Two dimensional (2D) reduced graphene oxide (RGO)/hexagonal boron nitride (h-BN) based nanocomposites as anodes for high temperature rechargeable lithium-ion batteries

With lithium-ion (li-ion) batteries as energy storage devices, operational safety from thermal runaway remains a major obstacle especially for applications in harsh environments such as in the oil industry. In this approach, a facile method via microwave irradiation technique (MWI) was followed to prepare Co3O4/reduced graphene oxide (RGO)/hexagonal boron nitride (h-BN) nanocomposites as anodes for high temperature li-ion batteries. Results showed that the addition of h-BN not only enhanced the thermal stability of Co3O4/RGO nanocomposites but also enhanced the specific surface area. Co3O4/RGO/h-BN nanocomposites displayed the highest specific surface area of 191 m2/g evidencing the synergistic effects between RGO and h-BN. Moreover, Co3O4/RGO/h-BN also displayed the highest specific capacity with stable reversibility on the high performance after 100 cycles and lower internal resistance. Interestingly, this novel nanocomposite exhibits outstanding high temperature performances with excellent cycling stability (100% capacity retention) and a decreased internal resistance at 150 °C.

www.nature.com/scientificreports www.nature.com/scientificreports/ followed by hydrothermal microwave irradiation (CEM One touch technology MARS 6) at 180 °C, 900 W and 150 psi for 15 min. The material was washed and centrifuged (HERAEUS -LABOFUGE 400 Centrifuge) and left for drying overnight at 60 °C.

Synthesis of Co 3 o 4 nanoparticles, Co 3 o 4 /RGO, and Co 3 o 4 /h-BN nanocomposites. To synthesis
Co 3 O 4 , 0.2 M of cobalt (II) acetate tetrahydrate was dissolved in distilled water (30 ml) followed by stirring for 30 min. Then, the mixture was exposed to hydrothermal treatment under microwave irradiation (CEM One touch technology MARS 6) at 180 °C, 900 W and 150 psi for 59 min. For the synthesis of Co 3 O 4 /RGO, 50 mM of cobalt (II) acetate tetrahydrate Co (C 2 H 3 O 2 ) 2 was dissolved in 30 ml of graphene oxide (2 mg/ml) and then the same procedure was followed as for the pure Co 3 O 4 . Similarly, the same procedure as the Co 3 O 4 /RGO was followed to synthesis Co 3 O 4 /h-BN but in this case h-BN (2 mg/ml).

Synthesis of Co 3 o 4 /RGO/h-BN nanocomposites.
A 2:1 ratio of Co 3 O 4 /RGO to h-BN was used to synthesize the Co 3 O 4 /RGO/h-BN nanocomposites. Co 3 O 4 /RGO (60 mg) was dissolved in 30 ml of ethanol followed by 30 min stirring. Then, 15 ml of h-BN (2 mg/ml) was added to the solution. The mixture was then exposed to hydrothermal treatment under microwave irradiation (CEM One touch technology MARS 6) at 180 °C, 900 W and 150 psi for 59 min.
Material characterization. Structural studies of the prepared nanocomposites were analyzed using X-ray diffraction (XRD) on a Rigaku MiniFlex600 X-ray diffractometer using Cu Kα radiation at a scanning rate of 2°/min. Chemical composition studies were performed on a Thermo Scientific Nicolet-iS10 Fourier transform infrared spectroscopy (FT-IR). Thermal gravimetric analysis (TGA) of the nanocomposites was recorded from ambient to 550 °C under N 2 atmosphere at a heating rate of 5 °C/min on a STA7200 thermal analysis system. Differential Scanning Calorimeter (DSC) DSC7020 was equipped to measure the DSC thermal studies and was performed under N 2 atmosphere at a heating rate of 10 °C/min from room temperature to 500 °C. Further spectral studies were performed for the obtained nanocomposites using an iRaman plus Raman spectroscopy. Morphological studies were conducted by Zeiss MERLIN Field Emission Scanning Electron Microscope (FE-SEM) field emission scanning electron microscopy and JEOL JEM-2100 F Transmission electron microscopy (TEM). Surface elemental analysis was performed using X-ray photoelectron spectroscopy (XPS) studies on ESCALABMK II X-ray photoelectron spectrometer with Mg-Ka radiation. The specific surface area was calculated using a Brunauer-Emmett-Teller (BET) method via N 2 adsorption-desorption measurement on Micromeritics ASAP 2020. The pore size distribution was obtained by the Barrett-Joyner-Halenda (BJH) method.

Electrochemical measurements.
To study the electrochemical performances, 80 wt.% of the active material (nanocomposites), 10 wt.% of conductive agent-carbon black (Sigma-Aldrich) and 10 wt.% of binding agent-polyvinylidene fluoride (Sigma-Aldrich) was dissolved in 1:1 ethanol: dimethylsulfoxide (Sigma-Aldrich) solvent. Homogenous slurry was obtained and was then coated on copper foil substrates and allowed to dry at 80 °C under vacuum. The working electrode was weighed which was then cut to form disks of 15 mm. The electrode was then assembled to CR2032 coin-type cell in an argon-filled glove box. Lithium metal was used as a counter electrode with polypropylene membrane separator (Celgard 2325) and 1 M LiPF 6 in EC + DMC + DEC (1:1:1 vol.%) was used as the electrolyte. Galvanostatic charge/discharge curves were obtained using BST8-5A-CST eight-channel battery analyzer and Gamry 3000 electrochemical working station in a potential window of 0-3 V at different current densities. Gamry 3000 electrochemical working station was also used to measure cyclic voltammetry (CV) at a scan rate of 50 mV/s in a voltage range of 0-3 V and electrochemical impedance spectroscopy (EIS) by applying a perturbation voltage of 10 mV in the frequency range of 1 Hz and 100 kHz. Furthermore, testing at high temperature was performed via a bomb calorimeter vessel in which the positive and negative terminals of the battery were connected to the two electrodes of the vessel. The coin cell was put inside the vessel at the desired temperature for several hours to ensure that the temperature is stable. Electrochemical characterizations including galvanostatic charge/discharge, CV and EIS were then performed by connecting the bomb calorimeter vessel to the Gamry potentiostat.

Results and Discussion
XRD studies were performed to investigate the phase structure of Co 3 O 4 , Co 3 O 4 /RGO, Co 3 O 4 /h-BN and Co 3 O 4 / RGO/h-BN as illustrated in Fig. 1(A). The diffraction peaks of Co 3 O 4 nanoparticles can be indexed to the pure Co 3 O 4 with normal cubic spinel structure (JCPDS card no. 76-1802). However, the increase in the peak intensity of (111) reveals that the Co 3 O 4 nanoparticles are exposed of (111) planes, where the surface is mainly composed of Co 2+ active sites. To prepare metal oxides like Co 3 O 4 , a basic medium is required, however, during our preparation of Co 3 O 4 the only source of OH − is from the hydrolysis of acetate anions. Thus, the H 2 O amount used was not sufficient for 0.2 M of cobalt acetate. This low amount of hydroxyl anions leads to the formation of Co 3 O 4 nanocubes with (001) exposed planes 23,24 . Co 3 O 4 /RGO nanocomposites also show similar peaks of Co 3 O 4 with normal cubic spinel structure. However, the XRD doesn't show an obvious peak of RGO, it could be due to Co 3 O 4 particles anchored on the surface of RGO and prevents exfoliated RGO sheets from face-to-face restacking following the reduction process. This confirms that there is an interconnection between Co 3 O 4 and graphene. Furthermore, the low amount of RGO and low diffraction intensity of graphene compared to Co 3 O 4 , can cause the RGO peak to disappear 25,26 . For Co 3 O 4 /h-BN, the very well crystalline peak at 26° is attributed to the (002) diffraction peak of h-BN 27 Fig. 1  www.nature.com/scientificreports www.nature.com/scientificreports/ curves displayed in Fig. S3. The TGA curve can be divided into three regions based on the weight losses, it was found that in the first region that is below 100 °C, only 1% decomposition is observed for the Co 3 O 4 /RGO nanocomposite that is due to the release of absorbed water. In the second region that is between 100 and 300 °C, Co 3 O 4 /RGO showed a 4% loss around 250 °C while Co 3 O 4 /h-BN showed a weight loss of only 2% around 225 °C. These losses are due to the decomposition of epoxy and hydroxyl groups found in RGO and hydroxyl groups in h-BN. This shows that the GO oxygen containing group were almost all reduced to RGO during the reduction process. After 250 °C, both Co 3 O 4 /RGO and Co 3 O 4 /h-BN display a gradual mass loss, which is due to the further removal of functional groups. At 550 °C, the TGA curves showed an overall weight loss of 10% for Co 3 O 4 /RGO and a weight loss of 3-6% for the h-BN based Co 3 O 4 nanocomposites. On the other hand, Co 3 O 4 showed only ~3% loss observed around 300 °C and remained stable up to 550 °C. For Co 3 O 4 /RGO/h-BN nanocomposites, a weight loss of only 0.5% was observed around 100 °C which is due to moisture. In the region between 100 °C and 300 °C, a weight loss of 1.5% was observed due to the reduction of oxygen containing functional groups from both RGO and h-BN. After 300 °C, a gradual decrease in mass loss was observed with a total weight loss of 6.5% at 550 °C due to the further removal of functional groups. It is noteworthy that the weight percentage of RGO in Co 3 O 4 /RGO nanocomposites was higher than in Co 3 Fig. 1(D). The DSC curve of Co 3 O 4 nanoparticles has three major endothermic peaks. The first endothermic peak at around 105 °C is due to the evaporation of adsorbed water. The endothermic peak at 238.30 °C can be associated with the decomposition of the cobalt precursor. The third endothermic peak observed at 357.42 °C is attributed to the further decomposition of intercalated CH 3 COO-groups. The addition of RGO to Co 3 O 4 , converted the second endothermic peak to an exothermic peak at 217. 12 Fig. 2(a). The Raman spectrum of Co 3 O 4 nanoparticles showed five Raman bands at around, 182 cm −1 (F 2g ), 478 cm −1 (E g ), 521 cm −1 (F 2g ), 611 cm −1 (F 2g ) and 689 cm −1 (A 1g ) which confirms the spinel structure of Co 3 O 4 . The Raman mode at 689 cm −1 (A 1g ) corresponds to characteristics of the octahedral sites whereas 182 cm −1 (F 2g ), 478 cm −1 (E g ), 521 cm −1 (F 2g ) and 611 cm −1 (F 2g ) are attributed to the combined vibrations of the tetrahedral site and octahedral oxygen motions 30 . For Co 3 O 4 /RGO nanocomposites, in addition to the five Co 3 O 4 Raman bands, extra two clear peaks at 1327 cm −1 and 1584 cm −1 are detected, which corresponds to the D and G peaks of graphene 31,32 , respectively, as shown in Fig. 2(b). The D band is attributed to the breathing mode of k-point phonons of A 1g symmetry and the G bands to the E 2g phonon of sp 2 C atoms. In addition, the D band results from the sp 3 edge defects, disorder and loss of graphitic structure. These results are strong evidence that supports the existence of both RGO and Co 3 O 4 in www.nature.com/scientificreports www.nature.com/scientificreports/ the prepared composites 33 . For Co 3 O 4 /h-BN, in addition to the five Co 3 O 4 Raman active modes, a single peak located at 1365 cm −1 corresponds to an in-plane vibration (E 2g ) and is analogous to the well-known G peak in graphene which confirms the existence of hexagonal boron nitride as illustrated in Fig. 2(c). In the case of Co 3 O 4 / RGO/h-BN (Fig. 2(d)), an intense and sharp peak at 1365 cm −1 is observed in the spectrum which is close to the characteristic Raman peak of bulk h-BN materials. The spectra of Co 3 O 4 /RGO/h-BN exhibits two peaks at 1344 and 1589 cm −1 , which indicates the presence of RGO sheets in the nanocomposites. However, D band of graphene is almost negligible indicating the lower number of defects in the graphene structure 34 Fig. 3(a). FE-SEM images for Co 3 O 4 /RGO nanocomposites show Co 3 O 4 nanoparticles residing on the reduced graphene oxide sheets with homogenous dispersion as shown in Fig. 3(b). Similarly, for the Co 3 O 4 /h-BN nanocomposites, cube like Co 3 O 4 nanoparticles residing on the surface of h-BN sheets was observed as displayed in Fig. 3(c). On the other hand, cubic nanoparticles incorporated in transparent h-BN sheets and RGO sheets were observed for the Co 3 O 4 /RGO/h-BN nanocomposites as shown in Fig. 3(d-f) More detailed structures of the prepared nanocomposites were observed using high resolution TEM (HR-TEM). The presence of Co 3 O 4 incorporated onto the graphene sheets and h-BN sheets is seen. In addition, to morphology, the d-spacing of the nanostructures can also be obtained using HR-TEM. The higher resolution TEM of pure Co 3 O 4 nanoparticles revealed that it is cubic as shown in Fig. 4(a) which is consistent with the FE-SEM result and further supports the cubic morphology. Furthermore, HR-TEM images were used to indicate that the RGO and BNNS consisted of a few atomic layers as shown in Fig. 4(b-d). Structural studies using XRD and morphological studies using FE-SEM and HR-TEM support each other and confirm the successful www.nature.com/scientificreports www.nature.com/scientificreports/ preparation of Co 3 O 4 /RGO/h-BN nanocomposites that consists of highly ordered and crystalline cubic structure incorporated into the conductive RGO matrix with a thin layer of thermally stable h-BN sheets. This unique structural features could help promote smooth channels for ion diffusion and alleviate the volume expansion that could occur during the charge and discharge process and inhibit the deterioration of the stable crystal structure 36 .
XPS was also performed to study the surface elemental composition of the prepared nanocomposites. The XPS spectra of Co 3 Fig. 5. Two major peaks corresponding to Co 2p 3/2 (785.18 eV) and Co 2p 1/2 (760.49 eV) are observed which corresponds to a typical Co 3 O 4 phase with both Co 2+ and Co 3+ cations 37 Fig. S3(b) can be deconvoluted to two peaks at 530.4 eV and 533.15 eV which is attributed to oxygen lattice atoms and adsorbed water, respectively 41 .

O 4 , Co 3 O 4 , Co 3 O 4 /h-BN, and Co 3 O 4 /RGO/h-BN are shown in
The specific surface areas were measured using the BET method as shown in Fig. S2 and summarized in Table S1. The specific surface area of pure Co 3 O 4 is 41 m 2 /g. Co 3 O 4 /RGO nanocomposites resulted in a higher BET surface area of 54 m 2 /g which could be due to the incorporation of graphene sheets into Co 3  To test the electrochemical performance, all the four Co 3 O 4 based composites were galvanostatically charged and discharged for 100 cycles. Figure 6(a)   www.nature.com/scientificreports www.nature.com/scientificreports/ also displayed high specific capacity initially of 317 mAh/g which decreased dramatically to 14.5 mAh/g after the 100 th cycle which is slightly higher than Co 3 O 4 which shows the potential of h-BN in increasing the performance of Co 3 O 4 in terms of specific capacity although Co 3 O 4 /h-BN displayed the lowest specific surface area. Certain modifications on the preparation of Co 3 O 4 /h-BN nanocomposites and understanding the chemistry behind the two in terms of interaction and so forth can make them potential candidates in li-ion battery energy storage devices. CV of the pure Co 3 O 4 , graphene based Co 3 O 4 nanocomposites, h-BN based Co 3 O 4 nanocomposites, and Co 3 O 4 /RGO/h-BN were performed in a potential voltage window between 0 and 3.0 V at a scan rate of 10 mV/s to further evaluate the electrochemical properties as illustrated in Fig. 6(b). A cathodic peak appeared at about 0.52 V for all the nanocomposites which is a result of the reduction of Co 3 O 4 to Co metal, formation of clusters between Co and Li 2 O and formation of the solid electrolyte interphase (SEI) layer on the active materials 43 . As for the oxidation or the anodic scan, two peaks one at 2.25 V after the first cycle was observed which resulted from the reversible oxidation of metal Co to Co 3 O 4 44,45 . The oxidation and reduction peaks are also attributed to the insertion/extraction of lithium into/from graphene. The overlap of the second and the third cycle is an indication of an enhanced cycling performance 46 . Interestingly, CV results are in line with the galvanostatic charge/ discharge studies as Co 3 O 4 /RGO/h-BN displayed the highest current. EIS was also plotted for all the Co 3 O 4 based nanocomposites performed after 100 charge-discharge cycles as shown in Fig. 6(c,d). From the Nyquist plots, it was observed that there is a depressed semicircle in the high-frequency region because of the resistance caused by the SEI layer (R SEI ). In the medium frequency region, a broad arc is observed which is attributed to the Li + charge-transfer resistance (Rct) on the electrode/electrolyte interface and an inclined line in the low frequency because of Warburg resistance (W). A lower resistance curve was observed for the Co 3 O 4 /RGO/h-BN nanocomposite when compared to the other nanocomposites. The smaller semicircle in the high-medium frequency region of Co 3 O 4 /RGO/h-BN nanocomposites is an indication of a faster charge-transfer and smaller internal electrochemical resistance while at the low frequency region, the steeper tail of Co 3 O 4 /RGO/h-BN nanocomposites means lower ion diffusion resistance and enhanced mass transport 47,48 thus an enhanced electrochemical performance which is in agreement with the previous electrochemical studies.
Thus, to further evaluate their electrochemical studies of the Co 3 O 4 /RGO/h-BN nanocomposites, cyclic performance was conducted for another 100 charge/discharge cycles as shown Fig. 7(a) and galvanostatic charge/ discharge curves for the first 5 cycles in Fig. 7(b). Cycling performance showed that the coulombic efficiency for the discharge/charge curve of Co 3 O 4 /RGO/h-BN nanocomposites was 100% for another 100 cycles with 186.58/187.40 mAh/g charge/discharge capacities after a total of 200 cycles. These results suggest that the as prepared Co 3 O 4 /RGO/h-BN can retain its initial morphology with stable reversibility on the high performance after 200 cycles. The galvanostatic charge/discharge of Co 3 O 4 /RGO/h-BN nanocomposites for the 1 st 5 cycles displayed in Fig. 7(b) shows the initial charge/discharge capacities of 284.80/275.90 mAh/g and minor irreversible capacities losses from the 2 nd to the 5 th cycle which is estimated to be only 8%.
The high electrochemical performance of Co 3 O 4 /RGO/h-BN nanocomposites is due to the synergistic effect of Co 3 O 4 , RGO, and h-BN. The abundant electrochemical active sites of cubic Co 3 O 4 can help adsorb more li-ions and shorten the electron/ions transfer path. RGO could act as a conductive carbon matrix and thus facilitate electron transfer. Whereas, the introduction of the thermally stable 2D h-BN could help maintain the structure of RGO by preventing it's restacking. Moreover, it is reported that the addition of heteroatom (e.g., S, N, and P) to a carbon matrix could help for rapid ion insertion/de-insertion and fast electron transportation 49,50 .
The electrochemical performance of Co 3 O 4 /RGO/h-BN was then studied at a higher operating temperature of 150 °C. If we have a closer look at the thermal studies conducted mainly the TGA and dW/dT curves shown in Fig. 1(C) and Fig. S1 it is apparent that Co 3 O 4 /RGO/h-BN nanocomposite was the most thermal stable. In addition, thermal studies for the nanocomposites were also performed when they were formed into a composite with activated carbon and PVDF as shown in Fig. S4. From the curves, no degradation was observed until 150 °C while www.nature.com/scientificreports www.nature.com/scientificreports/ degradation peaks appeared at around 160 °C which is due to PVDF thus further confirming the thermal stability of the prepared nanocomposites and have the tendency to operate well at 150 °C.  Fig. 8(b,c). It's noteworthy that a higher current density is required at high temperatures. An increase in the specific capacity was observed with cycle when the operating temperature was increased to 150 °C. From Fig. 8(c) it can be observed that an initial charge/discharge capacity of 280.21/286.04 mAh/g was obtained followed by a steady and gradual increase in capacity in the subsequent cycles. This increase upon cycling has been previously reported for carbon based metal oxides and could be due to several reasons such as electrochemical reversibility of the lithium oxide formation/decomposition and electrolyte reactions. Thus, a coulombic efficiency of higher than 100% was observed due to the reversible insertion of Li 2 O into the metal particles. This shows that high temperatures effectively promote decomposition of lithium oxide, electrolyte reactions, and the electrode material can be constantly activated which improves reaction kinetics with increased interfacial compatibility between the electrolyte and the electrode material [51][52][53][54] . Another contributing factor to the increase of capacity upon cycling could be due to the change in morphology of the composite with the increase of cycling number. The morphology of the composite would change to result in a structure that has a higher volume and surface area. This structural change increases the ability of the electrode material to adsorb and store a larger amount of Li + ions 55 . Recently, this phenomenon (the increase in capacity upon cycling) was also observed in sodium-ion batteries and was due to co-intercalation reaction where Na atoms that are aggregated and stored at a specific location inside a carbon material during discharge are released, creating a space for more Na atoms in the next charging process, consequently increasing the capacity with cycling and resulting in a coulombic efficiency greater than 100% 56 . Moreover, the superior li-ion performance in terms of specific capacity observed at higher temperatures when compared to room temperature could be also explained through Arrhenius equation, that is, since li-ion batteries undergo a chemical reaction, temperature increases the rate of the reaction which increases the chemical reaction leading to an increase in the specific capacity. Kilibarda, G. et al. described other two factors that could also influence the increase in specific capacity as the temperature increases which includes the electrical/ionic conductivity of the electrolyte, increases with increasing temperature due to the low viscosity at high temperatures which increases the specific capacity. The other factor could be due to the drop in the ohmic potential or the decrease in the internal resistance of the cell which is also confirmed by the EIS plot performed at 150 °C as shown in Fig. 6(d). The increase of temperature resulted in a decrease in the potential drop, which causes a potential increase of the cell and thus increasing the specific capacity 57 . Furthermore, in contrast to the EIS conducted at room temperature, at high temperatures an additional resistance around 8 ohms is experienced which rises from the electrolyte.

Conclusions
We have developed a novel Co 3 O 4 /RGO/h-BN nanocomposite using a facile MWI route as anodes for li-ion batteries. The prepared nanocomposites displayed high thermal stability making them good candidates in sustaining li-ion batteries thermal runaway. BET surface area showed that Co 3 O 4 /RGO/h-BN nanocomposites showed the highest specific surface area (191 m 2 /g) which is due to the high surface area of the well exfoliated graphene and h-BN sheets in addition to the synergistic effect between the two. Based on the advantages of their highest thermal stability and specific surface area, the Co 3 O 4 /RGO/h-BN anodes had the highest electrochemical response in terms of charge/discharge capacity, cyclic performance (100% capacity retention) and 100% coulombic efficiency even at high temperatures of 150 °C for 100 cycles. No short circuits, thermal runaway events or capacity decay was observed which makes Co 3 O 4 /RGO/h-BN nanocomposites potential candidates to be utilized as safe energy storage devices and especially in applications that operate at high temperatures. Specifically, can be utilized in industry to measure the 'downhole' pressure in oil wells and other oil applications where high temperatures are encountered ranging from 80 to 200 °C. In such high temperature environments, where batteries need to have an overall good electrochemical performance and should have the ability to operate safely over the wide range of high temperatures.

Data availability
All data generated or analysed during this study are included in this published article (and its Supplementary Information files).