The rapid development of computers, communications, and consumer electronics requires more advanced lithium-ion batteries and battery materials. LiCoO₂ is the first commercialized cathode material and has been put into the 3 C field for a long time. At present, LiCoO₂ still owns the highest volumetric energy density benefiting from its high density. In the past few decades, LiCoO₂ has gone through a long process of development. In the 1990 s, the practical reversible capacity of LiCoO₂ used in the first commercialized lithium-ion battery was 140 mAh · g⁻¹ (∼ 4.2 V versus Li/Li⁺), and now the value reaches as high as 170 mAh · g⁻¹ (∼ 4.4 V versus Li/Li⁺). There is still room for further increase of the useable capacity at higher charging cut off voltage (theoretical capacity of LiCoO₂ is 274 mAh · g⁻¹ for one mol of Li), however, the structural degradation during the phase transition process and interfacial instability between electrode and electrolyte are the main challenges to the successful application of high voltage LiCoO₂.^[1,2]

The phase transition of LiCoO₂ at 4.6 V (extraction of more than 0.5 Li⁺) causes severe changes in cell parameters. The phase transition from the hexagonal Li_xCoO₂ to the monoclinic CoO₂ can introduce more than 3.5% change along the c axis (compared with 2% change from LiCoO₂ to Li_0.5CoO₂).^[3] This process is accompanied with a series of negative effects, including the dissolution of Co into the electrolyte, more active anions, and accumulation of mechanical strains.^[4,5] As a result, the capacity will decay, oxygen will be released, and crack will be generated during battery cycle. Meanwhile, the formation of the unstable cathode electrolyte interphase (CEI) at high voltage can also affect the performance of LiCoO₂ at high cut-off voltage.^[6–8] Numerous works have been devoted to solving the problems mentioned above. Detailed technical strategies include doping elements into LiCoO₂ crystal structure, coating the LiCoO₂ particles or electrode with different compounds, and adding additives into electrolyte.^[9–11]

Doping or partial substituting of bare LiCoO₂ with different metal elements, such as Ti, Cr, Mn, Fe, and Zn, is known to be an effective way to enhance electronic conductivity and maintain structural stability.^[12–16] One example of the bulk doping method is the high capacity LiNi_xMn_yCo_{1 − x − y}O₂ solid solution cathode, which can be viewed as a co-doping of manganese and nickel in traditional layered LiCoO₂.^[17] On the other hand, surface oxide coating on LiCoO₂ was introduced by Cho et al. in 2000, which can well improve surface stability and capacity retention.^[18] After that, different compounds, such as oxide, phosphates, and fluorides, have also been used to improve the high voltage performances of LiCoO₂.^[19–21] After doping or coating treatment, LiCoO₂ has more than 800 cycles at 4.4 V (with graphite as the anode) at 25 °C, which has been successfully applied in 2014.

Although different strategies have been used to enhance the electrochemical performance of LiCoO₂, the cut-off voltage of LiCoO₂ has not exceeded 4.45 V so far. Efforts are still being made to explore the limit voltage of LiCoO₂, such as charging the cathode materials to 4.6 V with a capacity of 220 mAh·g⁻¹ (extraction about 0.8 Li⁺). Considering the vital importance of the surface treatment on the battery performance at high voltage, different techniques have been carried out. Up to now, solid phase mixing, wet chemistry, and various film coating methods have been utilized to modify surface.^[22–24] However, with the increase of the cut-off voltage, nearly all coating materials cannot deliver satisfactory performances. Therefore, more effective surface treatment methods are desired. Typically, cathodes materials contain 3d transition metal (TM) elements, such as Mn, Ni, and Co, and strong catalytic effect could trigger chemical or electrochemical side reactions of electrolyte on the surface. As a result, a buffer layer that consists of non-3d TM oxides may be a promising choice. Moreover, when battery is charged to higher voltage, oxygen release always occurs, accompanying with 3d TM migration. The structural instability will further accelerate the oxygen loss from crystal lattice.^[25,26] It is presumed that 5d TM doping is likely to stabilize crystal structure and induce superior performance due to the larger radius and stronger chemical bonds.^[27] As a result, 5d TM doping and surface coating is a potential method to develop high voltage cathodes of next generation, but there is still a lack of comprehensive investigations.

In this work, 5d TM doping tungsten (W) has been chosen as the modification element to improve the performance of LiCoO₂ at high cut off voltage (4.6 V versus Li/Li⁺).

As shown in the experimental section, the samples added WO₃ with LiCoO₂ precursors are described as doped materials. Similarly, the samples added WO₃ with first sintered LiCoO₂ are defined as coated materials. The SEM images of the bare and W modified LiCoO₂ are shown in Figs. 1(a)–1(e). The particle size of the W-doped LiCoO₂ is slightly smaller than that of the bare LiCoO₂. Figures 1(b) and 1(c) show the SEM images of W-doped LiCoO₂ particles with clean and smooth surface, implying that W has been successfully doped into the LiCoO₂ lattice. Figures 1(d) and 1(e) show the surface morphology of W-coated LiCoO₂ samples. Nano-sized dot-like species, which are identified as the WO₃ coating material, can be clearly seen and are uniformly distributed on the surface of LiCoO₂ particles. Note that we refer to WO₃ coated LiCoO₂ as W-coated LiCoO₂ for simplicity. With the increase in the amount of the coating material, the surface of LiCoO₂ demonstrates more complete coverage. It should be noted that the XRD patterns shown in Fig. 2 demonstrate that all five samples are the pure LiCoO₂ phase (S.G. R-3 m). Due to the small amount of the dopant, tungsten modification does not cause structural variation or introduce obvious impurity phase apparently. These results indicate that doping and coating samples exhibit a consistent crystal structure but present different surface structures. The difference in the electrochemical performance of doping and coating materials can be directly related to the surface structure due to different synthesis processes.

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (color online) SEM images of bare and W modified LiCoO2 materials. (a) Bare LiCoO2; (b) 0.1 wt% doped LiCoO2; (c) 0.3 wt% doped LiCoO2; (d) 0.1 wt% coated LiCoO2; (e) 0.3 wt% coated LiCoO2.

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. (color online) XRD patterns of different LiCoO2 materials, (a) bare material, (b) 0.1 wt% doped material, (c) 0.3 wt% doped material, (d) 0.1 wt% coated material, and (e) 0.3 wt% coated material. The insets show the magnified view of (003) peak of different LiCoO2.

X-ray photoelectron spectroscopy (XPS) measurements are performed to investigate the surface chemical states of the bare and W modified LiCoO₂. Figure 3 shows the spectra of C 1s and O 1s in different materials. For easy interpretation, the assignments of each XPS peak are summarized in Table 1. The signal from lattice oxygen and Li₂CO₃ residuals dominate the landscape of O 1s spectra of the bare LiCoO₂. Li₂CO₃ signal can also be found in C 1s spectrum, indicating the existence of Li₂CO₃ on bare LiCoO₂. Li₂CO₃ can often be found on layered oxide cathode materials due to the chemical reaction with moisture and CO₂ in the air during storage. After W doping, the peak at ∼ 532 eV in O 1s spectra demonstrating the presence of Li₂CO₃ or O–C=O type groups becomes weak. At the same time, peak at ∼ 289.5 eV in C 1s spectra is also weak, confirming less Li₂CO₃ accumulated on the surface. However, W coating seems unable to suppress the accumulation of Li₂CO₃ on the surface of LiCoO₂, as significant signal belonging to Li₂CO₃ can be observed in O 1s spectra of surface coated LiCoO₂. Dahn et al. have suggested that the surface species caused by air or moisture exposure could cause the poor performance of LiCoO₂ at high voltage.^[28] The XPS results are consistent with the different electrochemical performances between bare and doping/coating LiCoO₂, as discussed below. With more Li₂CO₃ on the surface, poorer electrochemical performances can be detected in W-coated LiCoO₂.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (color online) The O 1s (left column) and C 1s (right column) XPS spectra of different LiCoO2, (a) and (f) 0.3 wt% doped LiCoO2, (b) and (g) 0.1 wt% doped LiCoO2, (c) and (h) bare LiCoO2, (d) and (i) 0.1 wt% coated LiCoO2, (e) and (j) 0.3 wt% coated LiCoO2.

Table 1.

Table 1.

Table 1. Summary of XPS peak assignments. . Binding energy/eV Peak Assignments ∼ 529.7 O 1s lattice oxygen in LiCoO2 ∼ 532.0 carbonate or O–C=O ∼ 284.4 C 1s C–C ∼ 289.5 Li2CO3

Table 1. Summary of XPS peak assignments. .

XPS mapping experiments are also performed to uncover the detailed element distributions of W, Co, and O on the surface of LiCoO₂ particles. As shown in Fig. 4, the XPS mapping can be achieved by software Avantage, which presents element spatial distributions from W, Co, and O separately in each map. The strongest peak is selected as the identification standard for the three elements. For bare LiCoO₂, the XPS mapping demonstrates symmetric distribution of Co (green dots) and O (red dots). The relative spatial distribution of W (blue dots), Co (green dots), and O (red dots) can be viewed as the indicator of uncovered LiCoO₂. The XPS mapping results of W-doped and W-coated materials are shown in Fig. 4. It is obvious that with the increase of the modification element, the surface areas of the exposed pure LiCoO₂ become smaller. Moreover, combing with the XPS spectra of W 4f in Fig. 5, the coated materials present higher W concentration compared to the doped ones despite the consistent nominal W stoichiometry. The valence of W in both doped and coated materials is W⁶⁺, confirmed by the peak position of W 4f. The XPS mapping results verifies the distinguishable W distribution in samples from different synthesis methods.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (color online) XPS mapping on the surface areas of different materials. The binding energy at 35 eV (W, blue dots), 780 eV (Co, green dots), and 529 eV (O, red dots) are chosen as the characteristic peaks for each element.

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. (color online) The W 4f spectra of different LiCoO2 powders, (a) bare material, (b) 0.1 wt% doped material, (c) 0.3 wt% doped material, (d) 0.1 wt% coated material, and (e) 0.3 wt% coated material.

In order to further verify the spatial elemental distribution, depth profiles of W are characterized. A raster mode of SIMS provided by the Hiden SIMS is used, and a bigger ion beam current (100 nA) is applied to reflect the depth profile more accurately. Figure 6(a) shows the depth profile of tungsten on different LiCoO₂ derivatives, which indicates the W concentration distribution from surface to inner core of LiCoO₂ particles. It is obvious that the tungsten is enriched on the W coated LiCoO₂. The W-doped LiCoO₂ also shows a decay of the W signal, indicating a relatively higher surface concentration of W than bulk as well. The schematic diagram of the mechanism of two different surface treatment methods is shown in Fig. 6(b).

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. (color online) (a) The depth profile of tungsten abundance at the interface of different LiCoO2 materials. (b) The schematic diagram of the mechanism of two different surface treatment methods.

Electrochemical impedance spectroscopy (EIS) is carried out to investigate the internal resistance, as shown in Fig. 7. Overall, the impedance spectra of all cells with different LiCoO₂ materials show gradual increase upon electrochemical cycling. For each individual curve, three regions can be clearly identified, including a high frequency semicircle (R_SEI), a mediate frequency semicircle (R_CT), and a slope line at low frequency (W). The initial resistance of the doped LiCoO₂ is smaller than that of the bare LiCoO₂, while the resistance of the coated LiCoO₂ shows the highest resistance value (Fig. 7(a)). Upon electrochemical cycling, the increase of the impedance becomes more significant for the cell with W coated LiCoO₂ (Fig. 7(b)). The highest impedance of the coated material demonstrates that a large amount of W accumulation on surface hinders the Li diffusion across the interface. While doping can effectively lower the impedance and favor the Li⁺ migration, the effect of foreign atoms introduction strongly relies on the detailed synthesis method. With W doping into crystal lattice, especially accumulating at particle surface, LiCoO₂ surface area may well form lithium tungsten oxide compounds. As has been reported, lithium tungsten oxides, such as Li₂WO₄ and Li₂ W₂O₇, exhibited good lithium ion conductivity.^[29] And Hayashi et al. further proved that lithium tungsten oxides deposited on positive active materials could decrease the interfacial resistance of LiCoO₂.^[30] As a result, the formation of lithium tungsten oxide compounds on particle surface could enhance the transportation of lithium ion on the surface area. As for the coated materials, the WO₃ is enriched on the surface of LiCoO₂. The lithium intercalation potential of WO₃ is lower than 3.0 V (1.589 V versus Li/Li⁺), so the surface enriched WO₃ cannot participate into the Li transmission process and further hinders the transport of lithium ion.^[31,32]

Fig. 7.

	Figure Option ViewDownloadNew Window
	Fig. 7. (color online) The electrochemical impedance spectra of LiCoO2/Li battery measured after (a) 1st cycle and (b) 50th cycle. The spectra were collected at 25 °C between 1 MHz and 1 mHz.

Figure 8(a) shows the cycle performance of bare and W modified LiCoO₂ (1 wt% and 3 wt%) with different contents of W. It is very interesting that the capacity of LiCoO₂ has increased gently after minute quantities of W modified. This phenomenon is consistent with the results of Ti-doped LiCoO₂. The ionic radius of W⁶⁺ (0.6 Å) is similar to that of Ti⁴⁺ (0.605 Å), so they may express similar effects on the performance of LiCoO₂.^[13] The values of the typical capacity are shown in Table 2. The W doped LiCoO₂ materials show much improved cycle performances than bare LiCoO₂. 70.1% of initial capacity can be retained for 0.3 wt% W-doped LiCoO₂ after 100 cycles in the charge–discharge voltage range of 3.0 V–4.6 V, while only 55.7% for the bare one (compare with the 2nd cycle). In contrast, W coating does not benefit battery performances, as both the initial capacity and cycle stability decrease for coated LiCoO₂. The detailed mechanism of different electrochemical behaviors can be attributed to the modification of the surface properties of LiCoO₂ active particles. Figure 8(b) shows the charge–discharge curves of 0.1% W-doped sample at 1st and 200th cycles. The polarization of the battery increases dramatically after cycling, which is corresponding to the result of EIS.

Fig. 8.

Figure Option ViewDownloadNew Window

Fig. 8. (color online) (a) The cycling performance and coulombic efficiency of LiCoO2/Li battery for the first 100 cycles. All batteries are activated by a cycle of low-rate (0.1 C, 1 C = 274 mAh · g−1) charge/discharge before the stability test at 1 C (1 C = 274 mAh · g−1) in the voltage range of 3.0 V–4.6 V (versus Li/Li+). (b) The charge–discharge curves of 0.1% W-doped sample at 1st and 200th cycles.

Table 2.

Table 2.

Table 2. Typical discharge capacity values of different LiCoO2 (in unit mAh · g−1). . Cycle numbers Bare LiCoO2 0.1% doped 0.3% doped 0.1% coated 0.3% coated 1st 220 223 220 224 220 2nd 192 202 211 205 205 50th 121 158 166 114 110 100th 107 146 148 98 101

Table 2. Typical discharge capacity values of different LiCoO2 (in unit mAh · g−1). .

It can be concluded from the above discussion that 5d TM elements, such as tungsten, could effectively improve the performance of LiCoO₂ at high voltage. However, how to use these elements to get the best results relies on the detailed synthesis method. A small amount of W can increase the discharge capacity of LiCoO₂ from 220 mAh · g⁻¹ to 224 mAh · g⁻¹, but a large amount of impurity element will certainly lead to a decline in specific capacity. Another notable question is the existent form of W in the materials. Whether tungsten ionic are present in crystal lattice or just enriched at the grain boundary still needs further research. First principles calculation can forecast the results from the view of thermodynamic stability. Maybe the ABF-STEM or synchrotron radiation imaging technology can help us to uncover the truth of these questions. Although in this work the performance of coating materials is not as good as the doping ones, surface coating is still an effective mean to enhance the electrochemical stability of cathode materials. On the other hand, the process conditions during synthesis have great influence on the properties of the material, so further explorations are still needed to confirm the conclusion in this work. At the same time, only W-doped materials are not good enough to improve the performance of high voltage LiCoO₂ to meet the requirement of practical applications. The Coulombic efficiency of W-modified materials is only 99.4%, far from the target of 99.9%. Various modification methods, such as ultrathin surface protect layer, homogeneous bulk doping, and concentration gradient design, are also needed to improve the performance. Further research may focus on the process optimization and new element exploration.

In summary, a small amount (0.1 wt% and 0.3 wt%) of tungsten-modified (doped and surface coated) LiCoO₂ is synthesized via a solid-state reaction method. The capacity retention of W-doped material is 72.3% after 100 cycles at 3.0 V–4.6 V, which is higher than that of the bare material (55.7%) and the coated one (47.8%). We find that W is enriched on the surface of LiCoO₂, especially in the coated materials. The W mainly concentrates on the surface of LiCoO₂ particles with a depth of 30 nm. It means that W is not easily to be doped into the materials uniformly. Trace W-modified materials can contribute to the specific capacity of LiCoO₂ (improved with 4 mAh · g⁻¹ of the 0.1 wt% modified material). The surface impurity (such as Li₂CO₃) on the active particles is also proved to be harmful to the electrochemical stability of LiCoO₂. This work highlights that the appropriate treated method is very important for the application of new modified element. The 5d TM elements may be promising candidates as the modified elements for the next generation high voltage cathode materials.