Self-Assembly of Bi2Te3-Nanoplate/Graphene-Nanosheet Hybrid by One-Pot Route and Its Improved Li-Storage Properties

A sandwich structured Bi2Te3-nanoplates/graphene-nanosheet (Bi2Te3/G) hybrid has been synthesized by a facile in situ solvothermal route and has been investigated as a potential anode material for Li-ion batteries. Bi2Te3 grows during the solvothermal process with the simultaneous reduction of graphite oxide into graphene. The in situ formation process of the hybrid has been investigated by X-ray diffraction and X-ray photoelectron spectra. The Li-storage mechanism and performance of Bi2Te3/G and bare Bi2Te3 have been studied by galvanostatic cycling and cyclic voltammetry. The Bi2Te3/G sandwich exhibits an obviously improved cycling stability compared to bare Bi2Te3. The enhancement in electrochemical performance can be attributed to the combined conducting, confining and dispersing effects of graphene for Bi2Te3 nanoplates and to the self-assembled sandwich structure.


Introduction
Although carbon based materials are presently the dominant anodes in commercial Li-ion batteries, there is still a great challenge to develop alternative anode materials in order to meet the requirement of high-performance Li-ion batteries [1,2]. Alloy materials have been regarded as one of the most promising anodes for the next-generation Li-ion batteries because of their safe operating voltage and OPEN ACCESS the large energy density, especially the large volumetric energy density due to the high density [3,4]. Among various elements that can form Li-alloys, Sb has received a special interest because of its appropriate alloying/de-alloying voltage (around 0.8 V vs. Li/Li + ) and relatively large Li-storage capacity (660 mAh g −1 ) with the formation of Li 3 Sb composition [5]. However, the volume change is over 200% during the conversion from Sb to Li 3 Sb [6]. The volume change causes the cracking and crumbling of the particles, resulting in the loss of physical contact between the active particles and the current collector and the consequent capacity loss upon cycling.
Many measures have been taken to alleviate the volume changes for Sb during alloying/de-alloying processes. Previous work [7][8][9][10] showed that the cycling stability of metallic Sb can be improved by using Sb-based intermetallic compound MSb x , where M is an electrochemically inert element and acts as the buffering matrix. The improvement in electrochemical performance could also be realized by using nanostructured materials [11][12][13][14] due the large surface area, short Li-ion diffusion path and increased mechanical strength. However, the nanoparticles tend to aggregate, also leading to the disconnection of active material with the current collector. An effective approach to overcome this problem is to immobilize the nanoparticles on a matrix. Carbon-based materials [15][16][17][18] are considered as the ideal matrices since they are electrically conductive and electrochemically active in addition to the buffering effect for the volume change.
Graphene, a flat monolayer of sp 2 -bonded carbon atoms [19], was also the ideal matrix because of its appealing properties such as high electronic conductivity [20], high specific surface area [21] and high mechanical strength [22]. Our previous work showed that the cycling stability of some Sb-based intermetallic compounds could be improved by forming nanocomposites with graphene [23,24]. Similar to Sb, Bi in the same group can also electrochemically store Li by forming a Li 3 Bi composition [5,[25][26][27]. So far, the effect of graphene on the electrochemical performance of Bi-based anodes has not been reported yet. Herein, an in situ solvothermal route was used to prepare the Bi 2 Te 3 -nanoplate/graphene-nanosheet (Bi 2 Te 3 /G), where Te can also reversibly store Li by forming Li 2 Te [27]. The results showed that an obvious improvement in cycling stability of Bi 2 Te 3 could be realized by constructing a Bi 2 Te 3 /G hybrid with a sandwich structure. Figure 1a shows the XRD patterns of Bi 2 Te 3 /G and bare Bi 2 Te 3 . For comparison, the standard diffraction peaks of Bi 2 Te 3 are also given. All the diffraction peaks can be indexed to hexagonal Bi 2 Te 3 (JCPDS No. 89-2009) for both Bi 2 Te 3 /G and bare Bi 2 Te 3 . The diffraction peaks of graphene that should appear at around 2θ = 25 °C cannot be detected, suggesting that the restacking of graphene sheets after reduction was inhibited by uniformly loading Bi 2 Te 3 plates between the graphene sheets. The graphene content is roughly estimated to be 11.5 wt% by the carbon content analysis.

Results and Discussion
To check the reduction status of GO during the solvothermal reactions, C1s XPS spectra of GO and Bi 2 Te 3 /G are analyzed as shown in Figure 1b. The spectra can be fitted by four peaks for different forms of carbons: non-oxygenated carbon (C-C, 285.6 eV and C=C, 284.8 eV), carbon in C-O bonds (286.3 eV), carbonyl carbon (C=O, 287.6 eV) and carboxylate carbon (O-C=O, 289.0 eV) [28]. After the solvothermal reactions, the peak intensity of the oxygenated carbons (C-O, C=O, O-C=O) shows an obvious decrease, indicating the reduction of GO into graphene. Figure 1c gives the O1s XPS spectra of the hybrid. The broad peak at 532.2 eV is assigned to the residual oxygen-containing groups such as -COOH and -OH [29], which agrees with the result of C1s spectra. From the XRD and XPS analyses, it can be concluded that a Bi 2 Te 3 /G hybrid has formed with the simultaneous formation Bi 2 Te 3 and reduction of GO during the one-step solvothermal process.  Figure 2a shows the SEM image of bare Bi 2 Te 3 prepared by the solvothermal route without adding GO in the precursors. Generally, the sample exhibits a hexagonal plate shape with a size of around 0.5-3 µm. However, smaller sized broken plates with an irregular shape can also be observed. The thickness of the plates is roughly estimated to be 50-100 nm from some vertically (or partially vertically) aligned plates denoted by the arrows. Figure 2b shows the SEM image of the Bi 2 Te 3 /G hybrid. It is obvious from the transparent graphene that Bi 2 Te 3 plates are located between the graphene sheets. The size of the Bi 2 Te 3 plates is below 2 µm. A three-dimensional (3D) sandwich structure is thus constructed by alternatively arranged graphene layer and Bi 2 Te 3 layer as clearly indicated by the arrows. The transparent nature of the graphene layer suggests that it is rather thin, composed probably of single or few layers of graphene sheets.
The microstructure of Bi 2 Te 3 /G was further characterized by TEM as shown in Figure 2c. Note that nearly all of the Bi 2 Te 3 plates are anchoring on graphene without forming free Bi 2 Te 3 plates even after (c) vigorous ultrasonication. The wrinkles and folded edges of graphene indicate that the graphene layer is rather thin, which agrees with the SEM observation. The enlarged view of the circled part in Figure 2c is presented in Figure 2d. The overlapping of the Bi 2 Te 3 plates can be observed from Figure 2d, from which the transparent nature of the Bi 2 Te 3 plate is evident. From a vertically aligned plate, the thickness of the plate is estimated to be 15 nm, which is much smaller compared with that of bare Bi 2 Te 3 . The thickness of the Bi 2 Te 3 plates is in the range of 10−20 nm after observing some plates at the edges or defect sites of graphene. It is worth noting that the number of the Bi 2 Te 3 plates with a regular hexagon is much smaller compared with that of bare Bi 2 Te 3 plates. It is believed that the irregular shape and small thickness of the Bi 2 Te 3 plates in Bi 2 Te 3 /G are caused by the inhibited diffusion and re-crystallization of small sized Bi 2 Te 3 particles due to the pinning effect of the defects and the oxygen-containing groups in graphene [30]. The presence of the residual oxygen-containing groups can be confirmed by XPS shown in Figure 1b  In Figure 3, CV plots of Bi 2 Te 3 /G and Bi 2 Te 3 were measured to clarify the electrochemical reaction mechanism of Bi 2 Te 3 with Li. As seen in Figure 3a, the reaction Bi 2 Te 3 with Li exhibits a multi-step mechanism, evidenced by the multi-peak feature of the CV plots. During the first cathodic scan, three reduction peaks, located at 2.0, 1.25, and 0.6 V, respectively, appear as indicated by the black line in Figure 3a. The small peak at 2.0 V is related possibly to the insertion reaction of Li ion into layered Bi 2 Te 3 without significantly destroying its crystal structure [31]. The peak at 1.2 V corresponds to the lithiation reaction of Te by forming Li 2 Te (2Li + Te → Li 2 Te) [27], and the peak at 0.6 V is associated with the reaction of Bi with Li to form Li 3 Bi (3Li + Bi → Li 3 Bi) [5,26,27]. During the second cathodic scan, the peak at 2.0 V almost disappears, the peak at 0.6 V is shifted to a higher value of 0.7 V (A'), while the peak at 1.2 V is divided into two peaks, B' and C', located at 1.35 and 1.65 V, respectively. The division of the peak means that the reaction of Te with Li occurs via a two-step mechanism by forming LiTe 3 and Li 2 Te successively, both of which are stable phases at room temperature [32]. The lower peak potential in the first cathodic scan compared with those in the subsequent scans is due to the additional energy required for the displacement reactions of Bi 2 Te 3 upon Li uptake [4]. This also means that structure of Bi 2 Te 3 cannot be recovered after the anodic scan [27].
Note that three oxidation peaks (A, B, and C), at 1.0, 1.8 and 1.9 V, respectively, appear during the anodic scans, corresponding to the successive de-lithiation reactions of Li 3 Bi (peak A) and Li 2 Te (peak B and peak C). The CV plots of bare Bi 2 Te 3 are also given for comparison as seen in Figure 3b. The plots of bare Bi 2 Te 3 display a similar feature as those of Bi 2 Te 3 /G except that the peak intensity decreases more rapidly with scans, indicative of its lower reversibility. The relatively high reversibility of Bi 2 Te 3 /G can be attributed to the incorporation of graphene that improves the electrode kinetics and keeps the electrode integrity.  Figure 4a shows the chare-discharge curves of Bi 2 Te 3 and Bi 2 Te 3 /G at 50 mA g −1 at various cycles. Bi 2 Te 3 /G hybrid gives a first discharge (Li-uptake) capacity of 731 mAh g −1 and a first charge (Li-extraction) capacity of 481 mAh g −1 . The first irreversible capacity can be attributed to the reduction decomposition of the electrolyte and the formation of the solid state interface (SEI) layer. Successive potential plateaus can be seen during the charge-discharge cycling, which correspond to the current peaks in the CV plots. The appearance of the potential plateaus is due to the two-phase coexistence of original phase (Te or Bi) and the Li-alloys (Li-Te or Li-Bi alloys). As shown in the figure, bare Bi 2 Te 3 exhibits lower charge-discharge capacities compared with Bi 2 Te 3 /G. The theoretical maximum capacity of Bi 2 Te 3 is 401 mAh g −1 , related to the formation of Li 2 Te and Li 3 Bi.
Considering the fact that graphene itself shows a relatively low capacity of around 300 mAh g −1 , as reported in our previous work [33], the theoretical capacity of Bi 2 Te 3 /G should be lower than that of bare Bi 2 Te 3 . The high yieldable of Bi 2 Te 3 /G is possibly due to the synergistic effect between Bi 2 Te 3 and graphene. On one hand, as discussed above, the presence of graphene can restrain the crystal growth of Bi 2 Te 3 , which is favorable for rapid Li ion diffusion within the active material. On the other hand, the separator effect of Bi 2 Te 3 plates for graphene sheets maximizes the exposure of the graphene surface to the electrolyte, especially for the edges and vacancies on graphene. The edges and vacancies can supply additional sites for Li storage [34]. Note that Bi 2 Te 3 /G shows a progressive polarization upon cycling, but the extent is smaller than bare Bi 2 Te 3 owing to the presence of the conductive graphene. Figure 4b compares the cycling stability between bare Bi 2 Te 3 and Bi 2 Te 3 /G charged at 50 and 200 mA g −1 and discharged at 50 mA g −1 . Note that Bi 2 Te 3 /G demonstrates an obviously slower capacity fade compared to bare Bi 2 Te 3 . After 50 cycles at 200 mA g −1 (about 0.5 C), a charge capacity of 158 mAh g −1 can be retained for Bi 2 Te 3 /G. In contrast, the charge capacity of bare Bi 2 Te 3 drops rapidly to 33 mAh g −1 . The Bi 2 Te 3 /G hybrid also shows improved cycling stability compared with Bi 2 Te 3 -graphite composite [27] and other Bi-based anodes [25,26]. The enhanced cycling stability is attributed to the buffering effect of graphene that alleviates the large volume changes and the confinement effect of graphene that restrains the aggregation of the Bi 2 Te 3 plates, in addition to offering effective conducting networks. The 3D sandwich structure also facilitates the Li-ion diffusion within the free space and across the electrode/electrolyte interface. In addition, the free space within the sandwich also affords additional room to buffer the volume changes. It should be stressed that the long-term cycling stability of the Bi 2 Te 3 /G hybrid is not satisfactory yet owing to the intrinsic large volume changes of the Bi 2 Te 3 during alloying/de-alloying processes. However, it is clear that graphene plays an important role in enhancing the electrochemical properties of Bi 2 Te 3 . It also should be noted that the large-scaled of Bi 2 Te 3 /G is not favored due to the toxicity, price, and resource scarcity of Bi and Te. Nevertheless, it shows promising application in micro-batteries due to the possibility of micro processing of graphene and Bi 2 Te 3 .

Preparation of Bi 2 Te 3 -Nanoplate/Graphene-Nanosheet (Bi 2 Te 3 /G) Nanohybrid
Graphite oxide (GO, 80 mg), prepared by a modified Hummer's method [35], was added into 40 mL of ethylene glycol (EG) and sonicated for 24 h to obtain a sufficiently exfoliated GO suspension. Then, 1 mmol of BiCl 3 , 1.5 mmol of Na 2 TeO 3 and 0.4 g of NaOH was added to the above suspension with sonication. After stirring for 12 h, the mixed solution was then transferred to a Teflon-lined stainless steel autoclave (100 mL) and heated in an electric oven at 180 °C for 24 h. The as-obtained product was collected by centrifugation, washed with deionized water and absolute ethanol for several times and dried at 40 °C under vacuum for 10 h. Bare Bi 2 Te 3 was also synthesized using a similar method without adding GO.

Materials Characterization
The X-ray diffraction (XRD) patterns were obtained on a Rigaku D/Max-2550pc powder diffractometer using Cu K α radiation (λ = 0.1541 nm). X-ray photoelectron spectra (XPS) were collected on a KRATOS AXIS ULTRA-DLD spectrometer with a monochromatic Al K α radiation (hv = 1486.6 eV). The morphologies of the products were observed by field emission scanning electron microscopy (FE-SEM) on a FEI-sirion microscope and transmission electron microscopy (TEM) on a JEM 2100F microscope. The carbon content analysis was conducted on a Flash EA 1112 tester.

Electrochemical Measurements
The electrochemical measurements of Bi 2 Te 3 /G and Bi 2 Te 3 were performed with a half-cell configuration using CR2025-type coin cells. The working electrodes were prepared by spreading a slurry composed of 75 wt% active material (Bi 2 Te 3 /G or Bi 2 Te 3 /G), 15 wt% poly(vinylidene fluoride) (PVDF) and 10 wt% acetylene black onto Ni foam followed by drying at 100 °C for 12 h under vacuum. The loading mass of the active material is around 2 mg. The cells were assembled in an Ar-filled glove box using Li foil as the counter electrode and Celgard 2300 membrane as the separator. The electrolyte used was 1 M LiPF 6 in ethylene carbonate (EC)/dimethyl carbonate (DMC) (volume ratio 1:1). The cells were galvanostatically charged-discharged at 50 mA g −1 between 0.005 and 2.5 V vs. Li/Li + on a Neware battery tester (Shenzhen, China). The specific capacity of Bi 2 Te 3 /G was calculated based on the total weight of Bi 2 Te 3 and graphene. Cyclic voltammetry (CV) tests were carried out on an Arbin BT2000 system at a scan rate of 0.1 mV s −1 between 0.005 and 2.5 V vs. Li/Li + . All of the electrochemical measurements were carried out at room temperature.

Conclusions
The Bi 2 Te 3 /G sandwich has been prepared by an in situ route with alternatively arranged Bi 2 Te 3 nanoplates and graphene nanosheets. The diffusion and re-crystallization of Bi 2 Te 3 in Bi 2 Te 3 /G are inhibited due to presence of oxygen-containing groups and defects on graphene, leading to more irregular shape and smaller thickness of Bi 2 Te 3 in Bi 2 Te 3 /G compared with those in bare Bi 2 Te 3 . Both Bi 2 Te 3 /G and Bi 2 Te 3 exhibit a multi-step lithiation mechanism. The Bi 2 Te 3 /G hybrid can yield a higher-than-the-theoretical charge capacity of mAh h −1 at 50 mA g −1 due to the synergistic effect between Bi 2 Te 3 and graphene. The Bi 2 Te 3 /G hybrid shows an obviously improved cycling stability compared with bare Bi 2 Te 3 . The improved cycling stability is attributed to the 3D sandwich structure, where graphene not only buffers the volume changes but also immobilizes Bi 2 Te 3 nanoplates, in addition to offering effective conducting networks. The unique sandwich structure also facilitates Li-ion diffusion at the electrode/electrolyte interface and within the free space of the sandwich.