Fei Li,
Jianzhong Jiang,
Xinjing Wang,
Fan Liu,
Jinzuan Wang,
Yanwei Chen,
Sheng Han* and
Hualing Lin*
School of Chemical and Environmental Engineering, Shanghai Institute of Technology, Shanghai 201418, China. E-mail: hansheng654321@sina.com; Tel: +86 13524694909
First published on 17th December 2015
Three-dimensional (3D) TiO2–graphene frameworks (TGFs) with macroporous architecture were fabricated through the in situ synthesis of TiO2 with the participation of graphene oxide followed by hydrothermal assembly. TGFs exhibited a 3D hierarchical porous architecture with mesopores (2.4 nm), macropores (10–30 μm) and a large specific surface area (196 m2 g−1), which not only provided contacts between the electrode material and the electrolyte but also increased the mass transport of Li-ions in the charge/discharge process. When it was used as a cathode material in Li-ion batteries, TGFs presented an excellent reversible specific capacity of 210 mA h g−1 at 100 mA g−1 and an outstanding reversible cycling stability (111 mA h g−1 after 500 cycles); even at the current density of 500 mA g−1, TGF also performed very well. This excellent electrochemical performance was attributed to the unique 3D hierarchical porous architecture and the synergistic effects of TiO2 and graphene in Li-ion storage and transport.
Graphene, with a two-dimensional (2D) single carbon atom layer, has attracted tremendous attentions because of its excellent electrical conductivity (∼15000 cm2 V−1 s−1), large specific surface area (∼2600 m2 g−1) and extraordinary mechanical properties.25,26 Furthermore, the flexible honeycomb structure provides conduction channels for electron transport and allows other nanoscale materials to fill the clearance space. Several TiO2/graphene nanocomposites (TGNs) that showed superior Li-storage performance have been reported.27,28 However, these TGNs reported were simply prepared by decorating TiO2 NPs or nanocrystals with specific morphologies onto the graphene or forming 2D TGN paper. Moreover, 3D graphene-based frameworks integrate individual 2D graphene sheets into continuous structures. Therefore, such 3D hybrids not only contain the ascendant properties of single component but also translate the intrinsic features into the macroscopic scale. In particular, the construction of 3D TiO2–graphene frameworks (TGFs) with macroporous architecture exhibit an interconnected graphene network to promote electrolyte infiltration, carriage of charge carriers and adjustment of electrode volume changes. Thus, TGF construction is a promising strategy to achieve highly reversible capacity and excellent rate performance for LIBs.29–31
In this study, we prepared a novel self-assembly approach to construct 3D macroscopic TGFs as anode materials for LIBs. The 3D TGFs were synthesised by a two-step procedure, in which TiO2 was first decorated on graphene oxide (GO) sheets under mild conditions (80 °C) with stirring and then through hydrothermal treatment to establish 3D macroscopic frameworks. The obtained 3D TGFs showed high specific surface area, numerous macropores with diameters to be dozens of micrometres, and a large aspect ratio. In particular, this unique hybrid 3D architecture effectively accommodated the excessive volume change during the lithiation–delithiation process and provided numerous multidimensional channels for Li-ion diffusing from the electrolyte to the electrode. Serving as the anode materials for LIBs, TGFs not only exhibited an excellent reversible capacity of 210 mA h g−1 with up to 100 charge–discharge cycles at 100 mA g−1 but also maintained specific capacities of 82 mA h g−1 at a ultrahigh current density of 5000 mA g−1.
The morphologies and microstructure of the as-prepared TGF were first investigated using FESEM and TEM (Fig. 2). The FESEM images of a cross-section of TGF (1:
1) exhibited a highly interconnected 3D TGFs, which contained uniform macropores with diameters ranging from 10 μm to 30 μm (Fig. 2a). The high-resolution FESEM images of TGF apparently exposed that TiO2 NPs were consistently loaded on both sides of the graphene sheets without free particles or unloaded graphene sheets (Fig. 2b). Elemental mapping images of TGF confirmed the presence of Ti and O components in TGF (Fig. 2c), which further revealed that TiO2 homogeneously adhered to the surface of graphene sheets. Moreover, the high crystallinity of TiO2 shown in the TEM images demonstrated that TiO2 NPs with diameters of 5–7 nm were deposited on the graphene sheet surface (Fig. 2d). As shown in Fig. 2e, the d-spacing of the lattice fringes was about 0.34 nm, which could be identified by the interlayer spacing of the (101) plane in anatase TiO2 crystal lattice in accordance with the results of the XRD measurements.33
![]() | ||
Fig. 2 (a) and (b) FESEM images of TGF (1![]() ![]() |
Powder XRD experiment was conducted to gain the internal structure of the TGF, RGO and GO (Fig. 3a). The obvious peaks at about 2θ = 25.1°, 37.5°, 47.8°, 53.9°, 62.3°, 69.1° and 74.8° could be attributed to the (101), (004), (200), (105), (213), (116) and (215), respectively, which were well-indexed to the anatase TiO2 (JCPDS card 75-1537), demonstrating that GO application didn't influence the crystal fabrication of TiO2. At 26.6°, no apparent diffraction peak was observed, suggesting that no further agglomeration of few-layered RGO sheets hindered by TiO2 NPs and that the graphene sheets were effectively separated and highly dispersed into the TiO2 matrix.34
The Raman spectra of TGF, TiO2 and GO were shown in Fig. 3b. The peaks at 153, 202, 394, 505 and 631 cm−1 were attributed to the Raman modes of Eg (1), Eg (1), B1g (1), A1g and Eg (3), respectively, showing increased broadening and systematic frequency shifts in comparison to bulk anatase TiO2. Such Raman signal broadening and frequency shifting trends on the dimension was less than the typical 20 nm, supported by the TEM image.35,36 The G band (about 1595 cm−1) identified with the in-plane vibration of sp2-hybridised carbon and the D band (about 1345 cm−1) originating from chaotic carbon owing to RGO.37 After hydrothermal treatment, the D/G intensity ratio of the TGF composite (ID/IG = 1.02) was larger than that of the GO (ID/IG = 0.95), which indicated a decrease in the average size of the sp2 domains and an increase in the disorder degree of the graphene.38
TGA (in air from 25–800 °C with a heating rate of 20 °C min−1) was performed to determine the chemical composition of TGFs (1:
2), (1
:
1) and (2
:
1) (Fig. 3c). As shown in the TGA results, the TGFs lost weight in three steps; first, between 25 and 100 °C; then, between 100 and 400 °C; and finally, between 400 and 550 °C. The weight loss (<5%) appeared below 150 °C was attributed to the evaporation of adsorbed water molecules. The major weight loss from 100 °C to 550 °C was owing to the graphene combustion. In particular, the TGF curve showed a significant weight loss at approximately 400 °C and a constant weight at above 500 °C, indicating that TiO2 could improve the thermal stability of the graphene by conglutinating on its surface. Based on the calculations, the TiO2 contents in TGFs (1
:
2), (1
:
1) and (2
:
1) were approximately 61.9%, 72.2% and 76.2%, respectively.
The porous features of TGF were further investigated using the N2 adsorption–desorption analysis (Fig. 3d). The BET surface area of TGF (1:
2), (1
:
1) and (2
:
1) nanocomposites were calculated to be as high as 220, 196 and 168 m2 g−1, respectively. The specific surface area of the TGF nanocomposites was mainly due to the contribution of graphene, which highlighted that the building up of 3D frameworks was an effective way to achieve a high specific surface area for hybrid materials. Furthermore, the adsorption–desorption isotherm of all the three samples exhibited the typical-type IV nitrogen adsorption branch with an H2 hysteresis loop, suggesting the existence of mesoporous structures in the hybrid.39 The total pore volume was 0.1288, 0.1174 and 0.1082 cm3 g−1 with an average pore diameter to be 2.34, 2.40 and 2.57 nm, respectively, based on the Barrett– Joyner–Halenda calculations. Such porous structure could not only provide extreme contacts between the electrode material and the electrolyte but also enhance the mass transport of Li-ions in charge/discharge process, which was crucial to the rate capability of LIBs.
X-ray photoelectron spectroscopy (XPS) was employed to gain the insight chemical bonding environment of Ti, O, and C atoms within TGF (take the TGF (1:
1) for example). The peaks of Ti, O, and C elements peaks could be observed in the full spectra of TGF (Fig. 4a). The binding energy (BE) peaks of Ti 2p1/2 and Ti 2p3/2 for TGF was positioned at 459.1 and 464.8 eV (Fig. 4c) with a peak separation of 5.7 eV. These peak locations and separation were in excellent agreement with the reported values for fully oxidized anatase TiO2 single crystals.40,41 Additionally, the components peaks of TGF at 530.3 eV (O–Ti), 531.7 eV (C
O) and 532.8 eV (C–O) could be divided from the O 1s spectra for the compositions (Fig. 4d).42 The C 1s XPS spectrum of GO (Fig. 4b) could be deconvoluted into three peaks (centered at 284.8 eV, 286.9 eV and 288.6 eV), which were ascribed to sp2 bonded carbon (C–C), epoxy/hydroxyls (C–O), and carboxyl (O–C
O), respectively, indicating the high percentage of oxygen containing functional groups.43,44 In comparison, in the C 1s XPS spectrum of the TGF (Fig. 4b), the peaks for C–C, C–O and O–C
O were much lower in intensity than those in GO, indicating the efficient deoxygenation and reduction during the hydrothermal process.45
![]() | ||
Fig. 4 (a) XPS spectrum of TGF and GO; (b) XPS C 1s spectrum of the TGF and GO; (c) Ti 2p and (d) O 1s XPS spectra of TGF. |
Cyclic voltammetry (CV) was performed at a scan rate of 0.1 mV s−1 in the voltage range of 1.0–3.0 V to investigate the electrochemical reactivity of the TGF (1:
1) hybrids, as shown in Fig. 5a. The CV curves clearly observed from the TGF (1
:
1) showed a pair of redox peaks, including a reduction peak at 1.72 V and an oxidation peak at 1.99 V, which were corresponding to the Li insertion/extraction in anatase TiO2 lattice.46,47 As it was showed in Fig. 5a, the cathodic and corresponding anodic peaks only displayed a small change during the three initial CVs, suggesting little capacity loss. This result indicated that the TGF (1
:
1) nanocomposite showed a higher reversibility in electrochemical reactions.
The electrochemical performances of TGF (1:
1) were evaluated by galvanostatic charge–discharge cycling in a cell using Li metal as the counter electrode, in the voltage range of 1.0–3.0 V. Fig. 5b exhibited the charge–discharge curves of TGF (1
:
1) at a current density of 100 mA g−1, in which discharge and charge potentials stabilised at ∼1.73 and ∼1.98 V appeared, and these curves were consistent with the CV results. The first discharge and charge cycle capacities were 457 and 236 mA h g−1, respectively, which corresponded to an irreversible capacity loss of 48%. This strong reduction capacity was mainly caused by the electrolyte decomposition on the new surface of the carbon material and the formation of a solid electrolyte interphase (SEI) layer on the electrode surface during the first discharge,24 which could also be confirmed by electrochemical impedance spectroscopy (EIS) (see details below). For the second cycle, the charge and discharge capacities were 247 and 222 mA h g−1, indicating a reduced irreversible capacity loss of only about 10%. Moreover, the whole cycle charge curve displayed a similar tendency, suggesting that the charge/discharge process was highly reversible.
The long cycling performance of TGF (1:
2), (1
:
1) and (2
:
1) electrodes was compared at a constant current density of 500 mA g−1 between 1.0 and 3.0 V (Fig. 5c). All three samples exhibited excellent stability during the cycle processes, and TGF (1
:
1) displayed the highest capacity. The first charge and discharge capacities of the TGF (1
:
1) electrode were 332 and 195 mA h g−1, respectively, followed by a slight decrease in the capacity during the initial 50 cycles. More importantly, a high capacity of 111 mA h g−1 was maintained until 500 cycles, implying that the long-term charge/discharge cycles did not severely spoil the structure of the TGFs and the high discharge capacity could be well preserved. Comparatively, the reversible capacities of TGF (1
:
2) and (2
:
1) after 500 cycles were 60 and 70 mA h g−1, which were lower than that of TGF (1
:
1).
The high-rate performance of the TGF electrode was also investigated at various current densities between 1.0 and 3.0 V to inspect the possibility for battery applications. As indicated in Fig. 5d, the first discharge–charge step delivered a specific discharge capacity of 462 mA h g−1. When cycled at 100 mA g−1, the TGF (1:
1) delivered a second discharge capacity of 270 mA h g−1 and dropped to near 222 mA h g−1 after 10 cycles. During the subsequent cycles, the reversible capacities of the TGF (1
:
1) electrode were 185, 150, 125 and 105 mA h g−1 at current rates of 200, 500, 1000 and 2000 mA g−1, respectively. Even at super high current densities of 5000 mA g−1, the homologous recharge capacities of the TGF (1
:
1) electrode retained 82 mA h g−1 with nearly 100% coulombic efficiency, which clearly demonstrated that the TGF (1
:
1) nanocomposite could tolerate varied discharge current densities. Such remarkable rate capability and high capacity of the TGF nanocomposite may be attributed to the high electronic conductivity of graphene and the mesoporous structures of the TGF nanocomposite.43 Moreover, after the high-current density measurements, the capacity of the TGF (1
:
1) nanocomposite at 100 mA g−1 could recover to the initial value (221 mA h g−1), indicating an excellent cycle performance. For comparison, the performance of TGF (1
:
2) and (2
:
1) electrodes were only 132 and 150 mA h g−1, respectively, at a current density of 100 mA g−1 after 80 cycles. The capacity of TGF (1
:
1) achieved in this study was superior to that of the previous reports, in which the TiO2/graphene and other TiO2/carbon composites owned a similar TiO2 weight content.
To gain an insight vision of the prominent electrochemical behavior of the TGF, electrochemical impedance spectroscopy (EIS) measurements of pure TiO2 and TGF electrode were performed (Fig. 6a). Nyquist plots showed only one semicircle in the high-frequency range and an inclined line in the low-frequency range for TGF before cycling and after the 10th cycle, which were assigned to the charge-transfer impedance in the electrode/electrolyte interface and the Li+ ion diffusion process, respectively. However, there were two semicircles and a line which could be found in the EIS profile indicating the formation of solid–electrolyte-interphase (SEI) films after the 1st cycle.48,49 After the 10th cycle, the diameter of the semicircle for TGF was much smaller than that for pure TiO2, which suggested that TGF possessed lower contact and charge-transfer resistances. This result could be further supported by simulating their kinetic parameters through the typical Randles equivalent circuit (Fig. 6b). In contrast, Rct value of the TGF electrode was 62.29 Ω which was significantly lower than those of pure TiO2 electrode (137.01 Ω). Furthermore, the linear Warburg regions at low frequency in Nyquist plots could be used to represent the Li-ion diffusion behavior and the diffusion coefficient of Li+ also calculated by the following formula (eqn (1)):48
![]() | (1) |
This journal is © The Royal Society of Chemistry 2016 |