Substitution of magnesium towards stabilizing low-nickel layered oxides for high voltage and cost-effective sodium-ion batteries

Abstract

The development and advancement of low-nickel layered oxides for cost-effective sodium-ion batteries are hindered by the lack of comprehensive studies on structural stability and the specific phase transition mechanisms during multiple irreversible phase transitions, especially under high-voltage conditions. Herein Mg substitution for Ni in O3–NaNi_0.25Fe_0.25Mn_0.5O₂ (NNFM) is proposed to mitigate the structural degradation under high voltage and long-term cycling. Through in situ XRD analysis, the complete structural evolution of NNFM and NMNFM under high-voltage conditions was revealed. Most importantly, it is revealed that Mg substitution suppresses the complex phase transitions of low-nickel cathodes under high voltage conditions and mitigates the phenomenon of phase transition hysteresis. NMNFM exhibits a high reversible capacity of 153 mA h g⁻¹ at 0.1C, decent capacity retention after 100 cycles and good rate capability. Last but not least, the fabricated hard carbon//O3-NMNFM full cell delivers an initial discharge capacity of 144 mA h g⁻¹ at 0.1C within a voltage range of 2.0–4.1 V and a capacity retention of 87.8% after 100 cycles.

---

Introduction

The rapid development of the renewable energy sector has led to the widespread commercialization of lithium-ion batteries (LIBs) in various applications.^1–3 However, the limited availability and high cost of lithium make LIBs less suitable for large-scale energy storage systems.^4,5 As a viable alternative, sodium-ion batteries (SIBs) have gained attention due to the similar electrochemical behavior of sodium (Na) and lithium (Li), coupled with the abundant and lower-cost sodium resources.^6–9 Although both energy storage systems operate based on parallel electrochemical reaction principles, the larger ionic radius of sodium (Li: 0.69 Å and Na: 0.98 Å) introduces several challenges, including more pronounced volumetric changes and slower ion transport and reaction kinetics.¹⁰ Sodium-ion batteries (SIBs) consist of four main components: the cathode, anode, separator, and electrolyte. The Na⁺ ions intercalate/deintercalate between the cathode and anode through the electrolyte to store or release energy.¹¹ Among the various SIB cathode materials, layered transition metal oxides (Na_xTMO₂) stand out due to their relatively simple synthesis processes and superior energy density compared to other cathode materials.^12–14 However, challenges such as limited cycling stability and low air stability under high-voltage conditions remain significant obstacles to their broader application.^15,16

In sodium-ion transition metal layered oxide cathodes, Ni is commonly incorporated into the transition metal layer of iron–manganese-based cathode materials to enhance capacity and improve structural stability.^17–20 O3–Na_x(Ni_1−y−zFe_yMn_z)O₂ has shown promise for application in full cells with hard carbon anodes due to their elevated sodium capacity.^21–23 Additionally, the increasing demand for nickel in the electric vehicle market will likely drive up the cost, making them less feasible for cost-effective sodium-ion batteries. To mitigate the adverse effects associated with nickel and further reduce the cost of layered oxide cathodes for sodium-ion batteries, it is imperative to further decrease the nickel content. However, the reduction in nickel content not only leads to a decrease in capacity but, more importantly, compromises the overall structural stability. In the Ni–Fe–Mn ternary system, a low nickel content corresponds to a higher proportion of Fe and Mn in the cathode material. When the Fe content in the transition metal layer is relatively high, the energy required for Fe migration to the sodium layer is significantly reduced, thereby greatly increasing the probability of irreversible Fe migration from the transition metal layer to the sodium layer.²⁴ The complex phase transitions of the O3 phase are primarily caused by the sliding of the TM (transition metal) layers and the changes in Ni and Mn transition metals, which exhibit the Jahn–Teller effect leading to a significant decline in structural stability.^24,25 Particularly under high-voltage conditions, multiple irreversible phase transitions occur, leading to particle fracture after extended cycling. Additionally, these irreversible phase transitions impair the sodium-ion diffusion kinetics, resulting in a decreased sodium-ion diffusion coefficient.¹⁹ This, in turn, causes phase transition lag during cycling, ultimately affecting the cycling performance.²⁶ Specifically, this manifests as a deterioration in their cycling stability and leads to poor rate performance. Therefore, it is critical to explore alternative cathode materials with superior electrochemical properties that minimize or eliminate the need for nickel to fulfill the cost effective property of sodium-ion batteries.

In recent years, various strategies, including element doping and substitution, have been employed to address this issue.²⁷ Guo et al. enhanced the electrochemical performance of the traditional O3-type NaNi_0.5Mn_0.5O₂ cathode material by doping with Ti⁴⁺ in place of Mn⁴⁺. They discovered that Ti substitution expands the interlayer spacing, which helps suppress the detrimental and irreversible multiphase transitions in O3-type NaNi_0.5Mn_0.5O₂ under high voltage conditions.²⁸ Sun et al. stabilized the layered structure of NaNi_0.25Fe_0.25Mn_0.5O₂ (NNFM) through the appropriate substitution of Li, attributing this stability to the stronger Li–O bonds.^29,30 Mg doping has demonstrated potential in addressing these challenges, particularly in terms of stabilizing the crystal structure and reducing complex phase evolutions.^31–37 Despite these advancements, to further minimize the nickel content in layered oxide cathodes, it is imperative to conduct additional research focusing on ultra-low nickel O3-type oxides, along with investigating the impact of substitutions on structural stability, particularly under high voltage conditions.

In this study, magnesium is used to partially substitute Ni in the transition metal layer of O3-type NaNi_0.25Fe_0.25Mn_0.5O₂ (NaMg_0.08Ni_0.17Fe_0.25Mn_0.5O₂, abbreviated as O3-NMNFM) to further minimize the nickel content to 0.17. Our meticulous analysis revealed that Mg doping significantly mitigated the phase transition lag and decreased the incidence of multiple irreversible phase transitions that were prominent in low nickel O3-layered oxide cathodes. This stabilizing effect profoundly enhanced both the kinetic and structural stability of the material under high-voltage conditions, offering invaluable insights into the design of more robust high-voltage sodium-ion cathode materials.

Results and discussion

As observed from the scanning electron microscopy (SEM) images in Fig. 1, both O3-NNFM (Fig. 1a) and O3-NMNFM (Fig. 1b) materials synthesized using a convenient high-temperature solid-state method exhibit irregular, stacked plate-like structures with particle sizes spanning from 2 to 4 μm. The elemental mapping from SEM energy-dispersive X-ray spectroscopy (SEM-EDS) further substantiates the uniform distribution of Na, Ni, Fe, Mn, and O elements in O3-NNFM, as well as Na, Mg, Ni, Fe, Mn, and O elements in O3-NMNFM (Fig. S1 and S2†). The quantitative analysis of elements in the material samples using inductively coupled plasma optical emission spectroscopy (ICP-OES) indicated that the stoichiometric ratios of Ni, Fe, and Mn in NNFM are 0.25 : 0.25 : 0.5, while the stoichiometric ratios of Ni, Mg, Fe, and Mn in NMNFM are 0.17 : 0.08 : 0.25 : 0.5 (Table S1†), which is consistent with the designed proportions in the experiment.

Fig. 1 Structural characterization of NNFM and NMNFM. SEM images of (a) NNFM and (b) NMNFM. Rietveld refinement of XRD patterns for (c) NNFM and (d) NMNFM. Schematic diagrams of the crystal structure for (e) NNFM and (f) NMNFM with Mg substitution.

The Rietveld-refined XRD patterns of NNFM and NMNFM shown in Fig. 1C indicate that both materials possess a rhombohedral symmetry structure (R3̄m space group) without impurity phases, consistent with the α-NaFeO₂ structure.³⁸ Based on the refined crystallographic data presented in Tables S2 and S3,† the unit cell parameters of the two materials are outlined as follows.

The cell volume of NNFM is 122.286 Å3 with the cell parameters of a = b = 2.9366 Å and c = 16.3697 Å, while the cell volume of NMNFM is 122.175 Å3 with the cell parameters of a = b = 2.9325 Å and increased c = 16.4034 Å. The notable increase in the c-axis lattice parameter is attributed to the larger ionic radius of Mg²⁺ (0.72 Å) compared to that of Ni²⁺ (0.69 Å), which results in an expansion of the unit cell dimensions. This expansion facilitates a wider ionic channel for Na⁺ ions, thereby enhancing ion transport kinetics.^35,37,39 Both materials exhibit an O3-type layered structure, characterized by alternating transition metal polyhedral layers and NaO₆ alkali metal layers, further confirming that Mg has partially substituted Ni at the 3b sites and mixed well with the transition metal layers (Fig. 1e and f). Additionally, to ascertain whether the discrepancy in electrochemical performance between the two cathode materials stems from variations in their specific surface area, an N₂ adsorption analysis was carried out utilizing a dedicated analyzer. The results revealed that the specific surface area of NNFM was 2.0271 m² g⁻¹, whereas that of NMNFM was 3.5028 m² g⁻¹ (as shown in Fig. S3†), indicating a minor difference. This observation implies that both materials possess comparable surface areas, hinting that the performance disparities are more likely attributed to other underlying factors.

To explore the optimal doping ratio, as shown in Fig. S4,† under high-voltage conditions (voltage window of 2–4.2 V), after two activation cycles at 0.1C, subsequent tests were conducted under conditions of 0.5C charging and 1C discharging (where 1C equals 150 mA h g⁻¹) to evaluate the electrochemical performance of the prepared samples: NNFM, NMNFM-0.05, NMNFM-0.08 (NMNFM), NMNFM-0.1, NMNFM-0.15, and NMNFM-0.2. The samples NNFM, NMNFM-0.05, NMNFM-0.08 (NNMFM), NMNFM-0.1, NMNFM-0.15 and NMNFM-0.2 deliver an initial discharge capacity of 126.7, 130.7, 127.7, 119.9, 83.6 and 76.2 mA h g⁻¹, respectively. It is evident that NMNFM-0.05 and NMNFM-0.08 exhibit higher discharge capacities during the first 50 cycles. After 100 cycles at 1C, the capacity retention rates of NNMF, NMNFM-0.05, NMNFM-0.08 (NNMFM), NMNFM-0.1, NMNFM-0.15, and NMNFM-0.2 are 48.8%, 73.3%, 82.1%, 78.2%, 67.4%, and 57.3%, respectively. With appropriate magnesium substitution, the capacity retention is significantly improved; however, the discharge capacities of NMNFM-0.15 and NMNFM-0.2 decrease considerably compared to that of NNFM, likely due to the introduction of a substantial amount of electrochemically inactive Mg²⁺. Among them, NMNFM-0.08 (NNMFM) emerges as the optimal composition, exhibiting a higher discharge specific capacity and excellent capacity retention.

The electrode performance of O3-NNFM and O3-NMNFM in sodium cells is compared in Fig. 2. As illustrated in Fig. 2a, under low-voltage conditions within a voltage window of 2–4 V, following two cycles of activation at a rate of 0.1C, subsequent tests were conducted under conditions of 0.5C charging and 1C discharging (where 1C equals 150 mA h g⁻¹). After 100 cycles, NMNFM exhibited a capacity retention of 99.4%, outperforming NNFM's retention of 93.9% by a margin of 5.5%. This underscores the excellent cycling stability of both materials at low voltages. Furthermore, the charge–discharge curves of both materials during the initial cycle displayed similar trends and plateau regions, with negligible differences (Fig. 2c). Furthermore, although NMNFM initially exhibited a lower capacity than NNFM under high-voltage conditions with a voltage window of 2–4.2 V using the same protocol, the capacity fade of NMNFM was significantly mitigated. After 100 cycles, NMNFM exhibited a superior capacity retention of 82.1% compared to that of NNFM with a capacity retention of 48.8%, as shown in Fig. 2b. The electrochemical performance of the NMNFM cathode is comparable to some classic O3-type cathode materials, as shown in Table S4.† These results underscore the exceptional cycling stability of NMNFM at higher voltages, which can be attributed to its enhanced structural stability, ultimately leading to improved electrochemical performance. This advancement is associated with the inclusion of magnesium, indicating that the incorporation of electrochemically inactive magnesium plays a crucial role in stabilizing the structure. Fig. 2d displays the initial charge–discharge curves of NNFM and NMNFM cells at a rate of 0.1C. It is evident that both curves exhibit similar plateaus within the voltage range of 2.9–3.1 V, featuring smooth charge–discharge profiles overall. Intriguingly, in the high-voltage range of 4–4.2 V, the discharge plateau of NMNFM partially diminishes, and its charge plateau is shortened compared to that of NNFM. This phenomenon indicates that Mg doping effectively inhibits specific phase transitions under high-voltage conditions³¹

Fig. 2 Electrochemical performance of NNFM and NMNFM. (a) Cycling performance of NNFM and NMNFM at 1C (150 mA h g⁻¹) under low voltage conditions, (b) cycling performance of NNFM and NMNFM at 1C (150 mA h g⁻¹) under high voltage conditions, and (c) the first charge and discharge curves of NNFM and NMNFM at 0.1C. The voltage range is within 2–4.0 V. (d) The first charge and discharge curves of NNFM and NMNFM at 0.1C, the voltage range is within 2–4.2 V, and (e) dQ/dV vs. voltage curves derived from the first charge–discharge profiles of NNFM and NMNFM at 0.1C. The voltage range is within 2–4.2 V. (f) Rate performance of NNFM and NMNFM.

Fig. 2e depicts the dQ/dV curves of the materials, displaying two distinct sets of redox peaks at corresponding positions. The smaller peak near 2.8 V (adjacent to the prominent peak) corresponds to the O3–O′3 phase transition. The pair of sharp redox peaks around 3.01/2.81 V vs. Na⁺/Na, which aligns with the broad plateau in the charge–discharge curves of both materials, signifies the O′3–P3 phase transition, where the O′3 phase persists for a fleeting moment.⁴⁰ For NMNFM, redox peaks are observed at 4.06/3.75 V vs. Na⁺/Na; however, at present, we are unable to conclusively determine whether these peaks represent a phase transition or a solid solution reaction. Research suggests that broader peaks in dQ/dV curves may correspond to more uniform sodium ion insertion and extraction, especially in the ion diffusion processes within polyhedral structures.^41,42 The absence of significant phase transitions or abrupt changes in reaction kinetics helps to minimize mechanical stress and structural strain on electrode materials during cycling.⁴³ By analyzing the dQ/dV curves at various cycle numbers, as depicted in Fig. S5,† it is observed that the redox peak near 4.1 V for NNFM exhibits a notably high initial intensity, which undergoes a significant reduction after cycling. The decrease in curve overlap signifies a deterioration in the reversibility of the electrochemical reactions at high voltage for NNFM, hinting at poor cycling stability under high-voltage conditions. In contrast, for NMNFM, the redox peaks exhibit a relatively smaller decrease across different cycle numbers, accompanied by a high degree of overlap. This suggests reduced polarization during the electrochemical reaction and implies that the reaction proceeds more smoothly. As accordingly shown in Fig. S6,† XRD characterization was conducted before and after cycling for both materials. With an increasing number of cycles, the peak position of the (003) plane for both materials shifted to the left, indicating an increase in the interlayer spacing according to Bragg's equation.^37,44 Notably, for NNFM, the disappearance of the (104) peak after 200 cycles suggests significant damage to its crystal structure. Fig. S7† presents SEM images of particles dispersed in carbon/PVDF before and after cycling. Upon examination, NNMFM particles exhibited only minor cracks after 100 and 200 cycles. In contrast, NNMF particles displayed extensive cracking after just 100 cycles, with noticeable particle breakage. These cracks further widened and the particles were more severely damaged after 200 cycles, indicating that the structure underwent mechanical stress-induced strain during cycling. Furthermore, studies on sodium-ion battery cathodes have shown that reducing the abruptness of electrochemical reactions can alleviate localized lattice stress. This reduction in stress helps preserve the structural integrity of the material throughout extended cycling, thereby minimizing mechanical degradation over time.⁴³

To further assess the electrochemical performance of the fabricated cathode, rate capability tests were conducted within a voltage range of 2–4.2 V at varying current densities. As shown in Fig. 2f, the reversible capacity of the NNMF cathode at high voltage, with a standard specific capacity of 1C = 150 mA h g⁻¹, is 153, 138, 119, 110, 102, 95, 86, and 65 mA h g⁻¹ at current rates of 0.1C, 0.2C, 0.5C, 1C, 2C, 3C, 5C, and 10C, respectively. In comparison, the Mg-doped NMNFM cathode shows enhanced rate performance, with reversible capacities of 162, 145, 128, 112, 106, 101, 95, and 89 mA h g⁻¹ at identical current rates. As shown in Fig. S8,† at a discharge rate of 10C, NMNFM achieves 54.9% of the 0.1C discharge capacity, compared to 42.4% for NNFM. This indicates that NMNFM exhibits superior rate capability.

To reveal the phase transition behavior during the structural evolution of the two cathodes during sodiation/desodiation and figure out the critical role of Mg substitution in O3-NNFM under high-voltage cycling, in situ XRD analysis was conducted over the initial cycles in the voltage range from 2–4.2 V at 0.1C. The XRD patterns and contour plots (Fig. S9† and 3a, respectively) clearly demonstrate that the NNFM and NMNFM cathodes undergo a similar phase transition pathway: O3 → O3 + O3′ → P3 → O3′ + O3 → O3 consistent with the aforementioned dQ/dV curves. It is noteworthy that during the first three charging cycles of NNFM at a high voltage of 4.2 V, the states were different. During the initial charging process, the material resided in the P3 phase, where the characteristic peak associated with the (104) plane was absent in the O-type structure. In the subsequent charging, corresponding to the transition from the O3 + O3′ to the P3 phase, the characteristic peak of the (104) plane progressively diminished. Conversely, during the third charging, which corresponds to the transition from the P3 phase to the O3′ + O3 phase, the characteristic peak of the (104) plane gradually reemerged, signifying the gradual restoration of the O-type structure. This observation underscores the limited reversibility of NNFM under high voltage conditions, which constitutes the fundamental factor influencing its cycling stability at elevated voltages.⁴⁵ At the commencement of the charging process for NNFM, as Na ions are incrementally extracted from the O3-type layered structure, the reflection angle of the (003) plane within the O3 phase undergoes a gradual shift towards lower angles. This shift arises from the escalating electrostatic repulsion between adjacent layers of oxygen ions, ultimately causing an expansion of the interlayer spacing. At a voltage around 3 V, a phase separation event occurs, evident through the bifurcation of the (003) peak into two broader peaks, which signifies the coexistence of the initial O3 phase and the newly emerged O′3 phase. Subsequently, both phases diminish, facilitating a transition to the hexagonal P3 phase. This transformation is marked by a substantial reduction in the intensity of the (104) peak of the O3 phase, accompanied by an enhancement of the (015) peak of the P3 phase. This structural change can be attributed to the sliding mechanism of the transition metal layers. Upon further charging to 4.2 V, only the P3 phase persists, with no additional peaks observed. This region is representative of the solid solution reaction involving the P3 phase. However, during the subsequent discharge process, the anticipated transition from the P3 phase to the O3 phase does not occur around 3.1 V. It is only upon recharge to approximately 3.01 V that the (104) peak of the O3 phase reappears, while the (015) peak of the P3 phase diminishes, indicating a complete reversion of the material to the O3 phase.

Fig. 3 Two-dimensional intensity contour map of in situ synchrotron X-ray diffraction (XRD, λ = 0.55941 Å) patterns for (a) NNFM during the first three electrochemical cycles and (b) NMNFM during the first two electrochemical cycles.

This delayed P3-to-O3 phase transition underscores a kinetic barrier, highlighting sluggish reaction dynamics. This kinetic hindrance is a significant factor impacting the electrochemical performance of the material. Conversely, in Mg-doped NMNFM, at an elevated voltage of 4.2 V, the material remains within the solid solution reaction of the P3 phase without undergoing any phase transitions, which notably enhances its cycling stability under high-voltage conditions. The phase transition observed during the initial charging stage of NMNFM mirrors that of NNFM. However, during the discharge process, at approximately 3.75 V, a solid solution reaction occurs in NMNFM. As the discharge proceeds to 3.1 V, the intensity of the (104) peak corresponding to the O3 phase intensifies, while the (015) peak of the P3 phase weakens. At this juncture, both the P3 and O3 phases coexist, with the O3 phase gradually gaining dominance as the discharge continues. Upon recharge to 3.04 V, the material transitions back to the hexagonal P3 phase and subsequently reverts to the O3 phase during the subsequent discharge process. This demonstrates that the phase transitions are fully reversible without any observed delays. This further validates that Mg doping alleviates the issue of delayed phase transitions at high voltage, enhances the reaction kinetics, and minimizes irreversible phase transformations.

As shown in Fig. 4c and d, the cyclic voltammetry (CV) curves of NNFM and NMNFM at scan rates of 0.1, 0.2, 0.4, 0.6, and 0.8 mV s⁻¹ within the voltage range of 1.5 to 4.2 V vs. Na⁺/Na are presented. Although similar paired redox peaks around 2.7/3.05 V and 3.75/4.07 V are observed in both NNFM and NMNFM electrodes, it can be observed that the peak for NMNFM at a high voltage of 4.17 V remains prominent and broadens with increasing scan rates. In contrast, for NNFM at a scan rate of 0.8 mV s⁻¹, the peak becomes less pronounced and approaches disappearance, further demonstrating the poor reversibility of NNFM under high voltage conditions. Since the peak current is proportional to the square root of the scan rate, the sodium ion diffusion coefficients for NNFM and NMNFM during charging were calculated using the Randles–Sevcik equation.⁴⁶

Fig. 4 (a) GITT curves of NNFM and NMNFM. (b) Sodium ion diffusion coefficients calculated from the GITT curves. (c and d) CV curves of NNFM and NMNFM at different sweep speeds. (e and f) Linear fitting of I_p corresponding to the square root of v^1/2 during charging and discharging.

The values obtained were 8.742 × 10⁻¹² cm² s⁻¹ for NNFM and 6.573 × 10⁻¹² cm² s⁻¹ for NMNFM during charging, while during discharging, the diffusion coefficients were 5.061 × 10⁻¹² cm² s⁻¹ for NNFM and 4.032 × 10⁻¹² cm² s⁻¹ for NMNFM. In comparison, NMNFM exhibits a higher apparent Na⁺ diffusion coefficient, indicative of superior kinetic properties. This finding is in accordance with the structural insights derived from the aforementioned in situ XRD results. Furthermore, the GITT curves of NNFM and NMNFM during the charge–discharge process at high voltage are shown in Fig. 4a. Leveraging Fick's second law, the sodium-ion diffusion coefficients (D_Na+) were computed using the appropriate formula, with the results displayed in Fig. 4b.⁴⁷ The figure illustrates that the diffusion coefficient of NMNFM at high voltage is substantially higher than that of NNFM. Additionally, NMNFM maintains a more stable sodium-ion diffusion coefficient throughout the process in comparison to NNFM. The single-step GITT process curves for both materials illustrate the average voltage polarization and ohmic polarization throughout the charging process. Specifically, NMNFM exhibits a lower average voltage polarization of 26 mV and an ohmic polarization of 5.3 mV throughout the entire charging process (Fig. S10a†). Conversely, NNFM demonstrates significantly higher average voltage polarization (83.1 mV) and ohmic polarization (9.1 mV) (Fig. S10b†). These data further confirm the superior Na⁺ diffusion kinetics of NMNFM during the electrochemical process. The Nyquist plot, along with the corresponding fitted equivalent circuit, is depicted in Fig. S11.† Upon examining the figure, it is evident that in both impedance components, the charge transfer resistance (R_ct) is the dominant factor, while the ohmic resistance is virtually negligible. This observation implies that the electrochemical reactions occurring at the electrode interface are primarily driven by charge transfer processes, indicating efficient ion transport and minimal resistive losses within the system. For the pristine states, where the battery has yet to undergo any charge–discharge cycles, Fig. S11a† illustrates that the charge transfer resistance (R_ct) values for NNFM and NMNFM are 392 Ω and 274 Ω, respectively. This clearly demonstrates that NMNFM exhibits a lower charge transfer resistance, facilitating faster charge transfer, which aligns with previous discussions on the electrochemical performance of these materials. After undergoing 100 cycles of electrochemical cycling, as depicted in Fig. S11b,† the R_ct values for both NNFM and NMNFM samples decreased, specifically from 392 Ω to 226.5 Ω and from 274 Ω to 188.5 Ω, respectively. This reduction suggests the occurrence of activation reactions during cycling, with minimal side reactions, further reinforcing the idea that magnesium doping optimizes the material and enhances its kinetic performance. This enhancement is attributed to the incorporation of Mg, which fosters better structural stability and enhances ion transport pathways. Consequently, issues related to phase transition lag and poor phase reversibility are mitigated, enabling smoother sodium-ion diffusion, as previously demonstrated in this study.

Typically, sodium layered oxides suffer from poor air stability, which can lead to issues in transportation and storage.⁴⁸ To assess air stability, the two samples, NNMF and NMNFM, were placed in a controlled humidity chamber with 80% humidity at a temperature of 25 °C for seven days, followed by drying in air for two days. As shown in Fig. S12,† some NNMF particles exhibit surface microcracks, and in some cases, particle damage with visible cracks, along with the presence of rod-like substances. Based on previous experimental studies, it can be inferred that these are small amounts of hydrolyzed products, specifically NaHCO₃, attached to the surface.⁴⁹ In contrast, the surface of NMNFM remains relatively smooth, though small particles adhere to its surface. These particles are identified as nanosheet-like NaHCO₃.⁵⁰ As shown in Fig. 5a, the XRD patterns of NMNFM after humidity treatment retained the primary peaks of the (003) and (104) planes with only slight shifts in the remaining peaks, while the characteristic peaks of NNMF disappeared after soaking, and new phases emerged, indicating the formation of common hydrated impurity phases due to the reaction of the O3 material with H₂O and CO₂. The commonly observed hydrated impurity phase is water-sodium manganese mineral, which has a structure very similar to P2–Na_0.67MnO₂, except that additional water molecules are present in the sodium layers and the interlayer spacing is wider. The phase with an interlayer distance of approximately 7.1 Å corresponds to birnessite, while with further water insertion, the phase with an interlayer distance of 9.1 Å is the buserite phase.⁴⁹Fig. 5a represents that NNFM corresponds to the buserite phase. Accordingly, as depicted in Fig. 5b and c, upon exposure to humidity, the capacity of the NNMF sample diminished from 160 mA h g⁻¹ to 78 mA h g⁻¹, whereas the NMNMF sample experienced a more modest decline from 150 mA h g⁻¹ to 96 mA h g⁻¹. Furthermore, the NMNMF sample retained 75% of its initial capacity after 100 cycles at a 1C rate. These findings suggest that the incorporation of suitable ions into the framework can markedly mitigate the chemical sensitivity of O3-type Na-based layered oxides to moisture and carbon dioxide. This, in turn, enhances structural stability, fosters higher retention of active sodium ions, and minimizes the formation of residual alkaline species.

Fig. 5 Electrochemical performance of NNMF and NMNFM after water immersion and subsequent drying. (a) XRD patterns, (b) charge and discharge curves at 0.1C, and (c) cycling stability at 1C. Electrochemical performance of the sodium-ion full battery coupling O3-NMNFM as the cathode and hard carbon as the anode. (d) Rate performance, (e) charge and discharge curves at 0.1C, and (f) cycling stability at 1C.

In addition to its outstanding performance in sodium half-cells, the full cell assembled with NMNFM and commercial hard carbon exhibits a high initial discharge capacity of 144 mA h g⁻¹ as shown in Fig. 5d in the voltage range of 2.0–4.1 V at a 0.1C rate. Notably, even at an ultra-high current density, the cell exhibits a reversible specific capacity of 91 mA h g⁻¹, indicating that the full battery system demonstrates excellent rate performance. Fig. 5e illustrates the first charge–discharge cycle at a rate of 0.1C; after 100 cycles, the capacity retention is 87.8% at 1C, as shown in Fig. 5f. The electrochemical performance of the NMNFM//HC full cell is comparable to that of some classic O3-type cathode material-based full cells, as shown in Table S5.†

Experimental

Materials synthesis

O3-type layered NaNi_0.25Fe_0.25Mn_0.5O₂ (NNFM) and NaMg_0.08Ni_0.17Fe_0.25Mn_0.5O₂ (NMNFM) were synthesized by a high-temperature solid-state reaction. 5% excess Na₂O₃ (99.5%, sigma-aldrich) was measured with stoichiometric amounts of spherical NiO (99.5%, RHAWN), Fe₂O₃ (99.5%, MACKLIN) and Mn₂O₃ (98%, MACKLIN) for NNFM and stoichiometric amounts of MgO (99.9%, RHAWN), spherical NiO, Fe₂O₃, and Mn₂O₃ for NMNFM. The precursor was thoroughly mixed by ball milling at 500 rpm for 6 hours in a sealed zirconia container using ethanol as the dispersant. After the milling process, the milling jars were removed and placed in an oven to dry the materials inside. The dried sample mixture was thoroughly ground using a mortar and pestle, and then pressed into pellets under a pressure of 11 MPa. The pellets were placed in a muffle furnace and heated at 900 °C for 12 hours in air, followed by natural cooling to approximately 100 °C. After calcination, the two materials were stored in an argon-filled glove box (H₂O, O₂ < 1 ppm) to prevent exposure to moisture until further use.

Materials characterization

Scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDS, MAIA3 LMH) was employed to characterize the microstructure and elemental distribution of the materials. The crystal structure of the cathode powder material was analyzed using X-ray diffraction (XRD) (Bruker D8 Advance) with a Cu radiation (λ = 1.5418 Å) source. The scanning range for the diffraction pattern was set between 10° and 80°, with a scanning speed of 10° min⁻¹. After cycling, the electrodes of the material were cleaned with DMC, dried, and then tested. In situ XRD measurements during electrochemical cycling were also carried out on the in situ coin cell using a Ag Kα1 radiation source (λ = 0.55941 Å) in the range of 2θ = 2–40°. Accurate structural parameters were obtained through analysis using the Rietveld refinement method combined with the GSASII (General Structure Analysis System) software. The specific surface area (SSA) was analyzed by using N₂ as the adsorbent on a Micromeritics analyzer (HPHC PM7240). The VESTA software was used to generate schematic diagrams of the crystal structures of the two materials.⁵¹

Electrochemical characterization

The cathode active material, conductive carbon black, and binder were mixed in a mass ratio of 8 : 1 : 1 and thoroughly ground. Subsequently, an appropriate amount of NMP solvent was added, and the mixture was stirred for approximately 12 hours using a magnetic stirrer to form a homogeneous slurry. The slurry was uniformly coated onto carbon-coated aluminum foil using a blade of appropriate thickness, followed by drying in a 60 °C convection oven for 6 hours. It was then transferred to a 70 °C vacuum oven for overnight drying. The entire electrode preparation process was conducted in a drying room. The electrolyte was composed of 1 M NaClO₄ dissolved in polycarbonate (PC) with 5% fluoroethylene carbonate (FEC) as an additive. The CR2032 coin cells were assembled in an argon-filled glove box (O₂ < 0.1 ppm and H₂O < 0.1 ppm), with a glass fiber separator (Whatman) placed between the cathode made of the active material and the metallic sodium anode. For the sodium-ion full cells, commercially available hard carbon was used as the anode, and capacity matching was conducted based on the reversible capacities of the cathode and anode. Specifically, the N/P ratio of the full cell was set to 1 : 1.2. Prior to the assembly of the full cell, the anode electrode was pre-sodiated. The half-cells and full cells were tested using a Neware battery testing system, where constant-current constant-voltage (CC-CV) charging and constant-current discharging tests were performed at room temperature. The charging and discharging current settings were based on real-life scenarios of slow charging and fast discharging, with charging at a rate of 0.5C and discharging at 1C (1C = 150 mA h g⁻¹). This testing method can accurately and reliably evaluate the electrochemical performance of the cathode material. Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) measurements were conducted using a Princeton electrochemical workstation.

Conclusions

In summary, a comprehensive investigation was conducted on the impact of Mg substitution for Ni on the structural evolution and electrochemical performance of O3–NaNi_0.25Fe_0.25Mn_0.5O₂ during charge and discharge processes. In addition to the pillar effect of inert element doping on structural stability, replacing Ni with Mg increased the interlayer spacing, enhanced the Na⁺ diffusion rate, and improved the rate performance. Careful in situ XRD analysis demonstrates that Mg substitution significantly reduces the bond length between transition metals and oxygen (TM–O) and decreases the thickness of the TMO₂ layers and helps to minimize the sliding of transition metal layers under high voltage conditions. Consequently, the NMNFM material realizes a high initial reversible capacity of 153 mA h g⁻¹ at 0.1C and a capacity retention of 82.1% after 100 cycles at 1C in the potential range of 2–4.2 V, which is 33.3% higher than that of NNMF. The fabricated O3-NMNFM//hard carbon full cell demonstrates a discharge capacity of 144 mA h g⁻¹ in the voltage range of 2.0–4.1 V at 0.1C and a high capacity retention of 87.8% after 100 cycles at 1C. Additionally, the replacement of Ni with Mg significantly improves the material's stability in air. After exposure to moisture, the reversible capacity of the NNMF sample decreased from 160 mA h g⁻¹ to 78 mA h g⁻¹, while that of the NMNNF sample only dropped from 150 mA h g⁻¹ to 96 mA h g⁻¹. Additionally, after 100 cycles at 1C, NMNNF retained 75% of its initial capacity, indicating superior stability compared to NNMF under humid conditions. This study provides new insights into enhancing the structural stability and electrochemical performance of O3-type layered oxide cathode materials for sodium-ion batteries under high voltage conditions.

Data availability

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

Author contributions

Yongliang Ma conducted the experiments and data analysis. Wei Tang, Chengyong Shu, Weibo Hua, Xiaohui Sun Jintian Luo and Kang Yang formulated experimental planning. Haihan Zhang, liang Xie and Zhengxin Huang assisted in the experimental process. The original draft was written by Yongliang Ma. The manuscript was reviewed and edited by Wei Tang.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The project was supported by the National Key R&D Program of China (2021YFB2400400), the National Natural Science Foundation of China (Grant No. 22379120 and 22409157), the Key Research and Development Plan of Shanxi Province (China, Grant No. 2018ZDXM-GY-135 and 2021JLM-36), the Higher Education Institution Academic Discipline Innovation and Talent Introduction Plan (”111 Plan”) (No. B23025), the “Young Talent Support Plan” of Xi'an Jiaotong University (71211201010723). and the Qingyuan City’s 2023 Provincial Science and Technology Innovation Strategy Special Project (2023DZX015).

Notes and references

1. J. U. Choi, N. Voronina, Y. K. Sun and S. T. Myung, Adv. Energy Mater., 2020, 10, 2002027.
2. A.-M. Desaulty, D. Monfort Climent, G. Lefebvre, A. Cristiano-Tassi, D. Peralta, S. Perret, A. Urban and C. Guerrot, Nat. Commun., 2022, 13, 4172.
3. G. Qu, J. Wang, G. Liu, B. Tian, C. Su, Z. Chen, J. P. Rueff and Z. Wang, Adv. Funct. Mater., 2018, 29, 1805227.
4. K. Chayambuka, G. Mulder, D. L. Danilov and P. H. L. Notten, Adv. Energy Mater., 2018, 8, 1800079.
5. N. Yabuuchi, K. Kubota, M. Dahbi and S. Komaba, Chem. Rev., 2014, 114, 11636–11682.
6. Y. Wang, J. Liu, B. Lee, R. Qiao, Z. Yang, S. Xu, X. Yu, L. Gu, Y.-S. Hu, W. Yang, K. Kang, H. Li, X.-Q. Yang, L. Chen and X. Huang, Nat. Commun., 2015, 6, 7401.
7. P. K. Nayak, L. Yang, W. Brehm and P. Adelhelm, Angew. Chem., Int. Ed., 2017, 57, 102–120.
8. S. Osman, S. Hu, Y. Wei, J. Liu, J. Xiao, W. Yao, C. Han, X. Guo, J. Liu and Y. Tang, Adv. Energy Mater., 2024, 2404685,  DOI:10.1002/aenm.202404685.
9. Z. Liu, J. Shen, S. Feng, Y. Huang, D. Wu, F. Li, Y. Zhu, M. Gu, Q. Liu, J. Liu and M. Zhu, Angew. Chem., Int. Ed., 2021, 60, 20960–20969.
10. Y.-F. Sun, Y. Li, Y.-T. Gong, Z.-X. Qiu, J. Qian, Y. Bai, Z.-L. Wang, R.-P. Zhang and C. Wu, Energy Mater., 2024, 4, 400002.
11. Z. Pan, H. Chen, Y. Zeng, Y. Ding, X. Pu and Z. Chen, Energy Mater., 2023, 3, 300054.
12. P. F. Wang, Y. You, Y. X. Yin and Y. G. Guo, Adv. Energy Mater., 2017, 8, 1701912.
13. M. H. Han, E. Gonzalo, G. Singh and T. Rojo, Energy Environ. Sci., 2015, 8, 81–102.
14. H. Zhang, L. Song, S. Lin, Z. Huang, C. Shu, Y. Ma, Z. Tang, X. Wang, W. Tang and Y. Wu, Energy Stor. Mater., 2024, 73, 103796.
15. K. Zhang, Z. Xu, G. Li, R. J. Luo, C. Ma, Y. Wang, Y. N. Zhou and Y. Xia, Adv. Energy Mater., 2023, 13, 2302793.
16. X. G. Yuan, Y. J. Guo, L. Gan, X. A. Yang, W. H. He, X. S. Zhang, Y. X. Yin, S. Xin, H. R. Yao, Z. Huang and Y. G. Guo, Adv. Funct. Mater., 2022, 32, 2111466.
17. E. Gonzalo, M. H. Han, J. M. López del Amo, B. Acebedo, M. Casas-Cabanas and T. Rojo, J. Mater. Chem. A, 2014, 2, 18523–18530.
18. L. Sun, Y. Xie, X. Z. Liao, H. Wang, G. Tan, Z. Chen, Y. Ren, J. Gim, W. Tang, Y. S. He, K. Amine and Z. F. Ma, Small, 2018, 14, 1704523.
19. F. Ding, C. Zhao, D. Zhou, Q. Meng, D. Xiao, Q. Zhang, Y. Niu, Y. Li, X. Rong, Y. Lu, L. Chen and Y.-S. Hu, Energy Stor. Mater., 2020, 30, 420–430.
20. Z. Liu, C. Peng, J. Wu, T. Yang, J. Zeng, F. Li, A. Kucernak, D. Xue, Q. Liu, M. Zhu and J. Liu, Mater. Today, 2023, 68, 22–33.
21. K. Kubota, S. Kumakura, Y. Yoda, K. Kuroki and S. Komaba, Adv. Energy Mater., 2018, 8, 1703415.
22. Q. Liu, Z. Hu, M. Chen, C. Zou, H. Jin, S. Wang, S. L. Chou and S. X. Dou, Small, 2019, 15, 1805381.
23. X. Liang, J. Y. Hwang and Y. K. Sun, Adv. Energy Mater., 2023, 13, 2301975.
24. X. Li, Y. Wang, D. Wu, L. Liu, S.-H. Bo and G. Ceder, Chem. Mater., 2016, 28, 6575–6583.
25. Z. Chen, Y. Deng, J. Kong, W. Fu, C. Liu, T. Jin and L. Jiao, Adv. Mater., 2024, 36, 2402008.
26. W. Hua, X. Yang, S. Wang, H. Li, A. Senyshyn, A. Tayal, V. Baran, Z. Chen, M. Avdeev, M. Knapp, H. Ehrenberg, I. Saadoune, S. Chou, S. Indris and X. Guo, Energy Stor. Mater., 2023, 61, 102906.
27. Y. You and A. Manthiram, Adv. Energy Mater., 2017, 8, 1701785.
28. P. F. Wang, H. R. Yao, X. Y. Liu, J. N. Zhang, L. Gu, X. Q. Yu, Y. X. Yin and Y. G. Guo, Adv. Mater., 2017, 29, 1700210.
29. S.-M. Oh, S.-T. Myung, J.-Y. Hwang, B. Scrosati, K. Amine and Y.-K. Sun, Chem. Mater., 2014, 26, 6165–6171.
30. X. Li, X. Shen, J. Zhao, Y. Yang, Q. Zhang, F. Ding, M. Han, C. Xu, C. Yang, H. Liu and Y.-S. Hu, ACS Appl. Mater. Interfaces, 2021, 13, 33015–33023.
31. K.-N. Jung, J.-Y. Choi, H.-S. Shin, H. T. Huu, W. B. Im and J.-W. Lee, Solid State Sci., 2020, 106, 106334.
32. K. Kubota, N. Fujitani, Y. Yoda, K. Kuroki, Y. Tokita and S. Komaba, J. Mater. Chem. A, 2021, 9, 12830–12844.
33. Z. Liu, J. Wu, J. Zeng, F. Li, C. Peng, D. Xue, M. Zhu and J. Liu, Adv. Energy Mater., 2023, 13, 2301471.
34. L. Wang, C. Zhang, L. Yang, S. Li, H. Chu, X. Li, Y. Meng, H. Zhuang, Y. Gao, Z. Hu, J.-M. Chen, S.-C. Haw, C.-w. Kao, T.-S. Chan, X. Shen, Z. Wang and R. Yu, ACS Appl. Mater. Interfaces, 2023, 15, 11756–11764.
35. P. F. Wang, Y. You, Y. X. Yin, Y. S. Wang, L. J. Wan, L. Gu and Y. G. Guo, Angew. Chem., Int. Ed., 2016, 55, 7445–7449.
36. X. Xu, S. Hu, Q. Pan, Y. Huang, J. Zhang, Y. Chen, H. Wang, F. Zheng and Q. Li, Small, 2023, 20, 2307377.
37. C. Zhang, R. Gao, L. Zheng, Y. Hao and X. Liu, ACS Appl. Mater. Interfaces, 2018, 10, 10819–10827.
38. N. Yabuuchi, H. Yoshida and S. Komaba, Electrochemistry, 2012, 80, 716–719.
39. Q. W. Chenglong Zhao, Z. Yao, J. Wang, B. Sánchez-Lengeling, F. Ding, X. Qi, Y. Lu, X. Bai, B. Li and H. Li, Science, 2020, 370, 708–711.
40. M. Yang, Z. Fan, J. Liu, Y. Ma, Y. Bai, X. Wang, H. Zhang, F. Hai, W. Hua, Y. Ma, Y. Wu and W. Tang, Energy Fuels, 2024, 38, 2453–2462.
41. K. Ando, T. Matsuda and D. Imamura, J. Power Sources, 2018, 390, 278–285.
42. J. Kim, W. Lee, J. Seok, S. Park, J. K. Yoon, S.-B. Yoon and W.-S. Yoon, Cell Rep. Phys. Sci., 2023, 4, 101331.
43. L. Sun, Z. Wu, M. Hou, Y. Ni, H. Sun, P. Jiao, H. Li, W. Zhang, L. Zhang, K. Zhang, F. Cheng and J. Chen, Energy Environ. Sci., 2024, 17, 210–218.
44. C. Zhou, L. Yang, C. Zhou, B. Lu, J. Liu, L. Ouyang, R. Hu, J. Liu and M. Zhu, ACS Appl. Mater. Interfaces, 2019, 11, 7906–7913.
45. K. Kubota, I. Ikeuchi, T. Nakayama, C. Takei, N. Yabuuchi, H. Shiiba, M. Nakayama and S. Komaba, J. Phys. Chem. C, 2014, 119, 166–175.
46. T. Huang, Q. Cao, B. Jing, X. Wang, D. Wang and L. Liang, Chem. Eng. J., 2022, 430, 132677.
47. F. Ding, C. Zhao, D. Xiao, X. Rong, H. Wang, Y. Li, Y. Yang, Y. Lu and Y.-S. Hu, J. Am. Chem. Soc., 2022, 144, 8286–8295.
48. A. Joshi, S. Chakrabarty, S. H. Akella, A. Saha, A. Mukherjee, B. Schmerling, M. Ejgenberg, R. Sharma and M. Noked, Adv. Mater., 2023, 35, 2304440.
49. W. Zuo, J. Qiu, X. Liu, F. Ren, H. Liu, H. He, C. Luo, J. Li, G. F. Ortiz, H. Duan, J. Liu, M.-S. Wang, Y. Li, R. Fu and Y. Yang, Nat. Commun., 2020, 11, 41467.
50. J. Lamb, K. Jarvis and A. Manthiram, Small, 2022, 18, 2106927.
51. K. Momma and F. Izumi, J. Appl. Crystallogr., 2008, 41, 653–658.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4se01730g

---