Energy infrastructure of the future seeks advanced lithium-ion batteries (LIBs) with higher energy and power density, longer cycle life and better safety1-3. Cathode active material typically constitutes the most in terms of weight and cost in LIBs4-6. It is a common observation that high-voltage and high-rate cycling triggers accelerated impedance growth, premature failure, and safety concerns7,8. Many efforts have been spent on exploring new cathode chemistry, introducing dopants into the bulk and at the surface, and designing nano-/micro-/hetero-structures9-14. Coating is a widely exercised method to improve cathode stability, which can work synergistically with other cathode modifications15-18. While a thin coating with high stability and catalytic inertness is helpful, it is often difficult to achieve 100%
coverage (typically by wet-chemistry methods or atomic layer deposition, followed by high-temperature annealing) in synthesis due to solid-on-solid wetting problems. It is also difficult for the coating to remain conformal without cracking during electrochemical cycling. These challenges set the motivation for the present work, which utilizes a simple liquid-solution method to construct high-quality cathode coating by reactive wetting with the oxide active material. Furthermore, we will show that the coating chosen, cobalt boride, has such strong affinity with the oxide particles that it infuses in the grain boundary (GB) network inside the secondary particle (Figure 3-1) in a “complete-wetting” fashion, and this GB infusion occurs facilely at room temperature due to the reactive wetting nature. The obtained coating and infusion network protect the secondary particle from cracking and exposing fresh surfaces to the electrolyte, which greatly enhances the cycle stability, especially at high rates.
We choose a Ni-rich layered cathode LiNi0.8Co0.1Mn0.1O2(abbreviated as NCM in this work; also known as NCM811 in the literature) as a model system to demonstrate the boride coating-plus-infusion strategy. During electrochemical cycling of NCM, a series of detrimental processes are known to take place, including phase transformation in the bulk and at the surface19, intergranular cracking of the secondary particles along GBs20,21, formation and growth of cathode electrolyte interphases (CEIs)22, and side reactions consuming precious liquid organic electrolytes that also generate gases and cause transition metal (TM) dissolution (which may further migrate and precipitate at the anode side and affect the anode stability)23-25. The above processes result in continuous impedance growth and rapidly degrade the full-cell performance, especially when operating under high-rate conditions. One key issue is the stability of surface oxygen, which becomes labile at high charge voltages and easily escape from the active material surface. Such surface oxygen loss not only oxidizes organic electrolytes and evolves gases, but also leads to cation reduction, cation densification (e.g. TMs occupying tetrahedral/Li-layer octahedral sites in layered oxides to become resistive spinel/disordered rock-salt structures)26and phase transformations near the surface of the oxide cathode, which may in turn initiate many other degradation processes in a chain-reaction manner27-29. In this sense, it is beneficial to construct a coating that binds
strongly with the surface oxygen of NCM to address the root-cause of its high-voltage instability, and such a coating, if found, could also mitigate the bulk phase transformations observed in Li-/Mn-rich cathodes considering the better preserved oxygen species30,31.
We now report a cobalt boride (CoxB) coating that stabilizes NCM and greatly improves its high- rate performance. This material CoxB was selected with the following considerations: (i) CoxB is a metallic compound that has no direct tie-lines with oxygen32and thermodynamically it would react with oxygen to form stable compounds such as B2O3, Co3O4 and Co4B6O13, implying strong reactivity between CoxB and the surface oxygen of NCM. (ii) CoxB has an exceptional oxidation resistance even at elevated temperatures (850-950oC)33. This means even though CoxB likes to react with oxygen, the reaction is kinetically self-limiting, likely due to the glass-forming ability of the B2O3-like product at the interface that forms a compact self-healing passivation layer34,35. Thus, while the reactive wetting ensures complete coverage of the surface and tight adhesion between the CoxB and NCM, it does not consume oxygen from the NCM lattice. The passivation layer would kinetically suppress oxygen penetration/loss through this coating layer, and the interfacial polyanionic borate glass also incorporates Li alkali metal that comes with the NCM, making itself a mixed ionic and electronic conductor (MIEC).
(iii) CoxB coating can be easily synthesized at room temperature36,37, which eliminates the necessity of any follow-up high-temperature treatments that may introduce additional complexity and/or defects to the already heavily optimized synthesis route of NCM. (iv) Cobalt boride has been used to coat metal parts to improve their corrosion and wear resistance, thus its mechanical properties should be good in the sense that it must not easily chip or fracture at nanoscale. Remarkably, we shall show that the as- synthesized CoxB coating not only completely covers the surface of NCM secondary particles, but also infuses into the GBs between primary particles with zero wetting angle, which we shall abbreviate as
“infusion” to distinguish it from typical surface coating. This is similar to the complete wetting of GBs by liquid metal (e.g., liquid Ga in aluminum GBs), and intergranular amorphous nanofilms in ceramics.
This infused nanostructure (see schematics in Figure 3-1) dramatically improves the rate capability and cycling stability of NCM, including under high-rate (up to 1,540 mA g−1) and high-temperature (45oC) conditions, by greatly suppressed intergranular cracking, side reactions and impedance growth.
Considering the similar crystal structure, redox-active transition metal species and/or microstructure, we believe the investigated CoxB infusion may be directly applied to many other cathode materials for LIBs, including LiNi1−x−yCoxMnyO2, LiNi1−x−yCoxAlyO2, LiCoO2and Li-/Mn-rich cathodes.
97
Figure 3-1.Schematic “coating-plus-infusion” microstructure where CoxB uniformly coats the surface of NCM secondary particle and infuses into grain boundaries between NCM primary particles.