Table 1
The detailed refined structural parameters for TNO and Ni0.05-TNO.
Product | a/Å | b/Å | c/Å | α/° | β/° | γ/° | Volume/Å3 | wRp | Rp |
TNO | 15.518 | 3.813 | 20.542 | 90 | 113.05 | 90 | 1117.59 | 0.062 | 0.045 |
Ni0.05-TNO | 15.566 | 3.818 | 20.559 | 90 | 113.24 | 90 | 1123.63 | 0.061 | 0.042 |
Firstly, we investigate the crystal structure of Nix-TNO (x = 0.03, 0.05, and 0.07). As shown in Fig. 1a, the diffraction peaks observed in Nix-TNO index well to monoclinic phase TNO (space: A/2m, PDF#72–0159), without impurity phases of TiO2 or Nb2O5. This indicates the successful synthesis of TNO and Ni doping does not change the space group. With Ni content increasing from 0 to 0.07, the position of the diffraction peak of the (020) crystal surface moves to smaller angles (Fig. 1b), which suggests an increase of interplanar spacing at a higher Ni doping content. Besides, with the increase of Ni2+ content, the diffraction peak intensity also decreases, indicating that Ni2+ doping can effectively reduce the crystallinity of the material, which might reduce the Li+ diffusion barrier[26]. However, excessive doping gives rise to the variation of crystal structure, as unveiled from the disappeared (11\(\stackrel{\text{-}}{\text{1}}\)) peak of Ni0.07-TNO in Fig. 1c. More specifically, Fig. 1d-e presents the refinement results of the XRD patterns for TNO and Ni0.05-TNO using the Rietveld method, accompanied by comprehensive structural parameters provided in Table 1. It is seen that Ni0.05-TNO (1123.63 Å3) has a larger volume of unit cells than TNO (1117.59 Å3). This can be attributed to the fact that Ni2+ has a larger ionic radius when compared to Ti4+ and Nb5+. The larger volume of the unit cell can provide a wider diffusion channel for Li+ diffusion (Fig. 1h and 1i).
Further, Raman spectroscopy was applied to get insight into the bonding characteristics of the crystal structures. As shown in Fig. 1f, the bands occurring at 896 and 1000 cm–1 correspond to the tensile vibration of NbO6[28]. The bands of 544 and 646 cm–1 refer to the vibration of the Ti − O bonds of TiO6[26]. As for the bands at 270 and 349 cm–1, they are associated with symmetric and asymmetric bending motions of O-Ti-O and O-Nb-O bridge bonds, respectively[7]. However, the Ti–O vibrational peak of Ni0.05-TNO exhibits a slight blue shift, providing evidence for changes in the internal chemical environment, maybe attributed to increased cation-oxygen bonding and the development of oxygen vacancies[3].
The morphologies and particle sizes of each sample are examined through scanning electron microscopy (SEM). As depicted in Fig. 2a-h, all samples present irregularly reunion particles, and the particle size ranges from 0.2 to 1 µm. Compared to the original sample, Nix-TNO particles are more dispersed. To further reveal the morphological structures of Nix-TNO, TEM characterizations were conducted (Fig. 3a-i). The diffraction patterns in Fig. 3b and 3f unveil that TNO and Ni0.05-TNO are typical single-crystals. The different orientations of the crystal surfaces can be seen from Fig. 3c and 3g. After further amplification and analysis, the specific parameters of the crystal surface can be obtained (Fig. 3d and 3h). From the lattice fringes of (400) and (600) crystal planes of TNO and Ni0.05-TNO, their interlayer spacings are 0.354 and 0.237 nm respectively, which agrees well with the findings from XRD analyses. The homogeneous distribution of Ti, Nb, O, and Ni can be seen from the TEM-EDS elemental mapping (Fig. 3i), suggesting the successful introduction of Ni2+ into TNO.
Then, XPS is utilized to investigate the surface composition and valence states of Ni0.05-TNO and TNO. As seen in Fig. 4a, the full XPS spectrum of Ni0.05-TNO indicates the existence of Ti, Nb, O, and Ni, revealing that Ni has been successfully doped in Ni0.05-TNO. This can also be verified from the characteristic peaks of Ni2+ in Ni0.05-TNO (Fig. 4b). Similar results can be observed from TNO excluding the absence of the Ni element. As seen in Fig. 4c-d, for the TNO, a couple of peaks located at 464.02 and 458.26 eV of the Ti spectrum originate from Ti 2p1/2 and Ti 2p3/2, suggesting the existence of Ti4+. The Nb spectrum of TNO at 209.55 and 206.8 eV are indexed into Nb 3d5/2 and Nb 3d3/2 of Nb5+, separately. Specifically, the peaks of both Ti 2p and Nb 3d shift towards lower binding energy positions following Ni doping, which unveils more Ti3+ and Nb4+ in Ni0.05-TNO[5]. The presence of Ti3+ and Nb4+ can be attributed to the charge distribution caused by Ni2+ doping[16]. From Fig. 4e, Ni0.05-TNO shows a broader oxygen-deficient region in comparison with TNO, unveiling the more missing lattice O in Ni0.05-TNO[29]. This can also be reflected from the electron paramagnetic resonance (EPR). As illustrated in Fig. 4f, both TNO and Ni0.05-TNO display an EPR signal at g = 2.0 (typical signal of oxygen vacancies) and Ni0.05-TNO demonstrates a stronger peak intensity than that of TNO, which indicates the existence of an increased number of oxygen vacancies in Ni0.05-TNO.
The cyclic voltammetry (CV) was used to investigate the kinetic properties during the corresponding redox process in the voltage range of 1.0‒3.0 V for the first three cycles with the sweep of 0.2 mV s–1. As shown in Fig. 5a and 5b, for each cell, the locations of cathodic peaks in the CV during the first cycle deviate from those observed in the subsequent two cycles, which is due to the change of electronic structure of TNO electrode induced by the distortion of TiO6 or NbO6 octahedra and the irreversible consumption of Li+ during the process of Li+ insertion[3, 30]. The CV curves consistently display substantial sharp redox peaks at 1.55‒1.75 V, associated with the Nb5+/Nb4+ redox couple, and small gentle peaks near 1.9 V, corresponding to the Ti4+/Ti3+ redox couple. The broad peak at 1.0‒1.5 V refers to the Nb4+/Nb3+ redox couple. Ni0.05-TNO exhibits reduced redox potential differences compared to pure TNO, demonstrating superior kinetic reversibility after Ni doping[39]. Figure 5c illustrates the charge and discharge profiles of the initial cycle of all samples at a constant current density of 0.1 C. The charge and discharge curves can be divided into three regions according to the inflection point, which are consistent with the redox changes observed during the CV experiments. The regions of A and C correspond to the solid solution of Li+ in TNO[32]. The plateau in region B involves a phase of two-phase transition corresponding to a redox reaction in 1.65v[27, 33]. In this region, a longer reaction plateau in Ni0.05-TNO implies more capacity contribution, most likely originating from the more active sites brought about by oxygen vacancies. It is seen that the discharge capacity and Coulombic efficiency of the doped sample at small Ni content (x = 0.03 and 0.05) are superior to those of the original sample. Ni0.05-TNO exhibits the most superior electrochemical performance, with the first discharge capacity and Coulombic efficiency of 306 mAh g–1 and 91.46%, respectively. This could be attributed to the fact that the doped samples exhibit smaller average particle sizes and higher (ionic and electronic) conductivities[31]. Thus, Ni2+ doping has a positive effect on the surface reaction kinetics of Li+.
Figure 5d illustrates the rate capabilities of Nix-TNO (x = 0, 0.03, 0.05 and 0.07) at varied current densities. The capacities of Ni0.05-TNO at 0.5, 1, 2, 5, 10, 20, and 30 C are 265.68, 251.16, 239.78, 223.61, 204.89, 185.63, and 181.76 mAh g–1, respectively, which significantly elevated in comparison to TNO (202.70, 189.90, 180.52, 170.09, 158.89, 147.43, and 138.62 mAh g–1). Moreover, the cyclic stability of Nix-TNO at a large current rate of 10 C was evaluated (Fig. 5e). It is seen that Ni0.05-TNO displays the highest cycling stability among the four samples. The reversible capacity of Ni0.05-TNO after 500 cycles is 146.19 mA h g–1, with a capacity loss of only 0.04% per cycle, while the reversible capacities of other Nix-TNO samples (x = 0, 0.03 and 0.07) are 85.31, 125.95, and 81.96 mAh g–1. The outstanding cyclic stability may be due to the wider Li+ diffusion pathway after Ni2+ doping, which helps stabilize the structure during repeated cycles[3]. The reason for the poor cycling performance of Ni0.07-TNO might be that excessive Ni2+ doping results in excessive lattice distortion (see Fig. 1c), which leads to a significant decrease in structural stability.
Table 2
Impedance parameters and Li+ diffusion coefficients of TNO and Nix-TNO
Product | Rct (Ω) | σ (Ω s–0.5) | D×10–14 (cm2 s–1) |
TNO | 168.00 | 132.64 | 3.51 |
Ni0.03-TNO | 55.35 | 96.32 | 32.3 |
Ni0.05-TNO | 51.42 | 110.85 | 37.5 |
Ni0.07-TNO | 69.15 | 37.99 | 20.7 |
In order to investigate the nature of ion transport during the charge and discharge processes, we further performed the impedance analysis test (A.C. impedance technique) on the battery. Figure 6 illustrates the Nyquist plots for both TNO and Nix-TNO after cycling three times at 0.1 C with the fitted equivalent circuit inserted. In Fig. 6a, a semicircular feature is present in the high-frequency region, which refers to the reactive process of electron transfer during Li+ desolvation/adsorption (Rct). The low-frequency region is an inclined straight line, the slope of this line is related to the diffusion of Li+ in the bulk crystals (Warburg resistance)[34, 35]. The specific fitting parameters can be seen in Table 2, the doped samples demonstrate much smaller Rct (55.35, 51.42 and 69.15 Ω for Ni0.03-TNO, Ni0.05-TNO and Ni0.07-TNO, respectively) than that (168.0 Ω) of the pristine samples. The reason might be that Ni2+ could optimize the electronic structure of TNO, which facilitates the internal electron transfer, thus contributing to the electronic conductivity [36]. This can be verified from the obtained conductivity via the four-probe conductivity test. The electronic conductivity (2.2 × 10–7 s cm–1) of Ni0.05-TNO is two orders of magnitude higher than that (1 × 10–9 s cm–1) of TNO [25]. These results unveil that Ni doping is effective in tuning the electronic conductivity of TNO.
Then, we explore the diffusion coefficient of Li+ via the impedance analysis. The Li+ diffusion coefficient is mainly related to the slope of the straight line in the low-frequency region, and the specific diffusion process can be obtained according to the following equation:
$${D}_{{\text{L}\text{i}}^{+}}=\frac{{R}^{2}{T}^{2}}{2{A}^{2}{n}^{4}{F}^{4}{C}^{2}{\sigma }^{2}}$$
1
,
In the given expression, R represents the gas constant, T stands for the absolute temperature, A denotes the surface area of the active electrode, n is the number of electrons per molecule, F represents the Faraday constant, C is the molar concentration of Li+, and σ represents the Warburg factor. σ can be estimated based on the plot of the linear fit between Z and the inverse square root of the lower angular frequency (Fig. 6b). It can be expressed by the following equation:
$${\text{Z}}^{{\prime }}\text{ = }{\text{R}}_{\text{s}}\text{+}{\text{R}}_{\text{ct}}\text{+}{\text{σω}}^{\text{-}\text{1/2}}$$
2
.
The Li+ diffusion coefficients of Nix-TNO and TNO are tabulated in Table 2. The Li+ diffusion coefficients of Ni0.03-TNO, Ni0.05-TNO and Ni0.07-TNO are 3.23×10–13, 3.75×10–13, and 2.07×10–13 cm2 s–1, respectively, which are markedly higher than that (3.51×10–14 cm2 s–1) of the pristine one. The Ni0.05-TNO displays the highest Li+ diffusion coefficient. The reason for this phenomenon may be that the lattice distortion caused by Ni2+ doping gives rise to an increase in the cell volume, which provides a wider channel for Li+ diffusion.
To further unveil the change of crystal structure during the (de)lithiation process of Ni0.05-TNO, the non-in situ XRD tests were conducted for the pole pieces by discharging or charging them to specific voltages at current densities of 0.2 C. Here, four points were respectively selected for the initial charging and discharging processes (Fig. 7a). As shown in Fig. 7b, all the samples match well with the monoclinic Ti2Nb10O29 and the crystal diffraction peaks are relatively sharp, which suggest a good stability of the crystal structure during the charging and discharging process. Figure 7c-d illustrates the enlarged (21\(\stackrel{\text{-}}{\text{5}}\)), (41\(\stackrel{\text{-}}{\text{1}}\)), and (020) crystal planes, and it is seen that the (020) crystal plane shifts to a low angle during discharging and returns to the initial position in the charging process. This unveils the good structural stability of Ni0.05-TNO during (de)lithiation. However, the lattice parameters of (21\(\stackrel{\text{-}}{\text{5}}\)) and (41\(\stackrel{\text{-}}{\text{1}}\)) crystal planes do not change significantly, which is due to that these crystal planes are not the preferential diffusion paths of Li+ in Ni0.05-TNO[37–39].