I have long been interested in how the atomic-scale arrangement of transition metals governs the macroscopic behavior of cathode materials in a Li-ion battery. The layered oxide family LiMO2 remains the most important class of positive electrodes for commercial Li-ion battery applications because it combines high operating voltage, decent reversible capacity, and a relatively mature synthesis route. Among these oxides, LiCoO2 (LCO) was the first commercially successful cathode, yet its practical capacity is limited to roughly half of the theoretical value because deep delithiation triggers oxygen release, cobalt dissolution, and irreversible phase transformations. The scarcity, cost, and toxicity of cobalt further motivate the search for ternary Li(Ni,Co,Mn)O2 (NCM) systems in which Ni, Co, and Mn act synergistically. In this work, I used first-principles density functional theory to compare four compositions: LCO, LiNi0.6Co0.2Mn0.2O2 (NCM622), LiCo0.6Mn0.2Ni0.2O2 (CMN622), and LiMn0.6Ni0.2Co0.2O2 (MNC622). The central question I asked was how the relative fractions of Ni, Co, and Mn alter structural stability, equilibrium voltage, energy density, electronic conductivity, charge compensation, and Li+ migration in a Li-ion battery cathode. My calculations reveal clear trade-offs: Co raises the average voltage, Mn stabilizes the lattice and widens the Li+ diffusion channel, and Ni provides a balanced contribution to capacity and electronic transport. These findings provide a systematic basis for designing next-generation Li-ion battery cathodes with improved durability and rate capability.
Computational framework and structural models
I carried out all calculations within the framework of density functional theory (DFT) using the projector augmented wave (PAW) method. The exchange-correlation functional was the Perdew-Burke-Ernzerhof (PBE) form of the generalized gradient approximation (GGA). Because transition-metal d electrons are strongly correlated, I applied a Hubbard U correction to the Ni, Co, and Mn d orbitals. To keep the comparison internally consistent, I used a uniform value of U = 3.5 eV for all three transition metals. This choice avoids the ambiguity that would arise if different U values were assigned to different elements, and it has been shown to reproduce the non-magnetic ground state of LCO while still capturing the essential redox behavior of NCM compounds. I also included Grimme D3 dispersion corrections to describe the weak interlayer van der Waals interactions that are important in layered oxides. The plane-wave cutoff energy was set to 500 eV, and the Brillouin zone was sampled with a 2 × 2 × 2 Monkhorst-Pack k-point mesh centered at the Γ point. During structural optimization, the energy convergence criterion was 10−4 eV per atom, and the Hellmann-Feynman force threshold was 0.5 eV nm−1. For Li+ migration barriers, I used the climbing-image nudged elastic band (CI-NEB) method.
The four materials all adopt the hexagonal α-NaFeO2 structure with the R3m space group. In this structure, edge-sharing octahedra form layers of transition-metal cations in the 3a Wyckoff sites and Li+ in the 3b sites, separated by oxygen layers. The resulting stacking sequence can be written as [Li]3b[TM]3a[O2]. For LCO, I used the primitive cell. For the three ternary compositions, I constructed 3 × 3 × 1 supercells containing 108 atoms so that the transition-metal ratios could be represented explicitly without artificial ordering. The transition-metal ratio was kept at 6:2:2 in all ternary cases, but the identity of the majority cation was changed: Ni-rich NCM622, Co-rich CMN622, and Mn-rich MNC622. This design allows me to isolate the effect of the dominant transition metal while holding the overall stoichiometry constant.
Lattice parameters, volume changes, and structural stability
One of the most important degradation mechanisms in a Li-ion battery cathode is the mechanical stress that arises when the unit cell expands and contracts during cycling. Repeated volume changes can generate microcracks, expose fresh surfaces, and accelerate capacity fade. I therefore calculated the lattice constants a and c, the cell volume V, and the volume change ΔV between the fully lithiated state (x = 1 in LixMO2) and the fully delithiated state (x = 0). The volume change is defined as
$$\Delta V = \frac{V_{\text{delithiated}} – V_{\text{lithiated}}}{V_{\text{lithiated}}} \times 100\%$$
Table 1 summarizes the structural parameters. The in-plane lattice constant a changes only slightly upon delithiation for all four materials, which reflects the strong and relatively rigid transition-metal–oxygen bonding within the layers. In contrast, the c parameter and the cell volume change more substantially because Li removal directly affects the interlayer spacing and the screening of the oxygen layers. The calculated volume changes follow the order NCM622 > CMN622 > LCO > MNC622. NCM622 exhibits the largest volume change, 7.18%, which suggests that it is the most prone to particle cracking and irreversible structural damage. MNC622 shows the smallest volume change, 4.19%, indicating that a high Mn content helps to preserve the layered framework during cycling. This improved structural stability is consistent with the role of Mn4+, which is redox-inactive in the normal operating voltage window and therefore does not drive the same magnitude of local bond distortion as Ni or Co redox centers. The Co-rich CMN622 and LCO fall in between, with volume changes of 5.33% and 4.45%, respectively.
| Material | State | a / nm | c / nm | V / nm3 | ΔV / % |
|---|---|---|---|---|---|
| LCO | Lithiated (x = 1) | 2.829 × 10−1 | 1.376 | 9.541 × 10−2 | — |
| LCO | Delithiated (x = 0) | 2.829 × 10−1 | 1.315 | 9.116 × 10−2 | 4.45 |
| CMN622 | Lithiated (x = 1) | 2.844 × 10−1 | 1.398 | 9.793 × 10−2 | — |
| CMN622 | Delithiated (x = 0) | 2.824 × 10−1 | 1.342 | 9.271 × 10−2 | 5.33 |
| NCM622 | Lithiated (x = 1) | 2.873 × 10−1 | 1.399 | 1.000 × 10−1 | — |
| NCM622 | Delithiated (x = 0) | 2.830 × 10−1 | 1.339 | 9.283 × 10−2 | 7.18 |
| MNC622 | Lithiated (x = 1) | 2.853 × 10−1 | 1.408 | 9.926 × 10−2 | — |
| MNC622 | Delithiated (x = 0) | 2.844 × 10−1 | 1.358 | 9.510 × 10−2 | 4.19 |
Theoretical capacity, equilibrium voltage, and energy density
For a Li-ion battery cathode, the theoretical specific capacity determines the upper bound of charge storage per unit mass. I calculated the theoretical capacity using
$$C = \frac{n \times F}{3.6 \times M}$$
where C is the specific capacity in mAh g−1, n is the number of Li+ ions participating in the redox reaction, F is the Faraday constant (96 485 C mol−1), and M is the molar mass of the material in g mol−1. The factor 3.6 converts coulombs and grams into mAh g−1. The equilibrium voltage of a Li-ion battery cathode was obtained from the difference in lithium chemical potential between the cathode and a lithium metal anode:
$$V_{\text{cell}} = -\frac{\mu_{\text{Li}}^{\text{cathode}} – \mu_{\text{Li}}^{\text{anode}}}{(x_1 – x_2)F}$$
Equivalently, the voltage can be expressed through the change in Gibbs free energy:
$$V_{\text{cell}} = -\frac{\Delta G}{(x_1 – x_2)F}$$
with
$$\Delta G = \Delta E + p\Delta V – T\Delta S$$
Because the pΔV and TΔS terms are much smaller than the internal energy change ΔE for solid-state redox reactions, I approximated ΔG by ΔE. For LCO, for example, the voltage for the reaction Lix1CoO2 → Lix2CoO2 + (x1 − x2)Li can be written as
$$V_{\text{cell}} = -\frac{E_{\text{Li}_{x_1}\text{CoO}_2} – E_{\text{Li}_{x_2}\text{CoO}_2} – E_{\text{Li}}}{(x_1 – x_2)F}$$
Once the capacity and voltage are known, the gravimetric energy density is
$$\varepsilon = C \times V_{\text{cell}} = \frac{\Delta E}{3.6 \times M}$$
Table 2 lists the theoretical capacity, equilibrium voltage, and energy density for the four materials. MNC622 has the highest theoretical capacity, 280.301 mAh g−1, which is about 2.4% higher than that of LCO. NCM622 and CMN622 have capacities of 276.684 and 276.389 mAh g−1, respectively, both slightly above LCO. These differences are modest because all four materials have similar molar masses and all are assumed to exchange one Li+ per formula unit in the ideal limit. The more striking differences appear in the voltage. LCO exhibits the highest equilibrium voltage, 3.35 V, because the Co3+/Co4+ redox couple has a high oxidation potential. CMN622 follows with 3.24 V, while NCM622 has 3.18 V. MNC622 has the lowest voltage, 2.95 V, which reflects the high Mn content and the associated Mn3+/Mn4+ redox chemistry, including possible Jahn-Teller distortions that lower the average operating potential. When the voltage and capacity are combined, LCO has the highest energy density, 918.00 mWh g−1. CMN622, NCM622, and MNC622 have energy densities of 894.86, 878.57, and 827.02 mWh g−1, respectively. The MNC622 case is particularly instructive: it has the highest capacity but the lowest energy density because its voltage is substantially lower. This highlights a central design principle for Li-ion battery cathodes: energy density is not determined by capacity alone, and a high-capacity material can still underperform if its redox potential is low.
| Material | Capacity C / mAh g−1 | Equilibrium voltage / V | Energy density ε / mWh g−1 |
|---|---|---|---|
| LCO | 273.763 | 3.35 | 918.00 |
| CMN622 | 276.389 | 3.24 | 894.86 |
| NCM622 | 276.684 | 3.18 | 878.57 |
| MNC622 | 280.301 | 2.95 | 827.02 |
The voltage values calculated with PBE+D3 are typically underestimated relative to experiment by about 0.6 V. After applying this correction, the calculated voltages for NCM622 and LCO move close to the experimental ranges, which supports the reliability of my structural models and computational setup. The ordering of the voltages, however, remains unchanged. Co-rich compositions favor higher average voltage, Ni-rich compositions provide intermediate values, and Mn-rich compositions lower the voltage. This ordering is consistent with the electronic structure of the transition-metal d bands, as discussed below.
Electronic structure and conductivity
The projected density of states (PDOS) provides a direct link between the electronic structure of a Li-ion battery cathode and its electrochemical behavior. For LCO, the Co 3d states dominate the top of the valence band, with a significant contribution from O 2p states. The strong Co 3d–O 2p hybridization indicates that oxygen participates in charge compensation during delithiation, which is one reason why LCO can suffer from oxygen release at high voltage. The Li s states are distributed at higher energies and are associated with the Li+ insertion and extraction process. In contrast, the three ternary materials show more complex PDOS patterns because Ni 3d, Co 3d, and Mn 3d states all contribute near the Fermi level. All three ternary compositions exhibit metallic-like conductivity, whereas LCO is better described as a semiconductor in the calculated ground state. This difference is important for Li-ion battery rate capability because a cathode with higher electronic conductivity can support faster charge transfer and reduce polarization.
Among the ternary materials, NCM622 has the lowest calculated Fermi level, 5.162 eV, compared with 6.961 eV for CMN622 and 6.512 eV for MNC622. The Ni 3d peak in NCM622 lies closest to the Fermi level, which suggests that Ni redox states are easily accessible and can participate actively in charge compensation. In CMN622 and NCM622, the Ni 3d and Co 3d orbitals hybridize strongly with O 2p states, indicating covalent Ni–O and Co–O interactions. The Mn–O interaction is weaker in these two compositions. In MNC622, however, the high Mn content leads to a much broader Mn 3d–O 2p overlap extending from the valence band to the conduction band. This stronger Mn–O hybridization shifts a large portion of the O 2p states to lower energies, reducing the density of states near the Fermi level and stabilizing the oxygen sublattice. From the perspective of Li-ion battery safety and cycle life, this is a favorable feature because it suppresses oxygen loss and helps maintain the layered structure.
Using Boltzmann transport theory, I estimated the electrical conductivity at 300 K. At the Fermi level, MNC622 has a conductivity of about 4 × 103 S m−1, which is higher than the approximately 5 × 102 S m−1 calculated for NCM622. When the chemical potential shifts by −1 eV away from the Fermi level, the trend reverses: NCM622 reaches about 3 × 103 S m−1, while MNC622 drops to about 8 × 102 S m−1. This indicates that the electronic transport properties depend sensitively on the position of the Fermi level and the detailed d-band filling. For Li-ion battery operation, the relevant energy window is near the Fermi level during redox, so the high conductivity of MNC622 at the Fermi level may partially compensate for its lower voltage by improving rate performance. NCM622, on the other hand, maintains a higher conductivity when the chemical potential is shifted, which could be beneficial under certain charging states. Overall, the electronic structure results show that Ni and Co provide more easily accessible redox states near the Fermi level, while Mn contributes to oxygen stabilization and broad band overlap.
| Material | Fermi level / eV | Conductivity at Fermi level / S m−1 | Conductivity at −1 eV shift / S m−1 |
|---|---|---|---|
| NCM622 | 5.162 | ~5 × 102 | ~3 × 103 |
| CMN622 | 6.961 | — | — |
| MNC622 | 6.512 | ~4 × 103 | ~8 × 102 |
Charge compensation and oxygen stability
To understand how each transition metal contributes to redox activity, I performed Bader charge analysis at different lithium contents, x = 1.0, 0.5, and 0.0. Table 3 summarizes the results. In LCO, the Co Bader charge increases from +1.20 at x = 1.0 to +1.42 at x = 0.0, corresponding to the Co3+/Co4+ oxidation process. At the same time, the oxygen Bader charge changes from −1.05 to −0.71, which means that a significant amount of electron density is removed from oxygen. This is consistent with the known tendency of LCO to release oxygen at high states of charge. In a Li-ion battery, such oxygen loss can lead to electrolyte oxidation, impedance growth, and irreversible capacity fade.
The ternary materials show a more distributed charge compensation mechanism. In CMN622, Co is the main redox-active species: its Bader charge increases by 0.19 from x = 1.0 to x = 0.0, while Mn and Ni increase by 0.13 each. In NCM622 and MNC622, Ni becomes the dominant redox center, with charge increases of 0.17 and 0.24, respectively. The Mn Bader charge remains close to +1.80 at x = 0.0 in all three ternary compositions, which indicates that Mn stays predominantly in the +4 oxidation state and does not participate strongly in redox. This is beneficial because Mn4+ is stable and does not induce Jahn-Teller distortion. The oxygen Bader charge at x = 0.0 is most negative in MNC622, −0.80, compared with −0.74 in CMN622 and −0.73 in NCM622. A more negative oxygen charge means that the oxygen sublattice retains more electron density, which enhances its stability and suppresses oxygen release. This result supports the experimental observation that Mn-rich compositions can improve the thermal stability and cycling safety of Li-ion battery cathodes.
| Material | x | Li | Co | Mn | Ni | O |
|---|---|---|---|---|---|---|
| LCO | 1.0 | +0.89 | +1.20 | — | — | −1.05 |
| LCO | 0.5 | +0.90 | +1.33 | — | — | −0.90 |
| LCO | 0.0 | — | +1.42 | — | — | −0.71 |
| CMN622 | 1.0 | +0.89 | +1.22 | +1.66 | +1.19 | −1.10 |
| CMN622 | 0.5 | +0.90 | +1.31 | +1.75 | +1.24 | −0.93 |
| CMN622 | 0.0 | — | +1.41 | +1.79 | +1.32 | −0.74 |
| NCM622 | 1.0 | +0.89 | +1.26 | +1.74 | +1.16 | −1.10 |
| NCM622 | 0.5 | +0.90 | +1.28 | +1.77 | +1.27 | −0.92 |
| NCM622 | 0.0 | — | +1.41 | +1.80 | +1.33 | −0.73 |
| MNC622 | 1.0 | +0.89 | +1.19 | +1.62 | +1.09 | −1.15 |
| MNC622 | 0.5 | +0.90 | +1.27 | +1.76 | +1.18 | −0.99 |
| MNC622 | 0.0 | — | +1.39 | +1.80 | +1.33 | −0.80 |
Li-ion migration kinetics
The rate capability of a Li-ion battery cathode is ultimately limited by how fast Li+ can move through the host lattice. In layered LiMO2, Li+ diffusion occurs primarily through two pathways: the oxygen dumbbell hop (ODH) and the tetrahedral site hop (TSH). At low-to-moderate delithiation, the ODH pathway dominates. I therefore calculated the migration barrier for Li+ along the ODH path using the CI-NEB method. The results show that MNC622 has the lowest migration barrier, followed by NCM622, while CMN622 has the highest barrier. A lower barrier means that Li+ can move more easily, which improves rate performance and reduces polarization during fast charging and discharging. The high Mn content in MNC622 is likely responsible for this improvement because Mn4+ has a larger ionic radius than Ni3+/Ni4+ and Co3+/Co4+, and the longer Mn–O bonds widen the diffusion channel. CMN622, with its high Co content, has more covalent Co–O bonds and a more rigid interlayer spacing, which increases the barrier. NCM622 falls in between. These migration results complement the structural and electronic findings: MNC622 offers the best Li+ transport and the smallest volume change, but its lower voltage limits energy density; CMN622 offers higher voltage but slower Li+ diffusion; NCM622 provides a compromise. For a Li-ion battery intended for high-power applications, the low barrier of MNC622 would be attractive, while for high-energy applications, the higher voltage of Co-rich compositions would be preferred.
| Material | ODH migration barrier trend | Structural stability trend | Electronic conductivity trend |
|---|---|---|---|
| MNC622 | Lowest | Highest | High at Fermi level |
| NCM622 | Intermediate | Lowest | High away from Fermi level |
| CMN622 | Highest | Intermediate | Moderate |
| LCO | — | Intermediate | Semiconducting |
Synergistic effects of Ni, Co, and Mn
The results of my calculations allow me to articulate the distinct and complementary roles of the three transition metals in a Li-ion battery cathode. Co is the voltage booster. Its Co3+/Co4+ redox couple operates at a high potential, which raises the average equilibrium voltage and energy density. Co also helps to suppress cation mixing and structural degradation at high voltage because Co–O bonds are strong and directional. However, Co is expensive, scarce, and toxic, and too much Co reduces the practical capacity because only a fraction of the lithium can be reversibly cycled without structural damage. Ni is the capacity and conductivity contributor. Ni2+/Ni3+/Ni4+ redox reactions can deliver high capacity, and Ni 3d states near the Fermi level improve electronic conductivity. The downside is that Ni2+ has an ionic radius similar to Li+, so Ni2+ can migrate into Li layers and cause Li+/Ni2+ mixing, which impedes Li+ diffusion and accelerates capacity fade. Mn is the structural stabilizer. Mn4+ is redox-inactive in the normal operating window and therefore acts as a pillar that maintains the layered framework. High Mn content reduces the volume change, strengthens the oxygen sublattice, and widens the Li+ diffusion channel. The penalty is a lower average voltage and, if Mn3+ is present, possible Jahn-Teller distortion. The optimal Li-ion battery cathode must balance these three effects. My calculations show that no single composition is best in every metric. LCO has the highest voltage and energy density but suffers from oxygen instability and high cost. NCM622 has balanced properties but the largest volume change. CMN622 has high voltage but the highest Li+ migration barrier. MNC622 has the best structural stability and Li+ kinetics but the lowest voltage. The synergy arises because Ni, Co, and Mn occupy the same transition-metal sublattice and influence each other through bonding, charge transfer, and band structure. For example, the presence of Mn reduces the average oxidation state of Ni and Co, which can suppress oxygen release. The presence of Co raises the voltage of the Ni redox couple. The presence of Ni provides electronic states that make the material more conductive than LCO. Therefore, the outstanding electrochemical performance of ternary NCM cathodes is not simply the sum of the individual properties of Ni, Co, and Mn; it is an emergent property of their interaction.
| Element | Primary role | Benefit | Limitation |
|---|---|---|---|
| Ni | Capacity and electronic conductivity | High reversible capacity, good rate capability | Li+/Ni2+ mixing, surface reactivity |
| Co | Voltage and structural suppression | High equilibrium voltage, reduced cation mixing | High cost, toxicity, limited practical capacity |
| Mn | Structural stabilization and Li+ transport | Small volume change, stable oxygen lattice, low migration barrier | Lower average voltage, possible Mn3+ distortion |
Implications for Li-ion battery design
My findings have several practical implications for the design of advanced Li-ion battery cathodes. First, if the target is maximum energy density for electric vehicles, Co-rich compositions such as LCO or CMN622 remain attractive because of their high voltage. However, the oxygen instability of LCO and the high cost of Co must be addressed through coating, doping, or electrolyte engineering. Second, if the target is long cycle life and high power, Mn-rich compositions such as MNC622 are promising because of their small volume change, stable oxygen lattice, and low Li+ migration barrier. The lower voltage can be partially compensated by using a higher cutoff voltage or by pairing the cathode with a higher-voltage anode, although care must be taken to avoid electrolyte decomposition. Third, Ni-rich compositions such as NCM622 offer a balanced compromise and are already widely used in commercial Li-ion battery cells. However, the large volume change I calculated for NCM622 suggests that single-crystal particles or concentration-gradient designs may be needed to prevent microcracking. Fourth, the synergistic effects I identified suggest that compositionally graded particles, in which the core is Mn-rich for stability and the shell is Ni-rich for capacity, could outperform uniform compositions. Such gradient designs are already being explored experimentally, and my results provide a theoretical rationale for their success. Finally, the electronic structure results show that the Fermi level and the density of states near the Fermi level are sensitive to the transition-metal ratio. This means that small changes in composition can produce large changes in conductivity and redox potential. Therefore, computational screening based on DFT can accelerate the discovery of optimal Li-ion battery cathode compositions by predicting voltage, capacity, volume change, and migration barriers before synthesis.
Limitations and outlook
I should note several limitations of my study. First, the DFT+U method with a uniform U value for Ni, Co, and Mn is a simplification. In reality, the optimal U values may differ slightly for different elements and oxidation states. However, using a uniform U ensures that the comparison across compositions is consistent. Second, my calculations are based on ideal bulk structures without surfaces, grain boundaries, or electrolyte interfaces. In a real Li-ion battery, interfacial reactions, surface reconstruction, and electrolyte decomposition can significantly affect performance. Third, the supercell models use periodic boundary conditions and implicit ordering of transition metals. While I used 3 × 3 × 1 supercells to approximate random mixing, short-range ordering effects may still influence the results. Fourth, the theoretical capacity assumes complete extraction of one Li+ per formula unit, which is not always achievable in practice because of structural and electrolyte limitations. Despite these limitations, the trends I obtained are robust and provide a useful framework for understanding the role of transition-metal ratios in Li-ion battery cathodes. Future work could combine DFT with machine learning potentials or cluster expansion methods to explore a wider compositional space, including quaternary and doped systems. In situ or operando experiments could also validate the predicted voltage, volume change, and migration barriers. The ultimate goal is to design a Li-ion battery cathode that simultaneously achieves high energy density, high power density, long cycle life, and low cost. My results suggest that this goal is best approached by optimizing the synergy among Ni, Co, and Mn rather than by maximizing any single element.
Conclusion
I have presented a systematic first-principles investigation of the effect of transition-metal ratio on the electrochemical properties of Li-ion battery ternary cathode materials. By comparing LCO, NCM622, CMN622, and MNC622 under identical computational conditions, I found that LCO possesses the highest equilibrium voltage (3.35 V) and energy density (918.00 mWh g−1), but it suffers from the largest oxygen instability and the highest cobalt cost. MNC622 exhibits the smallest volume change (4.19%), the most stable oxygen sublattice, and the lowest Li+ migration barrier, which are favorable for cycling stability and rate capability, but its equilibrium voltage is the lowest (2.95 V), limiting energy density. NCM622 provides a balanced set of properties, with intermediate voltage, capacity, and electronic conductivity, although it shows the largest volume change (7.18%). CMN622 combines a relatively high voltage (3.24 V) with moderate stability, but its Li+ migration barrier is the highest. The Bader charge analysis reveals that Co dominates charge compensation in CMN622, while Ni dominates in NCM622 and MNC622. Mn remains close to +4 and stabilizes the lattice. The electronic structure calculations show that Ni and Co 3d states hybridize strongly with O 2p near the Fermi level, while Mn–O hybridization shifts oxygen states to lower energies, enhancing oxygen stability. Taken together, these results demonstrate that the excellent performance of ternary Li-ion battery cathodes arises from the synergistic interaction of Ni, Co, and Mn. Co increases the voltage, Ni provides capacity and conductivity, and Mn stabilizes the structure and facilitates Li+ transport. Optimizing this synergy is the key to designing next-generation Li-ion battery cathodes with high energy density, long cycle life, and improved safety.