Jun Jina,
Xiao-Ning Rena,
Yi Lua,
Xian-Feng Zhenga,
Hong-En Wanga,
Li-Hua Chena,
Xiao-Yu Yanga,
Yu Li*a and
Bao-Lian Su*abc
aLaboratory of Living Materials at the State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, 122 Luoshi Road, 430070, Wuhan, Hubei, China. E-mail: yu.li@whut.edu.cn; baoliansu@whut.edu.cn
bLaboratory of Inorganic Materials Chemistry (CMI), University of Namur, 61 rue de Bruxelles, B-5000 Namur, Belgium. E-mail: bao-lian.su@unamur.be
cDepartment of Chemistry and Clare Hall, University of Cambridge, Cambridge CB21 EW, UK. E-mail: bls26@cam.ac.uk
First published on 18th July 2016
Hierarchically porous TiO2/carbon hollow spheres (TiO2/C-HS) have been designed and prepared through a facile one-pot template-free hydrothermal route using sucrose as a carbon source, TiO2 solid spheres as a TiO2 source and NH4F as a structure-directing reagent. The nanocrystal constructed hierarchically porous hollow spherical structure offers enough space for electrolyte penetration and storage and a short path length for Li+ diffusion and e− transport. The carbon layer on TiO2 surface improves its conductivity as well as the structure stability. As a result, such a special hollow structure with carbon layers exhibits enhanced lithium storage properties comparing with the solid spheres. The TiO2/C-HS anode exhibits discharge capacities of 286, 235, 197, 164 and 127 mA h g−1 at various rates of 0.2, 0.5, 1, 2 and 5C (1C = 168 mA g−1), respectively. A capacity of 175 mA h g−1 still remains after 200 cycles at 1C, demonstrating a very high lithium insertion coefficient of 0.52, a little higher than the theoretical value of 0.5. SEM, TEM, HRTEM and electrochemical impedance spectra (EIS) techniques have been utilized to understand the Li+ insertion process and structural stability. Our results reveal that the high electrochemical performance of the TiO2/C-HS anode can be attributed to the synergy of the hierarchically porous hollow structure, carbon layer and newly formed numerous ∼5 nm Li2Ti2O4 on the surface of the TiO2 nanocrystals.
TiO2 with various crystalline phases has been widely investigated as a anode material in LIBs.5–7 TiO2 is not only a fast and low voltage insertion host for Li, but also an abundant, low cost, and non-toxic electrode.8 Thus, during the Li insertion/extraction, TiO2 has strong oxidation capability, chemical and structural stability.9 The electrochemical performance of TiO2 strongly relies on its crystalline phases, crystallite size, crystal structure, specific surface area and thermal stability.10,11 Among the various polymorphs, anatase is generally considered as the most electro-active Li-insertion host.12–15
Mesoporous TiO2 electrode material, especially hollow spherical porous TiO2, with a high specific surface area would reduce the diffusion pathway for electronic and ionic transportation, provide more sites for Li+ insertion and allow facile Li+ diffusion at high charge/discharge rate,11,16–18 Hollow spherical porous structure with high surface area, low density, and high loading capacity can improve the reversible lithium storage capabilities.19 First, the hollow porous structure is often associated with abundant active sites for Li+ storage and effective short distance for Li+ diffusion, resulting in good rate performance. Second, the mesoporous channels in the shell provide continuous transporting path for Li+ diffusion.20–23 As a result, the cavities in hollow structure can provide extra sites for Li+ storage, enhancing the specific capacity. Further, the cavities in hollow structure can buffer against the volume change during the lithium insertion/extraction and aggregation of the electrode materials.24 In addition, conductive materials doping is also an effective path to enhance the electrochemical performance of TiO2.25–27 For example, carbon coating is a widely adopted approach to enhance the conductivity of TiO2 nanostructures.28–30
In this work, we have designed and prepared TiO2/carbon hollow spheres (TiO2/C-HS) by one-pot template-free strategy based on inside-out Ostwald ripening though a facile hydrothermal method using sucrose as carbon source, TiO2 solid spheres as TiO2 source and NH4F as structure-directing reagent. Our results show that the carbon layers (∼6 nm) have been successfully coated on TiO2 surface. As expected, such special hollow structure with carbon layers exhibits enhanced lithium storage properties, especially for high rate performance due to its high surface area, hollow structure, carbon layers and newly formed Li2Ti2O4 nanocrystals on TiO2 nanoparticles.
The morphologies of the samples are investigated by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Fig. S1† display the initial TSS sample, which are monodisperse nanospheres with sizes of 300–500 nm. After hydrothermal treatment of TSS spheres for 10 h, the product is solid spheres with rough surface (Fig. 2a), indicating the recrystallization of TiO2. Worm-like mesopores among the nanoparticles are obviously observed as shown in Fig. 2b. The nanoparticles have a size ∼10 nm. The lattice spacing (Fig. 2c) is measured of 0.35 nm, corresponding to the anatase (101) crystal plane.
![]() | ||
Fig. 2 The SEM, TEM and HRTEM images of (a–c) TiO2-MS, (d–f) TiO2/C-MS, (g–i) TiO2-HS and (j–l) TiO2/C-HS. |
When sucrose is added in the reaction system, the spherical morphology of the product almost remains unchanged (Fig. 2d and e). In the HRTEM image (Fig. 2f), crystalline planes of TiO2 (101) with a distance spacing of 0.35 nm is observed and an amorphous carbon layer is wrapped on TiO2 nanoparticles. And the color of the TiO2/C-MS becomes black (Fig. S2†). When NH4F is added in the reaction system, the product retains spherical morphology. However, its surface is much rougher than that of TiO2-MS obtained without NH4F (Fig. 2g) due to the F− directing the crystallites growth. The broken spheres indicate that the TiO2 sphere is hollow structure with uniform ∼50 nm shell. The darker shell in Fig. 2h verifies the nanoparticles constructed hollow structure with a shell thickness of ∼50 nm, in consistent with SEM observations. HRTEM image reveals that the TiO2 nanoparticles are in nano-spindle shape (Fig. 2i). The lattice spacing is measured to be 0.35 nm, corresponding to the anatase (101) crystal plane. At the beginning of reaction, the outer surface layer of TSS crystallizes due to its directly contacting with the surrounding solution until the equilibrium of reaction system. However, the amorphous core remains out of equilibrium and has a strong tendency to dissolve and recrystallize owing to its higher solubility and higher surface energies compared to the outer surfaces. The addition of F− ions can increase the dissolution rate in amorphous core, resulting to the occurrence of hollowing in central region,31–35 as illustrated in Scheme 1. Further, F− can easily transfer from the outside solution to the centre of the spheres through the previous formed pores in the surface and induce the dissolution of TiO2 and direct the recrystallizing process to from spindle shape nanoparticles.36,37
When both NH4F and sucrose are added in the reaction system, the prepared product becomes black powder, comparing to the white colors of TiO2-MS and TiO2-HS (Fig. S2†). The product remains spherical hollow structure (Fig. 2j). The TiO2/C-HS with a shell size of ∼50 nm still consist of nano-spindle crystals (Fig. 2k and l). The SAED pattern in Fig. 2k inset shows a typical diffraction pattern of polycrystalline with good crystallinity. In the HRTEM image (Fig. 2l), crystalline planes of TiO2 (101) with a distance spacing of 0.35 nm is observed and an amorphous carbon layer with a thickness of 6 nm is wrapped on TiO2 spindle nanoparticles. The as-deposited carbon layer forms a 3D interconnected network and helps to stabilize the TiO2/C-HS hollow structure and enhances its electron conductivity, as illustrated in Scheme 1. Thermogravimetric (TG) analysis is then used to determine the carbon content of TiO2/C-MS and TiO2/C-HS. It gives ∼6.0% and ∼7.5 wt% carbon for TiO2/C-MS (Fig. S3†) and TiO2/C-HS (Fig. S4†), respectively.
The specific surface areas and pore size distributions of TiO2-MS, TiO2/C-MS, TiO2-HS and TiO2/C-HS are measured by nitrogen gas adsorption and desorption (Fig. 3). The standard multipoint Brunauer–Emmett–Teller (BET) method is used to calculate the specific surface area. Pore size distributions are obtained from the isotherm adsorption branches based on Barrett–Joyner–Halenda (BJH) model. All the samples exhibit a type-IV N2 adsorption/desorption isotherm (Fig. 3a), indicating the presence of mesopores. The strong hysteresis is related to the capillary condensation associated with the pore channels. The BET specific surface areas of TiO2-MS, TiO2/C-MS, TiO2-HS and TiO2/C-HS are 156.1, 66.4, 196.7 and 77.7 m2 g−1, respectively. TiO2-HS demonstrates the highest surface area. This can be contributed to the adding of F−, which can not only increase the primary particle size but also increase the specific surface area.11 The average pore sizes for the samples are 8.6, 7.7, 7.5 and 7.5 nm, respectively. The pore sizes of TiO2/C-MS, TiO2-HS and TiO2/C-HS are a little decrease comparing to TiO2-MS due to the particles size increase. It is worth mentioning that the mesopores allow easier electrolyte penetration into every part of the hollow structure and provide more sites for lithium ions to enter and allow facile lithium diffusion at higher current densities during the electrochemical charge/discharge reaction.
![]() | ||
Fig. 3 (a) Nitrogen adsorption/desorption isotherms and (b) the pore size distribution plots of TiO2-MS, TiO2/C-MS, TiO2-HS and TiO2/C-HS. |
Our results indicates that TiO2-MS delivers a low capacity of 120 mA h g−1 after 200 cycles at a current density of 1C (1C = 168 mA g−1), as shown in Fig. S5.† After carbon coating, the electrochemical performance of TiO2/C-MS (134 mA h g−1 at 1C) is a little better than that of TiO2-MS, due to the existence of carbon enhancing the conductivity (Fig. S6†). However, both performances are lower than those of TiO2-HS and TiO2/C-HS (see below results). Therefore, we will focus on the electrochemical performances of TiO2-HS and TiO2/C-HS in the following text to show the advantages of hierarchically hollow porous structure and carbon coating.
Fig. 4 depicts the representative CVs of TiO2/C-HS at a scan rate of 0.5 mV s−1 for the first, second and third cycles. Two peaks can be observed at 1.56 V (cathodic sweep) and 2.17 V (anodic sweep) in the curves, in agreement with previous researches.43,44 The peak at 1.56 V corresponds to the phase transition from tetragonal anatase (I41/amd) to orthorhombic Li0.5TiO2 (Imma) when the insertion coefficient of Li+ reaches at ∼0.5.45 From the second cycle, the cathodic peak becomes a little smoother and shift to 1.62 V, suggesting a possible activating process.13 In the third cycle, the position retained unchanged while the intensity becomes weaker compared to those of the second cycle. This observation is quite unusual and important for the electrode. During the anodic sweep, the peak intensity becomes weaker upon cycling while the peak position remains unchanged, indicating the reversibility of lithium ion insertion/extraction process.
Fig. 5a shows the charge/discharge voltage profiles of TiO2-HS and TiO2/C-HS for the first cycle at a current rate of 0.2C (1C = 168 mA g−1). The profiles of TiO2-HS have two voltage plateaus at ∼1.73 V and ∼1.94 V during the discharge and charge processes respectively, with the potential interval of 0.21 V. However, the plateau voltage of TiO2/C-HS for discharge is ∼1.80 V and the plateau voltage of TiO2/C-HS for charge is ∼1.89 V. The potential interval between the two plateaus is 0.09 V, indicating a lower polarization comparing to TiO2-HS. This means that TiO2/C-HS can exhibit better electrochemical performance than TiO2-HS. The discharge curves (e.g. TiO2/C-HS) clearly show that the lithium insertion process can be divided into three stages. The first stage is fast drop in the potential from the open circuit voltage to a plateau of ∼1.80 V, with a capacity of 30 mA h g−1. The second stage is the process of lithium insertion into the vacancy sites of the TiO2 crystal structure, corresponding to the horizontal plateau region, with a capacity of 95 mA h g−1. These are related to the coexistence of two-phase mixture of tetragonal phase and the Li-rich orthorhombic phase. Moreover, the voltage plateau in the discharge curves of TiO2-HS and TiO2/C-HS have almost the same plateau capacities. The final stage is the gradual decrease of the voltage after the plateau stage, indicating further insertion of Li+ into the surface of the electrode material. The lithium insertion process delivers initial discharge capacities of 243 and 286 mA h g−1 for TiO2-HS and TiO2/C-HS, respectively. Subsequently, the lithium extraction process reverses charge capacities of 226 and 262 mA h g−1 for TiO2-HS and TiO2/C-HS, respectively. The irreversible capacity losses of TiO2-HS and TiO2/C-HS are 6.7% and 8.4%, which are much lower than those of other anatase TiO2 electrodes.43,46,47
The charge/discharge curves of TiO2/C-HS with various rates (0.2, 0.5, 1, 2 and 5C) for the first cycle are shown in Fig. 5b. It displays the capacities corresponding to voltage plateau decrease with the current densities increasing. Interestingly, the decreased capacities from the second stage are approximately equal to the total capacities decrease at 0.2, 0.5, 1, 2 and 5C rates. This may be attributed to the faster kinetic of the additional surface capacity than that of insertion capacity.22 The Li+ insertion into the surface of TiO2 (the third stage) is faster than the Li+ insertion into the TiO2 lattice (the second stage), due to the facility of lithium storage in TiO2 surface and the tardiness of Li+ diffusion into TiO2 lattice.
Fig. 5c compares the rate capabilities of TiO2-HS and TiO2/C-HS at various current densities. TiO2-HS has discharge capacities of 243, 193, 171, 145 and 88 mA h g−1 at rates of 0.2, 0.5, 1, 2 and 5C, respectively. And a capacity of 204 mA h g−1 is resumed when the rate backs to 0.2C. In contrast, TiO2/C-HS has discharge capacities of 286, 235, 197, 164 and 127 mA h g−1 at various rates of 0.2, 0.5, 1, 2 and 5C, respectively. Then a capacity of 213 mA h g−1 is resumed while the current density is set back to 0.2C. The differences between capacities of the TiO2-HS and TiO2/C-HS electrodes at relatively low rates (0.2, 0.5 and 1C) are very little. The discharge capacities of TiO2/C-HS are however higher than those of TiO2-HS at relatively high rates (2 and 5C). This can be attributed to the carbon layer in TiO2/C-HS, which is beneficial for electronic conductivity and structure stability, further improving the high rate capability of TiO2/C-HS.26
The cycle performances of TiO2-HS and TiO2/C-HS at a current density of 1C are displayed in Fig. 5d. When charged at a rate of 1C, a lithium ion battery requires ca. 1 h to approach its full charge capacity. TiO2-HS gives an initial discharge capacity of 207 mA h g−1 and a subsequent charge capacity of 194 mA h g−1, leading to an irreversible capacity loss of 6.3%. After more than 100 cycles, a reversible capacity of 152 mA h g−1 is retained. And the charge–discharge capacity is very stable. After 200 cycles, a reversible capacity of 151 mA h g−1 is still remained. This corresponds to a lithium insertion coefficient of 0.45. TiO2/C-HS exhibits an initial discharge capacity of 204 mA h g−1 and a subsequent charge capacity of 193 mA h g−1, leading to a high coulombic efficiency of 94.6%. When the TiO2/C-HS electrode is charged for more than 200 cycles, a capacity of 175 mA h g−1 is retained, leading to a high lithium insertion coefficient of 0.52, higher than many other TiO2 anodes.27,48,49 The high capacity and cycle performances of the TiO2-HS and TiO2/C-HS electrodes are attributed to the high surface area and the presence of mesopores. Therefore, they offer short path lengths for Li+ diffusion and e− transport as well as easy liquid electrolyte penetration. In particular, TiO2/C-HS demonstrates a better performance than TiO2-HS due to the increased conductivity after carbon coating.
Further, the electrochemical impedance spectroscopy (EIS) measurement is utilized to investigate the electrochemical kinetics of TiO2-MS, TiO2/C-MS, TiO2-HS and TiO2/C-HS anode materials. Fig. 6 presents the Nyquist plots and corresponding equivalent circuit of the four anode materials on the charge state after 100 cycles at 1C. All the impedance spectra are composed of a depressed semicircle in the high frequency region and an inclined line in the low-frequency region. The high frequency region of the semicircle is attributed to the charge-transfer impedance (Rct) in the electrode/electrolyte interface, while the low frequency region of the straight line corresponds to the Li+ diffusion into the anode material (Warburg diffusion). The TiO2-MS, TiO2/C-MS and TiO2-HS electrodes demonstrate Rct value of 43, 33 and 40 Ω, respectively. The TiO2-MS and TiO2-HS electrodes demonstrate similar Rct value. Most possibly, the solid sphere of TiO2-MS without carbon coating can not ensure the fully electrolyte permeation. The electrolyte permeation path length of TiO2-MS may be very close to the thickness of TiO2-HS hollow structure. After carbon coating, TiO2/C-MS demonstrates a lower Rct due to the enhanced conductivity and Li+ diffusivity. In particular, the TiO2/C-HS anode shows the lowest Rct value of 27 Ω, due to the existence of carbon layer in the TiO2/C hollow structure. The low charge-transfer impedance and high conductivity facilitate the charge transportation during the anode material, resulting in the better electrochemical performance.
![]() | ||
Fig. 6 Electrochemical impedance spectra of the TiO2-MS, TiO2/C-MS, TiO2-HS and TiO2/C-HS anodes on the charged state after 100 cycles at 1C. |
In order to understand the structure stabilities of TiO2-MS, TiO2/C-MS, TiO2-HS and TiO2/C-HS, we have carried out the postmortem studies. In this study, the four anode materials after 200 charge–discharge cycles at 1C are removed from the unit and immersed in the acetonitrile solution for a week to wash off the electrolyte. Fig. 7 presents the SEM images of TiO2-MS, TiO2/C-MS, TiO2-HS and TiO2/C-HS electrodes. It can be seen that the overall hollow spherical structure is generally retained, demonstrating the good structure stability of TiO2-MS, TiO2/C-MS, TiO2-HS and TiO2/C-HS. This leads to the long cycle performances and high rate performances of TiO2-HS and TiO2/C-HS. Fig. 8 shows the TEM, HRTEM and SAED images of the TiO2/C-HS anode material after 200 cycles at 1C. Fig. 8a clearly displays the well-kept hierarchically hollow structure after the electrochemical reaction, confirming the structural and electrochemical stabilities of the TiO2/C-HS materials. Interestingly, plenty of crystalline dots with ∼5 nm diameter are randomly distributed on the surface of TiO2/C-HS. The lattice spacing distances are measured to be 0.351 and 0.209 nm (Fig. 8b and c), corresponding to the anatase (101) and Li2Ti2O4 (400) crystal plane, respectively, according the previous reports45 and our recent studies.13,19,50 Fig. 8c insert shows the corresponding SAED pattern taken from one TiO2/C-HS sphere after 200 cycles. The electron diffraction patterns can be indexed to anatase and Li2Ti2O4, respectively, further verifying the existence of the Li2Ti2O4 nanocrystallites. Therefore, according to principal reaction mentioned above, the continuous lithium insertion in LixTiO2 leads to the atomic rearrangement to partially form the new Li2Ti2O4 nanocrystallites with one Li inserted per formula unit of TiO2, as schemed in Fig. 8d. The formation of Li2Ti2O4 nanocrystallites can further enhance the insertion capacity and electrochemical performance.13,14,19
![]() | ||
Fig. 7 SEM images of (a) TiO2-MS, (b) TiO2-HS and (c–d) TiO2/C-HS after 200 charge–discharge cycles at 1C. |
Footnote |
† Electronic supplementary information (ESI) available: SEM image of TSS; photographs of the three samples; TG and DSC curves of TiO2/C-HS; electrochemical properties of TiO2-MS and TiO2/C-MS. See DOI: 10.1039/c6ra14895f |
This journal is © The Royal Society of Chemistry 2016 |