.H. Han, P. Liu, et al., Adv. Energy Mater., vol. 9, art. No. 1902675 (2019) [83]. Copyright 2019 Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission. (c) Y.X. Song, Y. Shi, J. Wan, B. Liu, L.J. Wan, and R. Wen, Adv. Energy Mater., vol. 9, art. No. 2000465 (2020) [84]. Copyright 2020 Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission. (d) Reproduced with permission from Operando investigation into dynamic evolution of cathode–electrolyte interfaces in a Li-ion battery, Nano Lett., 19(2019), No. 3, p. 2037 [85]. Copyright 2019 American Chemical Society.

下载: 全尺寸图片 幻灯片

Hou et al. [83] in situ observed the formation, growth, and destruction of SEI during lithiation and delithiation under a high current density using a sub-nanoscale mass-sensitive Cs-corrected STEM. Fig. 14(b) shows a schematic of the electrochemical device set for in situ STEM observation of the lithium-ion liquid microbattery. The evolution of SEI film was observed under high angle annular dark-field (HAADF) and annual bright-field (ABF) imaging modes, revealing the dynamic evolution of organic and inorganic layers, respectively. The experimental results showed that the failure of the SEI film was related to the rapid dissolution of the inorganic layer in contact with the electrolyte in the broken SEI film. In addition, the crack in SEI film was caused by the change in electrode volume during the process of delithiation.

Through in situ optical imaging technology, Song et al. [84] observed that the dissolved polysulfide was decomposed, and gases were generated inside the solid-state electrolytes (SSEs) as the cycle number increased, resulting in the severe deformation of the sulfur cathode structure. In addition, the structural evolution and dynamic behavior of composite electrolytes were observed in real time at the nano/microscale through in situ AFM. The results showed that the interaction between the polymer networks and functional filler particles weakened after polysulfide dissolution in composite electrolytes, which significantly affected the mechanical stability of the polymer networks and further intensified the deformation of composite electrolytes (Fig. 14(c)).

Chen et al. [85] reported the dynamic evolution of CEI on the surface of LiNi_0.33Co_0.33Mn_0.33O₂ (LNMC) by using a carefully designed in situ surface-enhanced Raman spectroscopy (SERS). To enhance Raman scattering, the scientists deposited gold NPs with enhanced Raman spectral activity on the surface of the LNMC electrode without a binder in monolayer form (Fig. 14(d)). Therefore, in ethylene carbonate/dimethyl carbonate (EC/DMC)-based electrolytes, the chemical vibration patterns generated on the surface of the LNMC electrode had a high sensitivity for Raman detection. Owing to the strong SERS effect of gold monolayer NPs, ester and ether chains were observed in the CEI from the experiment. Through the analysis of dynamic spectral characteristics of the battery in the cycle process, the surface of the LNMC electrode usually formed ether and ester substances under low potentials, which proved that CEI contained organic compounds with ether and ester functional groups. In the process of continuous charge and discharge, the Raman spectrum of CEI presented a very evident dynamic change, which was consistent with the change in the LNMC band.

3.4   In situ characterization of electrolyte

3.4.1   Electrolyte components

Understanding the reaction and state of electrolytes during electrochemical cycles is always the key to exploring and designing devices. In recent years, numerous in situ characterization techniques, such as in situ Raman microscopy [86–87], in situ FTIR spectroscopy [88], in situ small-angle neutron scattering, in situ AFM-based imaging, and NMR spectroscopy [89], have been developed and applied to the study of electrolytes during electrochemical cycling.

In common electrochemical equipment, the whole operation process is accompanied by the transfer of Li ions. Electrochemical devices can be designed or studied by measuring the spatial Li-ion concentration. Some studies have achieved this goal with confocal Raman microscopy. Li ions lack optical properties that are easy to capture and can be analyzed by measuring the vibration signatures of closely related molecules. A distinct linear relationship exists between LiClO₄ concentration and the fraction of intensity on the sideband (Fig. 15(a)). By combining this high-resolution spatial technology with a simple microfluidic device, the diffusion coefficient of lithium ions in DMC in two different concentration regimes can be determined. Confocal Raman microscopy is intuitive and effective in the determination of lithium-ion transport. Moreover, the value of this approach can be observed in other concentration-dependent spectral electrolyte systems [90].

Figure  15.  (a) In situ Raman peaks and the intensity change of O–C–O deformation of DMC with the addition of LiClO₄ (i) (The molecular drawing above the main peak is of DMC, and the arrow indicates the molecular vibration being monitored; Inset: The fraction of intensity in the sideband, α, increases linearly with increasing LiClO₄ concentration, and the black line is a linear fit to the data); schematic of the microfluidic channel (top) and normalized concentration profiles from a flow experiment (bellow) (ii); Li-ion diffusion coefficient (iii). (b) SANS intensity as a function of the scattering vector q, first discharge time, voltage (i), and in situ XPS spectra of Li 1s for the 1 M LiTFSI/propylene carbonate (PC) electrolyte systems (ii). (c) Schematic illustrating the half-cell setup for charge-transport imaging (i) and its ionic current image (ii). (a) Reproduced with permission from J.D. Forster, S.J. Harris, and J.J. Urban, Mapping Li⁺ concentration and transport via in situ confocal Raman microscopy, J. Phys. Chem. Lett., 5(2014), No. 11, p. 2007 [90]. Copyright 2014 American Chemical Society. (b) Reproduced with permission from C.J. Jafta, X.G. Sun, G.M. Veith, et al., Energy Environ. Sci., vol. 12, 1866-1877 [91]. Copyright 2019 The Royal Society Chemistry. (c) Reproduced from J. Power Sources, 481, C.S. Jiang, Y. Yin, H. Guthrey, K. Park, S.H. Lee, and M.M. Al-Jassim, Local electrical degradations of solid-state electrolyte by nm-scale operando imaging of ionic and electronic transports, 229138, Copyright 2020, with permission from Elsevier [92].

下载: 全尺寸图片 幻灯片

The research of the evolution of surface chemistry and electrode microstructure during battery cycles is also desirable. Jafta et al. [91] suggested operando small-angle neutron scattering (SANS) to study the effect of different concentrations of electrolytes in SEI formation, pore filling, and carbon framework expansion on the ordered-mesoporous carbon electrode surface (Fig. 15(b)). In the 4 M electrolyte system, Li-rich reduction products formed at high potentials in the micropores, mesopores formed lithium salts quickly before the formation of carbonaceous products. The expansion of the carbon framework was affected by micropore filling and co-intercalation.

A half-cell device containing AFM probe/SSE/Li (Fig. 15(c)) was used to explain the mechanism of SSE degradation from the nanoscale ions and electron transport [92]. Using nm-scale operando imaging, the researchers demonstrated highly nonuniform ion transport, and electron leakage resulted in the degradation of SSE. SSE degradation can be inhibited by covering both sides of SSE with a thin polymer film because the reduction of ions was effectively alleviated despite the slight increase in electronic leakage.

3.4.2   Gas products of electrolyte decomposition

SEI plays a key role in determining the performance of batteries [93]. In addition, during the formation of SEI, the electrochemical reaction of the electrolyte solvent leads to substantial gas evolution, which also largely affects the electrochemical performance. The challenges of volume expansion, explosion, and other safety issues or performance degradation of batteries are largely caused by gas evolution. Thus, studying the composition and proportion of gases can solve a number of the current issues in batteries. To obtain more intuitive and accurate test results, scientists have developed numerous in situ characterization methods for evolving gases, for example, in situ pressure analysis [94], in situ DEMS [95], in situ gas-chromatography (GC) analysis, and in situ Raman spectroscopy [96].

Teng et al. [97] reported a system of GC analysis and thermal conductivity detector that was connected to a set of 900 mAh pouch cells and used for in situ testing of gas generation during battery cycles (Fig. 16(a)). The electrochemical stability of γ-butyrolactone (GBL) was remarkably better than those of other electrolytes used in this work. The electrolyte composed of GBL and ethyl methyl carbonate showed the least gas release in the first charging process of LIBs, which improved the safety and stability of the battery. In addition to quantitative analysis, the system determined the composition of the gases. In GBL-containing electrolytes, the CO and C₂H₄ generated from EC decomposition were evidently less than those in the EC-containing electrolytes.

Figure  16.  (a) Schematic of a gas analysis system for LBs (left) and compositions of gas released at different states of charge (right). (b) CO₂ evolution in cells at different temperatures (left) and CO₂ evolution with and without added Li₂CO₃ (right). (c) Schematic of the EC-MS setup, tandem mass spectra of CoPc-polysulfide intermediates, and comparison of the distribution of various polysulfides. (a) Reproduced with permission from X. Teng, C. Zhan, Y. Bai, et al., In situ analysis of gas generation in lithium-ion batteries with different carbonate-based electrolytes, ACS Appl. Mater. Interfaces, 7(2015), No. 41, p. 22751 [97]. Copyright 2015 American Chemical Society. (b) Reproduced from J. Power Sources, 159, Vetter, M. Holzapfel, A. Wuersig, W. Scheifele, J. Ufheil, and P. Novák, In situ study on CO₂ evolution at lithium-ion battery cathodes, 277-281, Copyright 2006, with permission from Elsevier [98]. (c) Z.Y. Yu, Y. Shao, L.P. Ma, et al., Adv. Mater., vol. 2022, art. No. 2106618 (2022) [100]. Copyright 2022 Wiley-VCH GmbH. Reproduced with permission.

下载: 全尺寸图片 幻灯片

In situ DEMS was used to study the effect of various battery parameters on CO₂ evolution during the cycling process [98]. High temperatures and battery voltages increased CO₂ production (Fig. 16(b)). Additives and lithium carbonate in the cathode electrode affected the release of CO₂, and the effect of lithium carbonate only occurred in the first cycle. The type of cathode active material was the main influencing factor for the amount of CO₂ evolved. Based on the DEMS method, Zeng et al. [99] evaluated the oxidation stability of the Layered lithium-rich manganese oxide in tris(2,2,2-trifluoroethyl) phosphate (TFEP)-based electrolytes, in which no CO₂ evolved throughout the charge/discharge process. MS has high sensitivity and specializes in distinguishing various components by mass to charge (m/z) ratio. They also reported the direct observation of polysulfides in a Li–S battery by combining in situ electrochemistry and MS (Fig. 16(c)) [100]. Several key polysulfide intermediates during sulfur redox have been identified, and their abundance distributions at different potentials enhanced the understanding of the shuttle effect, which largely determines the performance of Li–S batteries.

4.   Summary and perspective

In summary, numerous leading research groups have conducted pioneering work in the developments and applications of in situ characterization techniques in the past decades, including the design of in situ cells. On the basis of single in situ technology, the multi-modal coupling characterization of different techniques was developed and is expected to eliminate the temporal mismatch of various information. Nevertheless, breaking through the limitation of temporal resolution is a great challenge. Although it is still in its infancy, multi-modal coupling will become a growing trend in the future. In situ characterization reveals indispensable information on how electrode materials behave and how their structures evolute under working conditions. These insights enable researchers to correlate the electrochemical process with the composition, phase structure, geometric structure, etc., at the macro- and micro-scales, which are essential for understanding the failure mechanisms of batteries. Moreover, these insights are also helpful in electrode recycling and sheds light on the utilization of spent LIBs.

Despite the significant progress achieved in the mechanism analysis and electrode material design through in situ techniques, numerous questions remain to be answered in the pursuit of new in situ characterization methods. In the future, more efforts need to be devoted to the below directions, which will promote in situ characterization to further contribute to LBs.

First, although single characterization techniques can also provide abundant information, the combination of different characterization techniques always shows a greater power in obtaining a more comprehensive image of electrode material structure–performance correlation and battery failure mechanism. For example, although in situ MS characterization of the gas evolution can provide some understanding of the battery failure mechanism, researchers can reach a deeper understanding of the ongoing anode, cathode, and electrolyte material structure and surface/interface chemistry evolution in the battery system under operation conditions by correlating the information obtained from different techniques, thus offering better guidance for the design of electrode and electrolyte materials to improve battery applications. Second, scholars should aim for the development of a multi-modal coupling system that collects different modal information of various applicable objectives at multiple length scales, thus decoding diverse physical/chemical phenomena in batteries. For instance, bench-top XAS can be combined with our developed NMR–Raman coupling system, that is, XAS–NMR–Raman. Thus, XAS can be used to analyze the phase structure evolution of cathodes and provide fine insights into the local structure, whereas NMR and Raman can reveal the evolution information of the anode and electrolyte, respectively. Certainly, the interference in different modal techniques should be avoided by carefully designing the assembly site, adopted material, etc., thus ensuring that the obtained information is reliable. In addition, in situ electrochemical cells can be optimized to meet the requirement of multi-modal coupling characterizations, including the selection of transmission window, encapsulation and connection of cell package, and so on. Third, spatial–temporal matching information must be obtained for multi-model coupling characterization. To address the challenge of temporal resolutions, scholars should accurately control the characterization site and signal acquisition time of different techniques. Moreover, a compatible operating system and a data processing system need to be developed for thorough analysis. Currently, most in situ characterizations focus on thermodynamical properties in static studies. Hence, more attention can be paid to the kinetics analysis. The probing of non-equilibrium states can provide hints on the battery reaction feasibility. By improving the data-collection rate, more understanding of the electrode kinetics can be obtained through multi-modal coupling characterizations. Fourth, another interesting direction would be the in situ characterization of batteries under external stimuli, such as optical, thermal, and magnetic. Finally, investigations on batteries will certainly provide insights into other scientific fields, such as physical chemistry and analytical chemistry, which will, in turn, promote the battery field. Moreover, Advances in these techniques and knowledge can promote the development of other electrochemical energy storage systems.

A more exciting understanding of the battery failure mechanism and additional guidance on a better battery design can be obtained with the continuous efforts devoted to in situ characterization technique field. We hope that this review will stimulate exciting investigations and help to accelerate the commercialization of next-generation batteries.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (Nos. 21820102002, 21931012, 22111530178, 51932001, 51872024, and 51972305), the Cooperation Fund of the Dalian National Laboratory for Clean Energy（DNL), Chinese Academy of Science (CAS) (No. DNL202020), the National Key Research and Development Program of China (No. 2018YFA0703503), and the Scientific Instrument Developing Project of the Chinese Academy of Sciences (No. YZ201623).

Conflict of Interest

The authors have no conflicts to declare.