Figure 1 shows the TG-DTD curves of glucose and dicyandiamide in N2. It can be seen that the decomposition of glucose occurs from 210 to 600°C with two weight losses on the TG curve and two peaks on the DTG curve, producing a lot of water. With the increase of the temperature, there are four weight losses on the TG curve and four peaks on the DTG, corresponding to DCDA gradually condensing to melamine, tris-s-triazine, and C3N4 [34]. Above 700°C, C3N4 completely decomposes. The residual rate of glucose and dicyandiamide is 13.4 wt.% and 1.1 wt.%, respectively, indicating that the carbon content from DCDA can be ignored.
The carbon content was determined via TG measurements in air for the LZTO@C-N composites (Fig. 2a). The weight loss below 200°C is due to the evaporation of adsorbed water. The sharp weight loss from 200 to 500°C is assigned to the oxidation of carbon in air. It can be seen that the carbon content is 6.5 wt.%, 6.6 wt.% and 6.3 wt.% for LZTO@C-N-1, LZTO@C-N-2 and LZTO@C-N-3, respectively, further confirming that the amount of dicyandiamide has small effects on the carbon content of the composites. The N content is 1.3 wt.%, 2.2 w.% and 3.5 wt.% based on the element analysis results for LZTO@C-N-1, LZTO@C-N-2 and LZTO@C-N-3, respectively, increasing with the increase of dicyandiamide. The XRD patterns of the P-LZTO and LZTO@C-N anodes are exhibited in Fig. 2b. The diffraction peaks of the four samples can well be assigned to a cubic spinel structure of LZTO (JCPDS No. 86-1512), indicating that the N-doped carbon cannot affect the formation of pure LZTO. Although carbon exists in LZTO@C-N anodes, the diffraction peaks of carbon are not detected, showing that the carbon is amorphous or the carbon layer is thin. In addition, compared with P-LZTO, the intensities of the diffraction peaks for the LZTO@C-N anodes are low, owing to the carbon layer preventing the growth of LZTO grains, and then leading to poor crystallinities. Compared with P-LZTO, the magnified (311) plane shifts to the high angle for LZTO@C-N-1, and to the low angle for LZTO@C-N-3, and has no obvious shift for LZTO@C-N-2, respectively (Fig. 2c), indicating that the lattice constants decrease for LZTO@C-N-1 and increase for LZTO@C-N-3, and have no obvious change for LZTO@C-N-2. The lattice parameters calculated from the XRD data are listed in Table 1 for the four samples and further confirm the results above. Some Ti4+ ions (r = 0.061 nm) will be reduced to Ti3+ ions (r = 0.067 nm) during carbothermal processing, and then some oxygen vacancies will form to maintain the electrical neutrality, which could cause defects and/or distortions and lead to the shrink of the cell volume [35]. So, compared with P-LZTO, the cell volumes of LZTO@C-N-1 and LZTO@C-N-2 with low N content are slightly small. When the N content further increases, some N3− ions (r = 0.171 nm) may replace some O2− ions (r = 0.138 nm), and then the cell volume of LZTO@C-N-3 is larger than that of P-LZTO. The large change of lattice constants will reduce the structural stability of LZTO@C-N-3, and then worsen the electrochemical performance.
Table 1
Lattice parameters of P-LZTO and LZTO@C-N anodes
Samples
|
a = b = c (Å)
|
V (Å3)
|
P-LZTO
|
8.371(2)
|
586.64
|
LZTO@C-N-1
|
8.364(9)
|
585.32
|
LZTO@C-N-2
|
8.371(0)
|
586.59
|
LZTO@C-N-3
|
8.374(2)
|
587.26
|
The XPS spectra show that Zn, O, Ti, N and C elements exist on the surfaces of LZTO@C-N anodes (Fig. 3a). The C1s high-resolution XPS spectra are exhibited in Fig. 3b. The binding energies at 284.5, 285.1, 287.3, and 289.0 eV correspond to C-C/C = C, C-N, C = N and O-C = O [27, 28], respectively. The existence of C-N and C = N indicates that N has been doped into C, and then the electronic conductivity of the carbon material will be further enhanced. The N1s high-resolution XPS spectra include five peaks (Fig. 3c), namely pyridine N (397.7 eV), pyrrole N (398.6 eV), graphitization N (400.0 eV), N-Ti-O (400.9 eV) and N-Ti (396.0 eV) [26, 28]. The pyridine N and pyrrole N have high electrochemical activity and can produce more defects. The N-Ti may be caused by TiN compound [36, 37]. However, no diffraction peaks of TiN are detected from the XRD patterns of LZTO-C-N materials, which may be that the TiN is amorphous. The ratio of N-Ti increases with the increase of N content. The existence of excess TiN may inhibit the migration of Li+ ions [38]. The Ti 2p high-resolution spectra can be deconvoluted into five peaks (Fig. 3d). The peak around 459.1 eV is ascribed to N-Ti-O. The peaks around 465.0 eV and 459.0 eV correspond to Ti 2p1/2 and Ti 2p3/2 of Ti4+, respectively. The peaks of Ti 2p1/2 at 463.4 eV and Ti 2p3/2 at 457.7 eV are assigned to Ti3+, which may be that some Ti4+ ions are reduced to Ti3+ ions during carbothermal processing. The existence of Ti4+/Ti3+ is beneficial to the improvement of the intrinsic electronic conductivity of LZTO.
Figures 4a-d display the SEM images of P-LZTO and LZTO@C-N materials. It can be seen that the four samples are composed of nano-sized particles. With the increase of the N content, the particles gradually fuse together. The TEM images show that carbon exists in LZTO@C-N materials. The elemental mappings show that C, N, O, Ti and Zn can uniformly distribute in LZTO@C-N-2 (Fig. 5).
To understand the electrochemical reactions of LZTO, galvanostatic charge-discharge curves were recorded at 1 A g− 1 (Fig. 6a). For each sample, two discharge plateaus ca. 1.09 and 0.5 V are ascribed to the insertion of Li+ ions. One charge plateau ca. 1.46 V corresponds to the extraction of Li+ ions. The plateau ca. 0.5 V may be due to the multiple restoration of Ti4+ [39, 40]. The electrochemical reactions for insertion and deinsertion of Li+ ions correspond to the Ti4+/Ti3+ redox couple. The discharge specific capacities of P-LZTO, LZTO@C-N-1, LZTO@C-N-2 and LZTO@C-N-3 are 226.5, 267.8, 256.3 and 222.1 mAh g− 1 with the coulombic efficiency of 65.6%, 79.4%, 84.4% and 78.5% at the 1st cycle, respectively. It can be seen that the reversible specific capacities of LZTO are enhanced due to the reduction of the side reactions between LZTO and electrolyte for the existence of the N-doped carbon coating layer. The voltage difference between the charge and corresponding discharge plateaus is largest for LZTO@C-N-3, with the highest polarization, which may be due to the excess TiN inhibiting the diffusion of Li+ ions, in line with the XPS results. 120.1, 158.5, 163.2 and 113.8 mAh g− 1 are kept with the capacity retention of 64.1%, 68.4%, 70.2% and 60.9% for P-LZTO, LZTO@C-N-1, LZTO@C-N-2 and LZTO@C-N-3 after 500 cycles at 1 A g− 1, respectively (Fig. 6b). Compared with P-LZTO, the cycling performance of LZTO@C-N-1 and LZTO@C-N-2 is good, which may originate from the N-doped carbon coating layers with proper N content and then optimize the structure of LZTO. The severe capacity fade of LZTO@C-N-3 may be that the excess N reduces the structural stability, in line with the XRD results. Discharging at 2 A g− 1 and charging at 0.5 A g− 1, the discharge specific capacities are 219.7 and 214.8 mAh g− 1 for LZTO@C-N-2 at the 1st and 150th cycles, respectively, with small capacity fade (Fig. 6c). The P-LZTO and LZTO@C-N composites were cycled from 0.5 to 3 A g− 1, and then down to 0.5 A g− 1 to evaluate the rate capability (Fig. 6d). LZTO@C-N-2 shows the best rate capability, even with the capacities of 205, 190.3 and 168.4 mAh g− 1 at 2, 2.5 and 3 A g− 1, respectively. When the current density is over 1.5 A g− 1, the discharge specific capacities of LZTO@C-N-3 sharply fade, which may originate from the excess TiN inhibiting the fast migration of Li+ ions due to the high N content. Compared with the LZTO anodes reported in previous literatures, good rate capability is exhibited for the LZTO@C-N-2 composite (Table S1).
To further understand the different electrochemical behaviors of P-LZTO and LZTO@C-N composites, the electrochemical impedance spectra (EIS) were collected and the equivalent circuit model are presented in Fig. 7a. The EIS patterns are composed of a small intercept, a semicircle and a straight line. The small intercept in high frequency is assigned to the ohmic resistance including the contact resistance between electrode, separator and electrolyte (RE); the semicircle in high frequency represents the charge-transfer impedance (RCT) on the electrode/electrolyte interface and the straight line in low frequency is the Warburg resistance (W), representing the diffusion of Li+ ions into the bulk of the electrode material. The parameters calculated from the equivalent circuit model are listed in Table 2 for P-LZTO and LZTO@C-N composites. It is well known that small RCT is beneficial to the electrochemical performance of an electrode. Compared with P-LZTO, the RCT of the LZTO@C-N composites reduces due to the N-doped carbon coating. Among, the LZTO@C-N-2 anode has the smallest RCT of 52.1 Ω.
The diffusion coefficients of Li+ ions (DLi+) are calculated from the following equation according to the Warburg diffusion in the low frequency for the P-LZTO and LZTO@C-N anodes
$${\text{D}}_{\text{L}{\text{i}}^{\text{+}}}\text{=}\frac{\text{0.5}{\text{R}}^{\text{2}}{\text{T}}^{\text{2}}}{{\text{A}}^{\text{2}}{\text{F}}^{\text{4}}{\text{n}}^{\text{4}}{\text{C}}^{\text{2}}{\text{σ}}^{\text{2}}}$$
1
where R is the gas constant of 8.314 J mol− 1 K− 1; T is the room absolute temperature of 298.15 K; A is the surface area of the electrode for 1.13 cm− 2 in the work; n is the number of electrons transferred in the half reaction; F is Faraday constant of 96,485 C mol− 1; C is the concentration of Li+ ions in LZTO, and can be obtained from the following equation
the Warburg factor σ obeys the following relationship:
The relationship between Zre and ω−1/2 is exhibited in Fig. 7b. The lithium diffusion coefficients are 1.1×10− 10, 8.3×10− 10, 2.9×10− 10 and 1.5×10− 10 cm2 s− 1 calculated from Eqs. 1–3 for P-LZTO, LZTO@C-N-1, LZTO@C-N-2 and LZTO@C-N-3, respectively. It can be seen that N-doped carbon coating improves the DLi+ of LZTO. High DLi+ is beneficial to the rate capability of an electrode.
Table 2
Electrochemical impedance parameters obtained according to the equivalent circuit model for P-LZTO and LZTO@C-N electrodes
Samples
|
RE (Ω)
|
RCT (Ω)
|
DLi+ (cm2 s− 1)
|
P-LZTO
|
1.8
|
82.0
|
1.1×10− 10
|
LZTO@C-N-1
|
2.1
|
65.9
|
8.3×10− 10
|
LZTO@C-N-2
|
1.9
|
52.1
|
2.9×10− 10
|
LZTO@C-N-3
|
1.9
|
69.1
|
1.5×10− 10
|
If an electrode is to be commercialized, it is important to possess good electrochemical performance in wide temperature range. Figure 8a displays the cycling performance of P-LZTO and LZTO@C-N-2 at the high temperature of 55°C for 1 A g− 1. 282.5 and 290.3 mAh g− 1 are delivered at the 1st cycle and the capacities fade to 81.5 and 151 mAh g− 1 at the 150th cycle for P-LZTO and LZTO@C-N-2, respectively. It can be seen that the cycling performance becomes worse for the two electrodes at the high temperature due to the intensified side reactions between active materials and electrolytes. Compared with P-LZTO, the cycling performance of LZTO@C-N-2 is good at 55°C for the N-doped carbon coating layer reducing the side reactions. It is well known that the diffusion rate of Li+ ions reduces and then the capacities decrease at a low temperature. 200.1, 185.7, 175.5 and 168.9 mAh g− 1 are delivered for the LZTO@C-N-2 anode at 0.4, 0.6, 0.8 and 1 A g− 1 for 0°C, respectively (Fig. 8b). The values decrease to 160.2, 147.9, 146.3 and 141.9 mAh g− 1 for P-LZTO. The N-doped carbon coating improves the rate capability of LZTO@C-N-2 at the low temperature of °C. In addition, LZTO@C-N-2 has good cycling performance at 0°C. Cycling for 300cycles at 0.2 and 0.5 A g− 1, the retention is 89.1% and 88.4% (Fig. 8c), respectively.