Facile Synthesis of Nb-Doped CoTiO₃ Hexagonal Microprisms as Promising Anode Materials for Lithium-Ion Batteries

Abstract

Bimetallic oxides are demonstrated to show better electrochemical performance than single transition metal oxides. Recently, ilmenite-type transition metal titanate (MTiO₃, M = Fe, Co, Ni, etc.) is emerging as a promising anode for lithium-ion batteries (LIBs) due to its comparable theoretical capacity and small volumetric change during cycling. However, the practical electrochemical performance is still harmed by its poor electronic conductivity. Herein, for the first time, a Nb-doping strategy is adopted to modify CoTiO₃ hexagonal microprisms by a facile solvothermal method combined with an annealing treatment. Benefiting from the unique 1D morphology and the ameliorated conductivities induced by Nb-doping, the optimized Nb-doped CoTiO₃ anode exhibits an improved lithium-storage capacity of 233 mA h g⁻¹ at 100 mA g⁻¹ after 100 cycles and excellent rate capability, which are superior to that of pure CoTiO₃. This work sheds light on the potential application of titanium containing bimetallic oxide in the next-generation advanced rechargeable LIBs.

1. Introduction

Since the energy crisis and environmental issues are becoming more and more serious with the fast growth of industrial civilization, the importance of exploiting green energy gradually stands out. Correspondingly, the development of energy storage technology plays an important role in taking full advantage of renewable resources and green devices [1,2]. In particular, lithium-ion batteries (LIBs) have been predominant for decades and are still widely used in various kinds of electronic devices, such as smart phones and portable laptops, as well as electric automobiles [3,4]. Nowadays, the traditional anode material for commercial secondary batteries is mainly commercial graphite. However, the drawbacks of poor rate capability and the safety risks of generating lithium dendrites severely limits its applications in electric vehicles [5]. Thus, seeking for alternative anode materials with satisfactory merits of high reversible capacity and superior cycling stability at high current densities, as well as high security, is of great significance.

Benefiting from the unique conversion reaction induced by multi-electron transfer during Li⁺ insertion/exaction, transition metal oxides (TMO) capable of delivering a much higher theoretical Li⁺-storage capacity have been regarded as one of the promising anodes for LIBs [6]. However, the great volume change during repeated cycling still restricts its lithium-storage performance [7]. To enhance the electrochemical performance of TMO, common methods including compositing with carbonaceous materials and designing hierarchical nanostructures have been widely investigated over the years [8,9,10]. In addition, it is reasonable to presume that the construction of bimetallic TMO would be another effective strategy. By means of integrating two kinds of metal oxides, in which one component is equipped with high theoretical capacity while the other one exhibits good structural stability during Li⁺ insertion/exaction, the synergistic effect by taking advantage of both components is favorable to achieve superior Li-storage properties compared to that of single TMO [5,11]. Notably, titanium dioxide (TiO₂), with inherent superior structural stability (less than 4% volume variation during Li⁺ insertion/exaction) and high-safety (suitable voltage potential around 1.5 V vs. Li/Li⁺), is widely considered one of the promising anodes for LIBs; however, the even lower theoretical capacity of 335 mA h g⁻¹ than that of graphite (372 mA h g⁻¹) is still unsatisfactory [12,13].

Recent years have seen the rapid emergence of ilmenite-type transition metal titanates (MTiO₃, M = Fe, Co, Ni, etc.), owing to their good chemical, physical, and electrical properties and low cost. They present great promise in the applications of electrochemical energy storage (e.g., metal-ion batteries and capacitors) [14,15,16], photocatalytic degradation [17,18], electrocatalysis [19], and gas sensing [20], and other areas. As for use in LIBs, MTiO₃ has attracted researchers’ attention due to the high theoretical capacity (~500 mA h g⁻¹), as well as the small volumetric change during cycling [14]. As a type of titanium containing bimetallic TMO, MTiO₃ combines the advantages of TiO₂ that shows excellent cyclic stability with the other type of TMO that has a high lithium storage capacity with respect to TiO₂. Therefore, MTiO₃ shows great potential as an anode material for LIBs. However, the poor electronic conductivity is still a key obstacle that harms the overall electrochemical performance [21]. Therefore, it is urgent to develop effective methods to achieve superior electrochemical reaction kinetics. Previous works have concentrated on compositing MTiO₃ with a conductive carbon material to acquire satisfactory performance [22,23,24,25]. For example, L. Chen’s group has reported the successful synthesis of a perovskite CoTiO₃/graphene composite via ball-milling and subsequent high-temperature treatment [23]. The good conductivity and large surface of graphene jointly contributes to the improved discharge capacity of ~350 mA h g⁻¹ at 1 A g⁻¹. B. Ouyang and co-workers have investigated the lithium-storage performance of one-dimensional (1D) hexagonal microprisms of CoTiO₃/C, synthesized by a simple sol-gel method, which exhibited an excellent cycling stability at 5 A g⁻¹ for 1800 cycles [24]. The above examples suggest the potential of CoTiO₃/carbon composites for enhanced lithium storage. However, to our knowledge, although introducing heterogeneous atoms was proven as another effective approach to improve the electronic conductivity of electrode materials, the modification of CoTiO₃ by cation-doping for energy storage applications has scarcely been reported yet.

Herein, 1D Nb-doped CoTiO₃ hexagonal microprisms were rationally designed for the first time via a facile solvothermal method and subsequent calcination. When applied as an anode material for LIBs, a reversible capacity of 233 mA h g⁻¹ could be achieved at 100 mA g⁻¹, much higher than that of pure CoTiO₃ (174 mA h g⁻¹). The enhanced lithium-storage property could be attributed to the improved electronic and ionic conductivities resulting from Nb-doping, as well as the unique 1D morphology.

2. Results and Discussion

The preparation process of 1D CoTiO₃ microprisms with the color evolution during the reaction is clearly shown in Figure 1. In a typical synthesis, several raw materials were added to EG initially, and the mixture was stirred magnetically at room temperature to form a clear red solution. During the continuous strong agitation, EG gradually reacted with Ti⁴⁺ and Co²⁺ to form Co-Ti-EG precursors, and the clear red solution turned into a pale pink suspension simultaneously. After undergoing the solvothermal and subsequent calcination treatment, both Nb-doped CoTiO₃ and pure CoTiO₃ products were obtained.

The information on crystal structure could be revealed by XRD patterns, as illustrated in Figure 2a. It can be seen that both samples exhibit obvious and intense peaks, which correspond to the (012), (104), (110), (113), (024), (116), (214), and (300) planes of CoTiO₃ (PDF No. 15-0866), indicating the successful preparation of pure samples [26]. From the enlarged XRD patterns of the two strongest diffraction peaks, the (104) and (110) planes (Figure 2b), both diffraction peaks of S1 shift slightly toward a higher 2θ angle, which maybe arises from the substitutional doping of the Nb element. Due to the fact that the ionic radius of Nb⁵⁺ (64 pm) is smaller than that of Co²⁺ (74.5 pm) [27,28], the replacement of Co²⁺ by Nb⁵⁺ in the CoO₆ octahedral layered structure can cause crystal lattice shrinkage of CoTiO₃, leading to the increment in diffraction angles. The incorporation of Nb into the crystal structure of CoTiO₃ can be proven by Rietveld refinement of the powder XRD patterns (Figure S1 in the Supporting Information). From the refined lattice parameters listed in Table S1, it is found that the parameters of S1 are slightly different compared to those of S0. The decrease in the lattice parameters together with the shrinkage of the cell for S1 are in accordance with the aforementioned peak shift, which proves that Nb⁵⁺ is successfully incorporated into the CoTiO₃ framework with the ionic radius of Nb⁵⁺ being smaller than that of Co²⁺. Moreover, the average grain size calculated based on the Bragg equation is 64.8 nm for S0, whereas it is slightly reduced to 61.7 nm for S1, which can be attributed to the grain refinement effect of Nb-doping [29].

Raman analysis was also performed to study the structural properties of the samples (Figure S2). It is observed that S0 shows evident peaks located at 203, 232, 263, 331, 378, 453, 473, 604, and 690 cm⁻¹, consistent with the reported active modes of the ilmenite CoTiO₃ phase [30,31]. The strongest peak near 700 cm⁻¹ is the most typical vibrational mode of CoTiO₃ and arises from the symmetric stretching mode of CoO₆ octahedra. Compared to S0, all peaks of S1 exhibit a slight blue-shift to 205, 236, 266, 335, 382, 455, 475, 605, and 695 cm⁻¹. The Raman shift suggests that the lattice constant of CoTiO₃ decreases due to Nb-doping, which confirms from another view of point the substitutional doping of Co²⁺ by Nb⁵⁺.

In order to confirm the doping configuration for Nb-doped CoTiO₃, periodically repeated single CoTiO₃ crystal models were used. Two possible cases were taken into account, as shown in Figure 2c,d for the optimized results, which represent the replacement of Ti and Co sites by Nb atoms, respectively. Furthermore, the electronic energy (E₀) for these different structures was calculated by density functional theory (DFT) method to evaluate their structural stability (computational details can be found in the Supporting Information) [32,33,34,35]. According to the calculated results, when Nb atoms replace Co atoms, the E₀ is −256.79 eV, while the E₀ is −251.98 eV for the substitution of Ti by Nb atoms. Thus, the doping configurations of Nb atoms replacing Co sites are more likely to form in contrast to replacing Ti sites. The DFT calculation results are in good agreement with the XRD analysis, further confirming our conjecture that Nb replaces Co in CoTiO₃ crystal.

To visualize the morphology of the samples, SEM images for both S0 and S1 were captured, as shown in Figure 3a,b. In Figure 3a, regular CoTiO₃ hexagonal microprisms are observed clearly for S0. The similar shape of S1 (Figure 3b) proves that Nb-doping does not destroy the original morphology of CoTiO₃ hexagonal microprisms. Notably, the smaller size of S1 further verifies the positive refinement on grain size owing to the doping effect. The unique 1D structure with the smaller grain size is conductive to shorten the Li⁺ diffusion pathway. TEM examination was utilized to illustrate the detailed lattice information. Distinct lattice fringes with a spacing of 0.27 nm for S0 and 0.25 nm for S1 are shown in Figure 3c,d, which originate from the (104) plane of CoTiO₃ [36]. These results are consistent with the above XRD analysis. Furthermore, the HAADF-STEM image of S1 and the corresponding EDX elemental mappings (Figure 3e) reveal the uniform existence of Co, Ti, and O. Particularly, the homogeneous distribution of Nb confirms the successful Nb-doping in CoTiO₃, which favors the improvement of the electronic conductivity and, thus, the electrochemical performance.

XPS measurement was conducted to investigate the chemical bonding states of the samples. The full spectrum of S1 (Figure S3) demonstrates the occurrence of elements of Co, Nb, Ti, and O. For Co 2p high-resolution XPS (Figure 4a), two broad peaks occur at 781.3 and 797.2 eV accompanied by two satellite peaks that correspond to Co 2p_3/2 and Co 2p_1/2, respectively, confirming the existence of divalent cobalt ions [37]. As for the Ti 2p spectrum (Figure 4b), the two distinct peaks located at 458.2 and 463.9 eV stem from typical Ti⁴⁺ 2p_3/2 and Ti⁴⁺ 2p_1/2, respectively [38]. In addition, as for the O 1s spectrum (Figure 4c), there are two peaks centered at 530.0 and 530.8 eV that can be assigned to the characteristic lattice oxygen (O_La) and surface chemisorbed oxygen (O_Ads) resulted from Co-O-Ti and H-O bonds, respectively [39]. The XPS spectra of S0 (Figure S4) are similar to that of S1 except for the absence of Nb element. Furthermore, in terms of Nb 3d spectrum of S1 (Figure 4d), it can be deconvoluted into two peaks at 206.8 (Nb 3d_5/2) and 209.5 eV (Nb 3d_3/2), verifying the doping of Nb⁵⁺ in S1 [40,41].

The electrochemical properties were evaluated and the obtained results are shown in Figure 5. Figure 5a illustrates the cycling performance at 100 mA g⁻¹ of both S0 and S1, and that of S2 and S3 are also shown in Figure S5a. The initial discharge/charge capacity of S0 is 670.5/292 mA h g⁻¹, corresponding to an initial Coulombic efficiency (ICE) of 43.5%, while the counterpart values for S1 are 693.7/331.5 mA h g⁻¹ and 48.5%. The capacity loss during the initial cycles is related to the formation of a solid electrolyte interphase (SEI) film and the degradation of the organic electrolyte [42], and the higher ICE of S1 than that of S0 indicates the fewer side reactions after the Nb-doping modification. Except for the initial several cycles, the CE of both S0 and S1 reached nearly 100%. After cycling 100 times, S1 exhibits a high reversible capacity of 233 mA h g⁻¹; however, only 174 mA h g⁻¹ is delivered for S0. Notably, both S2 and S3 maintain a relatively high reversible capacity of 213 and 223 mA h g⁻¹, respectively, which is still superior to that of S0.

Furthermore, the rate capabilities of S0 and S1 at different current densities were investigated (Figure 5b). It can be clearly observed that the reversible capacity is inversely proportional to the current density, which is attributed to the sluggish diffusion of both lithium ions and electrons at high current densities. At various current densities of 100, 200, 400, 800, and 1600 mA g⁻¹, S1 exhibits an average reversible capacity of 278, 191, 144, 114, and 82 mA h g⁻¹, respectively, much superior to S0 with 236, 147, 118, 86, and 54 mA h g⁻¹. When the current density returns to 100 mA g⁻¹, the specific capacity of S1 recovers to 223 mA h g⁻¹, which is close to the previous specific capacity value at the same current density. Although the reversible capacities of 198 and 209 mA h g⁻¹ are delivered for S2 and S3 (Figure S5b), the lower values demonstrate that only with an appropriate Nb-doping content can CoTiO₃ display the optimal electrochemical performance. In contrast, a much lower capacity of only 185 mA h g⁻¹ is retained for S0, further confirming the positive effect of Nb-doping.

The electrochemical reaction kinetics were studied by CV tests. Figure 6a,b display the CV curves of the initial three cycles for S0 and S1 with voltages between 0.01 and 3.0 V at a scanning rate of 0.3 mV s⁻¹. As for S0 (Figure 6a), by a closer observation of the CV curve of the first cycle, there is a pair of cathodic/anodic peaks at approximately 1.3 and 2.0 V that could be ascribed to the Ti⁴⁺/Ti³⁺ redox reaction in CoTiO₃ upon Li⁺ insertion/extraction [43]. The sharp reduction peak at a voltage below 0.2 V can be assigned to the reduction of Co²⁺ to Co⁰ [44]. Especially, the irreversible reduction peak at 1.0 V during the first cathodic scan that disappears in the subsequent cycles is due to the formation of an SEI film [45]. In the first anodic scan, there are two broad oxidation peaks located at approximately 2.0 and 2.3 V, corresponding to the oxidation of Ti³⁺ to Ti⁴⁺ and Co⁰ to Co²⁺, respectively [43,46]. Furthermore, it could be observed that the CV curves of S1 (Figure 6b) are similar to that of S0 except the signal ascribed to SEI formation occurs at approximately 0.9 V. From the second cycle onwards, both samples exhibit two obvious reduction peaks at 0.2 and 1.75 V and one broad oxidation peak at 2.0–2.4 V and the curves are almost overlapped, indicating the good reversibility of the electrochemical reactions. The mutual support of bimetals with different electrochemical potentials is beneficial to the structural stability. When one phase undergoes Li⁺ insertion/exaction reaction, the other one remains as an inert matrix, thus, rendering CoTiO₃ with the superior structural integrity.

$$Z\prime =R_s+R_{ct}+\sigma {\omega }^{-\mathrm{1/2}}$$

(1)

$$D_{L{i}^{+}}=\frac{R^2{T}^2}{2A^2{n}^4{F}^4{C}^2{\sigma }^2}$$

(2)

To further explain the enhanced electrochemical performance of S1, an EIS measurement of both S0 and S1 was carried out. As illustrated in Figure 6c, all the Nyquist plots exhibit one semicircle in the high-medium frequency region followed by an inclined line in the low frequency region, which is associated with charge transfer and the lithium-ion diffusion process [47,48]. An appropriate equivalent circuit (the inset in Figure 6c, where R_s, ohmic resistance from electrolyte; R_SEI, resistance of SEI film; R_ct, charge transfer resistance) was proposed to fit the plots, and the obtained impedance parameters are listed in

Table 1. The impedance parameters obtained from EIS fitting.

Sample	Rs (Ω)	RSEI (Ω)	Rct (Ω)
S0	9.3	24.6	22.1
S1	11.1	26.7	8.0

. It could be read that the semicircle of S1 is smaller than that of S0, and the R_ct value of S1 is only 8.0 Ω, indicating its fast electron transport. Moreover, the Li⁺ diffusion coefficient D_Li⁺ could also be evaluated based on Equations (1) and (2) [49,50], where Z′ represents the real part of impedance; ω refers to the angular frequency in the low frequency region; T, absolute temperature; R, gas constant; n, the number of electrons transferred during redox process; F, Faraday’s constant; C, concentration of Li⁺; σ, Warburg factor. Notably, the value of σ can be obtained by linear fitting of the Z′ − ω^−1/2 plot in the low frequency region (Figure 6d). It is obvious that the slope of the line for S1 (326.8) is smaller than that for S0 (424.3), which is indicative of a smaller Warburg impedance of S1. Based on Equation (2), S1 possesses a larger lithium-ion diffusion coefficient than S0. Therefore, it could be concluded that the ameliorated electronic and ionic conductivities resulting from Nb-doping contribute to the improvement in the electrochemical performance of CoTiO₃.

3. Conclusions

In summary, 1D ilmenite-type transition metal titanates of Nb-doped CoTiO₃ hexagonal microprisms were successfully prepared via a facile solvothermal method combined with a subsequent calcination treatment. Multiple characterizations and theoretical calculations confirm the substitutional doping of Co²⁺ by Nb⁵⁺. Due to the Nb-doping effect, the grain size of CoTiO₃ was slightly reduced and the conductivities were improved. Together with the merit of a unique 1D structure, the Nb-doped CoTiO₃ microprisms exhibit enhanced lithium storage performance and rate capability. The facile preparation technique, as well as the resulting excellent electrochemical performance, demonstrates the potential of the Nb-doped bimetallic oxide of CoTiO₃ as a promising anode for LIBs.

4. Materials and Methods

4.1. Materials Preparation

The synthesis of Co_1−xNb_xTiO₃ precursor (x = 0, 0.02, 0.05, and 0.08): First, stoichiometric amounts of urea, cobalt acetate tetrahydrate, columbium pentachloride, and tetrabutyl titanate were successively dissolved in 60 mL of ethylene glycol (EG). Under magnetic stirring at room temperature, the transparent red solution gradually turned into a pale pink suspension. Subsequently, the light pink suspension was transferred into a 100 mL PTFE-lined stainless-steel autoclave and further kept at 120 °C for 6 h. After thoroughly centrifuging and washing with ethanol several times, a Co_1−xNb_xTiO₃ precursor powder pale pink in color was obtained.

The synthesis of Co_1−xNb_xTiO₃ (x = 0, 0.02, 0.05, and 0.08): The precursor powder was loaded into a porcelain boat and sintered in a muffle furnace at 600 °C for 5 h in air with a heating rate of 5 °C min ⁻¹. Finally, the collected powder with x = 0.05 was marked as S1, and other samples with different Nb content were designated as S2 (x = 0.02) and S3 (x = 0.08), respectively. For comparison, pure CoTiO₃ without the Nb addition (x = 0) was prepared under similar conditions and was named S0.

4.2. Materials Characterizations

The crystal structure of the as-prepared samples was measured via X-ray diffraction (XRD) on Rigaku Utima IV with Cu Kα radiation. The morphologies and structures were examined by both an ultra-high-resolution field emission scanning electron microscope (FE-SEM, Nova NanoSEM450) and a high-resolution transmission electron microscope (TEM, ThermoFisher TalosF200X G2). High angle annular dark field (HAADF) scanning transmission electron microscope (STEM) examination and energy-dispersive X-ray (EDX) analysis were carried out on the same TEM. Raman spectra were measured on a Renishaw confocal Raman spectrometer with a 633-nm wavelength laser source. X-ray photoelectron spectroscopy (XPS, ESCALAB XI+) was further performed to study the elemental chemical bonding state of the samples.

4.3. Electrochemical Tests

CR2025-type coin cells were utilized to investigate the electrochemical performance of CoTiO₃ anodes. The active material, acetylene black (AB), and polyvinylidene fluoride (PVDF) with a mass ratio of 8:1:1 were mixed to form a homogeneous slurry, followed by casting onto copper foil. After drying at 120 °C for 12 h under vacuum, small working electrode disks of 14 mm in diameter were obtained. The entire assembly process was implemented in a glovebox filled with ultra-pure argon with 1 M LiPF₆ dissolved in ethyl carbonate (EC) and diethyl carbonate (DEC) (volume ratio of 1:1) with 5% fluoroethylene carbonate (FEC) additive as electrolyte, Celgard 2325 as separator, and lithium foil as the counter electrode. As for electrochemical performance measurements, a Neware battery test system was adopted to evaluate the cycling performance as well as rate capabilities. Cyclic voltammetry (CV) was performed on an IviumStat electrochemical workstation and the voltage window was set as 0.01 to 3.0 V vs. Li/Li⁺. Electrochemical impedance spectroscopy (EIS) was carried out in the frequency range from 100 kHz to 0.1 Hz with a disturbance amplitude of 5 mV.