Adverse to beneficial: upcycling residual lithium compounds on LiNi_0.8Mn_0.1Co_0.1O₂ into a stabilizing Li_1+xMn_2−xO₄ interface

Abstract

Ni-rich layered oxide (NRLO) cathodes, though promising for next-generation high-energy Li-ion batteries (LIBs), suffer from both bulk and surface structural instability. The chemical reactivity of NRLO surfaces to moisture (H₂O and CO₂) is an industrial concern, as it leads to the formation of residual lithium compounds (RLCs) such as Li₂CO₃ and LiOH. The alkaline RLCs undergo parasitic reactions with electrolytes, forming a resistive layer on NRLO cathode surfaces and limiting electrochemical performance. This work presents an “adverse to beneficial” approach, converting surface RLCs on LiNi_0.8Mn_0.1Co_0.1O₂ (NMC811) into a high-voltage stable Li_1+xMn_2−xO₄-based interface. The chemically inert protective interface, formed by a simple wet-coating method, reduces surface side reactions with electrolytes, enhancing NMC811's cycle life by retaining 75% capacity after 300 cycles at a voltage range of (3.0–4.3) V vs. Li⁺/Li. The protective interface stabilizes the cathode surface, lowering the Li⁺ intercalation barrier and reducing the overpotential for the H2 → H3 phase transition. It also mitigates microcrack development and delays structural collapse. This surface modification enhances NMC811's stability at high voltages (4.5 V and 4.7 V vs. Li⁺/Li), improving its chemical stability and overall electrochemical performance.

---

1. Introduction

State-of-the-art Li-ion battery (LIB) technology is expected to dominate the portable electronics and electric vehicle markets for at least the next few decades.^1–3 Ni-rich layered oxide (NRLO) will be one of the most promising cathode materials for LIBs due to its high theoretical capacity (274 mAh g⁻¹) and operating voltage (4.3 V vs. Li⁺/Li).^4–7 Currently, LiNi_0.8Mn_0.1Co_0.1O₂ (NMC811) is projected as a prospective cathode material to achieve the 300 Wh kg⁻¹ energy target at the pack level of LIBs.^8,9 Unfortunately, the commercialization of NMC811 is obstructed by its high structural, chemical, and thermal instabilities.^10–12 The structural instabilities originate in both the bulk and the surface. The similar ionic radius of Li⁺ (76 nm) and Ni²⁺ (69 nm) is the source for the exchange of Li⁺ and Ni²⁺ positions between 3a and 3b sites. This is known as Li⁺/Ni²⁺ cation mixing (antisite disorder) and is a main culprit for bulk structural degradation.^13,14 It also influences the surface by undergoing an irreversible phase transition from the layered to spinel to rock-salt (NiO) phases at the near surface.

Chemical instability of NRLO surfaces stems from their sensitivity to moisture (H₂O and CO₂), leading to the formation of residual lithium compounds (RLCs) such as Li₂CO₃ and LiOH, which act as Li-containing surface impurities.^15,16 It forms during the synthesis and storage period, creating a huge industrial concern in dealing with NRLO cathodes.^17–20 Alkaline RLCs can depolymerize the polyvinylidene fluoride (PVDF) binder and cause slurry gelation. It reacts vigorously with carbonate-based electrolytes and undergoes gas evolution (O₂, CO₂, CO, etc.).^15,21 Besides, it significantly lowers the cycle life of the material and increases the safety risk of the battery. Hence, it is very important to remove RLCs from the NRLO surface in a large-scale preparation. Recent studies have focused on the removal of RLCs from NRLO surfaces, but this approach remains in its infancy in academic research. Among the most reported strategies, washing (water or acid), secondary annealing, surface coating, etc., has its own limitations.^16,22,23 Washing NRLO surfaces with water or acids increases their vulnerability, heightening the risk of electrolyte decomposition and subsequent gas evolution during cycling. Secondary sintering effectively removes RLCs but compromises structural integrity due to high-temperature (∼700 °C) annealing, making surface coating a more widely adopted strategy. However, the buried RLC layer in conventional surface coatings adversely affects the de/lithiation process, eventually deteriorating NRLO's electrochemical performance.

The optimal approach to address RLCs is their chemical conversion into a beneficial interface, a strategy we previously termed the ‘adverse to beneficial’ approach.²⁰ Park et al. converted the RLCs layer of NMC811 to a LiF layer by using NH₄F salt and reported an improved performance of 50% capacity retention after 500 cycles at 4.6 V vs. Li⁺/Li.²⁴ In our previous studies, we reported the chemical conversion of RLCs into hybrid inorganic layers containing Li₃PO₄/LiPO₃ and Li₃PO₄/Li_xBO_yF_z, improving the cycle life of pristine NMC811 from 39% to 70–85% (for coated one) after 300 cycles at 0.5C current rate and voltage window of (3.0–4.3) V vs. Li⁺/Li.^20,25 The Li_xPO_y (Li₃PO₄/LiPO₃-based composite) interface also assisted in fast cycling by retaining 60% capacity after 500 cycles at a 5C current rate.²⁰

Further, deep deintercalation is a must to practically realize the high theoretical capacity (274 mAh g⁻¹) of NMC811. It stretches the upper cut-off voltage beyond 4.3 V vs. Li⁺/Li. Interestingly, most of the studies have cut-off voltage limited to 4.3 V. The instability of the interface and the inferior stability of conventional carbonate-based electrolytes are the major bottlenecks for the limited cut-off voltage. Beyond 4.3 V, electrolytes decompose, and the interface undergoes severe unwanted side reactions with the electrolytes.^26,27 The presence of RLCs on the surface even worsens the interface at high voltage. Hence, this study aims to form a 4.7 V stable interface by chemically converting RLCs into a spinel-oxide-based interface. Spinel LiMn₂O₄, LiNi_0.5Mn_1.5O₄, etc., are known for their high voltage (>4.8 V) operation. Unlike NRLOs, these spinel oxides are chemically stable and do not undergo any moisture-sensitive reactions. In 2016, the Manthiram group reported Mn-rich spinel LiMn_1.9Al_0.1O₄ coating on LiNi_0.7Co_0.15Mn_0.15O₂ to provide chemical stability to the surface by preventing moisture attack.²⁸ In our group's previous report, we observed that a composite (Li₃PO₄, LiMn₂O₄, and LiMnPO₄) coating involving the spinel-oxide phase enables high-voltage (4.7 V) stability.²⁹ Herein, in this work, we report the chemical conversion of the RLCs present on the surface of NMC811 into an Mn-based spinel oxide interface by following a single-step wet coating method. The optimized coating precursor (1.5 wt% Mn(CH₃COOH)₂) converts the RLCs into a near-uniform ultrathin (∼10 nm) Li_1+xMn_2−xO₄-based interface. The pristine and Li_1+xMn_2−xO₄-based interface-modified NMCs are denoted in the manuscript as PNMC and LMONMC, respectively. The formed interface significantly lowers the polarization for high-voltage H2 → H3 (by 60 mV) phase transition, reduces the surface reactivity, improves cycle life, enhances Li⁺ diffusion by reducing Li⁺/Ni²⁺ cationic disorder, and assists in high voltage (4.5 V and 4.7 V vs. Li⁺/Li) stability of NMC811.

2. Results and discussions

2.1. Physical characterizations

2.1.1. Conversion of RLCs layer into Li_1+xMn_2−xO₄ based interface without altering the bulk structure. The crystal structures of PNMC and LMONMC are indexed to hexagonal α-NaFeO₂-type layered structure with R3̄m space group, as shown in the PXRD profile (Fig. 1a). The X-ray diffraction pattern of LMONMC shows no additional peaks or impurity phases, indicating that the coating process does not alter the bulk structure. The Rietveld refinement analysis of the PXRD patterns of PNMC and LMONMC is shown in Fig. 1b and c, respectively. The details of lattice parameters obtained from Rietveld refinement (Table 1) show a slight increase in the a and c lattice parameters. This is attributed to the post-calcination process.²⁵ An increased axial ratio c/a suggests improved Li⁺ mobility. The absence of other changes in the PXRD pattern confirms that the bulk structure remains unaltered after surface modification.

Fig. 1 Physical properties of the coating layer: (a) PXRD of PNMC and LMONMC; Rietveld refinement of PXRD patterns (b) PNMC and (c) LMONMC; XPS of LMONMC (d) Li 1s, (e) Mn 2p, (f) O 1s; (g) ATR-IR of PNMC and LMONMC; (h) EPR of PNMC and LMONMC; (i) pH measurements of PNMC and LMONMC.

Table 1 Rietveld refinement parameters for PNMC and LMONMC

	PNMC	LMONMC
a (Å)	2.8626	2.8740
c (Å)	14.0633	14.1999
Unit volume (Å^3)	99.8037	101.58
c/a	4.9127	4.9408

---

The chemical composition of the surface layer of both PNMC and LMONMC is studied by using XPS. The Li 1s, C 1s, and O 1s XPS of PNMC in the ESI Fig. S1† shows the presence of RLCs (Li₂CO₃ and LiOH) on the surface. The XPS (Li 1s (Fig. 1d), Mn 3p (Fig. 1d), Mn 2p (Fig. 1e), and O 1s (Fig. 1f)) of LMONMC shows the conversion of RLC into a Li–Mn–O-based layer. The Li 1s XPS is associated with a highly intense doublet of Mn 3p at 49.8 eV and 51.6 eV, corresponding to the 3+ and 4+ oxidation states of Mn, referencing the presence of Li_1+xMn_2−xO₄.³⁰ The presence of Mn³⁺ and Mn⁴⁺ is also evident from the Mn 2p_1/2 peaks (654.2 eV, 655.5 eV) and Mn 2p_3/2 peaks (642.7 eV, 644.4 eV) in Mn 2p XPS. Further, the presence of Li_1+xMn_2−xO₄ is also supported by the O 1s peak at 530 eV and the Li 1s peak at 54.2 eV.^31,32 The RLC components are also observed in the ATR-IR spectrum of PNMC (Fig. 1g). The –CO₃²⁻ functionality of Li₂CO₃ in PNMC is detected from the out-of-plane bending (δ_O–C=O) at 860 cm⁻¹ and asymmetric stretching (ν_C=O) of the C=O group at 1367 and 1418 cm⁻¹ and asymmetric stretching (ν_C–O) of C–O group at 1221 cm⁻¹.^33,34 The presence of –OH functionality of LiOH in PNMC is detected from the fingerprints of O–H bending mode (δ_O–H) at around 1738 cm⁻¹ and 2366 cm⁻¹ in the ATR-IR spectrum.²⁰ The absence of all these peaks in the ATR-IR spectrum of LMONMC (Fig. 1g) clearly supports the removal of RLCs by the mentioned surface modification. Further, the strong peak at ∼570 cm⁻¹ in LMONMC is the asymmetric stretching (θ_O–Mn–O) of O–Mn–O linkage from MnO₆ octahedra in Li_1+xMn_2−xO₄ and is absent in PNMC.³⁵ Basically, from the XPS and ATR-IR spectrum, it can be concluded that the –CO₃²⁻ and –OH functionalities of PNMC are substituted by O–Mn–O linkage in LMONMC.

EPR can be useful in detecting surface changes because of the introduction of paramagnetic Mn centers in the form of Li_1+xMn_2−xO₄, replacing diamagnetic LiOH and Li₂CO₃ residue. The formal oxidation states in PNMC can be written as following LiNi_0.1²⁺Ni_0.7³⁺Mn_0.1⁴⁺Co_0.1³⁺O₂. Here, Ni²⁺ (S = 1), Ni³⁺ (S = 1/2), and Mn⁴⁺ (S = 3/2) are paramagnetic, while Co³⁺ (S = 0) is diamagnetic. Hence, the observed EPR signal of PNMC in Fig. 1h is ascribed to Mn⁴⁺–Ni²⁺ exchange coupling interactions.³⁶ However, the signal of LMONMC is showing stark differences with respect to PNMC signal, which is utterly useful in assigning the phases of Li_1+xMn_2−xO₄ (0 ≤ x < 1/3) spinel that could not be detected by any other physical characterization techniques used in this work. The center of the spectrum demonstrating the Lorentzian lineshape is a typical signature of Li-rich monoclinic Li₂MnO₃ phase, which is reported to exist as <1 wt% impurity during the synthesis of LiMn₂O₄ spinel below 800 °C.³⁷ The presence of LiOH/Li₂CO₃ precursor in the RLC layer in slight excess than utilized Mn-acetate precursor explains the formation Li-rich LiMn₂O₃ phase along with pure phase LiMn₂O₄. We believe that the content of Li₂MnO₃ phase is restricted within ∼1 wt% of Li_1+xMn_2−xO₄ (∼99 wt% LiMn₂O₄), although it is very difficult to quantify the same because the overall content of transformed Li_1+xMn_2−xO₄ lies below 2 wt% compared to pristine NMC phase. Further, the broad wings in the LMONMC signal belong to the combination of spinel LiMn₂O₄ phase and pristine NMC phase. The lineshape becomes complex in LMONMC due to the exchange interactions among incorporated Mn sites, i.e., Mn³⁺ (S = 2) and Mn⁴⁺ (S = 3/2) in LiMn₁³⁺Mn₁⁴⁺O₄, and Mn⁴⁺ (S = 3/2) in Li₂Mn₁⁴⁺O₃. It is important to mention here that the hopping electron conduction effect of spinel LiMn₂O₄ phase does not disrupt much by the presence of significantly low content of insulating Li₂MnO₃ phase. Hence, quantifying the value of ‘x’ is of little importance here. EPR data confirms the successful conversion of the RLC layer into the Li_1+xMn_2−xO₄ phase and indicates the minor formation of the Li₂MnO₃ phase.

RLCs are alkaline in nature, allowing their qualitative presence to be directly detected through pH measurements. The pH value for PNMC is 11.2, which decreases to 9.4 for LMONMC (Fig. 1i). The lower pH value indicates the successful removal of the alkaline RLC layer by the Li–Mn–O-based surface layer.

2.1.2. Surface modification converts amorphous RLCs layer into a near-uniform crystalline Li–Mn–O-based interface. The secondary particles of both PNMC and LMONMC have spherical morphology with an average diameter of 10–12 μm, as shown in Fig. 2a and b. The magnified image (Fig. 2c and d) of the corresponding SEM images shows the primary particles with irregular morphologies. The edges of the primary particles in PNMC are not clear, and some residues are present on the surface. In contrast, the sharp edges and clean surfaces in LMONMC show the removal of the impurity residues. The nature (e.g., thickness, uniformity, and crystallinity) of the modified surface layer can be observed from the HRTEM image. The poorly crystalline non-uniform RLC layer is visible in the PNMC, as shown in Fig. 2e. The HRTEM image of PNMC (Fig. 2f) shows the crystalline bulk with an interplanar spacing of 0.475 nm belonging to the (003) crystal plane of NMC811 and an amorphous surface layer, signifying the RLCs. As the reaction between the bare NMC surface and moisture is uncontrollable, the thickness of the RLC layer varies depending on the exposure time of the material. However, it is clear from the HRTEM image that the RLC layer is poorly crystalline or close to an amorphous nature. The TEM image of LMONMC shows a uniform surface layer (Fig. 2g) with a thickness of ∼10 nm. The HRTEM images of the bulk ([1] marked in Fig. 2g) region are shown in Fig. 2h and i for better visibility. The FFT pattern (Fig. 2j) and corresponding profile plot (Fig. 2k) represent the interplanar spacing of 0.475 nm, indicating the core NMC811. Meanwhile, the HRTEM image of the surface ([2] marked region in Fig. 2g) is shown in Fig. 2l. The corresponding FFT pattern (Fig. 2m), inverse FFT (Fig. 2n), and profile plot (Fig. 2o) show an interplanar spacing of 0.478 nm belonging to the 111-crystal reflection of Li_1+xMn_2−xO₄. Although the targeted primary particle here shows a uniform surface layer, it is important to mention here that in some regions, variations of thickness and non-uniformity are observed. The more likely explanation for this observation is that the surface Li_1+xMn_2−xO₄ phase is formed by the chemical reaction between the non-uniform RLC layer and externally added Mn(CH₃COOH)₂ reactant. Hence, the deposited interface in some regions is bound to be non-uniform. The HAADF STEM-EDX mapping of all the elements in LMONMC shows the presence of all elements with an even distribution (Fig. 2p) all over the particles. The elemental distribution of PNMC (ESI Fig. S2†) shows an even distribution of all the elements.

Fig. 2 Morphology study: (a) SEM image of PNMC, (b) SEM image of LMONMC, (c) magnified view of SEM image of PNMC, (d) magnified view of SEM image of LMONMC; (e) TEM image of PNMC, (f) HRTEM image of PNMC; (g) TEM image of LMONMC, (h) HRTEM image of LMONMC, (i) HRTEM image of bulk LMONMC, (j) FFT pattern of image (i); (k) profile plot corresponding to bulk LMONMC; (l) HRTEM image of LMONMC focusing on surface region; (m) FFT pattern of image (l); (n) inverse FFT pattern of image (m); (o) profile plot corresponding to surface LMONMC; (p) STEM-EDX mapping of all the elements in LMONMC.

Microstructural observations confirm the conversion of the poorly crystalline RLC layer into a near-uniform, crystalline Li_1+xMn_2−xO₄-based surface layer.

2.2. Electrochemical performances

2.2.1. Li_1+xMn_2−xO₄-based interface influences cycle life and lowers the voltage decay. To find out the optimum coating amount electrochemically, we carried out the experiment with three different sets by varying the coating amount to 1 wt%, 1.5 wt%, and 2 wt% of manganese(II) acetate tetrahydrate precursor. The capacities with different cycle numbers of the three different sets are shown in the ESI Fig. S3.† The capacity retention values after 100 cycles for 1, 1.5, and 2 wt% are 75%, 94%, and 85%, respectively. Since 1.5 wt% is giving better performance, it is considered an optimum amount to convert all the RLCs into a Li_1+xMn_2−xO₄-based interface. Lower than 1 wt% precursor may be resulting into an incomplete conversion of RLCs, while the use of ≥2 wt% maybe leading to a precursor residue lying unreacted on the surface. Hence, the precise point of optimization lies between 1–2 wt% of the precursor, which can be determined by measuring the amount of the RLCs. Meanwhile, the accurate quantification of RLCs is a tedious task as the amount of RLCs varies vastly from sample to sample depending upon synthetic history, storage conditions, and mishandling. Therefore, the optimized wt% of the precursor is chosen based on the electrochemical performance. All the further studies are carried out with 1.5 wt% and denoted as LMONMC throughout the manuscript.

The comparison of cycle life performances between PNMC and LMONMC is shown in Fig. 3a. PNMC retains only 31% capacity after 300 cycles at 0.5C current rate and voltage window of (3.0–4.3) V, whereas LMONMC retains 75% capacity under similar conditions. The significantly poor performance in PNMC is due to severe parasitic reactions caused by the RLCs with electrolytes. The continuous vigorous reactions of RLCs with carbonate-based electrolytes result in electrolyte decompositions and unwanted growth of the cathode–electrolyte interface (CEI). The 1st cycle discharge capacities for PNMC and LMONMC are 174 and 176 mAh g⁻¹, respectively (Fig. 3b). Although the initial discharge capacities are almost similar, PNMC shows a peculiar behavior during initial charging. As shown in Fig. 3b, there is a sudden voltage shoot-up at around (3.9–4.0) V vs. Li⁺/Li for PNMC. The reason behind such behavior is the kinetically sluggish Li⁺ diffusion caused by the initial interactions of RLCs with electrolytes.^28,38 PNMC shows a drastic voltage decay starting from the 100th cycle and becomes severe with the progress of cycling (Fig. 3c). However, the Li_1+xMn_2−xO₄ interface lowers the surface side reactions with electrolytes and alleviates the voltage decay in LMONMC (Fig. 3d).

Fig. 3 Comparison of electrochemical performances between PNMC and LMONMC (a) cycle life (0.5C current rate, voltage window of (3.0–4.3) V vs. Li⁺/Li), (b) voltage profile; discharge capacity at different (1st, 50th, 100th, 150th, 200th, and 300th) cycles (c) PNMC and (d) LMONMC; (e) chronocoulometry measurements of PNMC vs. LMONMC; (f) coulombic efficiency vs. cycle number of PNMC and LMONMC; (g) post-cycling (after 300th cycles at current rate of 0.5C, (3.0–4.3) V vs. Li⁺/Li) XPS of LMONMC (Li 1s, O 1s, F 1s, P 2p); electrochemical impedance analysis between PNMC and LMONMC (h) at 1st cycle and (i) at 300th cycle; (j) post-cycle XRD after 300th cycle under voltage window of (3.0–4.3) V vs. Li⁺/Li.

2.2.2. Surface modification stabilizes the interface, facilitates charge transfer kinetics, and reduces the Li⁺/Ni²⁺ cation mixing. As the bare NMC surface undergoes severe side reactions with electrolytes, a chronocoulometry measurement is carried out to calculate the approximate accumulated charge on the interface. The measurement is done by holding the working electrodes at 4.7 V for 1.5 hours (Fig. 3e). The accumulated charge for PNMC (4.1C) is 1.6C higher than that of LMONMC (2.5C). The profiles for both PNMC and LMONMC look similar. There is a sudden exponential increase in the initial 15–20 minutes, followed by attaining the equilibrium. PNMC surpasses LMONMC in the initial charge accumulation since the vigorous side reactions of RLCs with electrolytes become more severe at the initial moments. In contrast, the protective Li_1+xMn_2−xO₄ layer on the surface of LMONMC quickly forms a stable interface and restricts the propagation of side reactions. This is also reflected in the improved coulombic efficiency of LMONMC (99.6%) over PNMC (96.5%), as shown in Fig. 3f. Further, the chemical compositions of the cathode electrolyte interface (CEI) also support the same. LiF and Li_xPO_yF_z are the usual CEI components in any LiPF₆-electrolyte system.³⁹ Besides LiF and Li_xPO_yF_z, the CEI of PNMC (ESI Fig. S4†) consists of Li₂CO₃ and LiOH. This might be the reason for the unstable interface and the huge charge accumulation at the electrode–electrolyte interface. On the contrary, the CEI of LMONMC (Fig. 3g) contains LiF, Li_xPO_yF_z (/Li₃PO₄), and Li_1+xMn_2−xO₄. This is the main reason for suppressing side reactions followed by a stable interface formation and, hence, less charge accumulation at the EEI.

The resistive film formed by the insulating RLCs layer hinders the Li⁺ migrations through the interface. The high charge-transfer resistance of PNMC compared to LMONMC in both the initial (Fig. 3h) and at the end of the 300th cycle (Fig. 3i) represents the same. The corresponding resistance values (R₁ = solution resistance, and R₂ = combination of both charge transfer and interface crossing resistance) and equivalent circuit are shown in ESI Table S1 and Fig. S5.† The Li⁺-ion conducting Li_1+xMn_2−xO₄ layer facilitates Li⁺ migrations, and it is also reflected in almost two times lower charge transfer resistance for LMONMC (∼48 Ω) compared to PNMC (∼100 Ω) at the 300th cycle.

The surface reactions also gradually affect the bulk structure of the material. As mentioned, the similar ionic radius of Li⁺ and Ni²⁺ induces Li⁺/Ni²⁺ mixing and is one of the major culprits for the degradation of NRLO cathodes. The major driving force for the migration of Ni²⁺ to the Li⁺ layer is the instability of Ni³⁺, followed by its reduction to Ni²⁺ by oxidizing lattice oxygen. It forms a rock-salt NiO (Fm3̄m) phase on the surface. The formation of this impurity phase starts at the surface, propagates towards the bulk with the progression of cycling, and simultaneously increases the Li⁺/Ni²⁺ cation mixing (antisite disorder). The exchange of Li⁺ and Ni²⁺ position influences the 003 and 104 crystal reflections and directly reflects on the intensity of (003) and (104) crystal planes. The I₍₀₀₃₎/I₍₁₀₄₎ is a direct indication of Li⁺/Ni²⁺ mixing. Interestingly, the Li⁺/Ni²⁺ cation mixing calculated from the post-cycling (after 300 cycles) XRD (Fig. 3j) shows significant differences for PNMC and LMONMC. The I₍₀₀₃₎/I₍₁₀₄₎ for PNMC is 1.0 (<1.2), signifying severe antisite disorder. In contrast, the I₍₀₀₃₎/I₍₁₀₄₎ for LMONMC is 1.37 (>1.2), indicating significantly less cation mixing. However, a more in-depth study is required to conclude the reason behind such improvement, which is not currently the major focus of this work. Nevertheless, the surface Li_1+xMn_2−xO₄ layer in LMONMC is probably restricting the propagation of the surface impurity phase, and that is directly reflected in low Li⁺/Ni²⁺ cation mixing. The migration of Ni²⁺ to the Li⁺ layer also restricts the Li⁺ diffusion through the bulk. To check the effect of surface modification on the Li⁺ diffusion, the diffusion coefficient (D_Li⁺) values are calculated from the Warburg region of the EIS data. The mathematical expression for the calculation of D_Li⁺ is shown in ESI,† and the values are shown in ESI Table S2.† The corresponding linear fit of Z′ (real part) versus ω^−1/2 for PNMC and LMONMC at the 1st cycle and 300th cycle are shown in Fig. 4a and b. The D_Li⁺ values demonstrate a one-order-of-magnitude improvement for LMONMC compared to PNMC, directly supporting reduced Li⁺/Ni²⁺ mixing due to the conductive Li_1+xMn_2−xO₄ layer.

Fig. 4 Linear fitting of Z′ versus ω^−1/2 for both PNMC and LMONMC during the (a) 1st cycle, (b) 300th cycle; (c) differential capacity vs. voltage profile for PNMC and LMONMC; post-cycling (after 300th cycles at the current rate of 0.5C, (3.0–4.3) V vs. Li⁺/Li) SEM image of (d) PNMC and (e) LMONMC; (f) voltage profile of PNMC and LMONMC at 4.5 V vs. Li⁺/Li; (g) comparison of cycle life of PNMC and LMONMC at a voltage window of (3.0–4.5) V vs. Li⁺/Li (0.5C current rate); (h) voltage profile of PNMC and LMONMC at 4.7 V vs. Li⁺/Li; (i) comparison of cycle life of PNMC and LMONMC at a voltage window of (3.0–4.7) V vs. Li⁺/Li.

Briefly, the RLCs in PNMC undergo severe side reactions with electrolytes and form an unstable-insulating interface with a high activation barrier for Li⁺ migrations. It also influences the high degree of antisite disorder and restricts Li⁺ diffusion through the bulk. In contrast, the chemically stable Li_1+xMn_2−xO₄ layer lowers the surface-side reactions, forms a stable interface, improves Li⁺ migrations through the interface, and lowers the antisite disorder, hence, comparatively improves Li⁺ diffusion.

2.2.3. Li_1+xMn_2−xO₄-based interface withstands micro-cracks formation and improves high-voltage stability. During cycling, NRLO cathodes undergo three different phase transitions – starting from H1 → M (3.76 V vs. Li⁺/Li), M → H2 (4.0 V vs. Li⁺/Li), and H2 → H3 (4.2 V vs. Li⁺/Li). The differential capacity (dQ/dV) profile in Fig. 4c clearly shows three different phase changes for PNMC and LMONMC. The Li_1+xMn_2−xO₄ interface in LMONMC lowers the polarization of phase transition for H1 → M by 37 mV, M → H2 by 65 mV, and H2 → H3 by 60 mV than the PNMC. The main reason for intergranular microcracks formation in NRLO cathodes is the H2 → H3 phase transition. The 4.2 V H2 → H3 phase transition is associated with the abrupt and anisotropic change in the unit cell. Since the primary particles of NMC811 are irregular polyhedrons, the anisotropic deformation is random. Again, the irregular arrangement of the primary particles generates a force by the expansion and contraction of two adjacent primary particles at the grain boundaries. This force at the grain boundaries causes intergranular cracks. Electrolytes can infiltrate through the microcracks, undergo severe reactions, and finally degrade the performance of the material. The post-cycling SEM image of PNMC (Fig. 4d) shows a huge crack formation from the center of the particles to the edges and its pulverization into multiple pieces. In contrast, LMONMC lowers the tendency for crack formation (Fig. 4e). Although some ultrafine cracks are there, the coating layer hinders the radial propagation of cracks and restrains the structure from collapsing. This can be attributed to the 60 mV reduction of the polarization for the H2 → H3 phase transition in LMONMC. It is worth noting that surface modification cannot solve microcrack formation solely because it starts at the core. However, it can reduce the tendency of microcrack formation, prevent further development of cracks, and delay structural collapse.

As mentioned, stretching the upper cut-off voltage beyond 4.3 V (vs. Li⁺/Li) increases the achievable capacity of NRLO cathodes. However, the chemical instability of Ni⁴⁺ at a highly delithiated state and its subsequent conversion to the NiO layer, severe microcracks formation, decomposition of electrolytes, etc., are some major issues associated with the high voltage operation. Since Li_1+xMn_2−xO₄ is known for its high voltage and chemical stability. So, to check its effect as a surface layer, we have expanded the voltage window to 4.5 V vs. Li⁺/Li and 4.7 V vs. Li⁺/Li. Widening the voltage window to 4.5 V increases the initial discharge capacity by almost 35 (±5) mAh g⁻¹ for both PNMC and LMONMC. The 4.5 V voltage profiles for both PNMC (215 mAh g⁻¹) and LMONMC (212 mAh g⁻¹) are shown in Fig. 4f. However, the long-term performance of both materials is very different. Fig. 4g shows the comparison of cycling performance for PNMC and LMONMC. After 130 cycles, PNMC and LMONMC retain 58% and 84% of initial capacity, respectively. A similar behavior is observed when the voltage window expands to 4.7 V. Expansion of the voltage window to 4.7 V increases the initial capacities by 50 (±10) mAh g⁻¹ for both PNMC and LMONMC. The voltage profiles for PNMC and LMONMC at 4.7 V are shown in Fig. 4h. The voltage profiles are identical in all three voltages (4.3, 4.5, and 4.7 V vs. Li⁺/Li), where the PNMC shows a voltage shoot-up around 3.9–4.0 V due to the resistive film formed by the RLCs layer. The performance of PNMC is worst at 4.7 V, where it loses almost 50% capacity after 50 cycles (Fig. 4i). In contrast, LMONMC still performs well at 4.7 V by retaining 77% capacity after 100 cycles (Fig. 4i). This can be attributed to the fact that the RLCs on the surface of PNMC undergo severe side reactions with electrolytes, which become more vigorous with increasing the upper-cut-off voltage. On the contrary, the chemically stable Li_1+xMn_2−xO₄ layer is still able to protect the interface from unwanted side reactions even at higher cut-off voltages.

It is worth noting that the instability of the system at high voltage (>4.3 V) arises from both the chemical instability of the Ni-rich cathode surface and the inferior stability of the carbonate-based electrolytes. At high voltage, the electrolytes decompose and deteriorate the performance of the system. Since the reactive RLCs are already present on the surface of PNMC, the degradation process is accelerated. However, the Li_1+xMn_2−xO₄ layer reduces the surface side reactions of LMONMC and results in better electrochemical performance even at higher voltage. In short, LMONMC significantly lowers the polarization for the H2 → H3 phase transition, suppresses the formation of microcracks, and enables high-voltage operation. A comparison of the electrochemical performances of our surface engineering approach with previously reported studies is provided in Table S3 of the ESI.†

3. Conclusions

NMC811 is projected to be the most promising cathode material for high-energy-density Li-ion batteries (LIBs). Surface chemical instability leading to RLC formation on NMC811 poses an industrial challenge. This study reports an ‘adverse to beneficial’ approach, eliminating RLCs by converting them into a high-voltage stable Li_1+xMn_2−xO₄ interface. The proposed method preserves the bulk structure without introducing impurity phases, as evidenced by XRD patterns. XPS, ATR-IR, and EPR analyses confirm the successful conversion of RLCs to the Li_1+xMn_2−xO₄ interface. Electron microscopy reveals the uniformity (∼10 nm thickness) and crystallinity of the interface. This approach enhances electrochemical performance, retaining 75% capacity after 300 cycles at (3.0–4.3) V vs. Li⁺/Li (0.5C rate). The improved performance is primarily attributed to reduced surface-side reactions with electrolytes, leading to stable interface formation at the cathode–electrolyte interface (CEI). This is evidenced by lower charge accumulation for LMONMC (2.5C) compared to PNMC (4.1C) at the electrode–electrolyte interface (EEI) and nearly halved charge transfer resistance for LMONMC after 300 cycles. LMONMC demonstrates reduced overpotentials for H1 → M, M → H2, and H2 → H3 phase transitions, with the 60 mV reduction in H2 → H3 transition potentially mitigating NMC particle pulverization. This modified surface layer enhances high-voltage protection, evidenced by LMONMC retaining 84% capacity after 130 cycles at 4.5 V and 77% after 100 cycles at 4.7 V, significantly outperforming PNMC. The reported method improves cycle life by reducing parasitic reactions, alleviating voltage decay, lowering Li⁺ intercalation barriers, reducing Li⁺/Ni²⁺ cation mixing, enhancing bulk Li⁺ diffusion, delaying microcrack formation, and improving interface stability at high voltages.

In short, a simple wet-coating method converts detrimental RLCs into a high-voltage, stable, protective Li–Mn–O interface, which chemically stabilizes the surface and improves the electrochemical performance of NMC811 electrodes.

4. Experimental section

The pristine NMC811 used here for surface modification was supplied by Tokyo Chemical Industry Co., Ltd (TCI, Japan, 99.9% purity). The availed material was directly used as an active material for reference without any further modification and is denoted here as PNMC. For surface modification, a different weight ratio (1 wt%, 1.5 wt%, and 2 wt% with respect to PNMC) of manganese(II) acetate tetrahydrate (Sigma-Aldrich) was dispersed in ethanol. Then, an appropriate amount of NMC811 powder (at least a minimum of 3 g) was added to the previous dispersion medium, followed by stirring for 24 hours. The ethanol was evaporated, and then the dried material was calcined at 600 °C for 5 hours under an air atmosphere. The calcined black mass was used as an active material to study electrochemical performance and for further characterizations. The optimized modified material is denoted here as LMONMC. Other experimental processes (electrode preparation, cell assembly, and physical characterizations) are presented in the ESI.†

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

Jyotirekha Dutta: conceptualization, methodology, investigation, performing experiments, data curation, formal analysis, visualization, figure formatting, writing – original draft. Shuvajit Ghosh: formal analysis, investigation, writing – review and editing. Vilas G. Pol: supervision, manuscript editing. Surendra K. Martha: supervision, funding acquisition, manuscript editing, resources, project administration, and communication.

Conflicts of interest

The authors declare that they have no known conflicts of interest.

Acknowledgements

Jyotirekha Dutta acknowledges DST-INSPIRE (code: IF200099) Govt of India for fellowship. Shuvajit Ghosh acknowledges CSIR, Govt, of India (file no. 09/1001(0067)/2019-EMR-I) for fellowship. Surendra K. Martha acknowledges SERB-IRHPA (file no. IPA/2021/000007), Govt of India, for financial support of the work and DST-SPARC, Govt of India for the collaborative work between IIT Hyderabad and Purdue University under the project: SPARC/2019-2020/P2846/SL.

References

1. J. B. Goodenough and K.-S. Park, J. Am. Chem. Soc., 2013, 135, 1167.
2. A. Manthiram, Nat. Commun., 2020, 11, 1550.
3. S. Ghosh, U. Bhattacharjee, S. Bhowmik and D. S. K. Martha, J. Nucl. Energy Sci. Power Gener. Technol., 2022, 4, 1.
4. J. U. Choi, N. Voronina, Y. Sun and S. Myung, Adv. Energy Mater., 2020, 10, 2002027.
5. J. Kim, H. Lee, H. Cha, M. Yoon, M. Park and J. Cho, Adv. Energy Mater., 2018, 8, 1702028.
6. A. Manthiram, J. C. Knight, S. Myung, S. Oh and Y. Sun, Adv. Energy Mater., 2016, 6, 1501010.
7. S. H. Jung, U. Kim, J. Kim, S. Jun, C. S. Yoon, Y. S. Jung and Y. Sun, Adv. Energy Mater., 2020, 10, 1903360.
8. S. Lyu, J. Yu, X. Guo, X. Bu, L. Sun, N. Li, H. Chen, W. Song, X. Yu and S. Jiao, Adv. Energy Mater., 2024, 2403270.
9. K. Märker, P. J. Reeves, C. Xu, K. J. Griffith and C. P. Grey, Chem. Mater., 2019, 31, 2545.
10. M. Jiang, D. L. Danilov, R.-A. Eichel and P. H. L. Notten, Adv. Energy Mater., 2021, 11, 2103005.
11. T. M. M. Heenan, A. Wade, C. Tan, J. E. Parker, D. Matras, A. S. Leach, J. B. Robinson, A. Llewellyn, A. Dimitrijevic, R. Jervis, P. D. Quinn, D. J. L. Brett and P. R. Shearing, Adv. Energy Mater., 2020, 10, 2002655.
12. C. Xu, P. J. Reeves, Q. Jacquet and C. P. Grey, Adv. Energy Mater., 2021, 11, 2003404.
13. J. Zheng, Y. Ye, T. Liu, Y. Xiao, C. Wang, F. Wang and F. Pan, Acc. Chem. Res., 2019, 52, 2201.
14. J. Li, G. Liang, W. Zheng, S. Zhang, K. Davey, W. K. Pang and Z. Guo, Nano Mater. Sci., 2022, 5(4), 404–420.
15. W. M. Seong, Y. Kim and A. Manthiram, Chem. Mater., 2020, 32, 9479.
16. W. M. Seong, K.-H. Cho, J.-W. Park, H. Park, D. Eum, M. H. Lee, I. S. Kim, J. Lim and K. Kang, Angew. Chem., Int. Ed., 2020, 59, 18662.
17. S. S. Zhang, Energy Storage Mater., 2020, 24, 247.
18. Y. Su, L. Li, G. Chen, L. Chen, N. Li, Y. Lu, L. Bao, S. Chen and F. Wu, Chin. J. Chem., 2021, 39, 189.
19. A. Chen, K. Wang, J. Li, Q. Mao, Z. Xiao, D. Zhu, G. Wang, P. Liao, J. He and Y. You, Front. Energy Res., 2020, 8, 593009.
20. J. Dutta, S. Ghosh, K. K. Garlapati and S. K. Martha, Small, 2024, 2405432.
21. A. Chen, K. Wang, J. Li, Q. Mao, Z. Xiao, D. Zhu, G. Wang, P. Liao, J. He, Y. You and Y. Xia, Front. Energy Res., 2020, 8, 593009.
22. C.-H. Jo, D.-H. Cho, H.-J. Noh, H. Yashiro, Y.-K. Sun and S. T. Myung, Nano Res., 2015, 8, 1464.
23. Y. Su, L. Li, G. Chen, L. Chen, N. Li, Y. Lu, L. Bao, S. Chen and F. Wu, Chin. J. Chem., 2021, 39, 189.
24. J. Park, Y. Kim, Y. Kim, J. Park, D. G. Lee, Y. Lee, J. Hwang, K.-Y. Park and D. Lee, Chem. Eng. J., 2023, 467, 143335.
25. J. Dutta, S. Ghosh, K. K. Garlapati and S. K. Martha, J. Power Sources, 2023, 587, 233717.
26. Y. Mao, X. Wang, S. Xia, K. Zhang, C. Wei, S. Bak, Z. Shadike, X. Liu, Y. Yang, R. Xu, P. Pianetta, S. Ermon, E. Stavitski, K. Zhao, Z. Xu, F. Lin, X. Yang, E. Hu and Y. Liu, Adv. Funct. Mater., 2019, 29, 1900247.
27. J.-M. Lim, T. Hwang, D. Kim, M.-S. Park, K. Cho and M. Cho, Sci. Rep., 2017, 7, 39669.
28. P. Oh, B. Song, W. Li and A. Manthiram, J. Mater. Chem. A, 2016, 4, 5839.
29. J. Dutta, S. Ghosh and S. K. Martha, ACS Appl. Mater. Interfaces, 2024, 16, 19720.
30. M. Florencia and R. Diego, Surface Chemistry and Lithium-Ion Exchange in LiMn₂O₄ for the Electrochemical Selective Extraction of LiCl from Natural Salt Lake Brines, 2016.
31. D. W. Shin, J.-W. Choi, W.-K. Choi, Y. S. Cho and S.-J. Yoon, Electrochem. Commun., 2009, 11, 695.
32. C. V. Ramana, M. Massot and C. M. Julien, XPS and Raman spectroscopic characterization of LiMn₂O₄ spinels, Surf. Interface Anal., 2005, 412–416 , https://analyticalsciencejournals.onlinelibrary.wiley.com/doi/abs/10.1002/sia.2022.
33. N. Tian, C. Hua, Z. Wang and L. Chen, J. Mater. Chem. A, 2015, 3, 14173.
34. A. Ran, S. Chen, M. Cheng, Z. Liang, B. Li, G. Zhou, F. Kang, X. Zhang and G. Wei, J. Mater. Chem. A, 2022, 10, 19680.
35. M. Helan and L. J. Berchmans, Mater. Manuf. Processes, 2011, 26, 1369.
36. J. C. Hestenes, R. May, J. T. Sadowski, N. Munich and L. E. Marbella, Chem. Mater., 2022, 34, 232.
37. R. Sun, P. Jakes, S. Eurich, D. Van Holt, S. Yang, M. Homberger, U. Simon, H. Kungl and R.-A. Eichel, Appl. Magn. Reson., 2018, 49, 415.
38. S. S. Zhang, Energies, 2012, 5, 5190.
39. S. Ghosh, J. Dutta, K. K. Garlapati, M. Parvin, C. Gupta, H. N. Dixit and S. K. Martha, J. Power Sources, 2024, 623, 235440.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta03286e

---