Atomic-scale understanding of non-stoichiometry effects on the electrochemical performance of Ni-rich cathode materials
Graphical abstract
Introduction
Profiting from their higher energy density, Lithium-ion batteries (LIBs) have been often regarded as one of the most important energy storage systems to be used in cellular phones, personal computers, digital cameras, energy storage grids, hybrid or full electric vehicles, etc. As the earliest and most widely commercialized cathode material for LIBs, LiCoO2 has dominated the market for the last two decades. However, due to its low reversible capacity of 140 mAh g−1 and the high cost of Cobalt mining, LiCoO2 cannot meet the rapidly growing market of LIBs and the increasing demand for higher energy and power densities in electronic devices and vehicles. For these reasons, Cobalt-free cathode materials have been actively explored, with Ni-rich layered oxides, noted as LiNiO2-based oxides, being often regarded as the next generation cathode materials [[1], [2], [3], [4]]. These oxides potentially offer a specific capacity as high as 200 mAh g−1, and an energy density over 800 Wh kg−1. Moreover, they are also promising candidates for ultrahigh-rate cathode materials, because of their small band gap of about 0.5 eV (electrical conductivity of 10−1 S cm−1) and a low Li ion diffusion barrier below 0.4 eV [[5], [6], [7], [8]].
In spite of the aforementioned advantageous properties of Ni-rich oxides, synthesizing stoichiometric phases has proven to be extremely challenging. Extra Ni ions occupying Li sites are widely observed, giving rise to the formation of non-stoichiometric Li1-xNi1+xO2 (x = 0.02–0.1) [[1], [2], [3]]. The extra Ni defect concentration normally ranges from 2% to 10%, depending on the synthesis temperature, oxygen pressure, oxygen flow rate, precursor ratios, etc. [[9], [10], [11]] Density-functional theory (DFT) calculations showed that extra-Ni-defects in Ni-rich oxides have small formation energies, between 0.5 and 0.8 eV, as reported by different groups [[12], [13], [14]]. From simulated defect concentration vs. temperature diagrams, Koyama et al. have suggested that those extra Ni ion defects are unavoidable during synthesis [12]. The low formation energy of such defects was ascribed to the similar ionic radii of Li+ (0.76 Å) and Ni2+ (0.69 Å). More recently, Chen et al. have proposed that extra-Ni-defects could be stabilized by 180° Ni-O-Ni superexchange interactions with nearby Ni ions [13].
The non-stoichiometry character of Ni-rich oxides shows detrimental effects on the cathode performance. The appearance of extra-Ni-defects in the Li layer reduces the theoretical capacity limit, since these Ni ions are electrochemically inactive and occupy reversible Li sites [15]. A high concentration of extra-Ni-defects also leads to more capacity loss during the first cycle of charge-discharge [11]. Besides, it also worsens the rate performance, although a unified understanding of this effect has not yet been obtained by the scientific community. Delmas et al. suggested that extra Ni ions in the Li layer can be oxidized to a trivalent state, leading to a local collapse of the inter-slab space and inhibiting Li re-intercalation [16]. Fu et al. attributed the reason to the blocking effect of extra Ni ions on the Li diffusion, observing how the rate performance of Ni-rich LiNiCoMnO2 (NCM) materials was improved after decreasing the amount of extra-Ni-defects [11]. On the other hand, theoretical work by Kang et al. showed that the effect of extra-Ni-defects is not localized. By lowering the overall Li slab distance, Li hopping is impeded due to the increase of the diffusion barrier. They thought that the electrostatic repulsion between Li and Ni ions in the Li layer might not play a determining role [8].
Regardless of the negative effects of non-stoichiometry on the theoretical capacity limit and rate performance of Ni-rich oxides, the beneficial aspect of helping to maintain their layered structure, thus obtaining better capacity retention, has recently attracted considerable attention, and has been used in practice to stabilize the cathode material [1]. Extra-Ni-defect is regarded as a “pillar” to stabilize the layered structure by suppressing undesirable phase reactions or preventing Ni interlayer migration during charge/discharge cycles [[17], [18], [19], [20]]. Cho et al. took advantage of such effect and developed a “pillar”-rich coating layer to stabilize Ni-rich layered oxides [21]. In a recent work by the same group, a Ni “pillar” was used to improve the structural stability of LiCoO2 by screening interlayer oxygen repulsion [22]. The “pillar effect” is based on the immobile nature of the extra-Ni-defects that are pinned to the Li layer during electrochemical cycling. However, this assumption is not consistent with other experimental observations that Ni ions show a high mobility within the Li layer and could then diffuse to surface regions or form NiO-like domain clusters. [15,23] There are no extensive kinetic studies on the mobility of such extra-Ni-defects, and the “pillar effect” remains poorly understood.
In order to improve the electrochemical performance, practically used LiNiO2 compounds are always doped (or forming alloys if the dopant concentration is larger than 10%) with one or multiple elements, resulting in the formation of Li[Ni1-yMy]O2 (y = 0.1–0.3, M = Mn, Al, Fe, etc.) compounds. The dopants have revealed a significant influence on the defect chemistry. A recent theoretical work by Hoang and Johannes has compared LiNi1/3Co1/3Mn1/3O2 and LiNi1/3Co1/3Al1/3O2 cathode materials, for which the defect formation energy difference can vary between 0.3 and 0.9 eV [19]. Our recent work on a series of doped Ni-rich NCM materials also shows that the defect formation energy sensitively depends on the doping species [24]. It was experimentally reported that the introduction of Co and Mn or Ti could inhibit and promote Ni occupation at Li sites, respectively [[25], [26], [27], [28], [29]]. Also, a Mg doping concentration below 20% resulted in improved stoichiometry, while a concentration higher than 20% led to a deteriorated stoichiometry [30]. However, due to the lack of systematic studies on doped Ni-rich layered oxides, the fundamental role of doping species and the corresponding rational design strategies to control the stoichiometry of LiNiO2 oxides through doping methods have not yet been well outlined.
In order to address the aforementioned controversies and questions, in this work we performed DFT calculations to conduct a systematic study at the atomic scale on the effects of non-stoichiometry (extra-Ni-defects) on Ni-rich layered oxides. The thermodynamic, kinetic and electronic features of extra-Ni-defects, as well as their correlations with defect formation, “pillar effect”, rate performance and capacity loss of Ni-rich oxides are investigated in detail. A wide range of dopants is also systematically investigated, from which the relation between dopant valence state and stoichiometry is discovered and explained. This work not only provides an in-depth understanding of non-stoichiometry in Ni-rich oxides, but also helps to rationally design optimized cathode materials for LIBs. Non-stoichiometry effects have also been observed in other important cathode materials, such as olivine LiMnSiO4, phosphate LiFePO4 or spinel LiMn2O4, and the results of current study will also provide a basis to understand them.
Section snippets
Computational methods
In this work, the Vienna ab initio simulation package (VASP) and the projector-augmented wave (PAW) method are used to perform all DFT calculations [[31], [32], [33]]. Exchange and correlation interactions are described by the generalized gradient approximation of Perdew-Burke-Ernzerhof (GGA-PBE) [34]. The plane wave cutoff energy is 500 eV and the structure optimization is performed with the criteria that the equilibrium force on each atom is less than 0.05 eV Å−1 and the total energy is
Stoichiometric LiNiO2
Before addressing the particularities induced by extra-Ni-defects in LiNiO2, we firstly determined the atomic and electronic structures of stoichiometric LiNiO2, as a benchmark of our study. In stoichiometric LiNiO2, Li and Ni ions occupy octahedral sites bound to six neighboring oxygen atoms, with Ni ions oxidized in a 3 + valence state. According to Crystal Field Theory, the five 3d orbitals of Ni3+ at octahedral site will split into 2-fold degenerate d (x2-y2) and d (z2-r2) orbitals
Conclusions
In this work, DFT calculations were used to systematically investigate the effect of non-stoichiometry (extra-Ni-defects) on the electrochemical performance of Ni-rich layered oxides. Our results show that extra-Ni-defects trigger a charge disproportionation reaction that reduces the amount of Jahn-Teller distortion of Ni3+, lowering the defect formation energy and, thus, increasing the extent of the low stoichiometry degree during synthesis. The extra-Ni-defects are immobile in nature due to
Acknowledgment
The authors acknowledge the Texas Advanced Computing Center (TACC) for providing computational resources. This work was supported by the International Energy Joint R & D Program (No. 20168510011350) of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant funded by the Ministry of Knowledge Economy, Korean government. This work is also supported by the L&F Co.’s World Class 300 Project of the Korea Institute of Advancement of Technology (KIAT) funded by the Ministry of
References (63)
- et al.
Energy Stor. Mater.
(2017) - et al.
Solid State Ionics
(1999) - et al.
Solid State Ionics
(2002) - et al.
Solid State Ionics
(2001) - et al.
Electrochim. Acta
(1999) - et al.
J. Power Sources
(1997) - et al.
Solid State Ionics
(1995) - et al.
J. Power Sources
(2017) - et al.
Solid State Ionics
(1996) - et al.
J. Power Sources
(2013)
J. Power Sources
Comput. Mater. Sci.
Comput. Mater. Sci.
Solid State Ionics
Angew. Chem. Int. Ed.
Sci. Technol. Adv. Mater.
J. Mater. Chem.
J. Phys. Chem. C
Phys. Rev. B
ACS Appl. Mater. Interfaces
Chem. Mater.
J. Mater. Chem.
J. Mater. Chem.
Phys. Chem. Chem. Phys.
Chem. Mater.
Chem. Mater.
Nano Lett.
Adv. Mater.
Nano Lett.
J. Mater. Chem.
Adv. Energy Mater.
Cited by (20)
Unified understanding and mitigation of detrimental phase transition in cobalt-free LiNiO<inf>2</inf>
2024, Energy Storage MaterialsElectrospinning-assisted construction of 3D LiFePO<inf>4</inf>@rGO/carbon nanofibers as flexible cathode to boost the rate capabilities of lithium-ion batteries
2023, Ceramics InternationalCitation Excerpt :Moreover, for cathode materials, the structure and components of the material, especially the active material, will play a decisive role in the capacity and cycle life of the material. Previous works have systematically studied the physicochemical properties of numerous cathode materials such as LiFePO4 [3,4], LiCoO2 [5,6], LiNiO2 [7], and LiMn2O4 [8–10], mainly involving crystal phase, microstructure, and energy storage mechanism. Among them, a high theoretical capacity of about 170 mAh g−1 for LiFePO4 makes it become a promising cathode material and has widespread applications.
Elevated stability of nickel-rich oxide cathode material with concentration gradient of transition metals via a novel size-controllable calcination method
2022, Journal of Alloys and CompoundsCitation Excerpt :In recent years, the research of nickel-rich layered lithium transition-metal oxide cathode material with high capacity has attracted extensive attention, considering the energy density of lithium-ion battery can be improved obviously by increasing the nickel content of the LiNi1−x-yCoxMnyO2 cathode material [1–6]. Scientists and engineers are stepping up research and development of layered lithium transition-metal oxide material with Ni content higher than 80%, however, with the increase of Ni content, the cyclic stability and thermal stability of the material decrease [7–11]. On one hand, the enrichment of high valence of nickel ion Ni4+ on the surface of the active material intensify the side reaction between the highly active nickel ions and the carbonate electrolyte [12–14].
Experimental and mechanism research of gradient structured LiNi<inf>0.8</inf>Co<inf>0.1</inf>Mn<inf>0.1</inf>O<inf>2</inf> cathode material for Li-ion batteries
2020, Solid State IonicsCitation Excerpt :In comparison, the peak positions in XRD patterns of cycled S7.5 agree well with it before cycling. It is evident that the crystal structure changes less and the structure is more ordered, which is conducive to Li+ insertion/extraction [31]. In contrast, the structure of cycled S0 material is damaged greatly, some additional diffraction peaks appear in the patterns.