A. Characterization of synthesized Li[Ni_1/3Co_1/3Mn_1/3]O₂

XRD patterns of Li[Ni_1/3Co_1/3Mn_1/3]O₂ annealed in air for 3 h with increasing temperature are shown in Fig. 1. All peaks from L333 were indexed on the basis of the hexagonal α-NaFeO₂ structure with the $R\bar 3m$ space group. Indexing of these diffraction patterns confirmed the formation of the layered L333 structure without any detectable impurities. All of the samples displayed doublet peaks for the (006)/(102) at 37° (2θ) and the (108)/(110) reflections at 65° (2θ), an indication of ordering of the hexagonal layered structures. ^{Reference Hwang, Tsai, Carlier and Ceder6}

FIG. 1.

XRD patterns of Li[Ni_1/3Co_1/3Mn_1/3]O₂ annealed in air for 3 h at (a) 750 °C, (b) 850 °C, (c) 950 °C, and (d) 1000 °C.

This reduction of cationic mixing enabled a more efficient intercalation and deintercalation of lithium ions inside the host matrix, leading to improved electrochemical performance. ^{Reference Reimers, Dahn, Greedan, Stager, Liu, Davidson and Vonsacken29}

It is also clear that with increasing annealing temperature, the peak-splitting became more pronounced, indicating an increased cationic ordering. Besides peak splitting, the ratio of the intensities of the (003) to (104) indices (i.e., I ₍₀₀₃₎/I ₍₁₀₄₎) is an indicator of cation mixing in materials with a layered structure. For example, a ratio of I ₍₀₀₃₎/I ₍₁₀₄₎ > 1.2 indicates lower cation mixing with highly ordered structures. ^{Reference Ding, Zhang, Jiang and Gao18} From Table I, we can clearly see that the ratios of I ₍₀₀₃₎/I ₍₁₀₄₎ at 750, 850, 950, and 1000 °C are 1.19, 1.28, 1.49, and 1.47, respectively, demonstrating an increase in cationic ordering with temperature and confirming results from the peak splitting analyses.

TABLE I.

Calculated lattice parameters.

SEM micrographs of specimens annealed in air at different temperatures are illustrated in Fig. 2. As expected, the micrographs show that the particle sizes increase with increasing annealing temperature. Confirmation of the particle crystallinity was demonstrated by TEM from previous work, ^{Reference Zhu, Vo, Li, Lu, Kinsinger, Xiong, Yan and Kisailus26} which revealed significant crystal growth of L333 between 750 and 1000 °C. Specifically, specimens heated to 750 °C demonstrated a nanocrystalline nature (i.e., crystal diameters less than 50 nm) as determined by analyses of the FWHM from XRD patterns. These crystals grow in size with increased temperature up to 166 nm by 850 °C, 458 nm by 950 °C, and eventually to 850 nm by 1000 °C (Table I). It is evident that annealing at lower temperatures [i.e., Figs. 2(a)–2(c)] yields a more uniform size distribution than at 1000 °C [Fig. 2(d)]. Thus, the annealing temperature not only affects the crystal size but also changes the particle size dispersity.

FIG. 2.

SEM micrographs of Li[Ni_1/3Co_1/3Mn_1/3]O₂ annealed in air for 3 h at (a) 750 °C, (b) 850 °C, (c) 950 °C, and (d) 1000 °C.

B. Electrochemical performance

Figure 3 shows representative CV curves of samples annealed under different temperatures. Each of these curves displays only one oxidation peak and reduction peak, corresponding to the charge/discharge reactions of the Ni²⁺/Ni⁴⁺ redox couple. ^{Reference Kim and Chung30,Reference Wu, Wang, Su and Chen31} The lack of a peak at 3 V indicates the absence of Mn³⁺ in the as-synthesized material. ^{Reference Gopukumar, Chung and Kim32,Reference He, Ma, Liao and Jiang33} The potential separations between the anode and cathode peaks are at 0.12, 0.21, 0.24, and 0.28 V for samples annealed at 750, 850, 950, and 1000 °C, respectively. Here, there is an increased peak separation, which indicates a decreased rate constant for charge transfer due to the increasing crystalline sizes for specimens annealed at 750–1000 °C. In addition, samples treated at 850 and 950 °C display the highest peak currents (0.19 A/g, and 0.16 A/g) compared to samples annealed at 750 and 1000 °C. It is clear that samples treated at 850 and 950 °C which show smaller voltage differences between the charge and discharge as well as higher peak currents indicate better electrode reaction kinetics and rate performance. ^{Reference Lan, Wang, Zhang, Zhang, Wu and Zhang34,Reference Zhu, Fiore, Li, Kinsinger, Wang, DiMasi, Guo and Kisailus35}

FIG. 3.

CV profiles of Li[Ni_1/3Co_1/3Mn_1/3]O₂ (annealed in air for 3 h at various temperatures) in the voltage range of 2.8–4.5 V at a scan rate of 0.1 mV/s.

The initial charge/discharge profiles of cells with different L333 cathode crystal sizes are illustrated in Fig. 4. The first discharge capacities are 155.1, 204.7, 191.6, and 139.3 mA h/g at 750, 850, 950, and 1000 °C, respectively. Here the discharge capacities are comparable or even better than those currently reported. ^{Reference Gao, Li, Liu, Pinto, Jiang and Yang36} The discharge plateaus for samples annealed at 850 and 950 °C are longer compared to the other two samples, indicating the increased kinetics of these cathodes with increased annealing temperatures.

FIG. 4.

Initial charge/discharge profiles of Li[Ni_1/3Co_1/3Mn_1/3]O₂ samples at 16 mA g⁻¹ (synthesized at different annealing temperatures) in the voltage range of 2.8–4.5 V.

Further cycling charge–discharge capacities were measured at 0.1C for 10 cycles followed by 10 cycles at 0.2C, 0.5C, 1C, and 2C. Figure 5(a) clearly demonstrates that the sample annealed at 850 °C delivers the highest discharge capacity. Although the discharge capacity inevitably decreases with increasing current density, cathodes annealed at 850 °C in air for 3 h delivered a discharge capacity of 119.9 mA h/g at 2C (Table II), comparable with a carbon coated L333 system. ^{Reference Hsieh, Mo, Chen and Chung37} By comparing the samples annealed at 950 and 1000 °C, which have similar cation ordering (I ₍₀₀₃₎/I ₍₁₀₄₎ = 1.49 and 1.47, respectively), the smaller crystal size (458 nm at 950 °C) displays a higher capacity than larger ones (850 nm at 1000 °C) due to the increased surface area and the decreased diffusion path for lithium ions.

FIG. 5.

Cycling performance of Li[Ni_1/3Co_1/3Mn_1/3]O₂ cathodes between 2.8 and 4.5 V (versus Li⁺/Li) (a) at different current rates (0.1C, 0.2C, 0.5C, 1C, and 2C); (b) discharge capacity at 0.1C over 50 cycles.

TABLE II.

Specific capacities for Li[Ni_1/3Co_1/3Mn_1/3]O₂ cathodes annealed at various temperatures for 3 h.

Further calculations (Table II) show that the capacity difference between samples annealed at 850 and 950 °C is smaller (∼15 mA h/g) at low current densities (from 0.1 to 0.5C), but increases to 30 mA h/g at 1C and 40 mAh/g at 2C. Therefore, the smaller crystal size is beneficial under high rates.

The similar phenomenon was also reported by the other researchers in their nanomaterial system. ^{Reference Wang, Ruiz, Guo, Favors, Bay, Ozkan and Ozkan38} This is because the increased surface area provides increased interfaces between the electrolyte and the active material, at the same time, the smaller particle sizes shorten the diffusion path of lithium ions from the surface to the interior of particles, therefore delivering high capacity, especially at high rates. Figure 5(b) displays the cycling stability at 0.1C for all samples. Clearly, the 850 and 950 °C samples carry higher discharge capacities with higher capacity retention (84% for 850 °C and 88% for 950 °C) compared with the other two samples (81% for 750 °C and 79% for 1000 °C) after 50 cycles.

C. Degradation

To investigate the degradation in performance for Li[Ni_1/3Co_1/3Mn_1/3]O₂ cathodes annealed at different temperatures, we disassembled the cells after 50 cycles and collected the cathode materials from the Al foil. The mixture was rinsed with NMP several times to remove the PVDF binder and the remaining organic solvents from the electrolyte. The material was subsequently dried in the vacuum oven for 2 days at room temperature and then subjected to XRD and SEM characterization. Figure 6 demonstrates the SEM micrographs of cathodes removed from the cells after 50 cycles. Based on analyses by EDS, smaller particles in each micrograph are mostly carbon that was mixed during the cell preparation.

FIG. 6.

SEM micrographs of Li[Ni_1/3Co_1/3Mn_1/3]O₂ cathode particles, annealed in air for 3 h at (a) 750 °C, (b) 850 °C, (c) 950 °C, and (d) 1000 °C, and removed from cells after 50 charge–discharge cycles. Yellow arrows in (b) show particles that appear banded due to strain.

The particles that were annealed at 750 and 850 °C did not show obvious structural changes [Figs. 6(a) and 6(b), respectively]. However, some evidence of banding was observed in a few particles [see yellow arrows in Fig. 6(b)]. Small cracks appeared in particles annealed at 950 °C [Fig. 6(c)] and the significantly larger cracks and fractured particles were observed in samples annealed at 1000 °C [Fig. 6(d)]. To study the crystal microstructure change, we conducted the further XRD analyses to calculate the lattice parameters and volume changes.

Figure 7 show the XRD patterns for noncycled Li[Ni_1/3Co_1/3Mn_1/3]O₂ samples (black line) and after 50 cycles (blue line). It can be seen that some of the peaks shifted to lower 2-theta values after cycling, indicating an increased lattice parameter and expanded volume. As discussed earlier, the peak splitting for the (006)/(102) and (108)/(110) reflections represents a well-ordered layered structure. After 50 cycles, the extent of splitting of peaks did not appear to change for specimens annealed at 750 and 850 °C. However, the intensity of the splitting decreased for the sample annealed at 950 °C [Fig. 7(c)] and was especially significant in the first peak splitting ((006)/(102) reflections). The sample annealed at 1000 °C [Fig. 7(d)] showed a further decrease in peak intensity, with almost no splitting observed. These analyses are consistent with our SEM observations.

FIG. 7.

Comparisons of cation ordering via XRD for Li[Ni_1/3Co_1/3Mn_1/3]O₂ samples annealed in air for 3 h at: (a) 750 °C, (b) 850 °C, (c) 950 °C, and (d) 1000 °C, before (black) and after (blue) cycling with enlarged area highlighted to show peak splitting.

Detailed lattice parameters are listed in Table III, from which we can see that the 1000 °C cathode shows the largest lattice parameter increase and volume expansion (2.15%), while samples annealed at 850 and 950 °C display minimal changes. Here, we also observed that the percent change in the “a” lattice parameter (i.e., <0.5%) is significantly less compared to those occurring in the “c” (∼1 or >1%), indicating anisotropic degradation in the L333 system. It is known that lithium ions move along the ab plane in L333 during lithiation and delithiation. Thus, we suspect that parameter c change in this crystal structure degradation is comparably a mild effect to the performance compared to changes along a and b. By combining the above SEM and XRD analyses with charge–discharge cycling performance, it is clear that smaller particle sizes will yield smaller structural changes and thus, better strain accommodation, which is also observed in LiCoO₂ cathodes as well as Si-based anodes. ^{Reference Liu, Zhong, Huang, Mao, Zhu and Huang39–Reference Zhao, Pharr, Vlassak and Suo42} According to Liu et al., ^{Reference Liu, Zhong, Huang, Mao, Zhu and Huang39} there exists a critical particle size of ∼150 nm below which cracking did not occur, and above which surface cracking and particle fracture took place upon lithiation. The unexpected surface cracking is attributed to the development of hoop tension in the surface layer, resulting from a unique lithiation mechanism in crystalline Si nanoparticles by means of motion of a two-phase interface. In the study of Zhao et al., ^{Reference Zhao, Pharr, Vlassak and Suo42} the inhomogeneous distribution of lithium in LiCoO₂ results in stresses that may cause the particle to fracture. This might also apply to the L333 system based on the similar results, which is currently under investigation. One exception to this was the sample annealed at 750 °C. In this sample, it is believed that reduced ordering in the noncycled layered structure (I ₍₀₀₃₎/I ₍₁₀₄₎ = 1.19) yielded poorer performance compared to samples annealed at 850 and 950 °C.

TABLE III.

Lattice parameters for Li[Ni_1/3Co_1/3Mn_1/3]O₂ cathodes after 50 cycles.

In Figs. 7(c) and 7(d), it is clear that the intensity of some high index/low intensity crystallographic reflections was reduced or eliminated after extended electrochemical cycling (i.e., 50 cycles). For example, we observed that the (009) reflection in the 1000 °C annealed specimen almost completely disappears after 50 cycles. Careful observation of particles from Figs. 6(b)–6(d) reveals that many of the fractures initiate in specimens annealed at lower temperature (i.e., 850 °C) and are significantly exacerbated in specimens annealed at higher temperature (e.g., 1000 °C). Since lithium intercalation and deintercalation occur between metal oxide layers, which are parallel to (00l) planes, there are stresses developed due to volumetric strain. Thus, cracks are likely to develop along these planes. In fact, Fig. 6(b) shows a few particles that are apparently fracturing along parallel planes. During this fracture, it is clear that the severe fragmenting due to strain in larger crystals annealed at higher temperatures has reduced the faceting, fractured specimens along low index planes and subsequently reduced the intensities of higher order crystal reflections.