Ex situ Raman mapping of LiMn2O4 electrodes cycled in lithium-ion batteries

Ex Situ Raman Mapping of LiMn2O4Electrodes Cycled in Lithium-Ion Batteries Dominika A. Buchberger,*Bartosz Hamankiewicz, Monika Michalska, Alicja Głaszczka, and Andrzej Czerwinski Cite This: ACS Omega 2024, 9, 30381−30391 Read Online ACCESS Metrics & More Article Recommendations * sı Supporting Information ABSTRACT: In this study, we focus on the large-scale ex situ Raman mapping of LiMn2O4(LMO) electrodes maintained at varying states of charge. A comprehensive statistical analysis has been conducted at an area of ca. 3660 μm2on more than 3100 collected spectra for each LMO electrode sample. High-definition ex situ Raman maps provide profound insight into the lithiation process, offering an additional perspective on the mechanism of LMO intercalation. These maps clearly depict the coexistence of two phases, with evident phase transitions and state-ofcharge gradients. The set of spectra with various state-of-charge has been successfully deconvoluted taking into account the two-phase character of the ongoing reaction. In addition, we performed the study on the samples operated for 50 cycles at the high C-rates and tracked their delithiation state and impurity formation. This technique serves as a complementary visualization and analytical tool alongside other bulk-type methods employed in battery diagnostics. Importantly, this ex situ Raman mapping approach is applicable to any electrode material exhibiting a Raman response. 1. INTRODUCTION A spinel lithium manganese oxide (LiMn2O4) material is one of the cathode structures applied in secondary lithium-ion batteries. Several important advantages make it a suitable alternative to other currently researched materials, such as layered transition metal oxides. These benefits comprise its low cost, easy and nontoxic preparation, high discharge potential (4 V vs Li+/Li), practical capacity of 120 mAh g−1, high-energy density, and low self-discharge. 1−3 However, LiMn2O4(LMO) experiences a capacity fading during cycling, which restrains its applicability in commercial lithium-ion batteries. 4,5 To understand this undesirable trend, structural investigations over cycling have to be carried out. In recent years, many bulkand surface-type analyses such as in situ and ex situ techniques were demonstrated. 6−13 Raman spectroscopy stands out as a significant structural technique, primarily owing to its remarkable attributes such as high sensitivity, noninvasive nature, and nondestructive characteristics. More recently, the combination of optical microscopy and Raman spectroscopy has provided an additional advantage of efficiently scanning large sample areas. This renders it an ideal tool for examining structural changes in electrode materials. One of the first Raman spectra interpretations of the LMO spinel structure can be found in previous works of Julien et al. 14−16 Since then, the history of in situ and ex situ Raman analyses on the Li+intercalation process in LMO microcrystals has continued over 20 years and resulted in several important reports. 17−28 The first in situ measurement of Li+electroinsertion into a Pt/λ-MnO2electrode in a 0.1 M LiCl aqueous solution was presented in 1998 and showed that in situ recorded spectra were very similar to their ex situ data. 17 One of the most important Raman studies on the LiMn2O4material was introduced by Ammundsen et al. in 1999. 18 Their theoretical calculations on expected Raman-active modes in LMO spinel structures coupled with subsequent experimental evidence validating the calculated Raman shift positions render this research one of the most significant and valuable contributions in the field. They performed additional ex situ Raman measurements on two electrochemically delithiated and half-lithiated materials: λ-MnO2and Li0.5Mn2O4. The theoretical calculation model was further developed by M.M. Sinha and H.C. Gupta in 2002 and added a new interpretation to the available experimental data. 29 The first in situ Raman research on the Li-ion cell prepared in an Ar atmosphere was performed by W. Huang and R. Frech. 19 A two-electrode cell setup composed of the LMO cathode and lithium metal was assembled using a 1 M LiClO4 Received: February 15, 2024 Revised: May 7, 2024 Accepted: June 18, 2024 Published: July 1, 2024 Article http://pubs.acs.org/journal/acsodf © 2024 The Authors. Published by American Chemical Society 30381 https://doi.org/10.1021/acsomega.4c01480 ACS Omega 2024, 9, 30381−30391 This article is licensed under CC-BY 4.0 Downloaded via TECHL UNIV OF OSTRAVA on April 2, 2025 at 09:40:19 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. Table 1. Literature Data on LiMn2O4Were Obtained Using Ex Situ and In Situ Raman Studies laser used (power) electrolyte working electrode electrochemical cell novelty measurement type year ref. Nd:YAG laser at 532 nm (1 mW) 0.1 M LiCl + 0.05 M borate buffer aqueous solution thin layer electrode Pt/λMnO2 three-electrode: aqueous the first in situ Raman study of electrochemical Li insertion in the MnO2spinel structure in aqueous solution. in situ and ex situ 1998 17 Nd:YAG laser at 532 nm (unknown) 0.01 M LiCl aqueous solution thin film electrode Pt/LMO & further Pt/λ-MnO2 two-electrode: aqueous theoretical calculation of Raman active phonon positions in LiMn2O4, Li0.5Mn2O4, and MnO2and experimental evidence. ex situ 1999 18 Ar laser: 514.5 nm line (20 mW) 1 M LiClO4in 1:1 wt EC:DMC standard electrode slurry onto Al meshes. two-electrode: split-type cell (Ar) the first in situ Raman studies of LMO electrodes during lithium intercalation in Ar-sealed Li-ion cell cycled between 4.6 and 2.1 V vs Li metal in situ 1999 19 Nd:YAG laser at 532 nm (∼8 mW) 1 M LiPF6in 2:1 wt EC/DMC LiMn2O4particle electrodes in Au foil two-electrode: split-type cell (Ar) this research exploits the capabilities of in situ Raman technique for the acquisition of time-resolved Raman spectra of single particles of LiMn2O4embedded in Au foil substrate electrodes as a function of the applied potential. in situ 2001 20 Nd:YAG laser at 532 nm (low) 1 M LiClO41:1 vol. EC:DEC microelectrode Pt/LiMn2O4 microparticle two-electrode: microelectrode (Ar) statistical analyses of the spectra in the range 15% < SOD < 45% showed to be consistent with the coexistence of two distinct phases of lithiated metal oxide and agreed well in situ XRD measurements. in situ 2003 21 Ar laser: 514.5 nm line (unknown) 1 M LiClO4in PC/ EC Li1−xMn2O4thin films produced by electrostatic spray deposition two-electrode: microelectrode (Ar) the first to report in situ Raman measurements of structural changes of an electrostatic spray deposited thin-film (without a binder or conductive agents) of Li1−xMn2O4during lithium insertion and extraction processes. in situ and ex situ 2003 22 Ar laser: 514.5 nm line (10 Wcm-2) unknown unknown (Li0.5Mn2O4sample) unknown (electrochemical Li extraction) the study of the local structure of various lithium manganese oxide stoichiometries using both the classical group factor analysis and a local environment model. ex situ 2003 23 Nd:YAG laser at 532 nm (3 mW) 1 M LiPF6in 1:1 vol EC/DMC LiMn2O4single-crystal microelectrode two-electrode: microelectrode (Ar) the contributions of the three crystallographic phases of LixMn2O40 < x< 1 as a function of the amount of Li+in the lattice derived from the optical data were consistent with those extracted from a coulometric analysis of the voltammetric curves. in situ 2005 24 unknown unknown LMO electrode from commercial cell two-electrode: standard commercial ex situ Raman study on the commercial LMO electrodes at their charged and discharged states. ex situ 2015 25 514 nm laser (∼0.23 mW) 1 M LiPF6in 7:3 wt EMC/EC LiMn2O4, CB, and PVdF (80:10:10 by weight) on Al foil (∼20 μm thick) two-electrode: pouchtype (Li as CE and Celgard separator) the correlation between the electrochemical data and structural responses of LMO electrodes (commercial LMO material) charged to high anodic potentials of 4.3 to 5.1 V vs Li ex situ 2017 26 488 nm laser (unknown) unknown LMO electrode from commercial cylindrical cells two-electrode: commercial cylindrical the understanding of delithiation and degradation paths by local CRM measurements of the LMO cathode material with different “state of charge” and “state of health” parameters. ex situ 2018 27 ACS Omega http://pubs.acs.org/journal/acsodf Article https://doi.org/10.1021/acsomega.4c01480 ACS Omega 2024, 9, 30381−30391 30382 solution in 1:1 wt EC:DMC. The cell was measured in both directions (over charge and discharge processes). The authors stated that during the in situ Raman study of the LMO electrode, the spectral changes indicated that a single-phase reaction is followed by a two-phase reaction and then by another single-phase reaction as Li+ions are deintercalated from the stoichiometric LiMn2O4. When the spinel LMO structure is further intercalated with Li+forming the Li-rich phase, the in situ Raman spectral changes are consistent with a two-phase reaction between cubic LiMn2O4and tetragonal Li2Mn2O4. In situ Raman measurements were continued between 2001 and 2005. Raman spectra of isolated single particles of LMO embedded in Au foils were recorded in situ in 1 M LiPF6in 2:1 wt EC:DMC using a Raman microscope by D. Scherson’s group. 20 Authors correlated their in situ spectroscopic data of a single LMO particle directly with its electrical state. Their other work was based on simultaneous Raman spectra and CV data acquisition from the LMO single particle microelectrode immersed in a 1 M LiClO41:1 vol EC:DEC solution. 21 Statistical analyses were conducted on the spectra within the range of 15% < SOD < 45% (SOD�state of discharge). The results were consistent with the coexistence of two distinct phases of lithiated metal oxide and well-agreed with in situ XRD measurements reported previously. 30−32 The first in situ Raman spectra collection on Li1−xMn2O4thin films produced by electrostatic spray deposition was recorded during a CV scan in 1 M LiClO4in PC:EC solutions. 22 Authors suggested that each value of xin Li1−xMn2O4possesses its own electronic band structure. In 2005, a very interesting and thorough in situ Raman study was performed on a LiMn2O4single-crystal microelectrode recorded as a function of potential vs Li in a 1 M LiPF61:1 vol EC:DMC solution. 24 Based on this study, it was clearly evident that two well-defined steps were involved in the charge process of the LMO material and were in agreement with the voltammetric characteristics The corresponding SOD (state of discharge) vs Eplots, which were obtained by integration of the linear scan voltammogram, were very similar in shape. A couple of the most recent ex situ Raman studies cover the study of commercial LMO electrodes at their charged and discharged states 25 and the correlation between the electrochemical data and structural responses of LMO electrodes charged to high anodic potentials of 4.3 to 5.1 V vs Li/Li+. 26 Lately, the Raman mapping was also employed to study the LMO grains in two lithiated states x= 0.1 and 0.4 before and after aging (discharged at 1 and 16 C, respectively) of the commercial LMO electrodes. 27 The authors stated that cycling leads to (1) the formation of the Mn3O4phase with its further dissolution in the electrolyte and (2) qualitative change in the lithiation process in cycled LMO cathodes with significant inhomogeneity of the formed lithiation state. The summary of the state-of-the-art in terms of in situ and ex situ Raman studies on LMO materials is presented in Tables 1 and S1 (extended). In this study, ex situ Raman mapping is applied to investigate LiMn2O4electrodes cycled in lithium-ion cells. The main novelty lies in the ability to analyze a high number of spectra from extensive Raman maps covering a substantial surface area (∼3660 μm2) of the electrodes exposed to various potentials vs Li. This study presents the first comprehensive investigation of the state of charge gradient among LMO particles within positive electrodes using Raman spectroscopy. Furthermore, the study was carried out on the samples operated at the high C-rates for 50 cycles, following the delithiation state and the formation of impurities. This ex situ Raman approach combined with microscopic imaging improves the understanding of Li intercalation processes in the real battery. Moreover, this measurement approach holds the potential for application to many other electrode materials in the future. 2. EXPERIMENTAL SECTION 2.1. Sol−Gel Synthesis of LiMn2O4.Lithium manganese oxide (LiMn2O4) powder was synthesized via the sol−gel method using citric acid (C6H8O7·H2O, 99.5%, Sigma-Aldrich) and acetic acid (C2H4O2, 99.5%, Sigma-Aldrich) as main and minor complexing agents. Lithium acetate dihydrate (CH3COOLi·2H2O, 97%, Fluka) and manganese acetate tetrahydrate ((CH3COO)2Mn·4H2O, 99%, Sigma-Aldrich) were used in a molar ratio of Li:Mn = 1:2. The salts were separately dissolved in deionized water, followed by mixing them together and stirring for a few hours. Later, citric acid was slowly added to the solution, followed by acetic acid addition. The ratio of metal to citric acid was 1:1, and the ratio of citric to acetic acids was 1:0.25, respectively. The solution was thoroughly stirred and slowly evaporated. When the solution became a viscous transparent gel, it was dried for a few hours at 150 °C in the air atmosphere and finally ground in an agate mortar to obtain a fine powder. This xerogel was calcined at 300 °C for 7 h to 700 °C for 5 h under air flow at a rate of 5 °C min−1. The as-prepared product will be labeled as LMO. 2.2. Structural and Morphological Analysis. A scanning electron microscope (Hitachi S5500, Hitachi High Technologies Corporation, Japan) with an accelerating voltage of 5 kV was used for the LMO sample morphology determination and the estimation of the particle size. A Siemens D-500 X-ray diffractometer with CuKαradiation (λ= 1.542 Å) was used to investigate the phase composition of the LMO powder. The Match! Three and FullProf software were used to calculate the crystal size (using the (111) plane) and lattice parameters. The diffraction pattern was recorded from 10°to 60°using a 0.002°step size (0.04°min−1). The Scherrer equation was used to estimate the crystal size. 33 VESTA software was applied for creating crystallographic structure schematics using .cif files obtained from the Crystallography Open Database. 34 The N2adsorption/desorption experiments were conducted on a Micromeritics ASAP 2060 apparatus at 77.349 K absolute temperature in the relative pressure range of 0.01−0.995 p(p0)−1. Adsorption/desorption isotherm analysis was performed on ASAP 2060 software by calculating the specific surface area using the Brunauer−Emmett−Teller (BET) method and the distribution of pores and their volumes using the Barrett−Joyner−Halenda approach for desorption curves. 2.3. Sample Preparation. Electrodes were prepared using a standard laboratory procedure by a doctor blade slurry coating method. The slurry has been formed by mixing 80 wt % LiMn2O4spinel, 10 wt % carbon VulcanXC72R, and 10 wt % PVdF binder (predissolved in NMP solution). A 200 μm applicator was used in order to receive a thin film coating with similar cathode loadings. The amount of electrode material on Al foil was about 2.5 mg cm−2. The round electrodes of 0.9 cm diameter were cut from the foil and pressed under 20 MPa pressure. Before the cells were assembled, the electrodes were dried at 120 °C in vacuum for 15 h. ACS Omega http://pubs.acs.org/journal/acsodf Article https://doi.org/10.1021/acsomega.4c01480 ACS Omega 2024, 9, 30381−30391 30383 2.4. Electrochemical Measurements. Electrodes containing LiMn2O4nanomaterials were cycled versus lithium in the Swagelok-type cells and either fully charged (4.50 V) or discharged (3.55 V) or kept on partially lithiated states in order to study the intermediate structures. The material was electrochemically tested using current densities of 1 to 30 C. For the Raman study, the electrodes were cycled 3 times at the 1C rate and kept at different potentials, namely, 3.55 (100% DoD), 4.01, 4.055, 4.15, and 4.5 V (100% SoC). Next, cells were disassembled, and electrodes were washed carefully in dimethyl carbonate and allowed to dry in an argon atmosphere before further measurements. Additionally, a series of LMO electrodes were cycled over 50 cycles at 1 C, 2 C, and 5 C-rates and kept at the charged state (4.5 V) for 1 h [constant current constant voltage (CCCV) method]. Similar to the previous set of samples, cells were disassembled, and electrodes were washed carefully in dimethyl carbonate and allowed to dry in an argon atmosphere before performing further measurements. All potentials are given relative to the lithium electrode (Li/ Li+) unless otherwise stated. 2.5. Raman Analysis. Raman measurements were performed using a Renishaw inVia confocal Raman microscope and a Nd:YAG laser (wavelength: 532 nm; maximum power: 17 mW). The Lorentz function was used to fit the spectral lines. The LMO electrode maps were measured at the area of ∼3660 μm2with a 1.1 μm step (3136 points). Each spectrum was collected for 40 s. The Raman mapping was applied to this study to increase the measurement statics. The intensity of the laser was decreased to avoid the decomposition of the spinel material (0.1 mW). The spectra deconvolution was performed using OriginPro software and the multipeak fitting method. The position of Raman lines originating from all three phases (LiMn2O4, Li0.5Mn2O4, and λ-MnO2) during the electrochemical reaction was estimated using the theoretical predictions calculated in the work of Ammundsen et al. using atomistic simulations. 18 3. RESULTS 3.1. Basic Structural Analysis of LiMn2O4Powder. SEM images of the LiMn2O4powder show well-defined, agglomerated crystals. During the synthesis, they create cuboctahedrons and truncated cuboctahedrons with characteristic triangular facets and plane growth marks close to the edges (Figure 1). These smooth triangular surfaces indicate a selective growth of the direction family ⟨111⟩(Figure 1B) (all 8 directions) resulting in the well-established (111) facets. Based on theoretical calculations reported previously, 35 it is predicted that the (111) plane makes the most energetically stable surface facet in the LiMn2O4spinel structure. Our experimental results agree well with those estimates that the predominant facet is the (111) plane (followed by two others: (100) and (110) surfaces expected at similar energies). We experimentally observe a cuboctahedral shape with predominant (111) facets, which must possess the lowest surface energies. The truncated cuboctahedrons possess, most probably, additional small (100) surface facets. The representation of the (111) plane in different crystallographic directions is shown in Figure 1D−F, indicating that the Li channels are well exposed for such a morphology. Additionally, SEM imaging reveals a particle size with a distribution of 10 to 400 nm (with an average of about 150 nm). BET analysis indicates that the specific active area of this LMO powder is 3.18 m2/g, whereas the diameters of the Figure 1. SEM images of the pure LiMn2O4material at two magnifications: 20.000×(A) and 100.000×(B). Schematic of the cubic crystal structure (Fd3mspace group) is inserted onto the B image to show the selective facet growth in LMO crystals. A characteristic triangular shape of the crystal facets indicates a family of {111} planes. The small areas of the (100) surfaces in truncated cuboctahedrons are also visible. (C) XRD pattern of as-synthesized LiMn2O4. (D−F) (111) plane presented in different crystallographic directions, where green balls represent Li ions, violet�Mn, and red�oxygens. ACS Omega http://pubs.acs.org/journal/acsodf Article https://doi.org/10.1021/acsomega.4c01480 ACS Omega 2024, 9, 30381−30391 30384 dominant pores are about 1.5 and 30 nm. The second pore size value comes from the empty spaces between cuboctahedral grains (Figure 1A, B). The XRD analysis (Figure 1C) shows that the as-synthesized LMO powder is a single phase of the Fd3mspinel structure (ICDD PDF 35-0782). All characteristic reflexes are indexed in Figure 1C. It is noteworthy that the most intense peak at 18.7° originates from the diffraction of (111) planes, which are represented morphologically through triangular facets on SEM images. The Scherrer formula allows for a calculation of the average crystal size, which is about 75 nm. This fits well with the previous findings of the electron microscopy measurements and suggests the agglomeration of grains into a polycrystal (double-grain and more) particle. Furthermore, the alattice parameter is about 8.246 Å, whereas the cell volume is 560.7 Å, which is consistent with the standard values: a0= 8.248 Å and V0= 561.1 Å3from the ICDD PDF card. 3.2. Electrochemical Testing. We conducted a series of chronopotentiometric experiments to explore the electrochemical behavior of LMO materials under different current rates (Figure S1). At a slower current rate of 1 C, the initial specific capacity was found to be 113 mAh g−1. To simulate rapid charge and discharge conditions, we measured the electrochemical response at higher current rates of 2, 5, 10, and 30 C and obtained specific capacities of 112, 110, 107, and 99 mAh g−1, respectively. After 100 cycles, the materials retained over 95% of their initial capacity. Figure 2A shows the charging curve with indicators showing the potentials at which samples were stabilized and collected for ex situ Raman investigation. 3.3. Ex Situ Raman Mapping on Electrodes at Different State-of-Charges. A thorough ex situ Raman mapping technique is employed to investigate the changes occurring in the LMO structure during delithiation. After completing 3 cycles, the electrodes are charged to various potentials relative to Li/Li+in order to explore different stages of lithiation. Specifically, we examine the beginning state (3.5 V), the first plateau (4.01 V), the middle state (4.05 V), the second plateau (4.15 V), and the fully charged state (4.5 V) of the electrodes. Figure 2B shows the crystallographic model of the charging process with evolution from the LiMn2O4crystal structure through the Li0.5Mn2O4phase to the final λ-MnO2 phase. The electrochemical reaction is described below: 2LiMn O (Li e 2Li Mn O ) 2Li 2e 4 MnO 2 4 0.5 2 4 2 + + + + + + Raman maps perfectly exhibit a stationary state at each potential (Figure 3). Only fully charged and discharged states show very homogeneous spectra over the entire measured area of the electrode. A full discharged state shows that the active electrode material returns to the LiMn2O4structure (purple area). A minority of crystallites create the Li-rich Li1+zMn2O4 type of the spinel structure (where zis the excess of Li ions) 23 since a fitting analysis shows that the most intense peak (A1g phonon mode) shifts to a higher frequency region (from 627 cm−1up to 635 cm−1) for those few spectra (Figure S2). This indicates a shortening of the Mn−O bond. It is known that in Li-rich LMO (Li1+zMn2O4compounds), Li ions occupy available 16d octahedral sites and cause a distortion of MnO6octahedra. 23,36 The Raman map of a fully charged electrode validates the λMnO2-type structure as its main phase. 23,37 The main line (A1g) is located at 588 cm−1followed by the small peaks at 495 cm−1(F2g) and 460 cm−1(Eg), as well as another minor line (F2g) at 640 cm−1. It is also worth mentioning that over map investigations, a very small amount (below 1%) of Mn3O4 compounds are also detected. 38 Based on a prior analysis of the powder LiMn2O4material, it is known that the impurity phases are postsynthetic undesirable residues and can be a partial Figure 2. (A) Charging curve showing points at which electrodes were disassembled. (B) Crystal structures showing the evolution of the LiMn2O4crystal through Li0.5Mn2O4and the final λ-MnO2phase. Figure 3. Raman maps and corresponding spectra showing clearly that Li deintercalation is a three-phase process in the Li1−xMn2O4 structure (colors on the maps correspond to the colors of spectra below them). The schematic pictures of the delithiation process are also presented at the top. ACS Omega http://pubs.acs.org/journal/acsodf Article https://doi.org/10.1021/acsomega.4c01480 ACS Omega 2024, 9, 30381−30391 30385 reason for the initial electrochemical capacity value lower than theoretically expected for the LMO material. The intermediate processes are more varied. At 4.01 V, spectra change and bands from 670 up to 580 cm−1start to rise (Figure 3�blue area and Figure 4). This structural change continues until reaching an intermediate state (4.05 V) where the bands above 600 cm−1decrease, while the band at ca. 592 cm−1takes the lead and becomes the most intense peak in the spectra representing a structure close to Li0.5Mn2O4and occupying most of the electrode area (green spots). The Li0.5Mn2O4crystal has Li ions in every second tetrahedral site of fully lithiated LiMn2O4(Figure 2B). Spectra around the half-charge state, which originate from Li1−xMn2O4as well as Li0.5−yMn2O4compositions (where xand yare above 0 and lower than 0.5), are also present and visibly vary from Li0.5Mn2O4. Since diverse spectra coexist within partially charged samples, it shows that LMO grains exhibit the Li intercalation gradient during charging, which can be caused by their differences in the crystal size and thus their different levels of delithiation. Further delithiation (at 4.15 V) during the second plateau demonstrates that the Raman spectra continue to evolve into a highly delithiated phase, although it is not yet a pure λ-MnO2-type structure. The fully delithiated phase (λ-MnO2structure) is present only at 4.5 V. The delithiation gradient through the charging process significantly shows that the electrochemical reaction inside grains appears at different speeds. The cause of this effect can be explained by the distribution of grain size within the electrode or the agglomerated structure of the material, as shown in the SEM images (Figure 1). Such a situation can cause the environment for the local overcharging (overpotential) of grains and thus successive material degradation over prolonged cycling. A comprehensive Raman spectra analysis is performed to track line intensity changes (Figures 4 and S3, S4). The selected spectra are normalized to the most intense peak, carefully fitted (Figure 4A), and for a good presentation of the most intense peaks during charging at different delithiation stages, each peak area is displayed as the percentage of the full area under the spectra (Figure 4B). The Raman fitting was performed considering phase transition stages and the possible overlapping of structures over cycling. The irreducible representations of all three structures are denoted by A (R) E (R) F (in) 3F (R) 2A (in) 2E (in) 4F (ir) 2F (in) (LiMn O ) 1g g 1g 2g 2u u 1u 2u 2 4 = + + + + + + + 3A (R) 3E(R) 3F(in) 6F (R, ir) (Li Mn O ) 1 1 2 0.5 2 4 = + + + A (R) E (R) F (in) 2F (R) 2A (in) 2E (in) 3F (ir) 2F (in) ( Mn O ) 1g g 1g 2g 2u u 1u 2u 2 4 = + + + + + + + where (R) represents Raman-active vibration, (ir) infraredactive vibrations, and (in) are inactive modes. 23,38 It means that there are 5 predicted Raman modes in LiMn2O4, 12 in Li0.5Mn2O4, and 4 in the λ-MnO2structure. These assumptions as well as previous calculations of Raman band positions 18 in those structures were considered during fitting. All Raman line positions with band assignments and band widths for those selected spectra of LixMn2O4across cycling are presented in Table S2. By such an analysis, the evolution of Raman lines can be tracked. It is visible that the A1g phonon mode (around 627 cm−1) of the fully lithiated structure of LiMn2O4decreases by about 11% upon the first charging plateau and follows the next intensity decrease within the intermediate F43mphase formation, whereas it completely vanishes for the highly delithiated λ-MnO2structure. Since 627 cm−1is not present in theoretical Raman modes of Li0.5Mn2O4, the existence of this line is the most probable due to the presence of the solid solution LiMn2O4−Li0.5Mn2O4within the grain giving the response from both phases. Upon charging, the F2g(3) line (583 cm−1) in the lithiated Fd3mstructure spits into two lines: at 597 (A1) and 580 cm−1(E). During further delithiation, the A1 mode of the F43mphase increases and slightly shifts to 589 cm−1to become a leading peak (A1g phonon mode) for the fully delithiated λ-MnO2structure (588 cm−1). The F2g(2) line (482 cm−1) transforms into the F2line (485 cm−1) in the F43mphase with almost unchanged intensity. After the halfcharge state, it changes back into the F2g(2) line for the delithiated Fd3m structure. The other F2line in the F43m phase (at 610 cm−1) increases over charging and start to decay for lithium concentrations x< 0.2. It is also noted that very small traces of lines transformed from the F43mphase are visible in the Raman fitting of the fully charged sample, and they might be a representation of not-fully charged particles Figure 4. (A) Fitting of the Raman spectra and (B) Raman peak area dependence on delithiation of the LixMn2O4structure, where xis the Li content (0 ≤x≤1), LMO-LiMn2O4and MO-λ-MnO2. ACS Omega http://pubs.acs.org/journal/acsodf Article https://doi.org/10.1021/acsomega.4c01480 ACS Omega 2024, 9, 30381−30391 30386 overlapping with the pure λ-MnO2phase even stronger, supporting the Li concentration gradients within the LMO grains at the particular state-of-charge of the electrode. Figure S3 shows the evolution of spectra from x= 0 to x= 1 in LixMn2O4. Horizontal and vertical profiles of the map display the represented Raman spectra at transition stages and normalized intensity changes across charging, respectively. Additionally, an integral intensity ratio between the A1g Raman mode of the lithiated Fd3mstructure (ca. 627 cm−1) to the sum of the A1mode of the partially delithiated F43mphase (ca. 592 cm−1) and the A1g phonon mode of the highly delithiated λ-MnO2structure (ca. 588 cm−1) is shown in Figure S4. It demonstrates that the charge state of the particle can be calibrated and tracked numerically, giving the Raman technique a significant advantage over other structural methods. 3.4. Ex Situ Raman Mapping on Electrodes Charged at Different C-Rates. The developed Raman mapping method was used to study electrodes charged at different Crates (1, 2, and 5 C). The resulting Raman maps, shown in Figure 5, illustrate the distribution of the highly delithiated structures after 50 cycles. Corresponding spectra are presented in Figure S5. The electrode cycled at a 1 C-rate exhibited the most uniform distribution. The position of the A1g phonon mode of the fully delithiated λ-MnO2structure (violet-blue area) ranges from 587 to 588.5 cm−1. Approximately 10% of the map area (green) represents grains of the slightly lithiated F43mphase structure, where the A1mode position is at ca. 590 cm−1. When cycling electrodes at the higher C-rate, the green to red areas indicate the A1mode (position 589−591 cm−1) of the F43mstructure rather than the A1g mode of the fully delithiated spinel. At the 2 C-rate, the map shows the areas of the A1mode at 591 cm−1. After cycling at the 5 C-rate, it is clear that most of the grains are not fully delithiated, despite using the CCCV method to prepare the samples, where a constant voltage was applied to stabilize the electrode’s potential. The results indicate a limitation in Li+diffusion within the spinel structure and an increase in interfacial resistance at the surface of the spinel crystal after extended cycling at a high rate. This process is visible even at lower Crates but becomes more pronounced at higher rates. The location of the undercharged areas shows that this effect appears in the middle of the agglomerates. It is less pronounced on the outer sides where the contact with the conductive carbon is higher. This is shown schematically in Figure 5, where Li+ions are still in the middle of the crystal for the 5 C rate charged electrode. Isolated particles with the LMO agglomerates will be highly influenced by this effect due to the electric resistance, lower charge transfer, and inefficient ionic mobility. This shows that the electrical contact of each LMO particle within the electrode layer is crucial for better performance at high-rate applications. Furthermore, we conducted a study on impurity formation, as shown in Figure 6. Cycled electrodes contain minor areas Figure 5. Raman maps of the electrodes charged (up to 4.5 V) after the 50th cycle, charge/discharge curves of the 1st, 2nd, and 50th cycles, and crystal schematics showing the influence of the C-rate from 1 to 5 C on the performance of the Li1−xMn2O4structure (Li mark in green). Black spots on the map represent conductive carbon. ACS Omega http://pubs.acs.org/journal/acsodf Article https://doi.org/10.1021/acsomega.4c01480 ACS Omega 2024, 9, 30381−30391 30387 with a combination of additional phases, namely, MnO2 (todorokite) and Mn3O4(Figure S6). 39 The concentrations of these mixtures after cycling at 1, 2, and 5 C-rates are ca. 2, 3, and 5%, respectively. These concentrations are higher than the initial concentration of Mn3O4detected in the pure electrodes (up to 1%). It is evident that the amount of these phases increases with an increased C-rate. This indicates that the delithiated spinel structure gradually transitions into other MnxOyphases after extended cycling, and the process is strengthened for higher C-rates. This is in agreement with previous TEM studies, where the transition from the λ-MnO2 structure into the Mn3O4phase was observed at 4.3 V. 40 The formation of the monoclinic MnO2(todorokite) might be a result of the Mn2+ dissolution from the tetrahedral sides of Mn3O4and the creation of the Mn-deficient decomposition product. The coexistence of both reconstruction phases (MnO2todorokite and Mn3O4) in the same Raman spots strengthens this conclusion. This effect may be due to the Li concentration gradient among LMO particles, resulting from the local underand overcharging occurring within the same electrode. As impurities are mainly found at the edges of the agglomerates, where particles have good electrical contact with conductive carbon, it is suggested that for higher C-rates, the crystal reconstruction may be caused by a locally higher potential (local overcharge) appearing on the particles located on the outer side of the agglomerate, inducing these structural reconstructions (Figure S7). 4. DISCUSSION Raman microscopy serves as an advantageous instrument for the examination of materials in Li-ion batteries. Its notable sensitivity facilitates in-depth insights into structural differences within the crystals under investigation. Spinel-based electrodes, monitored throughout the charging process, offer a robust model for such investigations. Raman mapping allows for the exploration of phase concentration changes during cycling, aiding in the assessment of potential factors influencing the electrochemical performance of the material. This technique presents a distinctive opportunity to gain a more comprehensive understanding of the underlying mechanisms governing reactions and capacity degradation and provides answers that are challenging to ascertain through electrochemical measurements alone or structural methods that use bulk-type measurements. When compared to other experimental techniques, Raman microscopy is widely employed in laboratories for several compelling reasons. It is cost-effective, highly sensitive, and capable of detecting even subtle structural changes and offers precise and statistically significant results as the Raman mapping area of the investigated electrode expands. For instance, the conventional X-ray diffraction (XRD) method is insufficiently sensitive for gathering such data, mainly due to bulk-type measurements and the need for extended data acquisition times, resulting in a lack of specificity when compared to Raman spectra. This limitation arises from the small quantity of active material applied in a laboratory-scale electrode, typically in the range of 2−5 mg for 10 mm electrodes, rendering standard characterization methods inadequate, as they fall far below the required detection limits. Raman mapping offers the unique ability to analyze electrode surfaces for their crystallographic homogeneity including the possible detection of impurity phases resulting from synthesis or electrochemical side reactions. It can be complemented by an in-depth investigation of the powder material to prevent the unwarranted overinterpretation of electrochemical data. In the context of the LMO material, Raman results unmistakably reveal an electrochemical phase transition reaction. The progression of peaks and their subtle shifts serve as precise indicators of the battery’s state-of-charge (SoC) and the coexistence of both Li-rich and Li-poor regions within partially charged electrodes. This underscores that Raman mapping stands as a more precise method for investigating the electrode’s SoC when compared to conventional single-spot Raman measurements. The lithiation gradient across the extensive surface area provides a distinctive perspective of the electrochemical reaction mechanism. The Li concentration gradient within the crystal was also observed during in situ TEM studies of the LMO nanowire. 41 During rapid charge and discharge, the nanowire exhibited distinct Li-rich and Li-poor phases separated by a transition region. This transition region reversibly moved along the nanowire to facilitate the transport of lithium ions. Our detected spectral gradient in the Raman maps of the partially lithiated LMO material additionally confirms the above-mentioned mechanism. Following this finding, we employed Raman mapping to investigate electrodes charged at different C-rates (1, 2, and 5 C) over 50 cycles, revealing distinct distributions of delithiated structures. Notably, the Raman maps varied significantly depending on the C-rate, with the 1 C-rate showing the most uniform distribution of the fully delithiated phase. At higher C-rates (2 and 5 C), the maps showed increased areas of the slightly lithiated F43mstructure, indicating limitations in Li+diffusion within the spinel structure and increased interfacial resistance. Furthermore, Raman mapping can be used as a very sensitive tool for the detection of impurity formation detection. We were able to observe additional phases including MnO2 (todorokite) and Mn3O4structures within cycled electrodes, with their concentrations increasing with higher C-rates. This suggests potential crystal reconstruction due to a locally higher potential (local overcharging) and underscores the importance of optimizing electrical contact within LMO agglomerates for enhanced performance in high-rate applications. 5. CONCLUSIONS Examining the electrode surface by Raman mapping provides the opportunity to monitor the lithiation state of particles at specific potentials vs Li. The evidence of the phase transition between LiMn2O4into λ-MnO2with a significant contribution of the Li0.5Mn2O4intermediate state is observed. Furthermore, mechanisms complementing electrochemical processes and comprehensive assignments of each detected phase are performed. In this study, we monitor changes in the F43m phase for all stages for lithium concentration higher than 0 and lower than 1. Raman mapping reveals a gradient distribution of Figure 6. Raman maps of impurities detected at 1, 2, and 5 C. Colored dots represent mixtures of MnO2(todorokite) and Mn3O4. ACS Omega http://pubs.acs.org/journal/acsodf Article https://doi.org/10.1021/acsomega.4c01480 ACS Omega 2024, 9, 30381−30391 30388 Li concentration within the particles at particular potentials, as indicated by variations in the Raman response within the same electrode. In addition, samples operated at high C-rates for 50 cycles were evaluated for their rapid delithiation capabilities and to track impurity formation over harsh cycling conditions. Remarkably, even small traces of impurity phases, typically below the detection limit of X-ray diffraction, can be detected. Analysis of differences in Raman peak intensities between particles can provide an analytical method to track structural changes and phase gradients within the electrode. This method underscores the notable influence of particle/crystal size on irregularities in the lithiation state and possible local overcharging causing material degradation. Due to its high sensitivity, spatial resolution, simplicity in sample preparation, and relatively short detection times, Raman spectroscopy proves to be an effective method for the exploration of Li-ion battery electrodes. Its utility extends beyond as-synthesized materials, thus enabling insights into structural changes, phase distribution, and impurities such as byproducts resulting from side reactions within the electrode layer. Ex situ Raman mapping of electrodes is a valuable tool for compositional analysis. It allows for the evaluation of the phase heterogeneity during cycling and provides an approximation of impurities and inactive component phase concentrations. This approach can readily find applications in industrial settings, such as investigating the reasons for capacity fading and determining electrode compositions. Among ex situ experimental techniques employed in lithium-ion battery research, Raman mapping is a suitable tool for comprehensive system analysis thanks to good resolution capabilities and the ability to provide statistically significant data. In this research, we highlight the wide-ranging potential of Raman mapping as a valuable analytical tool for structural and compositional determination, both in research and industrial applications. ■ASSOCIATED CONTENT * sı Supporting Information The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.4c01480. Detailed information about literature data on LiMn2O4 using ex situ and in situ Raman studies; Raman positions and band widths for selected spectra across LixMn2O4 cycling; electrochemical results of (A) specific capacity and (B) relative capacity over cycling of LiMn2O4; Raman spectra of the fully discharged sample showing the shift in A1g line from 627 to 635 cm−1; spectral map showing changes within delithiation of the LixMn2O4 material including profiles within Xand Y-axis; peak height ratio between the A1g Raman mode of the lithiated Fd3m structure and sum of T2g(3) of the same structure, A1mode at 590 cm−1from the F43mphase, and A1g phonon mode of the highly delithiated λ-MnO2 structure; Raman spectra from colored map areas (Figure 5) representing the position of the A1g and A1 modes of the fully delithiated λ-MnO2structure and the slightly lithiated F43mstructure, respectively; examples of the Raman spectra showing MnO2(todorokite) and Mn3O4compounds detected in the electrodes cycled at 1, 2, and 5 C; and crystal structure comparison between MnO2(todorokite), Mn3O4, and λ-MnO2(PDF) ■AUTHOR INFORMATION Corresponding Author Dominika A. Buchberger −Faculty of Chemistry, University of Warsaw, 02093 Warsaw, Poland; orcid.org/00000001-7617-6991; Email: [email protected] Authors Bartosz Hamankiewicz −Faculty of Chemistry, University of Warsaw, 02093 Warsaw, Poland Monika Michalska −Faculty of Materials Science and Technology, VSB-Technical University of Ostrava, 708 00 Ostrava-Poruba, Czech Republic; orcid.org/0000-00026120-4950 Alicja Głaszczka −Faculty of Chemistry, University of Warsaw, 02093 Warsaw, Poland Andrzej Czerwinski −Faculty of Chemistry, University of Warsaw, 02093 Warsaw, Poland Complete contact information is available at: https://pubs.acs.org/10.1021/acsomega.4c01480 Author Contributions D.A.B. conceived the original idea, designed the study, and gained funding. D.A.B. made important contributions to interpreting the results and conception. M.M. performed the synthesis of the LMO material and arranged SEM and XRD experiments. D.A.B. conducted a full structural and morphological analysis. B.H. and A.G. performed and analyzed electrochemical data. B.H. and A.G. prepared samples for Raman experiments. D.A.B. conducted Raman experiments, developed a fitting model of Raman spectra, and performed a full analysis of Raman mapping. D.A.B. performed crystallographic models of structures. D.A.B. and B.H. wrote the manuscript. D.A.B. and B.H. prepared visual graphs and tables. D.A.B. and A.C. supervised the study. All the authors revised and commented on the paper. All authors have given approval to the final version of the manuscript. Funding Homing program of the Foundation for Polish Science (POIR.04.04.00-00-5EC3/18-00). LIDER X program of The National Centre for Research and Development (grant no. LIDER/28/0148/L-10/18/NCBR/2019). Ministry of Education, Youth and Sports, Czech Republic (contract no. 8F21007). European Union under the REFRESH ((project no. CZ.10.03.01/00/22_003/0000048). Notes The authors declare no competing financial interest. ■ACKNOWLEDGMENTS D.A.B. thanks the support through the Homing program of the Foundation for Polish Science (POIR.04.04.00-00-5EC3/1800) cofinanced by the European Union under the European Regional Development Fund. B.H. thanks support through the LIDER X program of The National Centre for Research and Development (grant no. LIDER/28/0148/L-10/18/NCBR/ 2019). M.M. would like to acknowledge for the financing support through the Ministry of Education, Youth and Sports, Czech Republic (contract no. 8F21007) under the Visegrad Group-Japan 2021 Joint Call on “Advanced Materials”” in cooperation with the International Visegrad Fund, and by the European Union under the REFRESH - Research Excellence For REgion Sustainability and High-tech Industries (project no. CZ.10.03.01/00/22_003/0000048) via the Operational ACS Omega http://pubs.acs.org/journal/acsodf Article https://doi.org/10.1021/acsomega.4c01480 ACS Omega 2024, 9, 30381−30391 30389