Revisiting the initial irreversible capacity loss of LiNi_0.6Co_0.2Mn_0.2O₂ cathode material batteries

Abstract

Layered LiNi_0.6Co_0.2Mn_0.2O₂ (NCM622) attracts widespread attention primarily due to its potential for high energy density and moderate thermal stability. However, the low initial coulombic efficiency (ICE) of the material limits the maximum utilization of their capacity. The capacity loss in the first cycle occurs under 4.0V and keep almost constant are considered as common characteristics for NCM-based materials. A clear cognition on the initial capacity loss may light the way to improve the practical reversible capacity of NCM622 at 4.0V. Conducting operando X-ray diffraction during galvanostatic charge/discharge cycling at different temperature (25°C, 45°C and 60°C) and different current (0.1C, 0.01C, 1C=120 mA g⁻¹) in the voltage range of 2.7–4.0V, we find that only 8% of the measured initial irreversible capacity loss is associated with parasitic reactions that form cathode/electrolyte interface, and that the dominant contributors include the slow Li⁺ diffusion kinetics (∼46% contribution) and irreversible O3/H1-3 phase transition (∼46% contribution). This semi-quantitative study provides new insight on initial capacity loss, guiding further targeted modification and fully utilization of NCM622.

Introduction

Layered transition metal oxides, such as LiCoO₂, are the predominant cathode materials for Li-ion batteries (LIBs) that dominate portable applications, electric vehicles (EVs) and stationary energy storage, because of their high energy density and considerably long cycling life [1], [2], [3], [4]. Ni-rich layered cathodes have drawn great attention for application in automobile batteries, primarily driven by the relatively lower cost of nickel than cobalt and the potential unlock more reversible capacity by raising their upper cutoff potential (up to 4.5 V vs. Li⁺/Li). LiNi_0.6Co_0.2Mn_0.2O₂ (NCM622), LiNi_0.8Co_0.1Mn_0.1O₂ (NCM811), LiNi_0.8Co_0.15Al_0.05O₂ (NCA) and even Co-free layered cathodes are the current hotspots [[5], [6], [7], [8], [9]]. Researchers are committed to addressing the severe challenges to improve their production and application, including a sensitivity to the environment, great reactivity with the electrolyte, fatal thermal abuse, and these challenges become more serious with Ni content increasing [[10], [11], [12], [13], [14]]. Besides the effort mentioned above, researchers also try to increase the charge voltage for the layered cathode materials with medium high Ni content to obtain higher specific energy density. For example, 4.6V-NCM622 (NCM622 charged to 4.6V vs. Li⁺/Li) can deliver 10% higher theoretical specific energy density than 4.3V-NCM811, while the challenges need to be addressed are similar or less severe than that of 4.3V-NCM811 (see Fig. S1 for a comparison of specific energy density at the material level). Given that the capacity loss during the first cycle is around 10–30% [15,16], another opportunity to significantly increase the energy density of these cathodes can be achievable by mitigating the capacity loss in the first cycle [15].

The reported explanations for this initial loss are as follows: (1) parasitic electrochemical reactions occurring on the surface of the cathode materials with electrolyte; [17], [18], [19], [20] (2) loss of Li-site due to irreversible structural changes; [21], [22], [23] (3) slow kinetics for lithium intercalation [15,24,25]. It is widely accepted that the parasitic side reactions and phase transitions are absolutely irreversible, while the capacity loss due to Li⁺ diffusion kinetics is conditional, varying with conditioning temperature and applied current density. In fact, the large capacity loss in the first cycle was once reported by Chen et al., it mainly occurs under 3.8V and kept almost constant when the charging potential was below 4.0V (vs. Li⁺/Li). In addition, this loss was observed only during the first cycle and the subsequent cycles achieved close to 100% coulombic efficiency [25]. This indicated that the parasitic reactions on the interface between cathode and electrolyte might not the main contributor for the initial capacity loss at low potential range. Zhou et al. recently showed that the slow lithium kinetics at high lithium contents was highly related with the initial cycle capacity loss of NCM 811, as Li⁺ diffusion coefficient dropped nearly by 2, 3 orders of magnitude from Li_0.7Ni_0.8Co_0.1Mn_0.1O₂ to LiNi_0.8Co_0.1Mn_0.1O₂ during lithiation [15,16]. Increasing the temperature from 25°C to 45°C to accelerate Li⁺ diffusion, the capacity loss of NCM811 decreased from 12.1 to 8.4 mAh g⁻¹. Decreasing C-rates also resulted in less charge-discharge asymmetry with better Li⁺ intercalation kinetics and higher initial coulombic efficiency [26]. To clarify the onset potential of parasitic electrochemical reaction, Robert et al. studied operando differential electrochemical mass spectrometry (DEMS) and revealed that CO₂ gas is not continuously released from the very beginning of charging. CO₂ generation started at ∼4V (vs. Li⁺/Li) in the first charging, which corresponding to the oxidative decomposition of the carbonate electrolyte [27]. Meanwhile, Zeng et al. used a home-built high-precision leakage current measuring system and also clearly stated that no obvious electrochemical reaction and chemical reaction occurred when the cut-off potential was lower than 4.0V (vs. Li⁺/Li) [28]. Hence, the capacity loss caused by parasitic electrochemical reaction below 4V (vs. Li⁺/Li) can be ignored. It is of great significance to investigate the causes of initial capacity loss in the low voltage (≤4V, vs. Li⁺/Li), and the investigation may guide the upside space for layered cathode material with higher specific capacity.

Commercial NCM622 attracts many attentions due to the highest energy density, good rate performance [24,29] and moderate cost among the layered oxide cathode materials when the cut-off potential is raised to 4.6 V (vs. Li⁺/Li) or above. In this work, we choose commercial LiNi_0.6Mn_0.2Co_0.2O₂ (NCM622) in the voltage range of 2.7–4.0V (vs. Li⁺/Li) as a model system. The structural evolution of NCM622 at different temperature (25°C, 45°C and 60°C) and different rate (0.1C, 0.01C, 1C=120 mA g⁻¹) are investigated using operando X-ray diffraction. Combining these results, the correlation between structural evolutions and Li⁺ diffusion kinetics are revealed. Our results clearly demonstrate that 46% of the initial capacity loss of NCM622 is due to the slow Li⁺ diffusion kinetics, another 46% of the capacity loss is caused by the irreversible O3/H1-3 phase transition, and only 8% is attributed to the surface changes in the material and/or CEI formation during the first charging.