Carbon coated “zero-strain” Li₂TiSiO₅ nanowires for high-performance lithium storage

ABSTRACT

Li₂TiSiO₅ is a ‘zero-strain’ anode material with a large theoretical capacity, appropriate operation potential, and good cyclability. However, its low electronic conductivity negatively affects its practical capacity and rate performance. Herein, we improve the electrochemical performance of Li₂TiSiO₅ by carbon-coating and nanosizing dual modifications, and further explore its lithium-storage and ‘zero-strain’ mechanisms by in-situ XRD. Within 0.2–3.0 V, the carbon-coated Li₂TiSiO₅ nanowires deliver a large reversible capacity of 257 mAh g⁻¹, high rate performance with a 5000 mA g⁻¹ vs. 50 mA g⁻¹ capacity ratio of 71.6%, and excellent cyclability with 90.0% capacity retention at 5000 mA g⁻¹ over 4000 cycles. The special crystal structure of Li₂TiSiO₅ with electrochemical active TiO₆ octahedra surrounded by inactive SiO₄ and LiO₆ polyhedra can effectively strengthen the volume-buffering capability, resulting in the ‘zero-strain’ behaviour with a tiny lattice-volume change of only 0.41%. Therefore, this dual-modified Li₂TiSiO₅ material has great practicability for high-performance lithium storage.

KEYWORDS:

• Li₂TiSiO₅
• dual modification
• electrochemical property
• in-situ XRD
• anode

Introduction

Nowadays, lithium ion batteries (LIBs) are among the best power sources with good electrochemical performances, such as high specific energy and good cyclability [1–4]. They are widely applied in a variety of devices used in the fields of portable electronics, electric vehicles, aviation, and military equipment [5]. For LIBs, high-performance cathode and anode materials are highly desirable [6–9]. As the commonly used anode material, graphite has a stable electrochemical performance with a reversible capacity approaching to its theoretical capacity of 372 mAh g⁻¹. However, its operation potential is too low (∼0.1 V), which may cause the formation of lithium dendrites, resulting in short circuit [10–13]. Li₄Ti₅O₁₂ is another successfully commercialized anode material for LIBs. As a ‘zero-strain’ material, Li₄Ti₅O₁₂ has excellent cyclability and a safe operation potential of ∼1.5 V. Nevertheless, its theoretical capacity is too small (175 mAh g⁻¹), and its overly high operation potential has a negative impact on its energy density [14,15].

Li₂TiSiO₅, a recently-developed ‘zero-strain’ anode material, has a theoretical capacity of 315 mAh g⁻¹ based on the Ti³⁺/Ti⁴⁺ and Ti²⁺/Ti³⁺ redox reactions and an appropriate operation potential, which appropriately tackle the issues in graphite and Li₄Ti₅O₁₂ [16]. However, it suffers from a low electronic conductivity, resulting in the negative impacts on its rate capability and reversible capacity [17]. So far, however, only a few studies on the modifications of Li₂TiSiO₅ have been reported [16–18]. In addition, the crystal-structural evolution and ‘zero-strain’ mechanism of Li₂TiSiO₅ are still unclear and should be clarified.

In this work, we improve the electrochemical performance of Li₂TiSiO₅ by carbon-coating and nanosizing dual modifications, and investigate its mechanisms by an in-situ XRD combined with Rietveld refinements. Within 0.2–3.0 V, the resulting carbon-coated Li₂TiSiO₅ (C-Li₂TiSiO₅) nanowires show a reversible capacity as large as 257 mAh g⁻¹ with an appropriate operation potential of 0.83 V and high rate performance with a 5000 mA g⁻¹ vs. 50 mA g⁻¹ capacity ratio of 71.6%. The excellent cyclability with 90.0% capacity retention at 5000 mA g⁻¹ over 4000 cycles confirms the ‘zero-strain’ behaviour of Li₂TiSiO₅ with a tiny lattice-volume change of 0.41%. The ‘zero-strain’ behaviour is mainly rooted in the ultra-stable crystal structure of Li₂TiSiO₅ with electrochemical active TiO₆ octahedra surrounded by inactive SiO₄ and LiO₆ polyhedra, which can effectively enhance the volume-buffering capability during lithiation/delithiation. Therefore, the C-Li₂TiSiO₅ nanowires are suitable for high-performance lithium storage.

Experimental

Material synthesis

LiOH (99.9%, Macklin), C₁₆H₃₆O₄Ti (99.9%, Macklin), and C₈H₁₂O₈Si (95%, Macklin) in a molar ratio of 2 : 1 : 1 were dissolved in 10 mL ethanol. Then, 0.5 g polyvinylpyrrolidone (PVP, Mw = 1300000, Macklin) was dissolved, stirring for 2 h. The obtained electrospinning solution was transferred to a syringe with a 19 G needle. The provided voltage was 15 kV. The film collector was about 10 cm under the spinneret. The collected film was calcinated at 870°C for 4 h in argon with a ramping rate of 2°C min⁻¹, obtaining the C-Li₂TiSiO₅ nanowires.

Material characterization

The X-ray diffraction (XRD) test was conducted on an Ultra IV diffractometer to characterize the crystal structure of the C-Li₂TiSiO₅ nanowires, and the obtained XRD spectrum was Rietveld-refined with the aid of the free GSAS programme [19]. The field emission scanning electron microscopy (FESEM) test was performed on a JEOL JSM-7800F to collect morphological features. The energy-dispersive X-ray spectroscopy (EDX) test was conducted on an OXFORD X-Max to examine the element distributions. The transmission electron microscopy (TEM) was applied using a JEOL JEM-2100F. An ASAP 2460 surface area analyser was used to determine the Brunauer−Emmett−Teller (BET) specific surface area. The compositions and elemental valences were examined by X-ray photoelectron spectroscopy (XPS) characterizations on a PHI5000 Versaprobe III spectrometer.

Electrochemical measurement

To prepare the working electrodes, the C-Li₂TiSiO₅ nanowires, conductive carbon, and polyvinyl difluoride were well mixed into N-methyl-2-pyrrolidone in a mass ratio of 7 : 2 : 1. The obtained slurry was evenly spread on copper foils with a mass loading of ∼1.0 mg cm⁻² and dried in vacuum at 110°C for 10 h. In a glove box filled with argon, CR2032 coin-type half cells were assembled by using the C-Li₂TiSiO₅ working electrodes and lithium-metal anode. The electrolyte was a mixture of dimethyl carbonate, ethylene carbonate, and diethylene carbonate in equal volumes with 1 M LiPF₆. The separators were Whatman GF/D-1832 glass fibres.

The electrochemical performance and galvanostatic intermittent titration technique (GITT) experiments were carried out using a CT-3008 battery tester. Cyclic voltammetry (CV) was conducted using a CHI660E electrochemical workstation. All the electrochemical performances of the C-Li₂TiSiO₅ nanowires were investigated within 0.2–3.0 V.

In-situ XRD characterization

To study the crystal-structural change of the C-Li₂TiSiO₅ nanowires, an in-situ XRD half cell (Scistar LIB-XRD) equipped with a beryllium window was assembled [20–22]. The separator was Whatman GF/D-1832 glass fibre. The recorded in-situ XRD spectra were Rietveld-refined by the free GSAS programme.

Results and discussion

The XRD spectrum of the C-Li₂TiSiO₅ nanowires is shown in Figure 1a, and the Rietveld refined data are summarized in Table S1 and Table S2. Li₂TiSiO₅ exhibits a tetragonal crystal structure and P4/nmm space group [23]. The refined lattice constants of Li₂TiSiO₅ are a = 6.44059(53) Å, c = 4.40981(52) Å, and V = 182.925(46) ų. The crystal structure of Li₂TiSiO₅ (Figure 1b) contains electrochemical active TiO₆-octahedron chains, which are surrounded by inactive SiO₄-tetrahedron and LiO₆-octahedron chains along the c axis. Each TiO₆ octahedron corner-shares with four LiO₆ octahedra and four SiO₄ tetrahedra, and face-shares with four LiO₆ octahedra. As a result, interconnected channels with large sizes are formed. The alternate arrangement of the active and inactive chains with an interconnected polyhedron network plays an important role in accommodating the volume change, and significantly contributes to the crystal-structure stability [24].

Figure 1. Physico-chemical characterizations of the C-Li₂TiSiO₅ nanowires. a) XRD spectrum with Rietveld-refinement. b) Crystal structure. c) FESEM image. d) EDX mapping images. e) HRTEM image (inset: SAED pattern).

The FESEM image (Figure 1c) clearly shows a nanowire morphology with a diameter range of 100 − 200 nm, and the large BET specific surface area is 31.90 m² g⁻¹ (Fig. S1). The EDX mapping images (Figure 1d) exhibit the homogeneous distributions of Ti, Si, and O in the C-Li₂TiSiO₅ nanowires. The lattice spacing in the HRTEM image (Figure 1e) of the C-Li₂TiSiO₅ nanowires is determined to be 0.219 nm, corresponding to the (002) crystallgraphic plane of Li₂TiSiO₅. The regular diffraction spots (Figure 1e inset) reveal the single-crystalline characteristic of the Li₂TiSiO₅ primary particles and can be assigned to the P4/nmm space group. The content of coated-carbon in the C-Li₂TiSiO₅ nanowires is determined to be 9.43 wt% from the TGA result (Fig. S2).

The CV curves after the first cycle are well overlapped at 0.1 mV s⁻¹ (Figure 2a), indicating the high reversibility after the initial activation. The cathodic peak located at 0.66 V in the first cycle is resulted from the solid-electrolyte-interface (SEI) formation. The cathodic/anodic peak-pairs located at 0.23/0.69 V can be based on the Ti²⁺ ↔ Ti⁴⁺ redox reactions [16].

Figure 2. Electrochemical performance of the C-Li₂TiSiO₅ nanowires. a) CV curves for first four cycles at 0.1 mV s⁻¹, b) first four-cycle GCD curves at 50 mA g⁻¹, c) GCD curves at various current densities, d) rate performance, and e) cyclability of the Li₂TiSiO₅/Li half cell.

The galvanostatic charge – discharge (GCD) curves (Figures 2b,c) of the C-Li₂TiSiO₅/Li half cell at various current densities show very similar shapes, confirming the high reversibility during lithiation and delithiation after the first activation. At 50 mA g⁻¹, the reversible capacity of the C-Li₂TiSiO₅ nanowires is as large as 257 mAh g⁻¹, and the operation potential is averaging at an appropriate value of 0.83 V (Figure 2b). This material retains large reversible capacities of 257, 240, 235, 226, 216, 203, and 184 mAh g⁻¹ (Figure 2d) at 50, 100, 200, 500, 1000, 2000, and 5000 mA g⁻¹. When the current rate returns to 50 mA g⁻¹, the reversible capacity recovers to 246 mAh g⁻¹, showing its good electrochemical stability. The 5000 mA g⁻¹ vs. 50 mA g⁻¹ capacity ratio of 71.6% indicates its high rate performance. When the half cell is cycled at 5000 mA g⁻¹ over 4000 cycles (Figure 2e), the capacity retention is up to 90.0%, revealing its excellent cyclability. Therefore, the electrochemical performance of the Li₂TiSiO₅ material in this work is among the best results in the Li₂TiSiO₅ research field (Table S3).

The ex-situ XPS characterizations reveal that the redox mechanism of the C-Li₂TiSiO₅ nanowires is based on the Ti⁴⁺ ↔ Ti²⁺ redox reaction. The Ti-2p peaks of the pristine C-Li₂TiSiO₅ nanowires (Figure 3a) consist of Ti-2p_1/2 (464.3 eV) and Ti-2p_3/2 (458.6 eV) respectively, which correspond to Ti⁴⁺ [25]. When lithiated to 0.2 V (Figure 3b), the Ti valence is changed to a combination of Ti³⁺ (19%) and Ti²⁺ (81%) [26]. When delithiated to 3.0 V (Figure 3c), the Ti valence is recovered to a combination of Ti⁴⁺ (76%) and Ti³⁺ (23%). Therefore, the active Ti³⁺/Ti⁴⁺ and Ti²⁺/Ti³⁺ redox couples in Li₂TiSiO₅ are identified. The fact that a small portion of Ti³⁺ ions are not oxidated to Ti⁴⁺ during delithiation is reasonable since a small portion of inserted Li⁺ ions during initial delithiation are trapped in the Li₂TiSiO₅ lattice [27–29].

Figure 3. XPS examinations and GITT experiment of the C-Li₂TiSiO₅ nanowires. Ex-situ Ti-2p XPS spectra of a) pristine, b) lithiated, and c) delithiated states. d) CV curves at various scan rates. e) Calculation of b values. f) D_Li variations during lithiation and delithiation.

As the scan rate increases (Figure 3d), the small shifts of the CV peaks demonstrate the good electrochemical kinetics with low polarization. To further study the electrochemical kinetics, the recorded CV data is analysed based on the equation of I = av^b (Figure 3e). In this equation, I and v respectively represent the peak current and scan rate, and a and b are changeable parameters [30–32]. The b value should be in a range of 0.5 − 1, and the larger b value indicates the larger contribution of the capacitive behaviour. In the case of the C-Li₂TiSiO₅ nanowires, the b values are as large as 0.852 during lithiation and 0.911 during delithiation, suggesting that the capacitive behaviour remarkably contributes to the lithium storage, which can be mainly due to the large specific surface area of the C-Li₂TiSiO₅ nanowires.

To characterize the Li⁺ diffusivity of the C-Li₂TiSiO₅ nanowires, a GITT experiment is conducted at 0.1C within 0.2–3.0 V. Fig. S3a presents the GITT curves of the C-Li₂TiSiO₅/Li half cell after the first-cycle activation, and a single typical GITT step is clearly shown in Fig. S3b. According to the Fick’s second law, the apparent Li⁺ diffusion coefficients (D_Li) at different states can be determined according to Eq. (1) [33,34]:

(1) $D_{\mathrm{L}\mathrm{i}}=\frac{4}{\pi }{\left(\frac{m_{\mathrm{b}}{V}_{\mathrm{m}}}{M_{\mathrm{b}}S}\right)}^2{\left(\frac{\mathrm{\Delta }{E}_{\mathrm{s}}}{\tau \left(d{E}_{\tau }/d\sqrt{\tau }\right)}\right)}^2\left(\tau \ll \frac{L^2}{D_{\mathrm{L}\mathrm{i}}}\right)$(1)

In this equation, V_m represents the molar volume of Li₂TiSiO₅, m_b represents the real mass of the C-Li₂TiSiO₅ nanowires, M_b represents the molar mass of Li₂TiSiO₅, τ represents the pulse duration time, S represents the electrode surface area, L represents the electrode thickness, and ΔE_S and ΔE_τ are respectively the changes in the total potential and equilibrium potential during the current pulse. As recorded in Fig. S3c, the voltage during the single titration shows a linear relationship with τ^0.5. Consequently, Eq. (1) is simplified to Eq. (2):

(2) $D_{\mathrm{L}\mathrm{i}}=\frac{4}{\pi \tau }{\left(\frac{m_{\mathrm{b}}{V}_{\mathrm{m}}}{M_{\mathrm{b}}S}\right)}^2{\left(\frac{\mathrm{\Delta }{E}_{\mathrm{s}}}{\mathrm{\Delta }{E}_{\tau }}\right)}^2\left(\tau \ll \frac{L^2}{D_{\mathrm{L}\mathrm{i}}}\right)$(2)

ΔE_S and ΔE_τ are obtained by the GITT curves (Fig. S3b). The D_Li values of the C-Li₂TiSiO₅ nanowires are calculated and recorded in Figure 3f, reaching 2.58 × 10⁻¹² and 2.70 × 10⁻¹² cm² s⁻¹ during lithiation and delithiation, respectively, which are about three orders larger than that of Li₄Ti₅O₁₂ thin films [35]. Apparently, the large Li⁺ diffusion coefficients of the C-Li₂TiSiO₅ nanowires together with their remarkable capacitive behaviour undoubtedly contributes to their good rate performance.

In order to explore the lithium-storage mechanism of the C-Li₂TiSiO₅ nanowires, an in-situ XRD characterization is performed at 0.5C within 0.2–3.0 V [36–38]. The collected in-situ XRD spectra of the first lithiation and delithiation process are recorded in Figures 4a,b. During lithiation, all the XRD peak intensities are obviously decreased since the inserted Li⁺ ions decrease the crystal structure order [37], and the peak intensities almost recover during delithiation, which suggests the excellent structural reversibility and stability of the C-Li₂TiSiO₅ nanowires. The negligible peak shifts during lithiation and delithiation indicate the tiny lattice-constant changes (Figures 4c,d), revealing the Li⁺ intercalation mechanism and the ‘zero-strain’ behaviour of the C-Li₂TiSiO₅ nanowires.

Figure 4. In-situ XRD characterization of the C-Li₂TiSiO₅ nanowires. a) Contour in-situ XRD spectra and b) original in-situ XRD spectra. Enlarged contour in-situ XRD spectra within c) 20.0–20.4° and d) 34.4–34.7°. e) Variations in lattice constants of Li₂TiSiO₅.

The tiny and reversible lattice-constant changes during lithiation – delithiation are confirmed by refining the collected in-situ XRD spectra (Figure 4e). After lithiation at 0.2 V, the a and c values slightly increases by 0.22 and 0.10%, respectively. Consequently, the V value shows a similar variation trend with an increase of 0.41%. This V-value change is among the smallest percentages in the research field of lithium-storage materials (Table S4), thereby leading to the excellent cyclability stability of the C-Li₂TiSiO₅ nanowires.

To explore the reason for the ‘zero-strain’ behaviour of Li₂TiSiO₅, its crystal-structure-change details at delithiated and lithiated states are carefully analysed by refining the collected in-situ XRD data, and the obtained bond-length, atomic-parameter, and polyhedron-volume changes are listed in Table S5–S9. After the lithiation with the external Li⁺ ions from the electrolyte intercalated into the 2a and 4e sites, no movements of the Li (4d) and Si⁴⁺ ions are observed, while the active Ti ions show a slightly shift along the c axis. Consequently, the Ti – O (8i), Li (4d)–O (2c), and Si – O (8i) bond lengths increase, and the Li (4d)–O (8i) and Ti – O (2c) bond lengths decrease. The O²⁻ (8i) ions obviously shift to the inserted Li⁺ (2a) ions, and the O²⁻ (2c) ions shift to the inserted Li⁺ (4e) ions. The lithiation expands the active TiO₆-octahedron volume by 1.72% since the Ti⁴⁺ ions are converted to Ti³⁺/Ti²⁺ ions with larger ion sizes [39], but slightly expands the inactive LiO₆-octahedron volume by 1.16% and the SiO₄-tetrahedron volume by 0.91%. Clearly, the inactive LiO₆-octahedra and SiO₄-tetrahedra have excellent volume-buffering capability, effectively buffering the volume expansions of the surrounded TiO₆ octahedra. In addition, the interconnected polyhedron network can also effectively accommodate the expansions of LiO₆, TiO₆, and SiO₄ polyhedra. Therefore, the ‘zero-strain’ behaviour of Li₂TiSiO₅ is achieved.

Conclusions

The practical capacity and rate performance of Li₂TiSiO₅ are improved by carbon-coating and nanosizing dual modifications. The carbon-coated Li₂TiSiO₅ nanowires show wire diameters of 100 − 200 nm and carbon percentage of 9.43 wt%. The Ti⁴⁺ ↔ Ti²⁺ redox reaction provides a large reversible capacity of 257 mAh g⁻¹ within 0.2–3.0 V at 50 mA g⁻¹ and an appropriate average operation potential of 0.83 V. The fast Li⁺ transport and remarkable capacitive behaviour effectively contribute to the good rate performance with a 5000 mA g⁻¹ vs. 50 mA g⁻¹ capacity ratio of 71.6%. The excellent cyclability with 90.0% capacity retention at 5000 mA g⁻¹ over 4000 cycles is undoubtedly due to the ‘zero-strain’ behaviour of Li₂TiSiO₅. The obviously expanded active TiO₆ octahedra are surrounded by slightly expanded inactive LiO₆ octahedra and SiO₄ tetrahedra, resulting in a tiny lattice-volume expansion of only 0.41%. This work provides new insights for exploring high-performance lithium-storage materials and more ‘zero-strain’ materials.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/10667857.2023.2204292.