Al and cation vacancies promote reversibility of Cobalt-free layered cathodes

Background

Lithium-ion batteries with layered NCM cathodes can effectively alleviate the range limitations of electric vehicles, but the scarcity of cobalt inhibits the large-scale application of layered cathodes. Eliminating cobalt in layered cathodes is a necessary condition for achieving a breakthrough in the global application of electric vehicles. However, since Co plays a key role in suppressing Li/Ni co-dispersion, Co-free layered cathodes face challenges in attenuation/lithiation reversibility. Many metals have been proposed to replace cobalt in layered cathodes, and complex composition designs always prioritize suppressing Li/Ni co-dispersion.

Brief Introduction

Professor He Xiangming's team at Tsinghua University found that Al and cation vacancies can inhibit Li/Ni mixing in different ways, but under the same Li/Ni mixing conditions, the irreversible capacity is significantly different. Al reduces the lattice strain between the H2-H3 phases by 13.6%, but has an adverse effect on the reversibility of the structure. The combined effect of Al and cation vacancies not only reduces the H2-H3 lattice strain by 78%, but also greatly improves the reversibility of the structure. In addition, first-principles calculations show that cation vacancies can significantly reduce the lithium migration energy barrier by an order of magnitude. Through the synergistic effect of Al and cation vacancies, the reversibility of the newly designed cobalt-free cathode is significantly enhanced. This work was published in Advanced Energy Materials under the title "Promoting Reversibility of Co-Free Layered Cathodes by Al and Cation Vacancy".

Research highlights

Although Al can inhibit Li/Ni mixing, it is harmful to the structural reversibility.

Cation vacancy-assisted Al substitution significantly enhances the structural reversibility and Li conductivity.

Picture and text guide

Fig. 1 Structure and charge-discharge capability. XRD Rietveld refinement of NM (a), NMA (b), and NMA-δ (c). 0.05 C charge-discharge curves of NM (d), NMA (e), and NMA-δ (f).

Cobalt-free LiNi0.8Mn0.2O2 (NM), LiNi0.8Mn0.15Al0.05O2 (NMA), and LiNi0.8Mn0.1125Al0.05O2 (NMA-δ) were prepared by molten salt-assisted calcination. The compositions estimated by inductively coupled plasma emission spectroscopy (ICP) of the obtained materials were close to their designed chemical formulas (Table S1, Supporting Information), and the Mn deficiency was designed to generate cation vacancies in NMA-δ. The cation vacancies in the TM layers were verified by Rietveld refinement of high-resolution X-ray diffraction (XRD) data. When refining the NMA-δ crystal structure, all cation vacancies in the transition metal (TM) layers were filled with Li ions as the initial state, and then the Rietveld refinement procedure was run. Finally, the refined crystal structure showed that in NMA-δ, about 4% of the octahedral sites in the TM layer are vacant; the R factor is around 2% and the χ2 is close to 3, indicating that the refined crystal model has high reliability (Figure 1a-c and Tables S2-S4, Supporting Information). The presence of cation vacancies may be due to the large electrostatic repulsion generated by high-valent cations, which is kinetically unfavorable for the migration of Li ions to the TM layer. The Li5AlO4 impurity phase usually found in the products of typical coprecipitation methods was not detected in this work. The Li/Ni mixing ratios of NM, NMA, and NMA-δ calculated by Rietveld refinement are 6.42%, 3.9%, and 3.84%, respectively, which are consistent with the (003)/(004) trend in Figure S1d. The typical (003)/(104) ratio of commercial NCM811 is between 1.2 and 1.3. The (003)/(104) ratio of NM decreases to 1.18, indicating that the Li/Ni mixing is aggravated in the absence of Co. The (003)/(104) ratio of NMA rebounds to above 1.2, which strongly supports the view that Al can promote cation ordering of the layered structure just like Co. Notably, the (003)/(104) ratio of NMA-δ is 1.29, indicating that cation vacancies also contribute to structural ordering.

Li/Ni intermixing is considered to be the main structural barrier for reversible delithiation/lithiation of layered cathodes, which leads to large IRC (irreversible capacity) in the initial cycle. In the layered structure, Li ions can jump between two octahedral sites, oxygen dumbbells directly (ODH pathway) or intermediate tetrahedral sites (TSH pathway). Due to the stronger Ni-O bonds, Ni ions mixed in the Li layer narrow the ODH pathway and produce greater electrostatic repulsion for Li migration. In general, the lower the degree of Li/Ni intermixing, the better the conductivity of conventional NCM cathodes and the smaller the irreversible capacity.

In the charge-discharge tests at 0.05 C, the Co-free cathodes showed great differences in the reversibility of lithium ion decay/lithiation. As shown in Figure 1d-f, NMA-δ can provide a discharge capacity of 238 mAh/g, while NM and NMA are 187 and 202 mAh/g, respectively. The reduction of Li/Ni mixing may be one of the reasons for the enhanced reversible capacity of NMA-δ. At the beginning of charge and discharge, NMA-δ captured the rapid rise/fall of voltage, which may be due to more active oxygen on its surface, as shown in Figure S2 (Supporting Information). More importantly, although the Li/Ni mixing is very close (3.8 ~ 3.9%), the IRC of NMA and NMA-δ is significantly different. NMA-δ has the smallest IRC, which is only 4.72 mAh/g. While NMA has the largest IRC, which is 78 mAh/g.

Figure 2 Structural evolution and Li diffusion coefficient change. a-c) In situ XRD monitoring of PITT measurements on NM, NMA, and NMA-δ. d) Phase evolution diagram of the delithiation process. e) Calculation of the diffusion coefficient during the decay.

In order to reveal the structural evolution of lithium ions at different states of charge (SOC) and its influence on lithium ion migration, in situ XRD-monitored potentiostatic intermittent titration (PITT) measurements were performed. As shown in Figure 2a-c, during the decay process, the (003) peaks of the three cathodes first shifted down and then up, reflecting the well-known H1→M→H2→H3 phase transition of layered cathodes. At high state of charge, the (003) peak of NM becomes a little blurred, which is caused by the coexistence of residual H2 phase and H3 phase. It is worth noting that the (003) peak of NMA becomes significantly blurred at high state of charge, indicating that the residual H2 phase further increases, which is also confirmed by the (104) peak, which is clearly divided into two strong branches contributed by H2 phase and H3 phase.

In contrast, NMA-δ behaves completely differently, with the (003) and (104) peaks remaining clear and sharp at high SOC, indicating that the phase transition is quite simple and the two-phase area is greatly reduced. With decreasing potential, the phase transitions of the three cathodes are basically opposite. However, the diffraction peaks of NM and NMA are very different from the decayed states, especially the (104) peak is obviously faded and shifted to a higher degree (Figure 2a-c). Both the peak intensity decay and position shift are the result of irreversible structural changes during the embrittlement/lithiation process, which also leads to the increase of IRC in the initial cycle. It is worth noting that the NMA-δ diffraction peak shows almost perfect intensity and position symmetry during the delithiation/lithiation process, indicating that the structural reversibility is greatly enhanced. Although previous work has reported the enhancement of Li diffusion through lattice defect engineering, this is the first time that cation vacancies have been found to help enhance the structural reversibility of Co-free layered cathodes.

The phase evolution diagrams of the three cathodes are shown in Figure 2d, which are obtained by carefully examining the XRD spectra at 0.03 V intervals (Figure S3). Compared with NM, Al slightly prolongs the first two-phase NMA and NMA-δ regions (H1/M). The second two-phase region (M/H2) of NMA is comparable to that of NM, but the third two-phase region (H2/H3) of NMA is significantly extended to full SOC. In contrast, NMA-δ has the shortest two-phase region.

Based on the PITT data, the lithium diffusion coefficients (DLi) of the three cathodes were calculated and shown in Figure 2e. At low states of charge (3.6 to 4.0 V), the DLi of NMA-δ is significantly higher than that of the other two cathodes, and the DLi of NMA-δ is comparable to that of NM in this region. Both the lattice expansion caused by initial Li extraction and the Ni2+/3+ redox contribute to the rapid diffusion of Li around 3.8 V. The DLi of NMA-δ and NM are comparable in the range of 4.15 to 4.375 V, and NMA-δ again exhibits the highest DLi at higher states of charge. In contrast, NMA shows the smallest DLi from 4.0 to 4.6 V.

The migration of Li in the layered structure mainly adopts two pathways: ODH and TSH. The ODH pathway is always an effective pathway for Li migration, while the TSH pathway can be activated at 1/6 SOC (≈3.75 V) and above. The energy barrier for Li migration decreases almost linearly with the increase of the dimension of the Li space. As shown in Table S5 (Supporting Information), the c parameters of the three cathodes are similar in the initial state. Between 1/6 SOC and 1/3 SOC, although the c parameter of NMA-δ is comparable to that of NM, the DLi of NMA-δ is still much higher than that of NM, indicating that cation vacancies can greatly reduce the barrier to Li migration. Starting from 1/3 SOC (≈3.93 V), NMA enters the M/H2 two-phase region before NM and NMA-δ, and the DLi of NMA is smaller than that of NM. The dense Li space and M/H2 phase boundary are responsible for the delayed Li diffusion coefficient in NM. From 2/3 SOC (≈4.26 V) to FSC, both NM and NMA-δ experience a short two-phase region (H2/H3) and then enter the last single-phase region (H3), while NMA continues to experience the two-phase region (H2/H3) until the end of decay. At FSC, the c parameter (H3) of NM is comparable to that of NMA, but the c value of NMA-δ is the largest.