- Email: [email protected]
Highly-crystalline ultrathin gadolinium doped and carbon-coated Li4Ti5O12 nanosheets for enhanced lithium storage
Journal of Power Sources 295 (2015) 305e313
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Highly-crystalline ultrathin gadolinium doped and carbon-coated Li4Ti5O12 nanosheets for enhanced lithium storage G.B. Xu a, L.W. Yang a, b, *, X.L. Wei a, J.W. Ding a, J.X. Zhong a, P.K. Chu b, ** a b
Hunan Key Laboratory of Micro-Nano Energy Materials and Devices, School of Physics and Optoelectronics, Xiangtan University, Hunan 411105, China Department of Physics and Materials Science, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, China
h i g h l i g h t s Ultrathin Gd3þ doped and carbon-coated Li4Ti5O12 nanosheets are prepared. The nanosheets have single-crystal nature with a thickness of about 10 nm. The nanosheets have high electrical conductivity and large Liþ diffusion coefficient. The benefits stem from synergistic effects by ultrathin structure, doping and carbon coating. The nanosheets demonstrate superior lithium storage performance.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 April 2015 Received in revised form 7 June 2015 Accepted 25 June 2015 Available online xxx
Highly-crystalline gadolinium doped and carbon-coated ultrathin Li4Ti5O12 (LTO) nanosheets (denoted as LTO-Gd-C) as an anode material for Li-ion batteries (LIBs) are synthesized on large scale by controlling the amount of carbon precursor in the topotactic transformation of layered ultrathin Li1.81H0.19Ti2O5$xH2O (H-LTO) nanosheets at 700 C. The characterizations of structure and morphology reveal that the gadolinium doped and carbon-coated ultrathin LTO nanosheets have high crystallinity with a thickness of about 10 nm. Gadolinium doping allows the spinel LTO products to be stabilized, thereby preserving the precursor's sheet morphology and single crystal structure. Carbon encapsulation serves dual functions by restraining crystal growth of the LTO primary nanoparticles in the LTO-Gd-C nanosheets and decreasing the external electron transport resistance. Owing to the synergistic effects rendered by ultrathin nanosheets with high crystallinity, gadolinium doping and carbon coating, the developed ultrathin LTO nanosheets possess excellent specific capacity, cycling performance, and rate capability compared with reference materials, when evaluated as an anode material for lithium ion batteries (LIBs). The simple and effective strategy encompassing nanoscale morphological engineering, surface modification, and doping improves the performance of LTO-based anode materials for high energy density and high power LIBs applied in large scale energy storage. © 2015 Elsevier B.V. All rights reserved.
Keywords: Lithium ion battery Lithium titanium oxide Ultrathin nanosheets Lanthanide doping Carbon coating
1. Introduction The increasing demand for consumer electronics and green transportation spurs the development of intercalated compounds especially for high-performance lithium-ion batteries (LIBs) that
* Corresponding author. Hunan Key Laboratory of Micro-Nano Energy Materials and Devices, School of Physics and Optoelectronics, Xiangtan University, Hunan 411105, China. ** Corresponding author. E-mail addresses: [email protected] (L.W. Yang), [email protected] (P.K. Chu). http://dx.doi.org/10.1016/j.jpowsour.2015.06.131 0378-7753/© 2015 Elsevier B.V. All rights reserved.
can store and deliver energy efficiently [1e4]. Spinel Li4Ti5O12 (LTO) has been demonstrated to be promising anode materials in highperformance LIBs because of its flat charging/discharging plateau at a high potential of 1.55 V vs. Li/Liþ, zero-strain feature towards lithium insertion/extraction, and excellent environmental benignity [5e10]. Unfortunately, practical application of spinel LTO in hybrid electric vehicles and large-scale energy storage has been hampered because of the unsatisfactory rate capability which cannot meet the demand for a high power density due to kinetic problems associated with the poor electrical conductivity (ca.1013 S cm1) and small lithium diffusion coefficient (ca. 109 to 1013 cm2 s1) [8,11e13]. To overcome these hurdles, many
306
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
strategies have been made to ameliorate the rate capability of LTO, for instance, developing various nano/microstructured materials, elemental doping, and surface modification with highly conductive additives. Lou et al. have synthesized mesoporous LTO hollow spheres and observed relatively high rate performance of 104 mA h g1 at 20 C [14]. Isovalent and aliovalent doping (e.g., Br6þ, Zr4þ,V5þ, Zn2þ, Al3þ and Sc3þ) have been observed to increase the electronic conductivity of LTO [8,15e22]. Various carbonmodified spinel LTO composites have attained more than 90% of the theoretical capacity and superior rate capability [11,13,14,23e39]. In particular, by using flexible current collectors such as carbon nanotubes and fibers, flexible LTO composite electrodes have been obtained [11,40]. In addition, combining LTO with conducting metallic nanoparticles can also improve the rate performance of LTO electrodes [41e44]. These improvements are noteworthy, but still insufficient for power and flexibility-oriented applications. It is thus highly desirable to further address the conductivity deficiency of pristine LTO via new composition modulations and nanoengineering combination strategies [30]. Among the various architectures, two-dimensional (2D) nanosheets with a thickness below 10 nm hold great promise in highperformance LIBs because of the short paths enabling fast lithium ion diffusion, large exposed surface, as well as abundant lithium insertion channels [12,28,45e51]. Ultrathin LTO nanosheets are also believed to exhibit a pseudocapacitive effect as the interaction takes place on the surface thus leading to the improved rate capability [48,51]. On the other hand, the construction of a stable separating thin carbon coating around LTO has been reported to be another effective strategy to improve the electrochemical performance of LTO electrodes due to enhancement in the charge transfer kinetics in the LTO electrodes and interfacial stability between the LTO and electrolyte [27,29,50]. Moreover, according to literature, doping lanthanide (Ln3þ) ions (such as La3þ, Y3þ and Gd3þ) with lower valence state yields oxygen ion vacancies in LTO, which could behave as ionic charge carriers to enhance intrinsic conductivity, thereby improving long-term cyclability and rate capabilities [15]. Despite recent research on LTO NSs, LTO with carbon encapsulation, and Ln3þ doped LTO, there have been few studies on highlycrystalline Ln3þ doped and carbon-coated ultrathin Li4Ti5O12 (LTO) nanosheets. Compared to the conventional solid-state reaction, hydrothermal, or solegel method to prepare LTO nano/microstructures, large scale synthesis of highly-crystalline ultrathin LTO nanosheets with a thickness of 10 nm is quite challenging. In our previous reports, we found that using trivalent lanthanide (for example, Gd3þ) to substitute for quadrivalent titanium ions allows the final spinel LTO products to be stabilized, thereby preserving the precursor's sheet morphology via the topotactic transformation of ultrathin layered Li1.81H0.19Ti2O5$xH2O (H-LTO) at 700 C [52]. In this paper, by controlling the amount of carbon precursor (D(t)glucose monohydrate) in the topotactic transformation, large scale synthesis of ultrathin Gd3þ doped and carbon coated LTO nanosheets (denoted as LTO-Gd-C) with high crystallinity and a thickness below 10 nm is accomplished. Owing to the synergistic effects rendered by ultrathin nanosheets with high crystallinity, Gd3þ doping and carbon coating, the materials deliver superior electrochemical performance as anode materials in LIBs in terms of specific capacity, cycling performance, and rate capability. 2. Experimental details 2.1. Materials The synthesis was carried out using commercially available reagents. The GdCl3 6H2O was 99.99% pure and supplied by Sinopharm Chemical Reagent Company. All the other chemicals were
analytical grade and used as received without further purification. 2.2. Synthesis of pure Li4Ti5O12 (LTO) The pure Li4Ti5O12(LTO) was prepared using a hydrothermal method followed by calcination as reported by Wang et al. [12] In the typical synthesis procedures, 1.7 ml (5 mM) of tetrabutyl titanate (TBT) and 0.189 g of LiOH$H2O were thoroughly mixed in 20 ml of ethanol at room temperature. The solution was mixed completely with a magnetic stirrer in a closed container for 24 h and 25 ml of deionized water were added to the container. After stirring for 0.5 h, the solution was transferred to a 50 ml Teflonlined stainless autoclave and placed in an oven at 180 C for 36 h to obtain Li1.81H0.19Ti2O5$xH2O (H-LTO). The white H-LTO powder on the bottom of the reactor was collected, washed with ethanol 3 times, and dried at 80 C for 6 h. Finally, the white hydrothermal HLTO product was heated to 700 C for 6 h in a horizontal tube furnace in air to obtain the pure Li4Ti5O12. 2.3. Synthesis of Gd3þ doped LTO (LTO-Gd) nanosheets The experimental procedures described in Section 2.2 were used to prepare the H-LTO-Gd3þ precursor by doping with the appropriate amount of Gd3þ in the mixture consisting 1.7 ml (5 mM) of tetrabutyl titanate, 0.189 g of LiOH$H2O, and 20 ml of ethanol. The white hydrothermal H-LTO-Gd product was heated to 700 C for 6 h in a horizontal tube furnace in air to obtain Gd3þ doped LTO nanosheets. 2.4. Synthesis of ultrathin Gd3þ doped and carbon coated LTO nanosheets The ultrathin LTO-Gd-C nanosheets were prepared by controlling the amount of carbon precursor (D(t)-glucose monohydrate) used in the experimental procedures described in Section 2.3 for the LTO-Ln3þ doped LTO nanosheets. The H-LTO-Gd precursors were prepared using the experimental procedures described in Section 2.3. 200 mg of the H-LTO-Gd precursor were dispersed in 10 ml of ethanol under magnetic stirring to form Solution A. D(t)glucose monohydrate in the amount corresponding to 2.5, 5, or 10 wt% of the H-LTO-Gd precursor was dissolved in 5 ml of ethanol to produce Solution B. Solution B was added dropwise to Solution A under stirring, followed by slow evaporation at 80 C into a slurry. The slurry was heated under Ar/H2(5%) at 700 C for 6 h. The products with different nominal carbon contents were designated as LTO-Gd-C (2.5%), LTO-Gd-C (5%), and LTO-Gd-C (10%). For comparison, a composite corresponding to LTO-Gd-C (5%) without Gd3þ dopant was synthesized using H-LTO as the precursor and designated as LTO-C (5%). 2.5. Materials characterization The crystal structure of the samples was determined by powder X-ray diffraction (XRD, Rigaku, D/MAX 2500) using a copper Ka radiation source (l ¼ 0.154 nm). X-ray photoelectron spectroscopy (XPS) measurements were performed using an Al Ka source (Kratos Analytical Ltd., UK) and the binding energy of 284.8 eV for C 1s was used as calibration. The morphology and microstructure of the samples were characterized by field-emission scanning electron microscopy (FE-SEM, Hitachi, S4800) equipped with energydispersive X-ray spectroscopy (EDS) and transmission electron microscopy (TEM, JEOL 2100F) equipped with selected-area electron diffraction (SAED). The Raman spectra were recorded on a Renishaw InVia system with the excitation laser (l) of 532 nm. Thermal gravimetric and differential scanning calorimetry (TG-
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
DSC) analyses were carried out on a TGA 2050 thermogravimetric analyzer. 2.6. Electrochemical characterization The electrochemical tests were conducted on the two-electrode CR2032 type coin cells. The working electrodes were prepared by pasting a mixture of the active materials (LTO, LTO-Gd, LTO-C, LTOGd-C), carbon black, and polyvinylidene fluoride (PVDF) with a weight ratio of 80:10:10 onto a Cu foil with a thickness about 0.2 mm and an area about 2 cm2, which acted as the current collector. The electrodes were dried at 80 C for 6 h in air and 120 C in vacuum for another 12 h followed by pressing. Each electrode was weighed accurately on an electronic balance and the weight of the active materials was controlled to be 1e2 mg. The coin cells were assembled in an argon-filled gloved box with a metallic lithium foil as the counter electrode. The electrolyte was 1 M LiPF6 in ethylene carbonate (EC)/dimethyl carbonate (DEC) (1:1, volume ratio) and Celgard 2400 polypropylene was the separator. The charging/discharging experiments were performed at various current densities over a voltage range of 1e2.5 V (vs Li/Liþ) using a multi-channel battery test system (NEWARE BTS-610). Cyclic voltammetry (CV) was performed on an electrochemical workstation (CHI660D) in the voltage range of 1e2.5 V (vs Li/Liþ) at different scanning rates and electrochemical impedance spectroscopy (EIS) was conducted by applying a perturbation voltage of 10 mV in a frequency range of 100 kHz to 10 mHz on a CHI660D electrochemical workstation. 3. Results and discussion The ultrathin spinel LTO nanosheets are prepared by a facile one-pot hydrothermal method in conjunction with postcalcination [12] as schematically illustrated in Fig. 1a. The ultrathin spinel LTO nanosheets are formed by a hydrolysis-ion exchange-Ostwald ripening-phase transition accompanying oxolation and rearrangement of the TiO6 octahedron. Initially, amorphous hydrous titanium oxide (TiO2$nH2O) considered as condensing irregular TiO6 octahedrons, in which each Ti atom is surrounded by six O atoms including two longer (0.1980 nm) TieO bonds and four shorter (0.1934 nm) ones, are produced by hydrolysis of TBT in the LiOH$H2O aqueous solution. During the
Fig. 1. (a) Formation process of the Li4Ti5O12 nanosheets, (b) Effects of Gd3þ doping and carbon coating.
307
hydrothermal treatment in the LiOH solution, the longer TieO bonds are attacked by OH- ions and break, whereas the shorter ones are stronger and so irregular swelling occurs, leading to the formation of some fragments and delamination from TiO2$nH2O. The fragments are linked to each other to form planar intermediate via eOeTieOe ionic bonds. At the same time, some O atoms combine with H2O or H3Oþ via hydrogen bonds, followed by ion exchange between Liþ and Hþ. These planar intermediates then rearrange to form layers composed of [TiO6] octahedra with shared edges and vertices and with Liþ and H2O or H3Oþ intercalated between the layers forming planar C-base-centered orthorhombic HLTO seed nuclei. As the reaction proceeds, the planar seed nuclei grow continuously via Ostwald ripening to produce the layered HLTO. Topotactic transformation of the H-LTO nanosheets to spinel LTO is accomplished by post-calcination at 700 C. The heat treatment not only removes H2O, H3Oþ, and Li2O molecules, but also reconstructs the layers during which some [TiO6] octahedra retain the same morphology while others restructure to form [TiO4] tetrahedra. However elastic strain inevitably results from the mismatch of the lattice of the LTO and H-LTO phase during the topotactic reaction. When the strain is beyond a critical value, break-up of large nanosheets occurs with formation of small flakes, which facilitate second Ostwald ripening of LTO. As the reaction proceeds, the LTO nanosheets often become irregular and thick. However, when trivalent Ln3þ ions are introduced, they occupy the Ti4þ sites and promote the formation of holes in the H-LTO upon annealing accompanied by removal of H2O, H3Oþ, and Li2O molecules, which may be filled with electrons from the LTO nuclei. In order to establish charge balance, extra positive ions have to be introduced into the surface of the LTO nuclei (see Fig. 1b). Such a charge transfer enables strong electrostatic interactions between the LTO nuclei and the H-LTO nanosheets to produce polaronic lattice distortion which effectively buffer the elastic strain in topotactic transformation. On the other hand, owing to the chargemediated bond reinforcement, the LTO nuclei tend to maximize the contact area with the H-LTO nanosheets and form 2D nuclei with larger characteristic dimensions [53,54]. The larger 2D nuclei benefit preferential adsorption of the reconstructed molecules on the side plane and hence, second Ostwald ripening occurs along the direction of the normal surface being hindered. As a result, textural modification and break-up associated with elastic deformation is depressed notably and the ultrathin LTO nanosheets are stabilized and retain the precursor morphology.
Fig. 2. (a)e(c) XRD patterns of H-LTO, pure LTO, and Gd3þ-LTO, respectively.
308
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
Fig. 2a depicts the X-ray diffraction (XRD) patterns of the hydrothermal products, suggesting that they are crystals composed of layered H-LTO with a C-base centered orthorhombic lattice (JCPDS No. 47-0123). Transmission electron microscopy (TEM) (see Fig. 3a) indicates that the hydrothermal H-LTO consists of large flat nanosheets with a diameter larger than 500 nm with smooth edges (see Fig. S1). The higher magnification image of individual vertical nanosheets reveals that the thickness of the nanosheets is about 8 nm. Fig. 3b displays the corresponding SAED pattern acquired from a single nanosheet indicating that the H-LTO nanosheets are single crystal. The bright spots in the SAED pattern can be indexed to the (002), (020), and (022) planes of orthorhombic H-LTO from the [100] zone axis. The HR-TEM image of an individual nanosheet in Fig. 3c shows that the lattice spacing is about 0.19 nm that corresponds to the lattice spacing of the (020) plane in orthorhombic H-LTO. In conjunction with the SAED results, it can be inferred that single-crystal H-LTO nanosheets grow along the (100) facet. Fig. 2b shows the XRD pattern of the samples of pure LTO, which can be well indexed to spinel LTO (JCPDS No. 49-0207). No impurities are observed confirming complete conversion from H-LTO to spinel LTO. Fig. 3d depicts the representative TEM image of the samples obtained after calcination at 700 C. The essential features indicative of thickening, textural modification including angulation and irregular edge, and small flakes suggesting interlayer splitting when the H-LTO nanosheets are topotactically converted into spinel LTO, can be observed, which is consistent with the results reported in Ref. [13]. Fig. 3f shows the representative SAED pattern taken from an individual sheet suggesting high crystallinity. However, in the irregular and deformed LTO nanosheets, it is rather difficult to observe a definite and reproducible orientation relationship for the H-LTO by SAED as previously observed from other topotactic transformation systems [55]. As shown in Fig. 3e, the crystalline region with clear lattice fringes has an inter-planar spacing of approximately 0.48 nm corresponding to the (111) atomic planes of spinel LTO and in good agreement with the SAED patterns. Fig. 2c shows the XRD patterns of the LTO sample after
doping Gd3þ with nominal content of 10wt%. All the peaks can be indexed to the spinel structure LTO and no other peaks such as Gd2O3 and TiO2 can be detected. The results are consistent with those of the pure LTO sample. Broadened diffraction peaks are observed suggesting reduced average crystallite size. Fig. 3g shows the typical TEM image of the Gd3þ doped LTO sample and it retains the original sheet-like morphology of the H-LTO precursor. The representative SAED patterns and HR-TEM image in Fig. 3h and i suggest single crystallinity in the Gd3þ doped LTO nanosheets grown along the ð011Þ plane. Furthermore, a definite and reproducible orientation relationship with H-LTO is revealed by the SAED pattern of the Gd3þ doped LTO nanosheets and the reaction from HLTO to LTO is highly topotactic with the following relationship:
ð100ÞHLTO
. 011
ð020ÞHLTO
ð400ÞLTO and ð022ÞHLTO ð444ÞLTO
LTO
ð002ÞHLTO
ð333ÞLTO
Interestingly, the detectable Gd3þ, which is below the detection limitation of EDS measurement and can be confirmed by high resolution XPS (see Fig. S2) is far below the nominal content. The results indicate that it is difficult to dope Gd3þ into LTO host due to large difference of ionic diameters between Gd3þ and Ti4þ. In our experiments, we found that the contents of Gd3þ doping affect notably the nucleation and growth behavior of LTO nanosheets during topotactic transformation at 700 C. With different Gd3þ doping contents, the microstructures and electrochemical performances of the obtained LTO samples are different (see Fig. S3). Similar effect on size and surface texture of LTO nanosheets was also observed via doping other Ln3þ ions such as Y3þ and La3þ. The influence of Ln3þ doping on the morphology, growth mechanism and electrochemical performance is in progress. Herein, Gd3þ ion with nominal doping content of 10wt% is used as a typical example to improve lithium storage performance via Ln3þ doping and carbon coating. Subsequently, the mixture consisted of Gd3þ-doped H-LTO and
Fig. 3. Typical TEM and HR-TEM images of (a, b) H-LTO, (d, e) Pure LTO and (g, h) Gd3þ-LTO nanosheets as well as corresponding SAED patterns (c, f, i). “H” and “L” represent H-LTO and LTO, respectively.
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
D(t)-glucose monohydrate at a predetermined weight ratio of 5% is heated to 700 C to produce ultrathin Gd3þ doped and carbon coated LTO nanosheets (see Fig. 1b). For comparison, the LTO-C sample with a nominal 5 wt% carbon coating is prepared using HLTO as the precursor. The TEM observations (see Fig. S4) indicate that the carbon coating obtained from D(þ)-glucose monohydrate has the ability to stabilize the surface of the LTO nanostructures and preserve the precursor morphology during topotactic transformation. The results are consistent with those reported in carbon-encapsulated F-doped LTO with a ball-in-ball morphology [30] and self-supported LTO-C nanotube arrays [29]. The actual carbon coating on the LTO-C sample is about 5.1 wt% (see Fig. S4), which is smaller than that of the LTO-Gd-C sample to be discussed later. Nevertheless, the electrochemical performance of LTO-C is rather unsatisfactory due to the lack of synergistic interactions between Gd3þ and carbon coating to be discussed later. Raman scattering is employed to examine the microstructure of the carbon layer on the LTO-Gd-C (5%) sample and as shown in Fig. 4a, the two small bands at 1178 and 1493 cm1 are short-range vibration of sp3 carbon atoms commonly observed from amorphous carbonaceous materials [50]. The two main bands at around 1330 and 1592 cm1 can be designated as the D and G bands of sp2 carbon, respectively. The surface chemical composition and states of the LTO-Gd-C (5%) sample are determined by XPS. Fig. 4b and S7 reveal the presence of Li, Ti, O, C, and Gd and the results are in good agreement with the nominal atomic composition of the LTO-Gd-C sample. The weak Gd 3d peak in Fig. S5 suggests a small concentration of Gd3þ. The TG analysis (TGA) (Fig. 4c) reveals that the carbon content in the LTOGd-C (5%) sample is about 6.5 wt%. Fig. 4d shows the representative XRD patterns of the LTO-Gd-C (5 wt%) sample and the peaks can be indexed to a cubic spinel structure according to the Fd3m space group without impurities. The mean crystallite size of LTO-Gd-C calculated from the (111) diffraction peak by the Scherer equation is smaller than those of the pure LTO, LTO-Gd, and LTO-C. This decreasing trend indicates that pyrolyzed carbon also restrains the
309
growth of LTO nanoparticles in LTO-Gd-C thereby preserving the nanoscale advantage. Fig. 5a shows the typical TEM image of the LTO-Gd-C sample, demonstrating that the sample retains the sheetlike morphology of the H-LTO precursor. The representative SAED pattern and HR-TEM results acquired from individual nanosheets (see Fig. 5d and inset of Fig. 5b) suggest single crystallinity in the LTO-Gd-C nanosheets with the feature of topotactic transformation. Comparison of the TEM images from LTO-Gd (see Fig. 5c) and LTOGd-C nanosheets (see Fig. 5b) with typical HR-TEM images (see Fig. 5d) shows the presence of an amorphous carbon coating on the surface of LTO nanosheets. Well-defined lattice fringes with a separation of 0.48 nm in the HR-TEM image of the carbon-coated LTO nanosheets correspond to the spacing of the (111) planes in the spinel LTO. In addition, as shown in the inset in Fig. 5a, the typical vertical nanosheet about 10 nm thick can be observed. Atomic force microscopy (see the inset in Fig. 5d) reveals the 2D sheet-like feature of the LTO-Gd-C nanosheets with a thickness of about 10 nm. According to TEM and AFM, the thickness of the primary LTO in LTO-Gd-C nanosheets can be inferred to be less than 10 nm due to the existence of coated carbon. To determine the effects of the carbon content on the electrochemical performance, different carbon coatings are prepared by mixing the H-LTO-Gd precursor with different amounts of D(t)glucose monohydrate (2.5, 5 and 10 wt% of the H-LTO precursor). As shown in Fig. S6aec, the nominal 5 wt% carbon coating delivers the best electrochemical performance. Fig. 6a shows the CV curve of the LTO-Gd-C nanosheets at a scanning rate of 0.1 mV/s in the potential range between 2.5 and 1.0 V (vs Li/Liþ). The well-defined redox peaks at 1.73/1.44 V can be attributed to the redox reaction of Ti4þ/ Ti3þ, respectively, associated with lithium insertion/extraction in the spinel LTO lattice [29]. The strong and well-defined redox peaks of LTO indicate well-behaved electrode kinetics of the LTO-Gd-C nanosheets electrode. Fig. 6b depicts the chargingedischarging voltage profiles of the LTO, LTO-Gd, LTO-C (5%), and LTO-Gd-C (5%) electrodes for the 1st, and 100th cycles at a current rate of 10 C within
Fig. 4. (a)e(c) Raman spectrum, TG-DSC curves and survey XPS spectrum of the LTO-Gd-C nanosheets; (d) XRD patterns of LTO-Gd-C nanosheets together with other LTO-C (5%), LTO-Gd, and LTO reference samples.
310
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
Fig. 5. (a) Typical TEM image of the LTO-Gd-C nanosheets with the inset showing the enlarged image of the marked area; (b) and (c) High-magnification TEM images of LTO-Gd-C nanosheets and reference LTO-Gd nanosheets with the inset in (b) showing the SEAD patterns of the LTO-Gd-C nanosheets; (d) HR-TEM and AFM images of LTO-Gd-C nanosheets.
Fig. 6. (a) CV curves acquired at a scanning rate of 0.1 mV s1 in the voltage range of 1.0e2.5 V of LTO-Gd-C nanosheets; (b) Chargingedischarging voltage profiles of the LTO-Gd-C nanosheets and other reference materials for the 1st and 100th cycles at a current rate of 10 C; (c) Cycling performance and Coulombic efficiency versus cycle number at different current rates from 1 C to 30 C of the LTO-Gd-C nanosheets and other reference materials. (d) Cycling performance of the LTO-Gd-C nanosheets and other reference materials at the standard dischargingecharging rate of 10 C.
a cut off window of 1.0e2.5 V. All the electrodes exhibit similar charging and discharging flat plateaus at around 1.5 V resulting from a two-phase reaction during electrochemical lithium insertioneextraction according to the following equation [23,38].
Li4 Ti5 O12 þ xLiþ þ xe 4Li4þx Ti5 O12 ð0 < x < 3Þ
(1)
The results indicate that Gd3þ doping and carbon coating do not alter the electrochemical reaction of LTO. However, the magnified chargingedischarging voltage profiles (see Fig. S6d) of the four electrodes for the 1st and 100th cycles at a current rate of 10 C reveal that the LTO-Gd-C (5%) electrode has a flatter profile than the other reference materials. The polarization between the charging
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
and discharging plateaus are 283 and 477 mV for the 1st and 100th cycles, respectively, of the pure LTO electrode, which are larger than those of the LTO-Gd (149 mV and 130 mV) and LTO-C (5%) (160 mV and 175 mV) electrodes. The lower electrochemical polarization suggests improved reaction kinetics due to the special structure and higher surface electronic conductivity of the LTO-Gd nanosheets and LTO nanosheets with carbon coatings. Electrochemical polarization is further depressed (100 mV and 128 mV at the 1st and 100th cycles, respectively) in the LTO-Gd-C nanosheets, implying the synergistic effects of the Gd3þ doping and carbon coating on the reaction kinetics in the electrode. Fig. 6c shows the LTO-Gd-C nanosheets cycled at different current rates from 1 C to 30 C in the voltage range between 1.0 and 2.5 V. At 1 C (1 C ¼ 175 mAg1), the LTO-Gd-C electrode shows a discharge capacity of 213 mAh g1, which are higher than the theoretical capacity of 175 mA h g1. A similar phenomenon has also been observed in self-assembled TiO2egraphene hybrid nanostructures [56e58]. This suggests the existence of additional lithium storage sites in the LTO-Gd-C electrode. On one hand, an extended solid solution region for Liþ insertion into LTO was observed obviously after Gd3þ doping or carbon coating in chargingedischarging voltage profiles (see Fig. S7) starting from the open-circuit potential (about 2.3 V) in voltage to ~1.5 V, similar to that which has been reported previously for small particles such as TiO2 and Si, in which an additional surface lithium storage mechanism exists [59,60]. On the other hand, shallower slope below 1.5 V that is a typical feature of supercapacitor in chargingedischarging voltage profiles was observed obviously after carbon coating, implying higher capacitance due to the insertion of lithium ions into the surface layer of the LTO-C nanosheets [59]. Moreover, it is inferred that the synergistic effects of the Gd3þ doping and carbon coating play a key role in further enhance surface lithium storage, thereby giving rise to a high capacity of 213 mAh g1 for the LTO-Gd-C nanosheets since the chargingedischarging voltage profile of the LTO-Gd-C electrode demonstrates the most extended solid solution region from the open-circuit potential in voltage to ~1.5 V and the shallowest slope in below 1.5 V. As the rate increases from 1 to 5 and 10 C, the discharge capacity decreases from 213 mAh g1 to 182 and 171 mA hg1, respectively. At a high rate of 30 C, the capacity retains still 146 mAh g1. Furthermore, the Coulombic efficiency of the LTO-Gd-C nanosheets at large charging/discharging rates approach 100% for each cycle. For comparison, the capacities of LTO, LTO-C (5%), LTO-Gd at 30 C are 96 mAh g1, 104 mAh g1, and 120 mAh g1, respectively. The results reveal that lithium diffusion is efficient and responsible for the excellent rate capability of the LTO-Gd-C nanosheets. With regard to the cycling behavior, we examine the cycling performance of the LTO-Gd-C nanosheets at a high current density of 10 C. The LTO-Gd-C anode at a rate of 10 C reaches a stable capacity of 168 mAh g1 with the Coulombic efficiency approaching 100% after 20 cycles. It retains a capacity of 152 mAh g1 after 500 cycles with a capacity loss of only 9.2%, as shown in Fig. 6d. The cycling performance of the LTO-Gd-C nanosheets is superior to those of pure LTO, LTO-C, and LTO-Gd nanosheets. In addition, we compare the present data with those observed from other LTO-based electrodes including LTO nanowire arrays [33],TiO2-coated LTO [12], carbon-coated LTO [27], Cu-doped LTO-TiO2 nanosheets [47], N-doped carbon-coated LTO [39], mesoporous LTOC [36], nanostructured carbon-coated LTO [10,25], self-supported LTO nanosheets arrays [49], LTO-C nanotube arrays [29], carbon-encapsulated F-doped LTO [30], LTO/graphene nanosheets [35], LTO tailored by cathodically induced graphene [38], etc. The rate and cycling performance of the LTO-Gd-C nanosheets are comparable to the best results found in the literature according to Table SI, confirming the positive effects of Gd3þ doping and carbon coating.
311
The superior electrochemical performance of the LTO-Gd-C nanosheets can be attributed to enhancement of the transport kinetics in the electrode. Owing to the unique two-phase reaction during electrochemical lithium insertioneextraction shown in Eqn. (1), LTO materials with an excellent rate capability possess both improved electrical conductivity and lithium ion diffusion, which can be studied by electrochemical impedance spectroscopy (EIS). The LTO-Gd-C nanosheets and other reference electrodes are analyzed by EIS at open-circuit voltages in the native states (before cycling) and after different discharging and charging cycles. The Nyquist plots in Fig. 7aed reveal a common purely resistive response at high frequencies represented by the ohmic resistance (Rs) of the electrode and electrolyte, a semicircle due to the charge-transfer impedance (Rct) on the electrodeeelectrolyte interface in the high-to-middle frequency region, and an inclined straight line ascribed to the Warburg impedance (Zw) in the low frequency region. The EIS spectra are simulated by the Z-view software using the inset equivalent circuit model and the corresponding simulation parameters are presented in Table 1. One can find that the semicircular arc of the LTO-Gd-C nanosheets electrode is much smaller than that of other reference electrodes. EIS performed after cycling displays the same trend with a smaller arc in the Nyquist plot. The smaller semicircular arc is an indication of an overall smaller charge transfer resistance or more facile charge transfer process at the electrode/electrolyte interface for the LTO-Gd-C nanosheets electrode [38,58]and the results are consistent with the simulation results in Table 1. The lithium ion diffusion coefficient (DLi) can be evaluated by EIS according to the following equations [50,52].
DLi ¼
R2 T 2 2A2 n4 F 4 C 2 s2
Z 0 ¼ Rs þ Rct þ su1=2
(2)
(3)
where R is the gas constant, T is the absolute temperature, A is the surface area of the electrode, n is the number of electrons transferred in the half reaction of the redox couple, F is Faraday's constant, C is the concentration of lithium ions, and s is the Warburg factor. The value of s can be obtained from the slope of the lines 0 between Z and 61=2 as shown in Fig. 7eef. The DLi values of the electrodes can be calculated and the results after 2 and 100 cycles are presented in Table 1. The DLi values of the LTO-Gd-C nanosheets are larger than those of other reference electrodes suggesting that the LTO-Gd-C nanosheets possess faster solid-state lithium ion diffusion thus enabling fast migration of lithium ions to the interior of the LTO NSs and full utilization of the active materials. The results also confirm that Gd3þ doping and carbon coating cooperatively improve the solid-state lithium ion diffusion coefficient and electrical conductivity of the LTO-Gd-C nanosheets and relieve the polarization of the electrode to enhance the cycling and rate capability. The smaller Rct and larger DLi of the LTO-Gd-C nanosheets result from the special structure and Gd3þ doping. Firstly, using a small amount of trivalent Gd3þ to substitute for quadrivalent titanium ions in the LTO nanosheets not only allows the products to be stabilized to a thickness about 10 nm via topotactic transformation of the layered H-LTO, but also provide electrical donor to improve intrinsic conductivity of ultrathin nanosheets, when the incorporated Gd3þ ions occupy the 16d site in LTO matrix, thereby depressing the electrochemical polarization and charge transfer resistance and expediting lithium ion diffusion during lithium intercalation and de-intercalation. Secondly, complete encapsulation of the LTO nanosheets by a pervasive network of
312
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
0
Fig. 7. (a)e(d) Nyquist plots of the LTO-Gd-C nanosheets and other reference materials after different discharging and charging cycles. (e) and (f) Z plots against 61=2 at the low frequency region (10e0.1 Hz) for the LTO-Gd-C nanosheets and other reference materials after the second and 100th cycles of discharging and charging.
Table 1 Rct and Di values for the LTO, LTO-Gd, LTO-C and LTO-Gd-C electrodes after different cycles in half-cells according to Fig. 7. Compound
2nd LTO LTO-C (5%) LTO-Gd LTO-Gd-C (5%)
Di (cm2 s1)
Rct (U)
153 117 105.61 88.5
10th 126.7 97.1 95.1 80.2
30th 119.7 89.4 88.1 72.1
carbon from pyrolysis of D(t)-glucose monohydrate preserves the unique ultrathin architecture of the precursor during the transformation from H-LTO to LTO due to growth inhibition of encapsulated LTO during the high temperature treatment and also decreases the external electron transport resistance making it easier for lithium ions to traverse to the primary LTO matrix through the layer. Thirdly, the ultrathin nanosheets with singlecrystalline nature not only promote the kinetics of lithiation and delithiation due to the shorter Li-ion diffusion distances and avoiding the drawback of many grain boundaries in electrodes of
60th 103.4 70.8 68.8 52.1
100th 84.4 57.3 55.42 38.5
2nd
100th 12
4.6 10 1.47 1011 1.66 1011 6.03 1011
1.6 1013 5.23 1012 6.16 1012 1.83 1011
reported LTO nanostructure such as polycrystalline nanowires, polycrystalline nanotube and thick LTO nanosheets, but also produce larger discharge capacity since the storage limit at the 16c sites of spinel structure is exceeded and co-occupation of the 8a and 16c sites in the surface region is expected to be more significant in ultrathin LTO nanosheets where the specific surface area is larger [30], resulting in storage properties different from those of the bulk, polycrystalline nanowires, polycrystalline nanotube and thick LTO nanosheets. Consequently, superior electrochemical performances are demonstrated in the ultrathin LTO-Gd-C nanosheets.
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
4. Conclusion In summary, we have prepared highly-crystalline gadolinium doped and carbon-coated ultrathin Li4Ti5O12 (LTO) nanosheets with thickness below 10 nm on large scale by controlling the amount of carbon precursor in the topotactic transformation of layered ultrathin Li1.81H0.19Ti2O5$xH2O (H-LTO) nanosheets at 700 C. Gadolinium doping allows the spinel LTO products to be stabilized, thereby preserving the precursor's sheet morphology and single crystal structure with enhanced intrinsic electron conductivity. Carbon encapsulation serves dual functions by restraining crystal growth of the LTO primary nanoparticles in the LTO-Gd-C nanosheets and decreasing the external electron transport resistance. Owing to the synergistic effects rendered by ultrathin nanosheets with high crystallinity, Gd3þ doping and carbon coating, the obtained materials have excellent specific capacity, cycling performance, and rate capability as anode materials in LIBs. The desirable effects stem from the larger contact surface area with the electrolyte, shorter transport length of lithium ions, bigger lithium diffusion coefficient, and enhanced intrinsic conductivity. The strategy of depending on the heterovalent Ln3þ doping to control microstructure and properties is simple and effective. The results reveal the benefits of employing multiple modification strategies such as nanoscale and morphological engineering, surface modification, and doping to improve the performance of energy storage materials applied in high energy density and high power devices. Acknowledgments This work was financially supported by the Grants from National Natural Science Foundation of China (Nos. 11474242, 51272220, 11374252, and 51472209), the Program for Changjiang Scholars and Innovative Research Team in University (IRT13093), Guangdong e Hong Kong Technology Cooperation Funding Scheme (TCFS) GHP/ 015/12SZ (CityU 9440103), and City University of Hong Kong Applied Research Grant No. 9667085. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2015.06.131. References [1] M. Armand, J.-M. Tarascon, Nature 451 (2008) 652e657. [2] X. Chen, C. Li, M. Graetzel, R. Kostecki, S.S. Mao, Chem. Soc. Rev. 41 (2012) 7909e7937. [3] Y.G. Guo, J.S. Hu, L.J. Wan, Adv. Mater. 20 (2008) 2878e2887. [4] C. Liu, F. Li, L.P. Ma, H.M. Cheng, Adv. Mater. 22 (2010) E28eE62. [5] S. Ganapathy, M. Wagemaker, ACS Nano 6 (2012) 8702e8712. [6] G. Jeong, Y.-U. Kim, H. Kim, Y.-J. Kim, H.-J. Sohn, Energy Environ. Sci. 4 (2011) 1986e2002. [7] C. Ouyang, Z. Zhong, M. Lei, Electrochem. Commun. 9 (2007) 1107e1112. [8] T.-F. Yi, S.-Y. Yang, Y. Xie, J. Mater. Chem. A 3 (2015) 5750e5777. [9] X. Lu, L. Zhao, X. He, R. Xiao, L. Gu, Y.S. Hu, H. Li, Z. Wang, X. Duan, L. Chen, Adv. Mater. 24 (2012) 3233e3238. [10] H.-G. Jung, M.W. Jang, J. Hassoun, Y.-K. Sun, B. Scrosati, Nat. Commun. 2 (2011) 516. [11] N. Li, Z. Chen, W. Ren, F. Li, H.-M. Cheng, Proc. Natl. Acad. Sci. 109 (2012) 17360e17365. [12] Y.-Q. Wang, L. Gu, Y.-G. Guo, H. Li, X.-Q. He, S. Tsukimoto, Y. Ikuhara, L.-J. Wan, J. Am. Chem. Soc. 134 (2012) 7874e7879. [13] G.-N. Zhu, H.-J. Liu, J.-H. Zhuang, C.-X. Wang, Y.-G. Wang, Y.-Y. Xia, Energy Environ. Sci., 4 4016e4022. [14] L. Yu, H.B. Wu, X.W.D. Lou, Adv. Mater. 25 (2013) 2296e2300.
313
[15] Y.-J. Bai, C. Gong, N. Lun, Y.-X. Qi, J. Mater. Chem. A 1 (2013) 89e96. [16] Y.-J. Bai, C. Gong, Y.-X. Qi, N. Lun, J. Feng, J. Mater. Chem. 22 (2012) 19054e19060. [17] J.G. Kim, M.S. Park, S.M. Hwang, Y.U. Heo, T. Liao, Z. Sun, J.H. Park, K.J. Kim, G. Jeong, Y.J. Kim, ChemSusChem 7 (2014) 1451e1457. [18] C. Lin, B. Ding, Y. Xin, F. Cheng, M.O. Lai, L. Lu, H. Zhou, J. Power Sources 248 (2014) 1034e1041. [19] T.-F. Yi, Y. Xie, Q. Wu, H. Liu, L. Jiang, M. Ye, R. Zhu, J. Power Sources 214 (2012) 220e226. [20] Y. Zhang, C. Zhang, Y. Lin, D.-B. Xiong, D. Wang, X. Wu, D. He, J. Power Sources 250 (2014) 50e57. [21] H. Song, S.-W. Yun, H.-H. Chun, M.-G. Kim, K.Y. Chung, H.S. Kim, B.-W. Cho, Y.T. Kim, Energy Environ. Sci. 5 (2012) 9903e9913. [22] G. Du, N. Sharma, V.K. Peterson, J.A. Kimpton, D. Jia, Z. Guo, Adv. Funct. Mater. 21 (2011) 3990e3997. € blinger, D. Fattakhova-Rohlfing, T. Bein, Angew. [23] J.M. Feckl, K. Fominykh, M. Do Chem. Int. Ed. 51 (2012) 7459e7463. [24] X. Hao, B.M. Bartlett, Adv. Energy Mater. 3 (2013) 753e761. [25] H.-G. Jung, S.-T. Myung, C.S. Yoon, S.-B. Son, K.H. Oh, K. Amine, B. Scrosati, Y.K. Sun, Energy Environ. Sci. 4 (2011) 1345e1351. [26] H.-K. Kim, S.-M. Bak, K.-B. Kim, Electrochem. Commun. 12 (2010) 1768e1771. [27] B. Li, C. Han, Y.-B. He, C. Yang, H. Du, Q.-H. Yang, F. Kang, Energy Environ. Sci. 5 (2012) 9595e9602. [28] N. Li, G. Zhou, F. Li, L. Wen, H.M. Cheng, Adv. Funct. Mater. 23 (2013) 5429e5435. [29] J. Liu, K. Song, P.A. van Aken, J. Maier, Y. Yu, Nano Lett. 14 (2014) 2597e2603. [30] Y. Ma, B. Ding, G. Ji, J.Y. Lee, ACS Nano 7 (2013) 10870e10878. [31] H. Pan, L. Zhao, Y.S. Hu, H. Li, L. Chen, ChemSusChem 5 (2012) 526e529. [32] K.-S. Park, A. Benayad, D.-J. Kang, S.-G. Doo, J. Am. Chem. Soc. 130 (2008) 14930e14931. [33] L. Shen, E. Uchaker, X. Zhang, G. Cao, Adv. Mater. 24 (2012) 6502e6506. [34] L. Shen, C. Yuan, H. Luo, X. Zhang, K. Xu, Y. Xia, J. Mater. Chem. 20 (2010) 6998e7004. [35] L. Shen, C. Yuan, H. Luo, X. Zhang, S. Yang, X. Lu, Nanoscale 3 (2011) 572e574. [36] L. Shen, X. Zhang, E. Uchaker, C. Yuan, G. Cao, Adv. Energy Mater. 2 (2012) 691e698. [37] Y. Tang, F. Huang, W. Zhao, Z. Liu, D. Wan, J. Mater. Chem. 22 (2012) 11257e11260. [38] Y. Yang, B. Qiao, X. Yang, L. Fang, C. Pan, W. Song, H. Hou, X. Ji, Adv. Funct. Mater. 24 (2014) 4349e4356. [39] L. Zhao, Y.S. Hu, H. Li, Z. Wang, L. Chen, Adv. Mater. 23 (2011) 1385e1388. [40] N. Li, T. Mei, Y. Zhu, L. Wang, J. Liang, X. Zhang, Y. Qian, K. Tang, CrystEngComm 14 (2012) 6435e6440. [41] J.-G. Kim, D. Shi, M.-S. Park, G. Jeong, Y.-U. Heo, M. Seo, Y.-J. Kim, J.H. Kim, S.X. Dou, Nano Res. 6 (2013) 365e372. [42] M. Krajewski, M. Michalska, B. Hamankiewicz, D. Ziolkowska, K.P. Korona, J.B. Jasinski, M. Kaminska, L. Lipinska, A. Czerwinski, J. Power Sources 245 (2014) 764e771. [43] Z. Liu, N. Zhang, Z. Wang, K. Sun, J. Power Sources 205 (2012) 479e482. [44] W. Wang, Y. Guo, L. Liu, S. Wang, X. Yang, H. Guo, J. Power Sources 245 (2014) 624e629. [45] F. Bonaccorso, L. Colombo, G. Yu, M. Stoller, V. Tozzini, A.C. Ferrari, R.S. Ruoff, V. Pellegrini, Science 347 (2015) 1246501. [46] Y. Sun, S. Gao, Y. Xie, Chem. Soc. Rev. 43 (2014) 530e546. [47] C. Chen, Y. Huang, C. An, H. Zhang, Y. Wang, L. Jiao, H. Yuan, ChemSusChem 8 (2015) 114e122. [48] J. Chen, L. Yang, S. Fang, Y. Tang, Electrochim. Acta 55 (2010) 6596e6600. [49] S. Chen, Y. Xin, Y. Zhou, Y. Ma, H. Zhou, L. Qi, Energy Environ. Sci. 7 (2014) 1924e1930. [50] C. Han, Y. He, B. Li, H. Li, J. Ma, H. Du, X. Qin, Q. Yang, F. Kang, ChemSusChem 7 (2014) 2567e2574. [51] C. Lai, Y. Dou, X. Li, X. Gao, J. Power Sources 195 (2010) 3676e3679. [52] G.B. Xu, W. Li, L.W. Yang, X.L. Wei, J.W. Ding, J.X. Zhong, P.K. Chu, J. Power Sources 276 (2015) 247e254. [53] X. Shao, S. Prada, L. Giordano, G. Pacchioni, N. Nilius, H.J. Freund, Angew. Chem. Int. Ed. 50 (2011) 11525e11527. [54] N. Mammen, S. Narasimhan, S.d. Gironcoli, J. Am. Chem. Soc. 133 (2011) 2801e2803. [55] M. Figlarz, B. Gerand, A. Delahaye-Vidal, B. Dumont, F. Harb, A. Coucou, F. Fievet, Solid State Ionics 43 (1990) 143e170. [56] X. Xin, X. Zhou, J. Wu, X. Yao, Z. Liu, ACS Nano 6 (2012) 11035e11043. [57] S. Yang, X. Feng, K. Müllen, Adv. Mater. 23 (2011) 3575e3579. [58] S. Yu, L. Yang, Y. Tian, P. Yang, F. Jiang, S. Hu, X. Wei, J. Zhong, J. Mater. Chem. A 1 (2013) 12750e12758. [59] J. Wang, J. Polleux, J. Lim, B. Dunn, J. Phys. Chem. C 111 (2007) 14925e14931. [60] B. Wang, X.L. Li, B. Luo, L. Hao, M. Zhou, X.H. Zhang, Z.G. Fan, L.J. Zhi, Adv. Mater. 27 (2015) 1526e1532.
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Highly-crystalline ultrathin gadolinium doped and carbon-coated Li4Ti5O12 nanosheets for enhanced lithium storage G.B. Xu a, L.W. Yang a, b, *, X.L. Wei a, J.W. Ding a, J.X. Zhong a, P.K. Chu b, ** a b
Hunan Key Laboratory of Micro-Nano Energy Materials and Devices, School of Physics and Optoelectronics, Xiangtan University, Hunan 411105, China Department of Physics and Materials Science, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, China
h i g h l i g h t s Ultrathin Gd3þ doped and carbon-coated Li4Ti5O12 nanosheets are prepared. The nanosheets have single-crystal nature with a thickness of about 10 nm. The nanosheets have high electrical conductivity and large Liþ diffusion coefficient. The benefits stem from synergistic effects by ultrathin structure, doping and carbon coating. The nanosheets demonstrate superior lithium storage performance.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 April 2015 Received in revised form 7 June 2015 Accepted 25 June 2015 Available online xxx
Highly-crystalline gadolinium doped and carbon-coated ultrathin Li4Ti5O12 (LTO) nanosheets (denoted as LTO-Gd-C) as an anode material for Li-ion batteries (LIBs) are synthesized on large scale by controlling the amount of carbon precursor in the topotactic transformation of layered ultrathin Li1.81H0.19Ti2O5$xH2O (H-LTO) nanosheets at 700 C. The characterizations of structure and morphology reveal that the gadolinium doped and carbon-coated ultrathin LTO nanosheets have high crystallinity with a thickness of about 10 nm. Gadolinium doping allows the spinel LTO products to be stabilized, thereby preserving the precursor's sheet morphology and single crystal structure. Carbon encapsulation serves dual functions by restraining crystal growth of the LTO primary nanoparticles in the LTO-Gd-C nanosheets and decreasing the external electron transport resistance. Owing to the synergistic effects rendered by ultrathin nanosheets with high crystallinity, gadolinium doping and carbon coating, the developed ultrathin LTO nanosheets possess excellent specific capacity, cycling performance, and rate capability compared with reference materials, when evaluated as an anode material for lithium ion batteries (LIBs). The simple and effective strategy encompassing nanoscale morphological engineering, surface modification, and doping improves the performance of LTO-based anode materials for high energy density and high power LIBs applied in large scale energy storage. © 2015 Elsevier B.V. All rights reserved.
Keywords: Lithium ion battery Lithium titanium oxide Ultrathin nanosheets Lanthanide doping Carbon coating
1. Introduction The increasing demand for consumer electronics and green transportation spurs the development of intercalated compounds especially for high-performance lithium-ion batteries (LIBs) that
* Corresponding author. Hunan Key Laboratory of Micro-Nano Energy Materials and Devices, School of Physics and Optoelectronics, Xiangtan University, Hunan 411105, China. ** Corresponding author. E-mail addresses: [email protected] (L.W. Yang), [email protected] (P.K. Chu). http://dx.doi.org/10.1016/j.jpowsour.2015.06.131 0378-7753/© 2015 Elsevier B.V. All rights reserved.
can store and deliver energy efficiently [1e4]. Spinel Li4Ti5O12 (LTO) has been demonstrated to be promising anode materials in highperformance LIBs because of its flat charging/discharging plateau at a high potential of 1.55 V vs. Li/Liþ, zero-strain feature towards lithium insertion/extraction, and excellent environmental benignity [5e10]. Unfortunately, practical application of spinel LTO in hybrid electric vehicles and large-scale energy storage has been hampered because of the unsatisfactory rate capability which cannot meet the demand for a high power density due to kinetic problems associated with the poor electrical conductivity (ca.1013 S cm1) and small lithium diffusion coefficient (ca. 109 to 1013 cm2 s1) [8,11e13]. To overcome these hurdles, many
306
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
strategies have been made to ameliorate the rate capability of LTO, for instance, developing various nano/microstructured materials, elemental doping, and surface modification with highly conductive additives. Lou et al. have synthesized mesoporous LTO hollow spheres and observed relatively high rate performance of 104 mA h g1 at 20 C [14]. Isovalent and aliovalent doping (e.g., Br6þ, Zr4þ,V5þ, Zn2þ, Al3þ and Sc3þ) have been observed to increase the electronic conductivity of LTO [8,15e22]. Various carbonmodified spinel LTO composites have attained more than 90% of the theoretical capacity and superior rate capability [11,13,14,23e39]. In particular, by using flexible current collectors such as carbon nanotubes and fibers, flexible LTO composite electrodes have been obtained [11,40]. In addition, combining LTO with conducting metallic nanoparticles can also improve the rate performance of LTO electrodes [41e44]. These improvements are noteworthy, but still insufficient for power and flexibility-oriented applications. It is thus highly desirable to further address the conductivity deficiency of pristine LTO via new composition modulations and nanoengineering combination strategies [30]. Among the various architectures, two-dimensional (2D) nanosheets with a thickness below 10 nm hold great promise in highperformance LIBs because of the short paths enabling fast lithium ion diffusion, large exposed surface, as well as abundant lithium insertion channels [12,28,45e51]. Ultrathin LTO nanosheets are also believed to exhibit a pseudocapacitive effect as the interaction takes place on the surface thus leading to the improved rate capability [48,51]. On the other hand, the construction of a stable separating thin carbon coating around LTO has been reported to be another effective strategy to improve the electrochemical performance of LTO electrodes due to enhancement in the charge transfer kinetics in the LTO electrodes and interfacial stability between the LTO and electrolyte [27,29,50]. Moreover, according to literature, doping lanthanide (Ln3þ) ions (such as La3þ, Y3þ and Gd3þ) with lower valence state yields oxygen ion vacancies in LTO, which could behave as ionic charge carriers to enhance intrinsic conductivity, thereby improving long-term cyclability and rate capabilities [15]. Despite recent research on LTO NSs, LTO with carbon encapsulation, and Ln3þ doped LTO, there have been few studies on highlycrystalline Ln3þ doped and carbon-coated ultrathin Li4Ti5O12 (LTO) nanosheets. Compared to the conventional solid-state reaction, hydrothermal, or solegel method to prepare LTO nano/microstructures, large scale synthesis of highly-crystalline ultrathin LTO nanosheets with a thickness of 10 nm is quite challenging. In our previous reports, we found that using trivalent lanthanide (for example, Gd3þ) to substitute for quadrivalent titanium ions allows the final spinel LTO products to be stabilized, thereby preserving the precursor's sheet morphology via the topotactic transformation of ultrathin layered Li1.81H0.19Ti2O5$xH2O (H-LTO) at 700 C [52]. In this paper, by controlling the amount of carbon precursor (D(t)glucose monohydrate) in the topotactic transformation, large scale synthesis of ultrathin Gd3þ doped and carbon coated LTO nanosheets (denoted as LTO-Gd-C) with high crystallinity and a thickness below 10 nm is accomplished. Owing to the synergistic effects rendered by ultrathin nanosheets with high crystallinity, Gd3þ doping and carbon coating, the materials deliver superior electrochemical performance as anode materials in LIBs in terms of specific capacity, cycling performance, and rate capability. 2. Experimental details 2.1. Materials The synthesis was carried out using commercially available reagents. The GdCl3 6H2O was 99.99% pure and supplied by Sinopharm Chemical Reagent Company. All the other chemicals were
analytical grade and used as received without further purification. 2.2. Synthesis of pure Li4Ti5O12 (LTO) The pure Li4Ti5O12(LTO) was prepared using a hydrothermal method followed by calcination as reported by Wang et al. [12] In the typical synthesis procedures, 1.7 ml (5 mM) of tetrabutyl titanate (TBT) and 0.189 g of LiOH$H2O were thoroughly mixed in 20 ml of ethanol at room temperature. The solution was mixed completely with a magnetic stirrer in a closed container for 24 h and 25 ml of deionized water were added to the container. After stirring for 0.5 h, the solution was transferred to a 50 ml Teflonlined stainless autoclave and placed in an oven at 180 C for 36 h to obtain Li1.81H0.19Ti2O5$xH2O (H-LTO). The white H-LTO powder on the bottom of the reactor was collected, washed with ethanol 3 times, and dried at 80 C for 6 h. Finally, the white hydrothermal HLTO product was heated to 700 C for 6 h in a horizontal tube furnace in air to obtain the pure Li4Ti5O12. 2.3. Synthesis of Gd3þ doped LTO (LTO-Gd) nanosheets The experimental procedures described in Section 2.2 were used to prepare the H-LTO-Gd3þ precursor by doping with the appropriate amount of Gd3þ in the mixture consisting 1.7 ml (5 mM) of tetrabutyl titanate, 0.189 g of LiOH$H2O, and 20 ml of ethanol. The white hydrothermal H-LTO-Gd product was heated to 700 C for 6 h in a horizontal tube furnace in air to obtain Gd3þ doped LTO nanosheets. 2.4. Synthesis of ultrathin Gd3þ doped and carbon coated LTO nanosheets The ultrathin LTO-Gd-C nanosheets were prepared by controlling the amount of carbon precursor (D(t)-glucose monohydrate) used in the experimental procedures described in Section 2.3 for the LTO-Ln3þ doped LTO nanosheets. The H-LTO-Gd precursors were prepared using the experimental procedures described in Section 2.3. 200 mg of the H-LTO-Gd precursor were dispersed in 10 ml of ethanol under magnetic stirring to form Solution A. D(t)glucose monohydrate in the amount corresponding to 2.5, 5, or 10 wt% of the H-LTO-Gd precursor was dissolved in 5 ml of ethanol to produce Solution B. Solution B was added dropwise to Solution A under stirring, followed by slow evaporation at 80 C into a slurry. The slurry was heated under Ar/H2(5%) at 700 C for 6 h. The products with different nominal carbon contents were designated as LTO-Gd-C (2.5%), LTO-Gd-C (5%), and LTO-Gd-C (10%). For comparison, a composite corresponding to LTO-Gd-C (5%) without Gd3þ dopant was synthesized using H-LTO as the precursor and designated as LTO-C (5%). 2.5. Materials characterization The crystal structure of the samples was determined by powder X-ray diffraction (XRD, Rigaku, D/MAX 2500) using a copper Ka radiation source (l ¼ 0.154 nm). X-ray photoelectron spectroscopy (XPS) measurements were performed using an Al Ka source (Kratos Analytical Ltd., UK) and the binding energy of 284.8 eV for C 1s was used as calibration. The morphology and microstructure of the samples were characterized by field-emission scanning electron microscopy (FE-SEM, Hitachi, S4800) equipped with energydispersive X-ray spectroscopy (EDS) and transmission electron microscopy (TEM, JEOL 2100F) equipped with selected-area electron diffraction (SAED). The Raman spectra were recorded on a Renishaw InVia system with the excitation laser (l) of 532 nm. Thermal gravimetric and differential scanning calorimetry (TG-
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
DSC) analyses were carried out on a TGA 2050 thermogravimetric analyzer. 2.6. Electrochemical characterization The electrochemical tests were conducted on the two-electrode CR2032 type coin cells. The working electrodes were prepared by pasting a mixture of the active materials (LTO, LTO-Gd, LTO-C, LTOGd-C), carbon black, and polyvinylidene fluoride (PVDF) with a weight ratio of 80:10:10 onto a Cu foil with a thickness about 0.2 mm and an area about 2 cm2, which acted as the current collector. The electrodes were dried at 80 C for 6 h in air and 120 C in vacuum for another 12 h followed by pressing. Each electrode was weighed accurately on an electronic balance and the weight of the active materials was controlled to be 1e2 mg. The coin cells were assembled in an argon-filled gloved box with a metallic lithium foil as the counter electrode. The electrolyte was 1 M LiPF6 in ethylene carbonate (EC)/dimethyl carbonate (DEC) (1:1, volume ratio) and Celgard 2400 polypropylene was the separator. The charging/discharging experiments were performed at various current densities over a voltage range of 1e2.5 V (vs Li/Liþ) using a multi-channel battery test system (NEWARE BTS-610). Cyclic voltammetry (CV) was performed on an electrochemical workstation (CHI660D) in the voltage range of 1e2.5 V (vs Li/Liþ) at different scanning rates and electrochemical impedance spectroscopy (EIS) was conducted by applying a perturbation voltage of 10 mV in a frequency range of 100 kHz to 10 mHz on a CHI660D electrochemical workstation. 3. Results and discussion The ultrathin spinel LTO nanosheets are prepared by a facile one-pot hydrothermal method in conjunction with postcalcination [12] as schematically illustrated in Fig. 1a. The ultrathin spinel LTO nanosheets are formed by a hydrolysis-ion exchange-Ostwald ripening-phase transition accompanying oxolation and rearrangement of the TiO6 octahedron. Initially, amorphous hydrous titanium oxide (TiO2$nH2O) considered as condensing irregular TiO6 octahedrons, in which each Ti atom is surrounded by six O atoms including two longer (0.1980 nm) TieO bonds and four shorter (0.1934 nm) ones, are produced by hydrolysis of TBT in the LiOH$H2O aqueous solution. During the
Fig. 1. (a) Formation process of the Li4Ti5O12 nanosheets, (b) Effects of Gd3þ doping and carbon coating.
307
hydrothermal treatment in the LiOH solution, the longer TieO bonds are attacked by OH- ions and break, whereas the shorter ones are stronger and so irregular swelling occurs, leading to the formation of some fragments and delamination from TiO2$nH2O. The fragments are linked to each other to form planar intermediate via eOeTieOe ionic bonds. At the same time, some O atoms combine with H2O or H3Oþ via hydrogen bonds, followed by ion exchange between Liþ and Hþ. These planar intermediates then rearrange to form layers composed of [TiO6] octahedra with shared edges and vertices and with Liþ and H2O or H3Oþ intercalated between the layers forming planar C-base-centered orthorhombic HLTO seed nuclei. As the reaction proceeds, the planar seed nuclei grow continuously via Ostwald ripening to produce the layered HLTO. Topotactic transformation of the H-LTO nanosheets to spinel LTO is accomplished by post-calcination at 700 C. The heat treatment not only removes H2O, H3Oþ, and Li2O molecules, but also reconstructs the layers during which some [TiO6] octahedra retain the same morphology while others restructure to form [TiO4] tetrahedra. However elastic strain inevitably results from the mismatch of the lattice of the LTO and H-LTO phase during the topotactic reaction. When the strain is beyond a critical value, break-up of large nanosheets occurs with formation of small flakes, which facilitate second Ostwald ripening of LTO. As the reaction proceeds, the LTO nanosheets often become irregular and thick. However, when trivalent Ln3þ ions are introduced, they occupy the Ti4þ sites and promote the formation of holes in the H-LTO upon annealing accompanied by removal of H2O, H3Oþ, and Li2O molecules, which may be filled with electrons from the LTO nuclei. In order to establish charge balance, extra positive ions have to be introduced into the surface of the LTO nuclei (see Fig. 1b). Such a charge transfer enables strong electrostatic interactions between the LTO nuclei and the H-LTO nanosheets to produce polaronic lattice distortion which effectively buffer the elastic strain in topotactic transformation. On the other hand, owing to the chargemediated bond reinforcement, the LTO nuclei tend to maximize the contact area with the H-LTO nanosheets and form 2D nuclei with larger characteristic dimensions [53,54]. The larger 2D nuclei benefit preferential adsorption of the reconstructed molecules on the side plane and hence, second Ostwald ripening occurs along the direction of the normal surface being hindered. As a result, textural modification and break-up associated with elastic deformation is depressed notably and the ultrathin LTO nanosheets are stabilized and retain the precursor morphology.
Fig. 2. (a)e(c) XRD patterns of H-LTO, pure LTO, and Gd3þ-LTO, respectively.
308
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
Fig. 2a depicts the X-ray diffraction (XRD) patterns of the hydrothermal products, suggesting that they are crystals composed of layered H-LTO with a C-base centered orthorhombic lattice (JCPDS No. 47-0123). Transmission electron microscopy (TEM) (see Fig. 3a) indicates that the hydrothermal H-LTO consists of large flat nanosheets with a diameter larger than 500 nm with smooth edges (see Fig. S1). The higher magnification image of individual vertical nanosheets reveals that the thickness of the nanosheets is about 8 nm. Fig. 3b displays the corresponding SAED pattern acquired from a single nanosheet indicating that the H-LTO nanosheets are single crystal. The bright spots in the SAED pattern can be indexed to the (002), (020), and (022) planes of orthorhombic H-LTO from the [100] zone axis. The HR-TEM image of an individual nanosheet in Fig. 3c shows that the lattice spacing is about 0.19 nm that corresponds to the lattice spacing of the (020) plane in orthorhombic H-LTO. In conjunction with the SAED results, it can be inferred that single-crystal H-LTO nanosheets grow along the (100) facet. Fig. 2b shows the XRD pattern of the samples of pure LTO, which can be well indexed to spinel LTO (JCPDS No. 49-0207). No impurities are observed confirming complete conversion from H-LTO to spinel LTO. Fig. 3d depicts the representative TEM image of the samples obtained after calcination at 700 C. The essential features indicative of thickening, textural modification including angulation and irregular edge, and small flakes suggesting interlayer splitting when the H-LTO nanosheets are topotactically converted into spinel LTO, can be observed, which is consistent with the results reported in Ref. [13]. Fig. 3f shows the representative SAED pattern taken from an individual sheet suggesting high crystallinity. However, in the irregular and deformed LTO nanosheets, it is rather difficult to observe a definite and reproducible orientation relationship for the H-LTO by SAED as previously observed from other topotactic transformation systems [55]. As shown in Fig. 3e, the crystalline region with clear lattice fringes has an inter-planar spacing of approximately 0.48 nm corresponding to the (111) atomic planes of spinel LTO and in good agreement with the SAED patterns. Fig. 2c shows the XRD patterns of the LTO sample after
doping Gd3þ with nominal content of 10wt%. All the peaks can be indexed to the spinel structure LTO and no other peaks such as Gd2O3 and TiO2 can be detected. The results are consistent with those of the pure LTO sample. Broadened diffraction peaks are observed suggesting reduced average crystallite size. Fig. 3g shows the typical TEM image of the Gd3þ doped LTO sample and it retains the original sheet-like morphology of the H-LTO precursor. The representative SAED patterns and HR-TEM image in Fig. 3h and i suggest single crystallinity in the Gd3þ doped LTO nanosheets grown along the ð011Þ plane. Furthermore, a definite and reproducible orientation relationship with H-LTO is revealed by the SAED pattern of the Gd3þ doped LTO nanosheets and the reaction from HLTO to LTO is highly topotactic with the following relationship:
ð100ÞHLTO
. 011
ð020ÞHLTO
ð400ÞLTO and ð022ÞHLTO ð444ÞLTO
LTO
ð002ÞHLTO
ð333ÞLTO
Interestingly, the detectable Gd3þ, which is below the detection limitation of EDS measurement and can be confirmed by high resolution XPS (see Fig. S2) is far below the nominal content. The results indicate that it is difficult to dope Gd3þ into LTO host due to large difference of ionic diameters between Gd3þ and Ti4þ. In our experiments, we found that the contents of Gd3þ doping affect notably the nucleation and growth behavior of LTO nanosheets during topotactic transformation at 700 C. With different Gd3þ doping contents, the microstructures and electrochemical performances of the obtained LTO samples are different (see Fig. S3). Similar effect on size and surface texture of LTO nanosheets was also observed via doping other Ln3þ ions such as Y3þ and La3þ. The influence of Ln3þ doping on the morphology, growth mechanism and electrochemical performance is in progress. Herein, Gd3þ ion with nominal doping content of 10wt% is used as a typical example to improve lithium storage performance via Ln3þ doping and carbon coating. Subsequently, the mixture consisted of Gd3þ-doped H-LTO and
Fig. 3. Typical TEM and HR-TEM images of (a, b) H-LTO, (d, e) Pure LTO and (g, h) Gd3þ-LTO nanosheets as well as corresponding SAED patterns (c, f, i). “H” and “L” represent H-LTO and LTO, respectively.
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
D(t)-glucose monohydrate at a predetermined weight ratio of 5% is heated to 700 C to produce ultrathin Gd3þ doped and carbon coated LTO nanosheets (see Fig. 1b). For comparison, the LTO-C sample with a nominal 5 wt% carbon coating is prepared using HLTO as the precursor. The TEM observations (see Fig. S4) indicate that the carbon coating obtained from D(þ)-glucose monohydrate has the ability to stabilize the surface of the LTO nanostructures and preserve the precursor morphology during topotactic transformation. The results are consistent with those reported in carbon-encapsulated F-doped LTO with a ball-in-ball morphology [30] and self-supported LTO-C nanotube arrays [29]. The actual carbon coating on the LTO-C sample is about 5.1 wt% (see Fig. S4), which is smaller than that of the LTO-Gd-C sample to be discussed later. Nevertheless, the electrochemical performance of LTO-C is rather unsatisfactory due to the lack of synergistic interactions between Gd3þ and carbon coating to be discussed later. Raman scattering is employed to examine the microstructure of the carbon layer on the LTO-Gd-C (5%) sample and as shown in Fig. 4a, the two small bands at 1178 and 1493 cm1 are short-range vibration of sp3 carbon atoms commonly observed from amorphous carbonaceous materials [50]. The two main bands at around 1330 and 1592 cm1 can be designated as the D and G bands of sp2 carbon, respectively. The surface chemical composition and states of the LTO-Gd-C (5%) sample are determined by XPS. Fig. 4b and S7 reveal the presence of Li, Ti, O, C, and Gd and the results are in good agreement with the nominal atomic composition of the LTO-Gd-C sample. The weak Gd 3d peak in Fig. S5 suggests a small concentration of Gd3þ. The TG analysis (TGA) (Fig. 4c) reveals that the carbon content in the LTOGd-C (5%) sample is about 6.5 wt%. Fig. 4d shows the representative XRD patterns of the LTO-Gd-C (5 wt%) sample and the peaks can be indexed to a cubic spinel structure according to the Fd3m space group without impurities. The mean crystallite size of LTO-Gd-C calculated from the (111) diffraction peak by the Scherer equation is smaller than those of the pure LTO, LTO-Gd, and LTO-C. This decreasing trend indicates that pyrolyzed carbon also restrains the
309
growth of LTO nanoparticles in LTO-Gd-C thereby preserving the nanoscale advantage. Fig. 5a shows the typical TEM image of the LTO-Gd-C sample, demonstrating that the sample retains the sheetlike morphology of the H-LTO precursor. The representative SAED pattern and HR-TEM results acquired from individual nanosheets (see Fig. 5d and inset of Fig. 5b) suggest single crystallinity in the LTO-Gd-C nanosheets with the feature of topotactic transformation. Comparison of the TEM images from LTO-Gd (see Fig. 5c) and LTOGd-C nanosheets (see Fig. 5b) with typical HR-TEM images (see Fig. 5d) shows the presence of an amorphous carbon coating on the surface of LTO nanosheets. Well-defined lattice fringes with a separation of 0.48 nm in the HR-TEM image of the carbon-coated LTO nanosheets correspond to the spacing of the (111) planes in the spinel LTO. In addition, as shown in the inset in Fig. 5a, the typical vertical nanosheet about 10 nm thick can be observed. Atomic force microscopy (see the inset in Fig. 5d) reveals the 2D sheet-like feature of the LTO-Gd-C nanosheets with a thickness of about 10 nm. According to TEM and AFM, the thickness of the primary LTO in LTO-Gd-C nanosheets can be inferred to be less than 10 nm due to the existence of coated carbon. To determine the effects of the carbon content on the electrochemical performance, different carbon coatings are prepared by mixing the H-LTO-Gd precursor with different amounts of D(t)glucose monohydrate (2.5, 5 and 10 wt% of the H-LTO precursor). As shown in Fig. S6aec, the nominal 5 wt% carbon coating delivers the best electrochemical performance. Fig. 6a shows the CV curve of the LTO-Gd-C nanosheets at a scanning rate of 0.1 mV/s in the potential range between 2.5 and 1.0 V (vs Li/Liþ). The well-defined redox peaks at 1.73/1.44 V can be attributed to the redox reaction of Ti4þ/ Ti3þ, respectively, associated with lithium insertion/extraction in the spinel LTO lattice [29]. The strong and well-defined redox peaks of LTO indicate well-behaved electrode kinetics of the LTO-Gd-C nanosheets electrode. Fig. 6b depicts the chargingedischarging voltage profiles of the LTO, LTO-Gd, LTO-C (5%), and LTO-Gd-C (5%) electrodes for the 1st, and 100th cycles at a current rate of 10 C within
Fig. 4. (a)e(c) Raman spectrum, TG-DSC curves and survey XPS spectrum of the LTO-Gd-C nanosheets; (d) XRD patterns of LTO-Gd-C nanosheets together with other LTO-C (5%), LTO-Gd, and LTO reference samples.
310
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
Fig. 5. (a) Typical TEM image of the LTO-Gd-C nanosheets with the inset showing the enlarged image of the marked area; (b) and (c) High-magnification TEM images of LTO-Gd-C nanosheets and reference LTO-Gd nanosheets with the inset in (b) showing the SEAD patterns of the LTO-Gd-C nanosheets; (d) HR-TEM and AFM images of LTO-Gd-C nanosheets.
Fig. 6. (a) CV curves acquired at a scanning rate of 0.1 mV s1 in the voltage range of 1.0e2.5 V of LTO-Gd-C nanosheets; (b) Chargingedischarging voltage profiles of the LTO-Gd-C nanosheets and other reference materials for the 1st and 100th cycles at a current rate of 10 C; (c) Cycling performance and Coulombic efficiency versus cycle number at different current rates from 1 C to 30 C of the LTO-Gd-C nanosheets and other reference materials. (d) Cycling performance of the LTO-Gd-C nanosheets and other reference materials at the standard dischargingecharging rate of 10 C.
a cut off window of 1.0e2.5 V. All the electrodes exhibit similar charging and discharging flat plateaus at around 1.5 V resulting from a two-phase reaction during electrochemical lithium insertioneextraction according to the following equation [23,38].
Li4 Ti5 O12 þ xLiþ þ xe 4Li4þx Ti5 O12 ð0 < x < 3Þ
(1)
The results indicate that Gd3þ doping and carbon coating do not alter the electrochemical reaction of LTO. However, the magnified chargingedischarging voltage profiles (see Fig. S6d) of the four electrodes for the 1st and 100th cycles at a current rate of 10 C reveal that the LTO-Gd-C (5%) electrode has a flatter profile than the other reference materials. The polarization between the charging
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
and discharging plateaus are 283 and 477 mV for the 1st and 100th cycles, respectively, of the pure LTO electrode, which are larger than those of the LTO-Gd (149 mV and 130 mV) and LTO-C (5%) (160 mV and 175 mV) electrodes. The lower electrochemical polarization suggests improved reaction kinetics due to the special structure and higher surface electronic conductivity of the LTO-Gd nanosheets and LTO nanosheets with carbon coatings. Electrochemical polarization is further depressed (100 mV and 128 mV at the 1st and 100th cycles, respectively) in the LTO-Gd-C nanosheets, implying the synergistic effects of the Gd3þ doping and carbon coating on the reaction kinetics in the electrode. Fig. 6c shows the LTO-Gd-C nanosheets cycled at different current rates from 1 C to 30 C in the voltage range between 1.0 and 2.5 V. At 1 C (1 C ¼ 175 mAg1), the LTO-Gd-C electrode shows a discharge capacity of 213 mAh g1, which are higher than the theoretical capacity of 175 mA h g1. A similar phenomenon has also been observed in self-assembled TiO2egraphene hybrid nanostructures [56e58]. This suggests the existence of additional lithium storage sites in the LTO-Gd-C electrode. On one hand, an extended solid solution region for Liþ insertion into LTO was observed obviously after Gd3þ doping or carbon coating in chargingedischarging voltage profiles (see Fig. S7) starting from the open-circuit potential (about 2.3 V) in voltage to ~1.5 V, similar to that which has been reported previously for small particles such as TiO2 and Si, in which an additional surface lithium storage mechanism exists [59,60]. On the other hand, shallower slope below 1.5 V that is a typical feature of supercapacitor in chargingedischarging voltage profiles was observed obviously after carbon coating, implying higher capacitance due to the insertion of lithium ions into the surface layer of the LTO-C nanosheets [59]. Moreover, it is inferred that the synergistic effects of the Gd3þ doping and carbon coating play a key role in further enhance surface lithium storage, thereby giving rise to a high capacity of 213 mAh g1 for the LTO-Gd-C nanosheets since the chargingedischarging voltage profile of the LTO-Gd-C electrode demonstrates the most extended solid solution region from the open-circuit potential in voltage to ~1.5 V and the shallowest slope in below 1.5 V. As the rate increases from 1 to 5 and 10 C, the discharge capacity decreases from 213 mAh g1 to 182 and 171 mA hg1, respectively. At a high rate of 30 C, the capacity retains still 146 mAh g1. Furthermore, the Coulombic efficiency of the LTO-Gd-C nanosheets at large charging/discharging rates approach 100% for each cycle. For comparison, the capacities of LTO, LTO-C (5%), LTO-Gd at 30 C are 96 mAh g1, 104 mAh g1, and 120 mAh g1, respectively. The results reveal that lithium diffusion is efficient and responsible for the excellent rate capability of the LTO-Gd-C nanosheets. With regard to the cycling behavior, we examine the cycling performance of the LTO-Gd-C nanosheets at a high current density of 10 C. The LTO-Gd-C anode at a rate of 10 C reaches a stable capacity of 168 mAh g1 with the Coulombic efficiency approaching 100% after 20 cycles. It retains a capacity of 152 mAh g1 after 500 cycles with a capacity loss of only 9.2%, as shown in Fig. 6d. The cycling performance of the LTO-Gd-C nanosheets is superior to those of pure LTO, LTO-C, and LTO-Gd nanosheets. In addition, we compare the present data with those observed from other LTO-based electrodes including LTO nanowire arrays [33],TiO2-coated LTO [12], carbon-coated LTO [27], Cu-doped LTO-TiO2 nanosheets [47], N-doped carbon-coated LTO [39], mesoporous LTOC [36], nanostructured carbon-coated LTO [10,25], self-supported LTO nanosheets arrays [49], LTO-C nanotube arrays [29], carbon-encapsulated F-doped LTO [30], LTO/graphene nanosheets [35], LTO tailored by cathodically induced graphene [38], etc. The rate and cycling performance of the LTO-Gd-C nanosheets are comparable to the best results found in the literature according to Table SI, confirming the positive effects of Gd3þ doping and carbon coating.
311
The superior electrochemical performance of the LTO-Gd-C nanosheets can be attributed to enhancement of the transport kinetics in the electrode. Owing to the unique two-phase reaction during electrochemical lithium insertioneextraction shown in Eqn. (1), LTO materials with an excellent rate capability possess both improved electrical conductivity and lithium ion diffusion, which can be studied by electrochemical impedance spectroscopy (EIS). The LTO-Gd-C nanosheets and other reference electrodes are analyzed by EIS at open-circuit voltages in the native states (before cycling) and after different discharging and charging cycles. The Nyquist plots in Fig. 7aed reveal a common purely resistive response at high frequencies represented by the ohmic resistance (Rs) of the electrode and electrolyte, a semicircle due to the charge-transfer impedance (Rct) on the electrodeeelectrolyte interface in the high-to-middle frequency region, and an inclined straight line ascribed to the Warburg impedance (Zw) in the low frequency region. The EIS spectra are simulated by the Z-view software using the inset equivalent circuit model and the corresponding simulation parameters are presented in Table 1. One can find that the semicircular arc of the LTO-Gd-C nanosheets electrode is much smaller than that of other reference electrodes. EIS performed after cycling displays the same trend with a smaller arc in the Nyquist plot. The smaller semicircular arc is an indication of an overall smaller charge transfer resistance or more facile charge transfer process at the electrode/electrolyte interface for the LTO-Gd-C nanosheets electrode [38,58]and the results are consistent with the simulation results in Table 1. The lithium ion diffusion coefficient (DLi) can be evaluated by EIS according to the following equations [50,52].
DLi ¼
R2 T 2 2A2 n4 F 4 C 2 s2
Z 0 ¼ Rs þ Rct þ su1=2
(2)
(3)
where R is the gas constant, T is the absolute temperature, A is the surface area of the electrode, n is the number of electrons transferred in the half reaction of the redox couple, F is Faraday's constant, C is the concentration of lithium ions, and s is the Warburg factor. The value of s can be obtained from the slope of the lines 0 between Z and 61=2 as shown in Fig. 7eef. The DLi values of the electrodes can be calculated and the results after 2 and 100 cycles are presented in Table 1. The DLi values of the LTO-Gd-C nanosheets are larger than those of other reference electrodes suggesting that the LTO-Gd-C nanosheets possess faster solid-state lithium ion diffusion thus enabling fast migration of lithium ions to the interior of the LTO NSs and full utilization of the active materials. The results also confirm that Gd3þ doping and carbon coating cooperatively improve the solid-state lithium ion diffusion coefficient and electrical conductivity of the LTO-Gd-C nanosheets and relieve the polarization of the electrode to enhance the cycling and rate capability. The smaller Rct and larger DLi of the LTO-Gd-C nanosheets result from the special structure and Gd3þ doping. Firstly, using a small amount of trivalent Gd3þ to substitute for quadrivalent titanium ions in the LTO nanosheets not only allows the products to be stabilized to a thickness about 10 nm via topotactic transformation of the layered H-LTO, but also provide electrical donor to improve intrinsic conductivity of ultrathin nanosheets, when the incorporated Gd3þ ions occupy the 16d site in LTO matrix, thereby depressing the electrochemical polarization and charge transfer resistance and expediting lithium ion diffusion during lithium intercalation and de-intercalation. Secondly, complete encapsulation of the LTO nanosheets by a pervasive network of
312
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
0
Fig. 7. (a)e(d) Nyquist plots of the LTO-Gd-C nanosheets and other reference materials after different discharging and charging cycles. (e) and (f) Z plots against 61=2 at the low frequency region (10e0.1 Hz) for the LTO-Gd-C nanosheets and other reference materials after the second and 100th cycles of discharging and charging.
Table 1 Rct and Di values for the LTO, LTO-Gd, LTO-C and LTO-Gd-C electrodes after different cycles in half-cells according to Fig. 7. Compound
2nd LTO LTO-C (5%) LTO-Gd LTO-Gd-C (5%)
Di (cm2 s1)
Rct (U)
153 117 105.61 88.5
10th 126.7 97.1 95.1 80.2
30th 119.7 89.4 88.1 72.1
carbon from pyrolysis of D(t)-glucose monohydrate preserves the unique ultrathin architecture of the precursor during the transformation from H-LTO to LTO due to growth inhibition of encapsulated LTO during the high temperature treatment and also decreases the external electron transport resistance making it easier for lithium ions to traverse to the primary LTO matrix through the layer. Thirdly, the ultrathin nanosheets with singlecrystalline nature not only promote the kinetics of lithiation and delithiation due to the shorter Li-ion diffusion distances and avoiding the drawback of many grain boundaries in electrodes of
60th 103.4 70.8 68.8 52.1
100th 84.4 57.3 55.42 38.5
2nd
100th 12
4.6 10 1.47 1011 1.66 1011 6.03 1011
1.6 1013 5.23 1012 6.16 1012 1.83 1011
reported LTO nanostructure such as polycrystalline nanowires, polycrystalline nanotube and thick LTO nanosheets, but also produce larger discharge capacity since the storage limit at the 16c sites of spinel structure is exceeded and co-occupation of the 8a and 16c sites in the surface region is expected to be more significant in ultrathin LTO nanosheets where the specific surface area is larger [30], resulting in storage properties different from those of the bulk, polycrystalline nanowires, polycrystalline nanotube and thick LTO nanosheets. Consequently, superior electrochemical performances are demonstrated in the ultrathin LTO-Gd-C nanosheets.
G.B. Xu et al. / Journal of Power Sources 295 (2015) 305e313
4. Conclusion In summary, we have prepared highly-crystalline gadolinium doped and carbon-coated ultrathin Li4Ti5O12 (LTO) nanosheets with thickness below 10 nm on large scale by controlling the amount of carbon precursor in the topotactic transformation of layered ultrathin Li1.81H0.19Ti2O5$xH2O (H-LTO) nanosheets at 700 C. Gadolinium doping allows the spinel LTO products to be stabilized, thereby preserving the precursor's sheet morphology and single crystal structure with enhanced intrinsic electron conductivity. Carbon encapsulation serves dual functions by restraining crystal growth of the LTO primary nanoparticles in the LTO-Gd-C nanosheets and decreasing the external electron transport resistance. Owing to the synergistic effects rendered by ultrathin nanosheets with high crystallinity, Gd3þ doping and carbon coating, the obtained materials have excellent specific capacity, cycling performance, and rate capability as anode materials in LIBs. The desirable effects stem from the larger contact surface area with the electrolyte, shorter transport length of lithium ions, bigger lithium diffusion coefficient, and enhanced intrinsic conductivity. The strategy of depending on the heterovalent Ln3þ doping to control microstructure and properties is simple and effective. The results reveal the benefits of employing multiple modification strategies such as nanoscale and morphological engineering, surface modification, and doping to improve the performance of energy storage materials applied in high energy density and high power devices. Acknowledgments This work was financially supported by the Grants from National Natural Science Foundation of China (Nos. 11474242, 51272220, 11374252, and 51472209), the Program for Changjiang Scholars and Innovative Research Team in University (IRT13093), Guangdong e Hong Kong Technology Cooperation Funding Scheme (TCFS) GHP/ 015/12SZ (CityU 9440103), and City University of Hong Kong Applied Research Grant No. 9667085. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2015.06.131. References [1] M. Armand, J.-M. Tarascon, Nature 451 (2008) 652e657. [2] X. Chen, C. Li, M. Graetzel, R. Kostecki, S.S. Mao, Chem. Soc. Rev. 41 (2012) 7909e7937. [3] Y.G. Guo, J.S. Hu, L.J. Wan, Adv. Mater. 20 (2008) 2878e2887. [4] C. Liu, F. Li, L.P. Ma, H.M. Cheng, Adv. Mater. 22 (2010) E28eE62. [5] S. Ganapathy, M. Wagemaker, ACS Nano 6 (2012) 8702e8712. [6] G. Jeong, Y.-U. Kim, H. Kim, Y.-J. Kim, H.-J. Sohn, Energy Environ. Sci. 4 (2011) 1986e2002. [7] C. Ouyang, Z. Zhong, M. Lei, Electrochem. Commun. 9 (2007) 1107e1112. [8] T.-F. Yi, S.-Y. Yang, Y. Xie, J. Mater. Chem. A 3 (2015) 5750e5777. [9] X. Lu, L. Zhao, X. He, R. Xiao, L. Gu, Y.S. Hu, H. Li, Z. Wang, X. Duan, L. Chen, Adv. Mater. 24 (2012) 3233e3238. [10] H.-G. Jung, M.W. Jang, J. Hassoun, Y.-K. Sun, B. Scrosati, Nat. Commun. 2 (2011) 516. [11] N. Li, Z. Chen, W. Ren, F. Li, H.-M. Cheng, Proc. Natl. Acad. Sci. 109 (2012) 17360e17365. [12] Y.-Q. Wang, L. Gu, Y.-G. Guo, H. Li, X.-Q. He, S. Tsukimoto, Y. Ikuhara, L.-J. Wan, J. Am. Chem. Soc. 134 (2012) 7874e7879. [13] G.-N. Zhu, H.-J. Liu, J.-H. Zhuang, C.-X. Wang, Y.-G. Wang, Y.-Y. Xia, Energy Environ. Sci., 4 4016e4022. [14] L. Yu, H.B. Wu, X.W.D. Lou, Adv. Mater. 25 (2013) 2296e2300.
313
[15] Y.-J. Bai, C. Gong, N. Lun, Y.-X. Qi, J. Mater. Chem. A 1 (2013) 89e96. [16] Y.-J. Bai, C. Gong, Y.-X. Qi, N. Lun, J. Feng, J. Mater. Chem. 22 (2012) 19054e19060. [17] J.G. Kim, M.S. Park, S.M. Hwang, Y.U. Heo, T. Liao, Z. Sun, J.H. Park, K.J. Kim, G. Jeong, Y.J. Kim, ChemSusChem 7 (2014) 1451e1457. [18] C. Lin, B. Ding, Y. Xin, F. Cheng, M.O. Lai, L. Lu, H. Zhou, J. Power Sources 248 (2014) 1034e1041. [19] T.-F. Yi, Y. Xie, Q. Wu, H. Liu, L. Jiang, M. Ye, R. Zhu, J. Power Sources 214 (2012) 220e226. [20] Y. Zhang, C. Zhang, Y. Lin, D.-B. Xiong, D. Wang, X. Wu, D. He, J. Power Sources 250 (2014) 50e57. [21] H. Song, S.-W. Yun, H.-H. Chun, M.-G. Kim, K.Y. Chung, H.S. Kim, B.-W. Cho, Y.T. Kim, Energy Environ. Sci. 5 (2012) 9903e9913. [22] G. Du, N. Sharma, V.K. Peterson, J.A. Kimpton, D. Jia, Z. Guo, Adv. Funct. Mater. 21 (2011) 3990e3997. € blinger, D. Fattakhova-Rohlfing, T. Bein, Angew. [23] J.M. Feckl, K. Fominykh, M. Do Chem. Int. Ed. 51 (2012) 7459e7463. [24] X. Hao, B.M. Bartlett, Adv. Energy Mater. 3 (2013) 753e761. [25] H.-G. Jung, S.-T. Myung, C.S. Yoon, S.-B. Son, K.H. Oh, K. Amine, B. Scrosati, Y.K. Sun, Energy Environ. Sci. 4 (2011) 1345e1351. [26] H.-K. Kim, S.-M. Bak, K.-B. Kim, Electrochem. Commun. 12 (2010) 1768e1771. [27] B. Li, C. Han, Y.-B. He, C. Yang, H. Du, Q.-H. Yang, F. Kang, Energy Environ. Sci. 5 (2012) 9595e9602. [28] N. Li, G. Zhou, F. Li, L. Wen, H.M. Cheng, Adv. Funct. Mater. 23 (2013) 5429e5435. [29] J. Liu, K. Song, P.A. van Aken, J. Maier, Y. Yu, Nano Lett. 14 (2014) 2597e2603. [30] Y. Ma, B. Ding, G. Ji, J.Y. Lee, ACS Nano 7 (2013) 10870e10878. [31] H. Pan, L. Zhao, Y.S. Hu, H. Li, L. Chen, ChemSusChem 5 (2012) 526e529. [32] K.-S. Park, A. Benayad, D.-J. Kang, S.-G. Doo, J. Am. Chem. Soc. 130 (2008) 14930e14931. [33] L. Shen, E. Uchaker, X. Zhang, G. Cao, Adv. Mater. 24 (2012) 6502e6506. [34] L. Shen, C. Yuan, H. Luo, X. Zhang, K. Xu, Y. Xia, J. Mater. Chem. 20 (2010) 6998e7004. [35] L. Shen, C. Yuan, H. Luo, X. Zhang, S. Yang, X. Lu, Nanoscale 3 (2011) 572e574. [36] L. Shen, X. Zhang, E. Uchaker, C. Yuan, G. Cao, Adv. Energy Mater. 2 (2012) 691e698. [37] Y. Tang, F. Huang, W. Zhao, Z. Liu, D. Wan, J. Mater. Chem. 22 (2012) 11257e11260. [38] Y. Yang, B. Qiao, X. Yang, L. Fang, C. Pan, W. Song, H. Hou, X. Ji, Adv. Funct. Mater. 24 (2014) 4349e4356. [39] L. Zhao, Y.S. Hu, H. Li, Z. Wang, L. Chen, Adv. Mater. 23 (2011) 1385e1388. [40] N. Li, T. Mei, Y. Zhu, L. Wang, J. Liang, X. Zhang, Y. Qian, K. Tang, CrystEngComm 14 (2012) 6435e6440. [41] J.-G. Kim, D. Shi, M.-S. Park, G. Jeong, Y.-U. Heo, M. Seo, Y.-J. Kim, J.H. Kim, S.X. Dou, Nano Res. 6 (2013) 365e372. [42] M. Krajewski, M. Michalska, B. Hamankiewicz, D. Ziolkowska, K.P. Korona, J.B. Jasinski, M. Kaminska, L. Lipinska, A. Czerwinski, J. Power Sources 245 (2014) 764e771. [43] Z. Liu, N. Zhang, Z. Wang, K. Sun, J. Power Sources 205 (2012) 479e482. [44] W. Wang, Y. Guo, L. Liu, S. Wang, X. Yang, H. Guo, J. Power Sources 245 (2014) 624e629. [45] F. Bonaccorso, L. Colombo, G. Yu, M. Stoller, V. Tozzini, A.C. Ferrari, R.S. Ruoff, V. Pellegrini, Science 347 (2015) 1246501. [46] Y. Sun, S. Gao, Y. Xie, Chem. Soc. Rev. 43 (2014) 530e546. [47] C. Chen, Y. Huang, C. An, H. Zhang, Y. Wang, L. Jiao, H. Yuan, ChemSusChem 8 (2015) 114e122. [48] J. Chen, L. Yang, S. Fang, Y. Tang, Electrochim. Acta 55 (2010) 6596e6600. [49] S. Chen, Y. Xin, Y. Zhou, Y. Ma, H. Zhou, L. Qi, Energy Environ. Sci. 7 (2014) 1924e1930. [50] C. Han, Y. He, B. Li, H. Li, J. Ma, H. Du, X. Qin, Q. Yang, F. Kang, ChemSusChem 7 (2014) 2567e2574. [51] C. Lai, Y. Dou, X. Li, X. Gao, J. Power Sources 195 (2010) 3676e3679. [52] G.B. Xu, W. Li, L.W. Yang, X.L. Wei, J.W. Ding, J.X. Zhong, P.K. Chu, J. Power Sources 276 (2015) 247e254. [53] X. Shao, S. Prada, L. Giordano, G. Pacchioni, N. Nilius, H.J. Freund, Angew. Chem. Int. Ed. 50 (2011) 11525e11527. [54] N. Mammen, S. Narasimhan, S.d. Gironcoli, J. Am. Chem. Soc. 133 (2011) 2801e2803. [55] M. Figlarz, B. Gerand, A. Delahaye-Vidal, B. Dumont, F. Harb, A. Coucou, F. Fievet, Solid State Ionics 43 (1990) 143e170. [56] X. Xin, X. Zhou, J. Wu, X. Yao, Z. Liu, ACS Nano 6 (2012) 11035e11043. [57] S. Yang, X. Feng, K. Müllen, Adv. Mater. 23 (2011) 3575e3579. [58] S. Yu, L. Yang, Y. Tian, P. Yang, F. Jiang, S. Hu, X. Wei, J. Zhong, J. Mater. Chem. A 1 (2013) 12750e12758. [59] J. Wang, J. Polleux, J. Lim, B. Dunn, J. Phys. Chem. C 111 (2007) 14925e14931. [60] B. Wang, X.L. Li, B. Luo, L. Hao, M. Zhou, X.H. Zhang, Z.G. Fan, L.J. Zhi, Adv. Mater. 27 (2015) 1526e1532.