Optimizing Electrochemical Performance in Sodium-Ion Batteries using O3-type Na_0.90Cu_0.22Fe_0.30Mn_0.48O₂ and Hard Carbon

The identification of compatible electrolytes is crucial for augmenting the cycle life of SIBs. Besides offering ionic conductivity, electrolytes aid in the formation of consistent, stable Solid Electrolyte Interface (SEI) films that thwart the continuous degradation of the electrode or electrolyte at extreme potentials. To locate the most suitable electrolyte for NCFMO cathode materials, tests were conducted on NCFMO half cells employing five distinct electrolyte compositions. The precise compositions of the five electrolytes are displayed in Table I.

The results of the rate and cycling performance tests using five electrolytes for NCFMO cathode materials are shown in Fig. S1 and Table SII. The maximal initial discharge specific capacity of 101.9 mAh·g⁻¹ was accomplished using electrolyte A, presumably owing to the substantial dielectric constant and proficient film-forming capacity of ethylene carbonate (EC) in conjunction with the utilization of propylene carbonate (PC) to facilitate ion pair and ionic aggregate formation, thereby enhancing the aggregate performance of the electrolyte.

With a high content of PC, electrolyte B demonstrated a propensity for poor cycle stability of the cell, attributed to the heightened viscosity of PC and challenges in forming dense interfacial films, thereby resulting in a capacity retention merely 56.4%.

Beyond the solvent, the variety of electrolyte salt proves instrumental in modulating electrochemical performance. Electrolyte C, despite having a similar composition, delivered inferior rate performance and cycling performance in comparison to electrolyte A. This underscores that NaClO₄ can offer superior capacity and cycling performance compared to the electrolyte salt NaPF₆.

It has been postulated that by-products emanating from the reaction between linear carbonate and sodium metal electrodes might migrate to the anode during cycling, and these migrated by-products could experience oxidative decomposition at the anode. ²⁸ Therefore, the inclusion of solvents EMC and DMC in electrolytes D and E likely react with the cathode, and their poor film-forming capacity and solvation of electrolyte salts result in diminished initial discharge specific capacities of 94.1 and 93.2 mAh·g⁻¹ when tested with electrolytes D and E, respectively.

Nonetheless, the selection of electrolytes should not solely consider the cathode material test results in isolation. Instead, it is imperative to also take into account the compatibility with the anode material.

The electrochemical attributes of the O3-type NCFMO material were scrutinized at 25 °C using electrolyte B, as presented in Fig. 3. Figure 3a delineates the constant-current charge/discharge curves of the NCFMO cathode material at 0.1C within the voltage scope of 2 ∼ 4 V. The results reveal an initial discharge specific capacity of approximately 102 mAh·g⁻¹, and the congruence of the charge/discharge curves across the initial three cycles, which implies a highly reversible sodium ion intercalation/deintercalation process.

Zoom In Zoom Out Reset image size

The rate performance and cycling stability of the NCFMO cathode material were examined, as depicted in Figs. 3b–3e. Figure 3b demonstrates the discharge specific capacities of NCFMO electrodes at various rates, ranging from 0.1C to 10C, amounting to 103, 86, 76, 58, and 48 mAh·g⁻¹. Figures 3c–3e exhibit the cycling stability of NCFMO half cells at differing rates, with specific conditions and outcomes tabulated in Table II. NCFMO exhibits relatively robust cycling stability across diverse rates, with a capacity retention of 80.5% over 200 cycles at 1C and a Coulombic efficiency approaching 100% throughout the cycles. There is a 46.5% retention over 500 cycles at 5C, with capacity decay partially attributable to the structural instability of sodium ions during recurrent deintercalation at elevated rates, culminating in irreversible capacity loss.

The electrochemical properties of the NCFMO material were assessed by varying the voltage range, as demonstrated in Fig. S2. With a voltage scope of 2.5 ∼ 4.05 V and a rate of 0.1C, the initial discharge specific capacity can achieve 99.4 mAh·g⁻¹. Following 100 cycles, the capacity retention rate is 83.7%, indicating exceptional cycling stability. All electrochemical assay outcomes denote that the Cu/Fe/Mn transition metal layer enhances not only the stability but also the reversibility of the material. ²⁹

The majority of SIBs performance optimization research has been executed using half cells, and such conclusions might not translate to full cell applications due to significant differences between the sodium metal counter electrode and the carbon anode. Initially, the capacity balance between the cathode and anode was evaluated, as this factor is crucial for maximizing the energy density of the full cell. The NCFMO/HC full cell capacity ratio (referred to as P/N) was varied from 0.8 to 1.2, followed by electrochemical testing.

The ratio of cathode and anode capacity in a full cell is limited by their weight (their ratio will be abbreviated as P/N), and P/N can be calculated by the following equation:

$$\begin{eqnarray}&&\displaystyle \frac{P}{N}=\displaystyle \frac{{m}_{p}\times {C}_{P}}{{m}_{N}\times {C}_{N}}\end{eqnarray} \tag{ 1 }$$

P is the capacity of the cathode, N is the capacity of the anode, m_P is the mass of the cathode material, m_N is the mass of the anode material, C_P is the first discharge specific capacity of the cathode material at 0.1C, and C_N is the first charge specific capacity of the anode material at 0.1C.

The findings are presented in Figs. 4a–4b and Table III. The optimal first discharge specific capacity of 80.2 mAh·g⁻¹ was obtained with a P/N of 0.9. Post 200 cycles at a 1C, the maximum capacity retention of 84.3% was recorded for the full cell with a P/N of 0.9. This illustrates that a minor surplus capacity on the anode side (5% ∼ 10% in the LIBs industry) can mitigate non-uniform current distribution, thereby preventing local sodium metal deposition. Nevertheless, an excessively small P/N may instigate irreversible capacity loss, leading to limited cathode utilization and reduced battery energy density, corroborating previous studies. ¹⁷

Zoom In Zoom Out Reset image size

Subsequently, the impact of the electrolyte on the NCFMO/HC full cell's electrochemical performance was assessed. Four electrolytes (B, C, D, and E) were used to assemble the full cell, and charge-discharge tests were performed. The results, illustrated in Figs. 4c–4d and Table IV, revealed that the highest first-cycle specific capacity and capacity retention after 200 cycles were obtained with electrolyte B, implying that sodium-ion electrolytes with high Propylene Carbonate (PC) content may enhance the cycling stability of the battery, consistent with earlier findings. ³⁰ Electrolyte E underwent continuous electrochemical decomposition during testing, which could be attributed to the poor film formation capability of the Dimethyl Carbonate (DMC) solvent and its dissolution capacity for electrolyte salts. The relatively high potential of the HC anode during the end of the charging cycle might inhibit the formation of a stable organic layer of Ethylene Carbonate (EC) or other esters, thus reducing the capacity retention of full cells employing electrolytes C and E. Furthermore, the presence of Ethyl Methyl Carbonate (EMC) solvent in electrolyte D has a tendency to interact with the cathode, consequently degrading the overall electrochemical performance of the full cell. Considering these observations, electrolyte B emerges as the preferred choice for NCFMO/HC full cell.

The full cell's electrochemical performance was assessed at 25 °C employing the optimal cathode-to-anode capacity ratio (P/N = 0.9) and electrolyte B. As depicted in Fig. 5a, even at a high discharge rate of 10C, the NCFMO maintains a specific discharge capacity of 35 mAh·g⁻¹ in the full cell, demonstrating admirable rate capability. Following 200 cycles at 0.5C, and 500 cycles at 1C and 5C, the respective capacity retention were 79.6%, 56.8%, and 56.5%.

Zoom In Zoom Out Reset image size

Despite the generally satisfactory performance of the full cell, there remains room for enhancement. Several potential constraints to its performance are worth mentioning. First is the oxidative decomposition of the electrolyte during the charge/discharge process, as discussed earlier. The second factor is the formation of the Solid Electrolyte Interface (SEI) and the cycling stability of the anode material when paired with NCFMO. The third aspect concerns the degradation of the cathode material, which can lead to the dissolution of transition metals and subsequent loss of accessible Na⁺. In a full cell, the quantity of Na⁺ is fixed, offering no additional Na⁺ to offset capacity degradation during side reactions and the irreversible intercalation/deintercalation of Na⁺ (such as SEI formation and irreversible reactions). ^25,34,35 These factors collectively contribute to the irreversible capacity fade in the full cell.

To ascertain the origins of the full cell's capacity fade, we carried out XRD and SEM tests on the cathode and anode electrode materials post 200 cycles at 1C. Figure S5 displays the XRD patterns for the cathode/anode electrode sheets (designated NCFMO-200th and HC-200th, respectively) following 200 cycles. Several diffraction peaks emerge at equivalent positions for the cathode and anode electrodes, which could be ascribed to a significant presence of amorphous sodium compounds on both electrode materials. NCFMO-200th exhibits more pronounced characteristic peaks at 23° and 29°. The 23° peak is likely due to water molecules from the atmosphere incorporated into the MO2 layer. ³⁶ The diffraction peak at 16° aligns with the (003) crystal plane of the material, with the peak intensity markedly amplified post 200 cycles compared to the pre-cycling stage. The (104) _O3 peak at 42° in NCFMO-200th vanishes, while new peaks appear at 43° and 45°, signifying a phase transition from O3 to P3. ³⁷ Moreover, the novel peaks at 47° and 48° might result from irreversible structural alterations prompted by excessive sodium extraction due to the charging cut-off voltage exceeding 3.5 V. ³⁸ The characteristic peak around 26° of the HC-200th electrode is attributed to the 002 crystal plane of the carbon material. The intensity of this peak correlates with the crystal's lamellar structure, with a more intact structure leading to a stronger peak. ³⁹ Since the active material, following the completion of the cycle, is directly coated on aluminum foil for XRD testing, an Al peak appears at 65° on the XRD pattern. Additionally, several unidentified peaks are present in the XRD pattern, warranting further investigation.

Figure S6 presents SEM images of the cathode and anode electrode materials following 200 cycles at 1C, coupled with Energy Dispersive Spectroscopy (EDS) analysis for examining the electrode sheet's material microregion composition post-cycling. As seen in Figs. S6a–S6b, minor fracturing is observable in the cathode electrode material particles, which could contribute to its capacity fade, while the small particles adhering to the surface of the negative electrode material might result from precipitated sodium. From Figs. S8c–S8d and Tables SIII–SIV, it can be inferred that the post-cycling NCFMO electrode material's metal ion content (e.g., Na, Cu, Fe, and Mn), as well as the HC electrode's carbon element composition and content, essentially aligns with the pre-cycling electrode material. The presence of Na and Cl elements in the HC electrode arises from irreversible sodium intercalation and a minor amount of NaClO₄ electrolyte adhered to the electrode.

Considerable performance disparities exist between half cells and full cells, as demonstrated in Fig. 6, where the charge-discharge curves for the NCFMO layered oxide material in the half-cell considerably deviate from those in the full cell. When employing sodium metal as the negative electrode in the half cell, the sodium-ion loss across the entire system is virtually negligible, thus obscuring any significant reflection of sodium-ion consumption at the interface. In stark contrast, the full cell, with its limited Na⁺ content, experiences a noticeable irreversible depletion of active sodium due to interfacial and side reactions. Consequently, from a theoretical standpoint, the half-cell ought to outperform the full cell in terms of cycling performance and specific capacity. However, the heightened reactivity of sodium metal to the electrode in the half-cell predisposes it to side reactions and sodium dendrite formation, ⁴⁰ hindering a critical evaluation of the material's cycling stability. Therefore, assessing electrodes/electrolytes via full cells is indispensable for the development of practical SIB materials.

Zoom In Zoom Out Reset image size

In addition to these factors, the choice of conductive additives and binders also exerts an influence on the full cell's electrochemical performance. The three most commonly utilized conductive additives in laboratory-scale systems are Carbon Black (CB), Acetylene Black (AB), and Ketjen Black (KB). Binders aid in the formation of a protective layer on the electrodes, curtailing the continuous growth of Solid-Electrolyte Interphase (SEI) films. The binders predominantly reported in current studies encompass polyvinylidene fluoride (PVDF), sodium carboxymethylcellulose (CMC), and sodium alginate (SA). To optimize the benefits of varying electrode materials and enhance the overall performance of the battery, it becomes necessary to select conductive additives and binders that are complementary to the electrode materials' properties.

The overpotential during charge and discharge cycles was ascertained using the Galvanostatic Intermittent Titration Technique (GITT), allowing for further computation of the apparent diffusion coefficient and irreversible heat of Na⁺. GITT is a cyclic process involving alternate pulsing and relaxation stages. Here, the pulsing refers to the stage where current flows, while relaxation denotes the stage without current flow. A constant current is applied to the cell at a specific rate, and after a certain duration, the current is halted, transitioning into a relaxation period, during which the electrode potential drifts towards a new equilibrium value. ⁴¹ This sequence continues until a predetermined voltage is attained, concurrently recording voltage fluctuations during both the constant current and relaxation periods. The test conditions adopted in this study include: each charging or discharging step lasts 22 min or until the cutoff voltage is reached, each relaxation step lasts 60 min or until |dV/dt| < 1 mV·min⁻¹, and the overpotential is captured for the steady-state phase of the relaxation period at the conclusion of each pulse. The apparent diffusion coefficient of Na⁺ is computed utilizing the following equation: ⁴²

$$\begin{eqnarray}&&{D}_{N{a}^{+}}=\displaystyle \frac{4}{\pi \tau }{\left(\displaystyle \frac{{m}_{B}{V}_{M}}{{M}_{B}S}\right)}^{2}{\left[\displaystyle \frac{{\rm{\Delta }}{E}_{s}}{{\rm{\Delta }}{E}_{\tau }}\right]}^{2}\left(\tau \ll \displaystyle \frac{{L}^{2}}{D}\right)\end{eqnarray} \tag{ 2 }$$

where D_Na ⁺ symbolizes the sodium ion diffusion coefficient (in cm²·s^–1), m_B represents the mass of the active material (in g), V_M is the molar volume of the electrode material (in cm³·mol^–1), M_B is the relative molar mass of the material (in g·mol⁻¹), S is the effective surface area of contact between the electrode and electrolyte (in cm²), τ is the time, ΔEs represents the change in cell voltage during charging (discharging), ΔEt is the alteration in voltage at resting to equilibrium, and L is the thickness of the electrode. Figure 7a illustrates the GITT curves and computed diffusion coefficients for the inaugural charge and discharge of the NCFMO/HC full cell. The Na⁺ diffusion coefficient fluctuates throughout the charging and discharging process with an average value of 6.002 × 10^–11 cm²·s^–1. This variation in the Na⁺ diffusion coefficient might be attributed to the complex kinetics of electrochemical transformations. ⁴³

Zoom In Zoom Out Reset image size

Figure 7b illustrates a typical potential response during the GITT assessment, where a transient current application results in a voltage increment over time. Subsequent interruption of the current triggers an abrupt voltage drops, ascribed to ohmic loss and charge transfer resistance, designated as η_IR . Subsequently, the voltage decay rate diminishes over time, indicative of the attainment of an equilibrium state, which corresponds to the mass transfer limiting process. The voltage shift ensuing from this process is represented as η_D . The irreversible heat is quantified via the following equation: ⁴⁴

$$\begin{eqnarray}\begin{array}{rcl}{q}_{irrev}\left(x,I\right) & = & \displaystyle \frac{I}{V}\eta \left(x,I\right)\\ & = & \displaystyle \frac{I}{V}\left[{\eta }_{IR}\left(x,I\right)+{\eta }_{D}\left(x,I\right)\right]\\ & = & {q}_{IR}\left(x,I\right)+{q}_{D}\left(x,I\right)\end{array}\end{eqnarray} \tag{ 3 }$$

Figures 7c–7d depict the values of η_D and η_IR at diverse states of charge (SOC) and states of discharge (SOD) during the charge/discharge cycles of the NCFMO/HC full cell. During the charging process, η_D and η_IR exhibit positive values, switching to negative during discharge. Initial charging typically results in lower η_D and η_IR values. Charging and discharging at 0.1C, η_D values ascend at the onset of charging and at the termination of discharging, while η_IR undergoes modest alterations. This suggests that η_D contributes more to the overpotential η than η_IR , primarily because the irreversible heat generated due to mass transfer overpotential constitutes the bulk of total irreversible heat. This observation aligns with preceding studies. ⁴⁵

Figure 8 presents the irreversible heat generation of the NCFMO/HC full cell at 0.1C, reaching its zenith at q_D at the commencement of charging. This is attributable to the large Na⁺ concentration in HC at the beginning of charging, leading to elevated diffusion resistance. As charging proceeds, Na⁺ diffusion resistance decreases, causing a gradual reduction in q_D . Throughout the 0.1C charging and discharging process, the cell's q_IR remains relatively stable, which could be ascribed to the constancy in material resistance to the flow of charge carriers during the intermediate stages of charging and discharging. Collectively, the irreversible heat generated during charging surpasses that during discharging.

Zoom In Zoom Out Reset image size

The reversible heat of the battery was quantified by examining the entropy changes during the charge and discharge processes using a variable temperature open-circuit voltage test. ⁴⁶ Initially, the battery was charged at 20 °C using a 0.1C. Following the attainment of specific states of charge (SOC, ranging from 10% to 100%), the variable temperature test was conducted. After reaching equilibrium potential at the designated temperature (20 °C) for a period of 2 h, the temperature was adjusted to the next interval (25 °C), this process was cycled three times. The battery was then fully charged and subsequently discharged to various states of discharge (SOD, ranging from 10% to 100%). The above temperature-time schema was replicated, and the equilibrium potential was documented, enabling the calculation of the entropy coefficient (∂E/∂T) of the battery. The reversible heat could then be computed using the following equation: ²⁶

$$\begin{eqnarray}&&{\rm{\Delta }}S=nF{\left(\displaystyle \frac{\partial E}{\partial T}\right)}_{p}\end{eqnarray} \tag{ 4 }$$

$$\begin{eqnarray}&&{q}_{rev}\left(x,I\right)=\displaystyle \frac{I}{V}\displaystyle \frac{T{\rm{\Delta }}S\left(x\right)}{nF}\end{eqnarray} \tag{ 5 }$$

$$\begin{eqnarray}&&{q}_{IR}\left(x,I\right)=\displaystyle \frac{I}{V}{\eta }_{IR}\left(x,I\right)\end{eqnarray} \tag{ 6 }$$

$$\begin{eqnarray}&&{q}_{D}\left(x,I\right)=\displaystyle \frac{I}{V}{\eta }_{D}\left(x,I\right)\end{eqnarray} \tag{ 7 }$$

Reversible heat is generated due to entropy variations during electrode reactions. An increase in entropy (positive temperature coefficient) causes the system to consume equivalent energy, resulting in an endothermic effect. Conversely, entropy reduction (negative temperature coefficient) incites an exothermic effect. As demonstrated in Fig. 9, during the charging process, reversible heat is lower at minimal SOC, rising towards the culmination of the charging process. During discharge, reversible heat is diminished at low and high SOD, and it escalates in the midway of discharge. The rate of reversible heat generation during discharge exceeds that during charging at an identical current, suggesting a greater influence of reversible heat on the discharging process. Figure 10 illustrates the generation of reversible heat in the NCFMO/HC full cell at 0.1C across different SOC, depicted as color-coded contour plots with positive and negative rates for charging and discharging, respectively. The rate of reversible heat generation escalates with an increase in the rate at a constant SOC. The cells display exothermic effects during both charging and discharging, with a peak value of q_rev observed in the middle of discharge and a minimum value of q_rev detected during the pre-charging period.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table V encapsulates the individual and total heat generation of the cell at 25 °C at varying SOC/SOD at a 0.1C charge/discharge rate, along with the calculated proportion of reversible and irreversible heat generation rates, as visualized in Fig. 11. Herein, the SOC or SOD is designated as the X-axis, while the proportion of each component's heat generation rate is represented on the Y-axis. Notably, the maximal proportion of q_D is approximately 78% when the cell undergoes charge and discharge at a 0.1C. Both the charging and discharging processes encompass instances of reversible heat absorption (negative q_rev ) and heat release reactions (positive q_rev ), whereas the irreversible heat consistently exhibits a heat-release reaction, with q_D accounting for the largest share in irreversible heat generation.

Zoom In Zoom Out Reset image size

Throughout the charging process, irreversible heat generation constitutes approximately 80% of the total heat when SOC is less than 50%, rendering the impact of reversible heat on the total heat relatively insignificant. As the SOC surpasses 50%, the influence of irreversible heat on the total heat intensifies, leading to purely exothermic reactions within the total heat. This rise in temperature necessitates particular attention towards cooling and thermal management during charging.

Contrastingly, during discharge, the reversible heat generation is negative and, when SOD exceeds 50%, the absorbed heat is insufficient to counterbalance the released heat, thus calling for a focus on the latter stages of the cell discharge. The prevailing conclusion is that the irreversible heat generated by ohmic resistance and charge transfer plays a dominant role when the full cell undergoes charge and discharge at lower rates. With reversible heat generation being positive during charging and negative during discharge, this culminates in a higher total heat generation during charging, hence, warranting special attention to thermal safety during charging.