Exploring the Properties of Disordered Rocksalt Battery Cathode Materials by Advanced Characterization

Cation-disordered metal oxides as cathode materials for Li ion batteries have been overlooked from early studies due to to the restriction of Li ion diﬀusion, leading to poor electrochemical performance. However, the discovery of a new disordered rocksalt (DRX) structured material Li 1.211 Mo 0.467 Cr 0.3 O 2 with a high capacity of > 260 mAh g − 1 at 0.05 C opened new research prospects in this emerging ﬁeld and established DRX materials as a promising alternative with wider choices of transition metal elements compared with currently widely used layered cathode materials. Some of the major obstacles of the DRX materials include 𝜸 -LiFeO 2 type cation short-range-order that impedes Li ion diﬀusion, irreversible oxygen loss, and transition metal dissolution, which also present challenges for appropriate characterization techniques. Several performance optimization strategies have been employed, including ﬂuorine incorporation, high entropy modiﬁcation, and surface coating. This review article focuses on advancements in characterization techniques to uncover underlying mechanisms of Li ion diﬀusion and degradation of the DRX cathode materials to address the abovementioned challenges and provide inspiration for future studies of this class of materials.


Introduction
Although the energy density of Li ion batteries (LIBs) has been steadily increasing, there exists a broad variety of electrode materials that are not fully explored yet.4][5] Unlike conventional intercalation cathode materials with an -NaFeO 2 -type structure where 2D Li transport occurs between TM layers, the mixing of cations (Li and TM) is observed in the cation DRX type cathodes, which adopt a -LiFeO 2 -type structure [6] (Figure 1c).The DRX materials had long been neglected as a good cathode material, as cation mixing was believed to restrict effective Li transport, and poor electrochemical performance. [7]owever, the early report of a high capacity (>260 mAh g −1 at 0.05 C) for the DRX Li 1.211 Mo 0.467 Cr 0.3 O 2 (LMCO) material [7] unlocked the potential research space.Later studies show some of the DRX materials possess higher theoretical capacities (up to 425 mAh g −1 ) and energy densities (up to 1200 Wh kg −1 ) compared to existing ordered cathode materials, e.g.≈220 mAh g −1 and 800 Wh kg −1 for LiNi x Mn y Co z O 2 (NMC) owing to the energy storage contributions from excess Li and both TMs and O-redox that increase capacity, [3,4] making DRX materials a promising alternative to layered and olivine structured cathode materials.
Among all rocksalt-typed crystals, there exists stoichiometric octahedral holes (Oh) and tetrahedral holes (Td) in the lattice (Figure 2).For Li migration process in the conventional layered crystal structure, Li ions alternate through Oh to Td to Oh, known as the o-t-o diffusion.In DRXs, 5 different environments can be found according to different cation mismatches and arrangements, 0-TM, 1-TM, 2-TM, 3-TM, and 4-TM local environments, corresponding to Li 4 , Li 3 TM, Li 2 TM 2 , LiTM 3 , and TM 4 tetrahedral clusters, respectively. [2]Effective Li transportation is mainly via 0-TM channels, i.e., through the connected Li 4 clusters to enable Li ion mobility. [2]The presence of 0-TM diffusion channel is a unique characteristic of DRXs, which is not found in the pristine layered materials as no cation mismatch is observed.Crystal structure and Li ion transport experiments and simulations  [8] 2021, American Chemical Society.b) The natural abundance of metals in the earth's crust and the price of each metal in USD kg −1 (shown in the blue figures on top of each bar).Adaptedwith permission [2][3][4][5] 2023, U.S. Geological Survey.c) Schematics of crystal structures for (I) the disordered rocksalt -LiFeO 2 structure in which all cation sites are equivalent, (II) the layered -NaFeO 2 structure, (III) the spinel-like low-temperature LiCoO 2 (LCO) structure, (IV) the -LiFeO 2 structure.Large empty circles indicate oxygen sites and small gray and black filled circles stand for lithium and TM sites, respectively.Adapted with permission. [6]2014, John Wiley and Sons.Reproduced with permission. [7]014, The American Association for the Advancement of Science.
results reveal that Li-excess above 1 stoichiometry level is required to transform DRX materials to high-capacity cathodes to unlock the 0-TM percolation, e.g. it was found that a ≈10% Liexcess is required to enable 0-TM percolating network for LMCO to achieve a high capacity of 327.5 mAh g −1 at 0.05 C. [7] The 0-TM channels are essential for fast Li diffusion, however, the number of 0-TM channels is insufficient to form a percolating network in a non-Li-excess DRX in most cases, resulting in slow Li diffusion and poor performance in the previous studies. [2,7]or multi-TM DRXs, only certain types of TMs contribute to the battery capacity, while metals with zero valence electrons are inactive. [9]The presence of inactive TMs (so-called compensator ions or d 0 metals, e.g.Ti 4+ , Nb 5+ , Ru 5+ , Mo 6+ ) are generally smaller in size and higher in charge, which could accommodate the distorted sites in the lattice and therefore stabilize the DRXs.DRXs with multiple TMs incorporated in the structure can be referred by their active metals.For instance, in Li 1.25 Nb 0.25 Mn 0.5 O (LNMO), [10] the contribution to capacity is based on Mn 3+ /Mn 4+ redox while Nb 5+ is the inactive d 0 metal which does not contribute to capacity, LNMO can thus be classified as a Mn-based DRX.
Density functional theory (DFT) calculations and Monte-Carlo (MC) simulations predicted a metal mixing compatibility map of high-entropy oxyfluoride cathodes which provides useful instructions for metal selection according to fundamental thermodynamics. [11]The synthesized material from 6 types of TMs (Li 1.3 Mn 2+ 0.1 Co 0.1 Mn 3+ 0.1 Cr 0.1 Ti 0.1 Nb 0.2 O 1.7 F 0.3 ) with a high configurational entropy demonstrated a desirable capacity of 307 mAh g −1 (955 Wh kg −1 ) at a low rate (20 mA g −1 ), with more than 170 mAh g −1 capacity retention after cycling at a high rate of 2000 mA g −1 . [12]The development of DRX materials also opens up compositional flexibility among earth-abundant TMs.
Currently, the performance of DRX materials is impeded by several technical challenges including low electronic conductivity of DRX particles which requires large amounts of C additives in electrode processing, [13] the lack of scalable synthesis routes to prepare high performance DRX cathodes with controlled morphologies, and the sloping voltage profiles of most DRX cathodes which creates challenges for system integration.
While there have been review articles on the DRX cathode materials, e.g.computational studies [9] and characterization techniques for batteries in general, [14] this review focuses on the recent advancements in the characterization techniques to address two major challenges of the newly emerging DRX cathode materials: (i) the presence of the -LiFeO 2 -type short range order (SRO) of TM (Figure 1c) that impedes effective Li percolation in the lattice and cathode performance; [15,16,9] (ii) irreversible anionic redox and subsequent dissolution of TM cations into the electrolyte that lead to rapid reduction of battery cycling life.Both obstacles also impose significant challenges in characterization methods to study the associated underlying mechanisms as it is notoriously challenging to detect the low atomic mass Li and O which usually result in low signal-to-noise ratios. [17]This review first outlines the two significant challenges of the DRX cathode materials in Section 2, and then group various characterization techniques based on functionalities to tackle each challenge in Sections 3 and 4. Section 5 summarizes the development and applications of advanced characterization techniques that investigate the DRX materials, and provides an outlook of the aforementioned experimental approaches in the development of DRX cathodes.

Li Percolation
A recent study [15] uses DFT and MC simulations to study the local and macroscopic Li transport in two Li-rich DRXs: Li 1.2 Mn 0.4 Ti 0.4 O 2 (LMTO) and Li 1.2 Mn 0.4 Zr 0.4 O 2 (LMZO).The LMTO prefers forming Li 4 clusters over LMZO, indicating suppressing the -LiFeO 2 type SRO is critical in controlling Li transport environments in DRX which has a great dependence on the type of d 0 metals. [15]Various synthetic modifications have been investigated to alleviate the harmful impact of the -LiFeO 2 type SRO.
[20][21] The strong bonding tendency between Li and F maximizes the Li content around F, resulting in a greater number of Li-rich clusters, thereby suppressing the -LiFeO 2 type SRO.The orthorhombic-LiMnO 2 (o-LiMnO 2 ) exhibits similar structural motifs to Mn-based Li-rich cathodes, [22]  The polygons near the bars (triangle, square, and circle) specify where variations in the barriers come from (varying local environment, Li concentration, and Li slab space, respectively).Reproduced with permission. [22]2020, Royal Society of Chemistry.
the range of migration barriers in the model system of o-LiMnO 2 with and without F/Li-excess were calculated and compared with other layered, spinel and olivine systems as shown in Figure 3, this provides an overview of the typical migration barrier range for the materials with and without F/Li-excess and different types of cathode materials. [22]he impact of fluorination on SRO and Li transport in DRX materials was investigated using DFT-based MC simulations. [18]igure 4a shows the fraction of Li accessible by percolation in the M 2+ systems, for all four systems, the 0-TM percolating Li amount initially decreases as F is added, but then increases at high F content, suggesting a transition from "less percolating" to "more percolating" SRO with heavy fluorination. [18]Figure 4b shows the evolution of isolated Li4 as a function of fluorination, initially the amount of isolated Li increases, at the highest F-levels there are almost no isolated Li 4 clusters, indicating that most Li is in the percolating clusters. [18]Overall, a large amount of F substitution (>15%) generally improves 0-TM percolation because Li-rich clusters around F become increasingly connected. [18]At low (<15%) F concentration, dilute F ions attract Li around them, forming yet isolated rather than percolating Li-rich domains, [9] the effects of fluorination can be beneficial or disadvantageous depending on the intrinsic SRO in the unfluorinated oxide. [18]en the origin of poor Li transport in the unfluorinated oxide is cation SRO that disfavors 0-TM percolation, fluorination enhances Li transport by disrupting these local configurations and promoting the formation of connected Li-rich environments. [18]n the other hand, if the cation SRO in the pure oxide already exhibits good 0-TM Li percolation, [7] the Li redistribution caused by small amounts of F substitution can be detrimental to 0-TM percolation as it may interrupt the existing connections between local Li-rich domains.
The highly fluorinated composition space is restricted by the F solubility limit. [23]Research in understanding the F solubility in DRX materials and how it can be manipulated through metastable synthesis routes is beneficial to the development of stable, high capacity DRX materials.For example, first-principles calculations were used to investigate the thermodynamics of fluorination in cathode materials to determine the conditions under which bulk fluorination is possible, [23] and an experiment method of using a fluorinated polymeric precursor instead of the LiF precursor was developed during synthesis of a model Li-Mn-Nb-O (LMBO) system, and increased the F content from 7.5 at% to 10-12.5 at.%. [19] Several synthesis routes targeting a highly fluorinated DRX cathode Li 1.2 Mn 0.4 Ti 0.4 O 1.6 F 0.4 were also designed and tested. [24]Precursor combinations were screened   as a function of F content.Adapted with permission. [18]2020, John Wiley and Sons.
to raise the F chemical potential and avoid the formation of LiF which inhibits fluorination of the targeted DRX phase. [24]t low temperatures, MnF 2 reacts with the Li source, forming LiF, thereby trapping F in strong Li-F bonds, LiF can react with Li 2 TiO 2 to form Li 3 TiO 3 F; however, MnO is not easily incorporated into this DRX phase. [24]Higher temperatures increase F solubility, the volatility of LiF above its melting point (848°C) inhibits fluorination of the DRX phase. [24]Based on these findings, metastable synthesis techniques are suggested for future work on DRX fluorination. [24]ther potential methods include adopting the high-entropy alloys of multi-TM SRO [12] and Cl incorporation.For example, substitution of Cl in Li 1.225 Ti 0.45 Mn 0.325 O 2 (LTMO) was used to form Li 1.225 Ti 0.45 Mn 0.325 O 1.9 Cl 0.1 (LTMOCl), and experimental results showed better performance for LTMOCl compared with LTMO. [25][27] Electrochemical impedance spectroscopy (EIS) measurements showed the low-frequency linear slope for the LTMOCl electrode was greater than that of the LTMO electrode, indicating a higher Li diffusion rate for LTMOCl. [25]Figure 5 shows that Li diffuses via the o-t-o path, the diffusion barrier is influenced by the height (or size) of the intermediate tetrahedron.The larger ionic radius of Cl leads to an increased height of the intermediate tetrahedron in LTMOCl, thus reducing the energy barrier. [25]However, theoretical research in Cl substitution is less extensive than that in F substitution, the mechanism of how Cl influences Li migration may be materialsdependent and is not fully understood yet.
Measuring and visualizing SRO in the crystal structures can be challenging and typically requires advanced characterization techniques.Solid-state NMR (ssNMR) results suggested that the formation of the detrimental -LiFeO 2 type SRO can be explained by the local electroneutrality preference, [28] whereby each O 2− is coordinated octahedrally by six cations with a combined charge of 12 + .Taking Li 1.25 Nb 0.25 Mn 0.5 O 2 as an example, the available metal cations have charges of 1 + , 3 + , or 5 + (Li, Mn, and Nb, Figure 5. Schematic illustration of the local environment for Li migration in LTMO and LTMOCl.Adapted with permission. [25]2023, American Chemical Society. respectively), and local electroneutrality can only be achieved through the OLi 3 Mn 3 (similar to -LiFeO 2 ) or OLi 4 NbMn arrangements.The latter arrangement is better suited to the composition, with excess Mn 3+ ions accommodated through the former arrangement. [28]Both experimental and simulations results show that both ionic charge and size influence the stability of the DRX materials.On one hand, the high-valent TMs tend to repel each other and intimately mix with Li to keep local electroneutrality, thereby inhibiting Li segregation into Li 4 tetrahedra; on the other hand, the size mismatch between highvalent TMs and Li facilitates Li segregation to minimize strain, counteracting the charge effect. [15]For trivalent redox-active TMs where the average valence of the TMs is fixed and so is the electrostatic interaction strength between Li and TMs, the size effect becomes dominant, e.g.ions with large ionic sizes such as Sc 3+ (0.745 Å) and In 3+ (0.80 Å) are likely to impede Li + diffusion, whereas other ions with smaller ionic sizes such as Ga 3+ (0.62 Å) and Sb 5+ (0.60 Å) are likely to facilitate Li + efficient transport. [15]

Materials Degradation
During the charge and discharge process of DRX cathodes, two types of redox reactions occur, namely the cationic and anionic redox, both contribute to the overall charge/discharge capacities. [29]he cathode degradation is primarily governed by: (i) TM ion migration and (ii) oxygen evolution.The dynamic behavior of TM ions during cycling which includes their continuous migration, reduction, and dissolution lead to oxygen redox, especially the release of oxygen, and changes in the electronic and crystalline structure. [29]Another significant factor contributing to the cathode degradation is the oxidative decomposition of the carbonate solvent(s) in the electrolytes at high voltages that generates water and other protic species. [30]In the standard LiPF 6 -based electrolytes, they decompose the PF 6 − into PF 5 and HF, with further reactions lead to the formation of Lewis acidic OPF 3 and fluorophosphate salts such as PO 2 F 2 − . [31]The TM dissolution from the cathode materials has been reported to coincide with electrolyte oxidation, rationalized by the HF formation and acidmediated dissolution. [32]The dissolved TMs migrate through the electrolyte and subsequently to the anode where they deposit.This has been proposed to cause continuous decomposition of the solid electrolyte interphase (SEI) and electrolyte, increasing impedance. [33,34]or the current collector of the cathode, the electrochemical stability of Al stems from the presence of a thin (≈5-20 nm) nonporous native protective layer that consists mostly of Al 2 O 3 and traces of Al(OH) 3 . [35]This layer provides high corrosionresistance to the current collector up to ≈3.8 V versus Li/Li + in conventionally used organic battery electrolytes, [35][36][37][38] and the oxide film becomes susceptible to decomposition when this potential limit is exceeded or in contact with chemically aggressive electrolyte. [36,37,39,40]The electrochemical corrosion requires a contact of the metallic surface with an electrolyte solution, this may emerge due to inhomogeneity of the material and its environment, e.g.presence of impurities, grain boundaries, crystal planes, lattice defects and concentration gradient of a constituent in the solid phase. [41]r oxygen evolution, DFT is used to calculate density of states (DOS) and charge/spin density around oxygen ions in various local environments and suggests O oxidation occurs preferably along the Li-O-Li configuration, with orphaned hole states on O 2− . [42]Delocalized electron-hole states over time can condense to form molecular O 2 which is either trapped within the lattice or form O 2 gas that escapes from the lattice, leading to oxygen loss. [43]Ultimately, O-redox cathodes that form hole states on O in lattice that lead to oxidized oxygen in lattice and do not transform into O 2 are desirable to deliver reversible bulk oxygen oxidation and are reversible cathode materials. [43]Using transmission electron microscopy (TEM) and neutron diffraction, peroxo-like O 2 n− species with long O-O bonds are visualized, [44,45] while the DFT calculation predicts the formation of peroxides O 2 2− as an intermediate of the anionic redox. [46,47]The contribution to energy storage by other TM and O in DRX materials is shown by combined in situ XRD and electron energy loss spectroscopy (EELS) measurements. [10]he layer densification phenomenon [48] is one of the theories rationalizing the possible oxygen loss mechanism.Upon cycling, the observed TM-enrichment (reduction of Li-excess) leads to poorer 0-TM percolation on the particle surface, causing capacity fading and hysteresis.After the first charge, oxygen vacancies are generated on the surface, causing instability of highvalent TM cations, which are then reduced to a lower oxidation state. [49]The reduced cations may then migrate into the neighboring octahedral vacancy sites left by Li, resulting in a densified surface layer with fewer Li vacancies and lower TM oxidation states compared to that in the bulk. [49]After extensive cycling, the above densification-based degradation will be gradually extended into the bulk materials, causing capacity fading, the O 2 loss mechanism is also supported by the resonant inelastic X-ray scattering (RIXS) and NMR results [29] (details in Section 5.1.2),and neutron pair distribution function (PDF) [50] (details in Section 5.2.2).
A novel methodology for calculating the surface energy of DRX materials was developed. [51]Results from a previous cluster expansion model, trained using ab initio calculations as described, [18] are used as a basis for MC simulations to describe and equilibrate the disordered bulk structures.A robust ab initio calculation of surface energies for facets of the DRX material Li 2 MnO 2 F as a function of surface cation and anion decoration was performed. [51]The obtained energies were Boltzmann-weighted to obtain an average energy for each facet which were then used to calculate the equilibrium particle shape. [51]Li 2 MnO 2 F displays predominantly [100] type Li/F-rich facets, favoring a cubic particle shape. [51]A statistical analysis and subsequent linear regression indicated that stable surfaces were likely enriched Li and F which correlate with improved cycle life and Mn and O retention. [51]n unusual  phase transformation was discovered from disordered Li-and Mn-rich rock salts to a new  phase which displays a partial spinel-like ordering with short coherence length and exhibits high energy density and rate capability. [52]Unlike other Mn-based cathodes, the  phase exhibits almost no voltage fade upon cycling.The driving force and kinetics of this in situ cathode formation where immobile TMs, such as Ti 4+ , play an important role in regulating the driving force for the transformation and controlling the structure of the  phase. [52]esign guidelines were established for Li-and Mn-rich compositions that combine high energy, high rate capability, and good cyclability, thereby enabling Mn-based energy storage, future improvements may be possible by reducing the immobile cation content to increase the length of the 4-V plateau and by replacing Mn with a multi-electron redox center to increase the theoretical TM redox capacity. [52]he design space of compositional complex energy materials (CCEMs) is intricate due to higher dimensionality of compositional space, as well as convoluted interplay between long range order and short-range structures. [53]Progress of machine learning algorithms, the physical interpretation, and their potential applications in studying CCEMs has been reviewed. [53]To give a specific example, a machine-learning approach that predicts synthesis conditions using large solidstate synthesis data sets text-mined from scientific journal articles was developed. [54]Using feature importance ranking analysis, the optimal heating temperatures were discovered to have strong correlations with the stability of precursor materials quantified using melting points and formation energies (∆G f , ∆H f ). [54]This correlation between optimal solidstate heating temperature and precursor stability extends Tamman's rule from intermetallics to oxide systems, suggesting the importance of reaction kinetics in determining synthesis conditions. [54]On the other hand, heating times are shown to be strongly correlated with the chosen experimental procedures and instrument setups. [54]Using these predictive features, the authors constructed machine-learning models with good performance and general applicability to predict the conditions required to synthesize diverse chemical systems. [54]

Characterization Techniques to Address the Li Percolation-Based Challenges
Characterization techniques that are sensitive to Li include atom probe tomography (APT) [55] and 7 Li solid-state nuclear magnetic resonance (ssNMR) spectroscopy that allows the investigation of local environments of Li. [56] Other techniques that are used for studying Li percolation (sometimes include using focused ion beam, FIB, and ion etching depth profile to section the sample particles first) include neutron PDF (nPDF) for assessing the SRO structure in DRX materials; [57] X-ray absorption near-edge structure (XANES) analyzes the variations in the oxidation state of TM during Li intercalation and de-intercalation to measure the amount of Li in the material; [58] X-ray Compton scattering (XCS) that measures Li chemical stoichiometry of the electrode materials through the electron momentum ratios between the valence and core electrodes of the electrode material compounds; [59] and atomic-resolution high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) imaging to image the atomic arrangement as the intensity recorded is proportional to the square of the atomic number Z. [60] The combined technique of HAADF-STEM with electron ptychography offers the advantage of the possibility of direct imaging of Li ions, [61] where Li and O atoms are visible in the phase image. [61]Details are shown below.

Scanning Transmission Electron Microscopy (STEM)-Based Characterization for Analysis of Li and Other Cation SRO
STEM with atomic resolution Z-contrast imaging and electron energy-loss spectroscopy (EELS) has been applied for imaging and quantifying the atomic and electronic structures of the DRX materials. [62]SRO represents a systematic local bonding preference between certain ion species.The form of cation SRO is important since the (local) structure type and the degree of (dis)order affect both Li mobility and Li percolation. [9][66] Atomic-resolution HAADF-STEM imaging technique is complementary BF-TEM as it helps visualize the local structure, in which the image intensity recorded in HAADF-STEM mode correlates to the average atomic number Z along the atomic column and roughly proportional to Z 2 .In the case of Li 1.3 Nb 0.3 Mn 0.4 O 2 , [001], HAADF-STEM images reveal variations in intensity profile for the different atomic columns. [60]righter atomic columns, corresponding to the peaks with higher intensity in the profile, are enriched with heavier Nb atoms, while the less intensive columns are enriched with Mn.
Nevertheless, STEM probes only a small region of the crystal and cannot represent the overall information from the whole particle.Addressing these challenges not only requires experimental techniques but also analytical approaches to determine realistic atomic-scale models of SRO consistent with experiments.One methodology is cluster-expansion MC simulation which couples first-principles energy calculations of candidate SRO models with large-scale atomistic simulations. [67]A new methodology has been developed that combines MC simulations and PDF calculations to evaluate SRO in DRX materials [68] where MC simulations allow for the construction of an A 1+x B 1-x O 2 phase space across a range of stoichiometries, the structures in the phase space constitute a library of calculated PDFs, each pointing to a specific cation SRO motif.Using these motifs as structure identifiers, SRO is analyzed by reverse mapping their structures via the PDFs onto the phase space. [68]

X-Ray Absorption Spectroscopy (XAS) for Analysis of Li Intercalation and De-intercalation and Oxidation State Changes of Transition Metals
XAS includes a diverse family of spectroscopic techniques, such as XANES [69] and extended X-ray absorption fine structure (EXAFS). [70]Resonant inelastic X-ray scattering (RIXS) [71] is a combination of XAS and X-ray Emission Spectroscopy (XES), where a core electron is also excited by an incident X-ray photon and this excited state decays by emitting an X-ray photon to fill the core hole in XES.The XES high-resolution fluorescence detection technique measures the X-ray absorption spectrum via monitoring the intensity of a fluorescence line corresponding to a specific excited state decay process using a narrow energy resolution, making it possible to overcome some of the main limitations of conventional XAS. [72]While XANES encompasses an energy range of 50-100 eV in close proximity to the absorption edge, EXAFS covers a broader energy range, spanning several hundred eVs and commencing ≈30 eV beyond the absorption edge. [73]XANES analysis facilitates the determination of oxidation states and local symmetry changes of the specific absorbing element, while EXAFS provides quantitative information about the local atomic structure. [74]Synchrotron XAS is adopted for investigating both the sublattice structure and redox reactions in DRX materials, e.g. the oxyfluoride type DRX with a more sophisticated anionic sublattice. [75]This technique enables the simultaneous probing of local structure and chemical states with element-selectivity as the high intensity of synchrotron X-rays enables the detection of elements at extremely low concentrations to an extent. [76]s for cationic sublattice studies, the variations in the bulk oxidation state of TM in pristine and cycled samples can be assessed by analyzing the reflection energy of the XANES in TM K-edge, M-edge, and L-edge.XANES examine the unoccupied density of states and the measurements are sensitive to the local bonding environment, such as the number of valence electrons, their spin configurations, and the symmetry and coordination number of the TM compound structural unit cell.Upon oxidation/reduction, the peak is expected to shift to a higher/lower energy region; the energy position corresponding to the first derivative of the rising edge is commonly used as an index of the oxidation state. [19]Alterations in the in situ XANES profile shape can be compared subsequently to the one after cycling, which signify the modifications in the oxidation states and electronic configurations of the TM.On the study of Ti-doped LiFePO 4 cathode (LiTi 0.01 Fe 0.99 PO 4 ), alternations in the Fe valence during Li intercalation and de-intercalation were observed by the in situ Fe Kedge XANES measurements, [77] indicating the role of Fe ion as electron donor/acceptor during charge/discharge.
To probe the anionic sublattice in F-DRX, F K-edge, and O K-edge spectra have been collected on various pristine samples. [78,79]A typical F K-edge XAS spectrum is comprised of three main regions: the pre-edge, absorption edge, and postedge.These correspond to the transitions from the F 1s orbital to higher energy electronic states, such as TM 3d , F 2p , and F 4s4p .Ex situ F K-edge XAS was obtained for both the pristine and discharged Li 1.2 Ti 0.2 Mn 0.6 O 1.8 F 0.2 (LTMOF2.0)cathodes after 50 cycles with a probe depth of ≈5 nm. [79]The XAS profile exhibited sharper peaks and a newly emerging broad feature in the range of 710-730 eV, demonstrating the presence of LiF. [79]The results suggested cycling promoted the ionic character of F on the cathode surface, leading to atomic-level changes in both TM cationic sublattice and O/F anionic sublattice, as well as gradual formation and extension of domains resembling LiF on the sample surface. [79]

X-Ray and Neutron Scattering and Pair Distribution Function (PDF) for Analysis of Li SRO
Diffraction is a form of elastic scattering where the diffracted Xray photons and neutrons have similar energy as the incident photons and neutrons.Diffraction takes place in which the uniform spacing of crystal structure causes an interference pattern of the waves present in an incident beam of X-rays and neutrons.The X-ray and neutron diffraction give complementary informa-tion.X-rays interact primarily with the electron cloud surrounding each atom, so the contribution to the diffracted X-ray intensity is larger for atoms with a large atomic number Z than it is for atoms with a small Z. [80] On the other hand, neutrons interact with the nucleus of the atoms, and the contribution to the diffracted intensity is different for each isotope, it is often the case that low Z atoms contribute strongly to the diffracted intensity even in the presence of large Z atoms. [81]n recent years, neutron pair distribution function (nPDF) has become an increasingly popular method for probing the atomiclevel structure of disordered materials. [82]The nPDF method employs neutron scattering data to study materials structures at the atomic scale, generating functions that describe the probability of locating two atoms at a particular separation distance within the material. [83]In addition to its utility in investigating the nature of O-containing species generated during cathode degradation and the coordination around O in the local structure, the nPDF technique can also analyze the SRO present in crystalline materials.
The PDF method is sensitive to the chemical properties of the SRO materials [84,85] by measuring the intensity of high-energy X-rays, neutrons, or electrons scattered from a sample at various angles. [86]PDF analysis makes use of the entire spectrum, including both Bragg and diffuse scattering.The inspection of the diffuse scattering can provide information regarding the local structure. [87]In contrast to the refinement of structures based on conventional diffraction data where only primitive unit cells are typically considered, modeling PDF data often requires the use of large supercells to account for the wide range of coordination environments that may exist. [88] recent study [89] performed a combination of neutron scattering and first principles simulations techniques to model SRO in Li 1.3 Mn 0.4 Ti 0.3 O 1.7 F 0.3 (LMTOF).A variety of first principles simulations techniques were used and tested, and the accuracy of each method was validated from PDF collected using total neutron scattering (Figure 6). [89]The PDF contained several asymmetric and bimodal peaks at short distances which arise from changes in bond lengths associated with distinct local coordination environments, referred to as displacive SRO. [89]To fully reproduce the experimental data, configurational SRO was also considered as it sets the local coordination environment of each species. [89]The results showed coupling between configurational and displacive SRO gave the best agreement with the experimental data, showing by combining these first-principles techniques with PDF analysis can be a powerful tool to provide a quantitative description of SRO. [89]eutron diffraction has also been applied to Li 1.25 Ni 0.625 Mn 1.125 O 3 F to study its cubic structure and lattice parameter. [90]Sites occupied by cations were initially identified from the Fourier difference maps extracted from X-ray and neutron diffraction data, and site occupancies were then refined simultaneously against both diffraction data using a global composition restraint set to the nominal composition of this material. [90]The neutron diffraction data proved critical for correct assignment of cation occupancies on the tetrahedral 8b position. [90]The 8b site has a positive neutron scattering density overall, indicating partial occupancy with Ni that has a positive neutron scattering length compared to the negative neutron scattering lengths of 7 Li and Mn. [91]The local, medium, and average structures of pyrochlore Nd 2 M 2 O 7 were examined using In all cases, G(r) is plotted in terms of Å -2 .Adapted with permission. [89]2023, American Chemical Society.a combination of neutron diffraction and theoretical analysis method. [92]A systematic and comprehensive investigation evaluating the relationship between the local and average structure of two closely related pyrochlores was demonstrated. [92]Neutron diffraction and nPDF analysis proved that the high entropy oxides (HEOs) form as pyrochlore Imma structural units, local TiO 6 octahedral distortion and some degree of short-range Ti clustering were identified in the Nd 2 (Ti 0.2 Nb 0.2 Sn 0.2 Hf 0.2 Zr 0.2 ) 2 O 7+x compound, in contrast to the homogeneous random distribution of cations observed in Nd 2 (Ta 0.2 Sc 0.2 Sn 0.2 Hf 0.2 Zr 0.2 ) 2 O 7 . [92]eutron diffraction and total scattering were also combined to investigate a series of single-phase 10component compositionally complex fluorite-based oxides, [(Pr 0.375 Nd 0.375 Yb 0.25 ) 2 (Ti 0.5 Hf 0.25 Zr 0.25 ) 2 O 7 ] 1-x [(DyHoErNb)O 7 ] x , denoted as 10CCFBO x Nb. [93] A long-range order-disorder transition (ODT) at x = 0.81±0.01from the ordered pyrochlore to disordered defect fluorite was discovered. [93]Neutron diffraction results suggested that this ODT occurs via the migration of O anions from the position 48f to 8a, with a small final jump to the ODT; however, the 8a O occupancy changes gradually. [93]

X-Ray Fluorescence (XRF) Scattering and X-Ray Compton Scattering (XCS) for Quantifying and Mapping Li Concentration
XRF occurs when the inner shell electrons of atoms are excited by the incident X-ray photons and subsequently release X-ray photons when the system relaxes and the electron transition from the higher energy level to the vacant inner shell. [94]The emitted secondary X-ray photons have spe-cific energies that are fingerprints of the atoms from which the photons are originated, so XRF is used for elemental analysis. [94]XRF offers elemental detection limits in the low parts per million (ppm) range. [95]Both in situ [96] and ex situ [96][97][98][99] XRF were complemented with inductively coupled plasma (ICP) measurements for quantitative analysis of TM stoichiometric ratios in battery cathode materials. [100]In recent studies, XRF techniques are also used for characterizing F-incorporated DRX materials, [111,112,[114][115][116][117][118] which are often unsuitable for ICP measurement due to the high ionization potential of F. [101] Total reflection X-ray fluorescence spectrometry (TXRF) [102] and wavelength dispersive X-ray fluorescence (WDXRF) [103,104] have been established for trace element analysis, their range of elemental detection has been expanded to include those with an atomic number as low as 6 (carbon). [103,104]In synchrotron XRF, elemental mapping technique is a commonly used for determining the elemental distribution and can be carried out in 2D and 3D. [105,106]s a popular cathode coating method, atomic layer deposition (ALD) of TiO 2 has been performed on the surface of various cathodes.Synchrotron XRF imaging was used to further confirm the uniformity of the TiO 2 distribution on the electrodes (Figure 7). [107]-ray Compton scattering (XCS) occurs when the outer shell electrons of atoms are ejected by the incident X-ray photons and subsequently release X-ray photons with a different wavelength as the incident X-rays in an inelastic and incoherent process.[108] XCS is an unique technique for detecting the interactions between incident photons and valence electrons of atoms or  [107] 2019, American Chemical Society.
compounds which are involved in Li intercalation and extraction redox reactions in cathode materials. [94]A quantification method was developed to quantify changing Li stoichiometry in cathode materials (e.g., LiNi 0.8 Mn 0.1 Co 0.1 O 2 , [109,94,95] Li x Mn 2 O 4 , [110] Li x FePO 4 [111,112] ) during battery cycling through calculating the electron momentum density ratio between the valence and core electrons from the X-ray scattering energy spectrum obtained by XCS.Operando correlative imaging of combining X-ray computed tomography (XCT) [113] with XCS imaging (XCS-I) was developed to quantify and map spatially resolved Li chemical stoichiometry distributions of the cathode compound inside different 3D electrode microstructures [114,59] to study the effects of anisotropic microstructure on Li ion diffusivity (Figure 8). [94,95]

Solid State Magic Angle Spinning Nuclear Magnetic Resonance (ssMAS NMR) for Directly Probing of Li Distribution and SRO
In ssNMR spectroscopy, magic-angle spinning (MAS) technique is routinely used to produce better resolution NMR spectra, MAS NMR consists in spinning the sample usually at a frequency of 1-130 kHz at the magic angle  m ca.54.74°where cos 2  m = 1/3 with respect to the direction of the magnetic field. [115] 6/7 Li NMR, [116] 19 F NMR, [117] and 17 O NMR [118] are among the most popular solid state magic angle spinning nuclear magnetic resonance (ssMAS NMR) techniques in probing the local structure and dynamics of DRX materials, [119][120][121]28] e.g. 19 FNMR is helpful for understanding the local structure of oxyfluoride-type DRX materials, [122] and 17 O NMR has been applied to study a range of cathode materials.] Sharp 7 Li signals centered at ≈0 ppm correspond to Li nuclei in diamagnetic impurities, whereas the chemical shift of Li nuclei in the cathode structure is dominated by paramagnetic interactions with unpaired d electron spins of TM ions such as Mn 3+126 .In DRX structures, Li ions are coordinated to 12 nearestneighbor metal (M) ions (M = Li + , Mn 3+ , or Ti 4+ ) through a 90°Li-O-M bond angle and to 6 next-nearest-neighbor M ions through a 180°Li-O-M bond angle, leading to a large number of different Li environments and a broad spectrum.[126] Small ordered domains in the Li 1.25 Nb 0.25 Mn 0.5 O 2 DRX were discovered by 7 Li ssMAS NMR.[28] It explained the formation of small -LiFeO 2 SRO domains within the DRX structure partially by the local electroneutrality preference, resulting in the preferential formation of OLi 3 Mn 3 and OLi 4 NbMn clusters for slowly cooled samples compared with rapidly cooled samples after sintering during material synthesis. [28] ] This in turn resulted in a greater voltage fade and poorer capacity retention as the -LiFeO 2 domains have higher Li + diffusion barriers.[28] Conversely, in ordered metal arrangements, such as those found in spinel or layered rocksalt oxides, the number of distinct Li environments is significantly reduced, resulting in discrete 7 Li ss-NMR resonances.[127] Ex-situ 19 F ssNMR studies on Li 1.15 Ni 0.45 Ti 0.3 Mo 0.1 O 1.85 F 0.15 (LNF15) cathode were performed to track and analyze changes in the local F environments in the bulk and on the surface of cathode, enabling a detailed understanding of the cathode material structure. [128] Interpretation f SRO has been achieved by combining ssNMR, DFT calculations, and MC simulations (Figure 9).[121] It was found that the incorporation of fluorine in as-synthesized LNF15 resulted in the emergence of short-range patterns, where F atoms occupied Li-rich positions surrounded by five or six neighboring Li atoms.Furthermore, F-incorporation greatly enhanced the capacity contribution from the Ni 2+ /Ni 4+ pair, allowing F-bonded Ni ions to oxidize via a different mechanism, with the generated Ni 3+ intermediate requiring lower energy to oxidize compared to those in NiO 6 octahedra.[121] Unlike in oxides, where the oxidation of Ni 2+ to Ni 4+ is incomplete due to the overlap between the O and Ni valence states, this outcome suggests that fluorination could be an effective approach to enhance the utilization of the Ni 2+ /Ni 4+ redox reservoir.9] For example: Li-Mn-Ni-O-F compounds with varying stoichiometries, the observable paramagnetic 19 F NMR signals were attributed to the F nuclei surrounded by six Li atoms in their first coordination shell.On the other hand, paramagnetic species, such as Mn and V, were found to be located in their second and/or third metal coordination shells.[130] However, signals originating from F directly bonded to Mn or V were not observable due to the strong paramagnetic interactions.[130] The weak 19 F signal obtained for all the DRX compounds with the Li-Mn-Ni-O-F chemical space and different stoichiometries suggests that a significant fraction of F is bonded to at least one TM, consistent with the predicted behavior of the DRX materials.[130] In addition, high entropy DRX materials contain multi-cations (>3) in the structure, methods utilizing exotic nuclei, e.g.61 Ni, 59 Co, 53 Cr, 47/49 Ti in NMR would be an attractive field for future studies.

Focused Ion-Beam (FIB) in Time-of-Flight Secondary Ion Mass Spectroscopy (ToF-SIMS) and Atom Probe Tomography (APT) for Analysis of Li Concentration and Diffusion
ToF-SIMS and APT can detect light elements such as Li for studying DRX materials. [131]ToF-SIMS is a surface-sensitive method that involves the ejection of secondary species from a sample outermost surface caused by the bombardment of a primary ion beam, only a small portion of the total sputtered population is ionized and can be directed toward a mass analyzer where the mass of the particles is determined by measuring the time at which they reach the detector. [132]FIB and depth profiling were used to accompany ToF-SIMS by sectioning a particle, revealing the cross-section of the particle for studying the ion distribution within the particle. [133]The ionization process is significantly influenced by the surface chemistry of the solid and surface modifications resulted from the sputtering process, commonly referred to as the matrix effect. [134]ToF-SIMS has been used to provide information on the distribution of cathode active materials when accompanied with FIB and depth profiling, [135,136] the formation of SEI during charge discharge cycles, [137][138][139] and detecting degradation species of cathodes. [140]he use of an FIB scanning electron microscopy (SEM) in ToF-SIMS presents a promising approach for the detection and quantification of Li and other active elements in battery materials in both fully charged and fully discharged states, e.g.LiNi 0.5 Mn 0.3 Co 0.2 O 2 (NMC532) down to 20 nm spatial resolution, [133,141] as the detection of discrete packets of secondary ions with varying masses as a function of time at the detector enables SIMS to exhibit the sensitivity that surpasses the analytical capability of energy dispersive spectroscopy (EDX). [142]gure 8. Operando correlative imaging of combining XCT with XCS-I.3D volume rendering of a region of cathode extracted virtually from inside of an operating battery coin cell from the XCT data, showing a) vertically oriented pore arrays inside a cathode made by directional ice templating (DIT); and b) random pore orientations inside a cathode made by isotropic ice templating (IIT).XCS-I results showing voxel-by-voxel color maps of the lateral distribution of Li chemical stoichiometry in the middle depth region of the DIT cathode at c) charged to 4.3 V; d) discharged to 2.75 V; and of the IIT cathode at e) charged to 4.3 V; f) discharged to 2.75 V. Adapted with permission. [114]F NMR spectra at high states of charge corresponds to diamagnetic F environments seen in the computed models.d) Calculated and experimental redox mechanism suggesting that nickel ions in octahedral coordination, when bonded directly to F, undergo a distinct oxidation mechanism compared to those surrounded by six oxygens.In this case, the nickel ions form Ni 3+ intermediates instead of transitioning directly from Ni 2+ to Ni 4+ during oxidation.Adapted with permission. [121]018, American Chemical Society.
Consequently, SIMS can detect trace quantities of low-mass elements, e.g.Li on surfaces. [143]The resulting ions or speciesspecific nanoscale maps revealed both crystallographic orientation and chemical effects in primary grains of the cathode material. [143]Figure 10 shows that the Li ions became localized and exhibited constrained mobility at the interfaces between particle-matrix interfaces, as well as at grain boundaries. [141]The 6 Li and 7 Li isotopes tracer diffusion coefficients of LiMn 2 O 4 thin films (200-300 nm thick) were determined via isotope ion-exchange and SIMS line analysis. [144]The edge effects are localized fluctuations in signal intensity, may arise due to changes in the sputtering yield caused by the sample topography, and the matrix effects arise because each material exhibits a distinct emission response to the primary ion-beam sputtering, [133] both affecting the SIMS results.Experimental calibration is required for different materials to determine their corresponding Li concentrations.Nanoscale elemental distributions of LiMnO 2 and other electrode materials were also reconstructed using APT, [145,146] showing the 3D distributions of elements such as Li, Mn, and O. [146] In contrast to a homogeneous distribution of Li observed in a LiNi 0.5 Mn 1.5 O 4 cathode, research revealed the existence of a heterogeneous distribution of Li in the cycled spinel Li 1.2 Ni 0.2 Mn 0.6 O 2 cathode, indicating a pronounced loss of Li upon cycling. [147]APT provides 3D high resolution compositional mapping of elements including H, C, or Li with subnanometer resolutions where the surface atoms from a sample in the shape of a very sharp tip are evaporated by field effect (near 100% ionization) and the resultant ions are projected onto a position sensitive detector (PSD) that are identified by their massto-charge ratio using a ToF mass spectrometry. [148]

Characterization Techniques to Address the Materials Degradation Challenge
Characterization techniques that detect elemental mixing include HAADF-STEM that measures the changing distances of Adv.Funct.Mater.2024, 2308165

(12 of 26)
Figure 10.The mapping of FIB-SEM/ToF-SIMS, which involves the combination of secondary electron images obtained from FIB-SEM with three normalized elemental distribution maps of Li, Mn, and Co for the fully discharged (top row) and fully charged (bottom row) samples.The entire field of view measures 8 μm.Adapted with permission. [141]2015, Elsevier.
TM-TM and O-O for resolving structural degradation of the materials through O and TM migration; [61] combining XAS and scanning transmission X-ray microscopy (STXM) that acquires nm resolution mapping of chemical species within individual particles to study dissolution mechanisms; [149] and ToF-SIMS depth profiling and high-resolution imaging [150] and APT [151] that have been used to visualize the concentration and spatial distribution of degradation species on the surface of active cathode particles.
Characterization techniques focusing on studying O loss degradation mechanism include 17 O ssNMR that probes local chemical environment around O, [29] and combining O-K resonant mapping of resonant inelastic X-ray scattering (RIXS), nPDF and ssNMR that identify the O redox and trapped molecular O 2 in the materials. [50]Finally, the electrochemical mass spectrometry (EC-MS) monitors the formation of intermediate species and gas evolution during electrochemical reactions to uncover unexpected side reactions. [152]We show each of the characterization techniques and their complementary use in the following section.

STEM-Based Characterization for Probing Structural Evolution and O-Redox During Cathode Degradation
HAADF-STEM imaging was used for resolving the structural degradation in bulk cathode during cycling. [21]The development of various O-redox models has been heavily debated from computational chemistry, [42,[44][45][46][47] Figure 11 shows HAADF-STEM and fast Fourier transform (FFT) demonstrated structural transformation in DRX materials is attributed to irreversible O-redox and loss of O 2 during cathode degradation processes. [21]he RIXS and NMR results explain the O 2 loss mechanism at bulk and surface levels occurs through in-plane TM migration, and F-incorporation stabilizes DRX structures by surface O 2 suppression. [29]The unified O 2 loss mechanism applies to a range of layered rocksalt materials, with the trapped molecular O 2 in the bulk level supported by evidence from the nPDF. [50]s an innovative approach to characterize the anionic redox in rocksalt cathodes, electron ptychography was applied to directly image oxygen shifts in Li 1.2 [Ni 0.13 Mn 0.54 Co 0.13 ]O 2 . [153,154,61]Oxygen shifts were seen in both the bulk and surface, which is indicative of the unified O 2 -loss mechanism, [29] where solid O 2 molecules appear to be generated and trapped in the bulk during cycling.
EELS mapping can directly visualize the distribution of Mn oxidation states near the surface.Figure 12 shows the reduction of the Mn valence can be reflected by the shift of the Mn L3 peak to lower energies in EELS mapping.A uniform elemental distribution in the pristine particle was observed, and formation of nanoscale voids in the cycled one, as indicated by the HAADF-STEM images and corresponding EELS maps. [155]Recent work uncovers a heterogenous Li 1.68 Mn 1.60 O 3.7 F 0.3 (LMOF0.3)cathode, composed of three distinct domains of spinel-like, DRXlike, and layered-like phases, in contrast to the traditional singlephase DRX cathodes. [156]The nanocomposite material exhibited different average Mn oxidation states for the spinel and rocksalt regions, as indicated by the varying ratios of Mn-L 3 /L 2 peaks observed in the EELS analysis of the two distinct regions.Ti was doped to enhance the DRX content within the spinel-like matrix of the nanocomposite particle and increased capacity retention by ≈30% after 50 cycles due to Adv.Funct.Mater.2024, 2308165

(13 of 26)
Figure 11.Degradation of LiNi 0.5 Ti 0.5 Mn 1.5 O 4 (LTMO) and LiNi 0.5 Ti 0. 5Mn 1.5 O 4 F 0.2 (LTMOF) cathode materials has been studied by a combined technique of HAADF-STEM and FFT, where a rocksalt-to-spinel-like structural transformation was observed upon cycling.By filtering the structural information in the Fourier transform, a combination of the filtered images showed that the nanosized spinel-like domains were dispersed in the rocksalt matrix.Further analysis using atomic-scale HAADF-STEM imaging showed that the structural change occurs gradually and continuously upon cycling.A magnified HAADF-STEM image of a boundary between rocksalt and spinel-like regions in LTMOF after 50 cycles signifies a disordered spinel-like structure, where the Td-8a sites were occupied by both Li and TM atoms.EELS elemental maps and spectra of the mixed-phase structures in LTMOF after 20 cycles indicated that there was no elemental segregation within different phases and the TM valences were consistent across different phases at the same depth beneath the surface.Reproduced with permission. [21]2022, John Wiley and Sons.
the reduction of cooperative Mn 3+ Jahn-Teller distortion and O-loss.
Combining STEM with electron ptychography offers the advantage of direct imaging of Li ions, this is highly beneficial in studying cathode material degradation mechanisms in LIBs. [61]Electron ptychography can uncover the movement of both heavy and light atoms during electrochemical reactions to understand the atomic structure and calculating the projected atomic bonds of battery materials in relation to their redox mechanisms.Focused-probe electron ptychography was utilized to investigate the average projected O-O distance at various stages in the first charge-discharge cycle, revealing oxygen deformation in the lattice structure.The method was applied to obtain simultaneous HAADF and ptychographic phase images of pristine Li 1.2 Ni 0.13 Mn 0.54 Co 0.13 O 2 projected parallel to the layers, with Li and O atoms invisible in the HAADF but visible in the phase image, allowing for direct measurement of the projected distance (Figure 13). [61]The atom distance was derived from multiple measurements using line profiles, although the values can be distorted due to sample drift during the slow-speed acquisition of the 4D datasets, which are sensitive to both heavy and light elements. [157]To generate zero-drift HAADF images and obtain the zero-drift TM-TM distances, multi-frame fast scanning of the HAADF images followed by rigid and non-rigid registration were performed, which served as a reference to calibrate the values directly measured from the ptychographic phase images at each charge-discharge stage.These distances were then used as a reference to calibrate values measured from the ptychographic phase images at each charge-discharge stage, and to quantify the changes in layer spacing and distortion of O octahedra coordinating the TM species during the first cycle. [61]igh-resolution STEM imaging is susceptible to sample drift and beam damage, which are further amplified in 4D STEM due to the longer dwell time required. [158]To address these issues, a multi-frame fast acquisition approach was conducted, which fractionates the dose across several images and reduces sample damage. [61]The postprocessing of the multi-frames using rigid and non-rigid registration algorithms produced high signal-to-noise ratio and zero-drift HAADF images. [159]The TM-TM distances derived from the zero-drift HAADF images were then used to calibrate distances in the ptychographic phase image, and the difference ratio was used to calibrate the O-O distances.The changes in the statistically analyzed O-O projected distance can reveal how the O shifts and how the TMO 6 2 ) cathode particles.HAADF-STEM images and corresponding EELS maps of nanoscale surface regions in a) a pristine particle; and b) a cycled particle.Corresponding color maps of the Mn L3 peak energy shift for c) pristine; and d) cycled particles.Converted 3D surface plots on the right side.Line profiles for the energy shifts measured along the direction marked by the black dashed arrows in the 2D maps are plotted at the bottom.Reproduced with permission. [155]2024, John Wiley and Sons.octahedra distort at different stages of cycling in the bulk region.

Combinations of Resonant Inelastic X-Ray Scattering (RIXS), STXM, and XAS for Analysis of O-Redox
A spectroscopic study of the O K-edge was conducted to probe the electronic structure of O in the charged material following one day and one month of rest. [43]High-resolution RIXS and soft XAS were used in combination to measure the filled and empty O 2p states of the cathode materials, respectively. [43]Previous research shows that RIXS of the O K-edge can overcome some of the limitations of conventional O K-edge XAS due to higher sensitivity to unconventional states that are triggered at very high and low voltages, making it a complementary method for detecting lattice oxygen redox. [160]Unlike the conventional O-K XAS, the oxidized oxygen in the lattice can be distinguished from the stronger and broader signals associated with TM-O hybridization in the emission energy in RIXS. [161]IXS is a photon emission technique, it typically offers greater bulk sensitivity compared to X-ray photoelectron spectroscopy (XPS) . [162]O RIXS maps were obtained by collecting the emission spectrum at a series of excitation energies across the O Kedge.In the one-day sample, a feature at 528 eV was evident in the XAS. [163]At this excitation energy in the RIXS map, an intense energy loss feature centered 1 eV below the elastic line was observed, there was also evidence of emission intensity at 0-1 eV energy loss, suggesting there are electronic states in this band all the way up to the Fermi level, supporting that the new O 2p electronic states are delocalized. [43]For the charged cathode material following one month of rest, there was a decrease in the intensity of the 528 eV features in the XAS spectrum and RIXS map, and growth in intensity at 531.5 eV.The decrease of the feature at 528 eV and increase at 531.5 eV by RIXS and XAS reflect the decay of the oxidized oxide ions to trapped oxygen, partially contributing to oxygen loss. [43]The RIXS line scans reveal a broad peak at 7.5 eV and a series of sharp peaks between 0-2 eV, which indicate the formation of molecular O 2 gas through bulk oxygen oxidation and also contribute to oxygen loss. [164]IXS and XAS were complemented by solid-state 17 O MAS NMR which is a bulk average probe of the local chemical environment around O and quantifying the amount of O 2 present in the bulk. [29]Through enrichment cathode materials with NMRactive 17 O nuclei, a new, strongly paramagnetic chemical environment forms during O redox, and a sharp, defined sideband manifold indicated that O 2 molecules that were rigidly confined within the crystal structure. [124]Through combining ex situ RIXS and in situ nPDF, the presence of short 1.2 Å O-O distances in a positively charged O-redox cathode material were demonstrated, [50] closely matched with the bond length in molecular O 2 . [165]The detection of molecular O 2 trapped in the material via nPDF provided a valuable complement to X-ray spectroscopy studies, such as RIXS. [71]They provide an initial structural proof of trapped O 2 in Li-rich cathodes.The approximate quantity of detected molecular O 2 was calculated to make up ≈20% of the O atoms octahedra, with atom types in the brackets of the names indicating nearby species to the midpoint of the projected bond.D) shows line profiles across the red lines from the HAADF and WDD phase images, with the phase image pointing out the locations of Li and O where they were absent in the line profiles from the HAADF image.Adapted with permission. [61]2022, Elsevier.
present in the cathode. [30]Integration from interdisciplinary subjects is expected to offer RIXS even more powerful applications with developing machine learning RIXS imaging techniques to achieve RIXS measurements with desired spatial resolution, [166] as machine learning is capable of handling large volumes of high-dimensional and noisy data generated by synchrotron techniques.This approach will provide unique insights into the interactions among multiple coexisting redox-active species in a range of battery cathodes, including DRX. [166] Synchrotron-based XAS combined with STXM with highresolution mapping at the scale of 20 nm acquired 2D quantitative data regarding the distribution of chemical species. [167]he combined STXM-XAS approach provides valuable insights into the distribution of chemical species, including their elemental composition and oxidation state. [168]Furthermore, the combination of STXM with angle-scan tomography permitted 3D microscopic analysis. [168]Spatially resolved 2D-and 3D-STXM were applied to reveal the variation of charge heterogeneity and cationic redox pattern in the DRX material Li 1.3 Mn 0.4 Nb 0.3 O 2 [169]   and high-nickel, low-cobalt NMC. [149]3D full field transmission X-ray microscopy (TXM) technique permits a quantitative evaluation of the 3D charge distribution, accounting for 20 million of XAS data (with ≈30 × 30 × 30 nm 3 nano-domain resolution) (Figure 14). [149]The depth-dependent analysis indicated that the surface of the material was highly oxidized, which can be attributed to the presence of a core-shell charge distribution. [149]x situ analysis with STXM-XAS was conducted on a phosphinemodified Li 1.17 Mn 0.44 Ni 0.35 Co 0.04 O 2 cathode that revealed the presence of a distinct area with high O, F, and C content between the secondary particles due to the agglomeration of the binder and conductive carbon where the low atomic mass of the associated elements are usually challenging to be detected and visualized using conventional electron microscopy techniques. [170](e-f) measure 1 μm.The color scheme for the Ni K-edge absorption energies is as follows: blue corresponds to lower edge energy, while red indicates higher edge energy.Adapted fwith permission. [149]2020, Springer Nature.

ssMAS NMR for Analyzing O-Redox
understanding the anion redox mechanism to utilize the extra capacity from O-redox in DRX while suppressing the irreversible O 2 loss.The unified O-redox model [123][124][125]29,43,50] with the trapped molecular O 2 is a result of combined STEM, RIXS, nPDF, and ssNMR analysis, where 17 O NMR indicates that the molecules are physically trapped in a condensed, solid-like environment. In anothe study, the nature of the hole states on O 2− before the formation of molecular O 2 and the chemical environment around the oxide ions in Na 0.6 [Li 0.2 Mn 0.8 ]O 2 were probed using 17 O ss-MAS NMR with 17 O-labelled samples.[43] The pristine material showed two well-defined chemical environments attributed to oxide ions bonded to two Mn (O-Mn 2 ) and three Mn ions (O-Mn 3 ).[43] Fast-relaxation 17 O ssMAS NMR revealed the existence of a strongly paramagnetic chemical environment of O for the charged material, which increased in intensity over time, further proved the presence of molecular O 2 trapped within the bulk of the cathode.[43]

ToF-SIMS and APT for Detecting Degradation-Induced Species in Cathodes
Extensive cycling of cathode materials is known to result in structural degradation such as amorphization and void for-mation at the surfaces of particles, which impede Li transport and cause rapid capacity loss. [21][21] Previous research suggests that the chemical and structural stability near the surface of cathode materials is critical in determining cycling performance. [171,172]A comprehensive understanding of the cycling-induced chemical reconstruction and TM oxidation states evolution associated with the structural evolution near the surface of cathode materials remained limited, e.g. the effects of fluorination on the cathode electrolyte interphase (CEI) layer formation on the surface of cathode particles remain unclear.This requires characterization such as using ToF-SIMS which was used in the study of cycled DRX cathodes, Li 1.2 Mn 0.4 Ti 0.4 O 2 (LMTO), and observed the existence of a thick and unstable CEI layer on cycled particles due to surface oxidation and dissolution of TMs (Mn 3+ , Mn 4+ ), loss of CO 2 and O 2 , and crystal structural changes. [173]oF-SIMS depth profiling and high-resolution imaging were used to study degradation of Ni-rich layered TM oxide cathodes and spinel cathodes, e.g.NaNi 0.7 Mn 0.15 Co 0.15 O 2 (NMC701515) [150] and TiO 2 -coated Ni-rich layered cathode, [174] to visualize and monitor the concentration and spatial distribution  [174] 2020, Elsevier.
of degradation species on the surface of active cathode particles, including adsorbed species, identified as C 3 H 2 − , sodium residues, identified as NaC 2 O 2 − , and nickel oxide, identified as 62 NiO − for NMC701515 [150] and NiF 4 , CoF 4 , and MnF 4 for the TiO 2 -coated Ni-rich layered cathode (Figure 15). [174]The signal strength of the TM species was relatively lower for the coated sample owing to the effective TiO 2 coating.A significant accumulation of F − was found on the cathode surface and deeply into the bulk particles, indicating pronounced chemical reactions between the NMC701515 particles and the polyvinylidene fluoride (PVDF) binder in the cathode. [150]CEI on two LMNO spinel cathodes, LiMn 1.5 Ni 0.5 O 4 and Li 1.2 Mn 0.6 Ni 0.2 O 2 , and an ultra-low cobalt cathode content material, LiN i0.94 Co 0.06 O 2 , were studied using ToF-SIMS, showing all the three cathode materials were initially covered by carbonate and hydroxyl-based CEI that underwent a dynamic change in chemical composition due to continuous electrolyte decomposition, particularly at high voltage. [175]Combined systematic XPS and ToF-SIMS were used to investigate Li-and Mn-rich Li 1.2 Mn 0.6 Ni 0.2 O 2 (LMNO) cathode materials [176] and Li 6 PS 5 Cl-based NMC composite cathodes, [177] as TOF-SIMS has a very high sensitivity toward compounds with high ionization probabilities (several orders of magnitude compared with XPS) and can detect relative concentration changes in decomposition products. [178]The results show a non-uniform coverage of the CEI on the particles after 100 cycles. [176]able 1.A summary of the various techniques described above and their characteristics relevant to studying DRX materials.The APT technique was utilized to successfully investigate the formation of the CEI layer on the surface of a LiMn 2 O 4 (LMO) electrode. [179]The results revealed the presence of a mixed layered-and mosaic structure comprising two Mn-rich layers adsorbed onto the LMO surface and an external mosaic structure consisting primarily of LiF, Li 2 O, and Li-carboxylates.Decomposition of the Li salt from the liquid electrolyte on the electrode surface began even before charge discharge cycles as shown by cryo-APT, leading to newly created grain boundaries that facilitated structural changes of the electrode at the nanoscale. [151]s structural defects assist the nucleation of Li-rich phases in subsequent lithiation and accelerate electrolyte decomposition, these insights into the interactions between the nanoscale structure and electrolyte contribute to improve the understanding of self-catalyzed/accelerated degradation of electrode materials. [151]ryogenic specimen preparation, transport under ultra-high vacuum and air have been investigated in detail for analysis of the electrode materials. [180]

EC-MS for Studying Gas Evolution During Degradation
In EC-MS, electrochemistry is used to determine the concentration or chemical reactivity of a substance by measuring either the potential, charge, or current; and mass spectroscopy is applied to detect gaseous products generated in an electrochemical cell. [181]By utilizing a mass spectrometer, EC-MS can effectively monitor the formation of intermediate species in electrochemical reactions, identify and quantify reaction products, advance understanding of ion formation, and uncover unexpected redox reactions. [152]For example, EC-MS has been used for evaluation of electrolyte decomposition inside a battery.The mass spectrometric signals from EC-MS detected different mass/charge number (m/z) values from the cycled battery which indicated the formation of different gases and were characteristic of the electrolyte decomposition reactions. [182]At the beginning of charging of the battery, a peak corresponding to m/z = 2 appeared in the spectrum, indicating H 2 evolution, referring to a reductive gas formation reaction, [182] and this may occur due to trace amounts of H 2 O in the electrolyte. [183]Apart from H 2 , m/z = 26, 27, 28, and 44 also appeared at the beginning of the charge cycle, these values are referred to the formation of C 2 H 4 , CO, and CO 2 , respectively.Almost at the end of the 1st charge cycle, a peak at m/z = 44 appeared, probably due to oxidative decomposition of solvent molecules in the electrolyte at higher voltages. [182]Overall, the formation of both CO 2 and H 2 was observed during 20 cycles whereas C 2 H 4 was only detected in the 1st cycle, the evolution rate of H 2 remained constant, but that of CO 2 increased, this may be associated with a change in the electrode/electrolyte interfacial structure upon cycling. [182]he on-chip EC-MS method employs a microfabricated membrane chip to regulate volatile species transfer from an electrochemical cell to a mass spectrometer, allowing instantaneous gas exchange to simulate cross-talk phenomena. [184]This study is one of the pioneering applications of EC-MS in the investigation of cathode degradation in LIBs.Additionally, the study also introduced a new cell design to enable operando measurements of gas evolution and gain insights into these intricate mechanisms of the impact of TM dissolution on the stability of the anodic SEI by monitoring ethylene evolution, as well as conducting isotopic labeling studies to analyze the evolution and consumption of CO 2 generated from the cathode.Additionally, ex situ surface-sensitive analysis techniques such as XPS and ToF-SIMS were employed to obtain complementary understanding of the reactivity and chemistry evolution of LIBs during operation.The DRX cathodes follow similar redox mechanisms with O 2 evolution upon degradation.Operando EC-MS was also utilized to track the gas emissions of not only cathode materials but also solid-state electrolyte materials such as Li 7 La 3 Zr 2 O 12 (LLZO). [185]By utilizing proper controls, CO 2 and O 2 were identified as emanating from the interface between the electrolyte and cathode.The gas evolution was correlated with a significant increase in cathode interfacial resistance as observed through potential-resolved impedance spectroscopy.To gain a comprehensive understanding of the intricate reactivity and chemical evolution of LIBs during operation, a combination of complementary and correlative surface sensitive analysis techniques, such as XPS and ToF-SIMS, can also be employed.

Figure 1 .
Figure1.a) Comparing the specific energy and specific capacity of different cathode materials with those of DRXs.Adapted with permission.[8]2021, American Chemical Society.b) The natural abundance of metals in the earth's crust and the price of each metal in USD kg −1 (shown in the blue figures on top of each bar).Adaptedwith permission[2][3][4][5] 2023, U.S. Geological Survey.c) Schematics of crystal structures for (I) the disordered rocksalt -LiFeO 2 structure in which all cation sites are equivalent, (II) the layered -NaFeO 2 structure, (III) the spinel-like low-temperature LiCoO 2 (LCO) structure, (IV) the -LiFeO 2 structure.Large empty circles indicate oxygen sites and small gray and black filled circles stand for lithium and TM sites, respectively.Adapted with permission.[6]2014, John Wiley and Sons.

Figure 2 .
Figure 2. A) Octahedral to tetrahedral to octahedral (o-t-o) diffusion involves two tetrahedral paths that connect each pair of neighboring octahedral sites.(B to D) In the activated state, diffusion can occur through different channels depending on the number of TMs involved.The diffusion channels are categorized as follows: B) The 0-TM channel refers to diffusion where the activated state shares face with no octahedral TMs.C) The 1-TM channel corresponds to diffusion where the activated state shares a face with one TM.D) The 2-TM channel represents diffusion where the activated state shares a face with two TMs.Reproduced with permission.[7]2014, The American Association for the Advancement of Science.

Figure 3 .
Figure 3.Comparison between activation barriers of the model system (o-LiMnO 2 with and without F/Li-excess) via both oxygen dumbbell hop (ODH) and tetrahedron site hop (TSH) mechanisms with available migration barriers for ODH and TSH mechanisms for various delithiated states of different layered, spinel, and olivine systems.The barriers for ODH, TSH, and o-t-o paths are shown as turquoise, yellow, and magenta bars respectively.Double stars (**) next to the name of the structure indicate a disordered structure.In the case of Li 1+x Mn 2−x O 4 , only the Li-excess phase (x = 0.125) is disordered.The polygons near the bars (triangle, square, and circle) specify where variations in the barriers come from (varying local environment, Li concentration, and Li slab space, respectively).Reproduced with permission.[22]2020, Royal Society of Chemistry.

Figure 4 .
Figure 4. 0-TM percolation analysis and short-range order parameters for four types of Li 1.2 compounds obtained at 1873 K. a) 0-TM percolating Li, and b) amount of Li in isolated Li 4 as a function of F content.Adapted with permission.[18]2020, John Wiley and Sons.

Figure 6 .
Figure 6.nPDF analysis based on random ionic configurations.a) Refinement performed using an average structure where all atoms occupy ideal positions on the fcc lattice.b) Refinement of a special quasi-random structure (SQS) relaxed with DFT calculations.Blue dots represent experimental data (Obs) and red/orange lines represent calculated values (Calc).Green lines represent the difference between these two (Diff).Panels c-e) show enlargements around the first three peaks where calculations and experiments differ most.In all cases, G(r) is plotted in terms of Å -2 .Adapted with permission.[89]2023, American Chemical Society.

Figure 7 .
Figure 7. XRF elemental mapping of Ti, Ni, Mn, and Co for an ALD treated NMC electrode after 20 charge and discharge cycles.Adapted with permission.[107]2019, American Chemical Society.

Figure 9 .
Figure 9. a) The 19 F spin echo NMR spectra were acquired for LNF15 and LiF powders using 60 kHz MAS.The spin echo spectra display spinning sidebands, which are marked with asterisks.The isotropic spectrum of LNF15, obtained through 19 F projected magic angle turning phase-adjusted sideband separation (pj-MATPASS), is depicted with a gray shading.b) The typical fluorine (F) local environment in a rocksalt TM oxyfluoride compound is illustrated, showcasing the first three coordination shells (P1, P2, and P3) surrounding the central F atom. c) SRO in as synthesized LNF15 is illustrated by the frequency of F-site types through DFT and MC simulation.The observed emergence of new signals in the ex situ19 F NMR spectra at high states of charge corresponds to diamagnetic F environments seen in the computed models.d) Calculated and experimental redox mechanism suggesting that nickel ions in octahedral coordination, when bonded directly to F, undergo a distinct oxidation mechanism compared to those surrounded by six oxygens.In this case, the nickel ions form Ni 3+ intermediates instead of transitioning directly from Ni 2+ to Ni 4+ during oxidation.Adapted with permission.[121]2018, American Chemical Society.

Figure 12 .
Figure 12.Cycling-induced chemical reconstruction and Mn oxidate states evolution in Li 1.2 Mn 0.7 Nb 0.1 O 1.8 F 0.2 (LMNOF 0.2 ) cathode particles.HAADF-STEM images and corresponding EELS maps of nanoscale surface regions in a) a pristine particle; and b) a cycled particle.Corresponding color maps of the Mn L3 peak energy shift for c) pristine; and d) cycled particles.Converted 3D surface plots on the right side.Line profiles for the energy shifts measured along the direction marked by the black dashed arrows in the 2D maps are plotted at the bottom.Reproduced with permission.[155]2024, John Wiley and Sons.

Figure 13 .
Figure 13.The measurement of the projected atom distance from [010] monoclinic zone axis of Li 1.2 [Ni 0.13 Mn 0.54 Co 0.13 ]O 2 using A) HAADF and B) weak-beam dark-field (WDD) imaging in a 4D STEM dataset.The atomic model is superimposed, with purple representing TM, red representing O, and green representing Li.The scale bar in A and B is 1 nm.The illustration in (C) depicts the named atom distances in TMO 6 (purple) and LiO 6 (green)octahedra, with atom types in the brackets of the names indicating nearby species to the midpoint of the projected bond.D) shows line profiles across the red lines from the HAADF and WDD phase images, with the phase image pointing out the locations of Li and O where they were absent in the line profiles from the HAADF image.Adapted with permission.[61]2022, Elsevier.

Figure 14 .
Figure 14.3D TXM results illustrate the distribution and dimension of the high valence Ni domains.a) Distribution of Ni valence states in three dimensions; b) Representative area displaying the distribution of Ni valence states in three dimensions; c) Valence gradient of nanodomains in a 2D plane for rod-shaped NMC (Nickel Manganese Cobalt); d) 3D distribution of Ni valence states; e) Representative region depicting the 3D distribution of Ni valence states; f) 2D valence gradient of nanodomains for gravel-shaped NMC.The nanodomain valence gradient vectors are illustrated by black arrows, with the direction and magnitude of the vector represented by the arrow's orientation and length respectively.The scale bars in (a) and (d) measure 3 μm, while the scale bars in (b-c) and(e-f) measure 1 μm.The color scheme for the Ni K-edge absorption energies is as follows: blue corresponds to lower edge energy, while red indicates higher edge energy.Adapted fwith permission.[149]2020, Springer Nature.

Figure 15 .
Figure 15.Transmission electron microscopy (TEM) images of two cathodes, NMC523 and TiO 2 -NMC523, after undergoing 200 cycles within a voltage range of 2.7-4.5 V. a) shows the TEM image of the NMC523 cathode, while b) shows the TEM image of the TiO 2 -NMC523 cathode.The figure also includes depth profiles of NiF 4 − , CoF 4 − , and MnF 4 − collected on the NMC523 and TiO 2 -NMC523 cathodes in (c).Illustrative ToF-SIMS chemical mapping is presented in (d) and (e) for the NMC523 and TiO 2 -NMC523 cathodes, respectively, after 200 cycles.Adapted with permission.[174]2020, Elsevier.

Table 1
below summarizes the various techniques and the characteristics associated with each technique.DRX materials hold great promise as Li-rich cathode materials for rechargeable batteries.The wider choice of TM elements and higher capacities due to the utilization of both TM and Oredox have made DRX materials attractive in terms of sustainability and advancements in performance.However, the presence of -LiFeO 2 type cation SRO impedes Li ion diffusion, reducing achievable capacities and redox kinetics.The DRX materials are also prone to degradation through irreversible O 2 loss and TM dissolution.Strategies to overcome these barriers include F incorporation, high entropy (multiple TM) modification, and sur-face coating of the particles.This review focuses on different types of advanced characterization techniques and combining the different complementary characterization techniques for correlative studies to jointly advance the fundamental understanding of DRX materials and their degradation mechanisms to show the effects of the strategies on performance improvements, e.g.visualizing cation SRO, Li percolation, tracking cationic/anionic redox, probing degradation mechanisms, monitoring surface reactions, and detecting F environments.Specific perspective toward achieving deeper understanding of DRX materials using complementary characterization techniques is summarized in Figure 16 and listed as the following: i.The cation SRO has been visualized and quantified by STEM, nPDF, and ssNMR.These are key techniques to effectively provide deep understanding of Li concentration distributions, Li percolation, and the change of SRO in DRX upon addition/reduction of different cation types, as these mechanisms have only been previously studied by DFT and MC simulations, for limited compositional space.Future investigation on SRO visualization toward the addition/reduction of different redox centers can be an effective option to guide cathode compositional design as well as further understanding of structural evolution upon cycling.ii.Compared to the cationic SRO, there are fewer studies on the anionic SRO in DRX.STEM with ptychography has the power of analyzing anionic SRO, owing to its ability to visualize light elements; nPDF has already been applied to the study of anion ordering.Given that different types of anions ordering in DRX and different arrangements of Li-rich environments, and hence, effective percolation in the structure, more advanced characterization can be applied to the anionic SRO studies.

Figure 16 .
Figure 16.Connecting the challenges for DRX and future outlook for the applications of complementary characterization techniques. 4