Nitrogen and boron doped carbon layer coated multiwall carbon nanotubes as high performance anode materials for lithium ion batteries

Figure 1a shows the synthetic route for the fabrication of the final products, which is also described in detail in the Experimental section. Briefly, MWCNTs were first oxidized and subsequently functionalized by a thin layer of metal–organic framework complex (4,4′-bipyridine-Cu). After pyrolysis and leaching process, the resultant MWCNTs featured with rich defects and heteroatoms (N, B, O) were then obtained. The surface morphology and microstructure of the fabricated NBC@MWCNTs were examined by scanning electron microscopy (SEM) with different magnifications, as displayed in Fig. 1b,c. Accordingly to the top morphology, it can be clearly seen that the tubular structure is still retained after coating thin carbon layer. All carbon nanotubes are highly aggregated and overlapped with each other with very thin wall, which can provide plenty of room for the improvement of lithium ion transport properties. As presented in Fig. 1d, the energy-dispersive X-ray spectroscopy (EDX) elemental mappings results further confirm the presence of N, B, C and O elements in the materials. It is observed that both N and B are uniformly distributed nanotubes, revealing the successful doping of N and B.

Figure 1

Structure and composition of NBC@MWCNTs. (a) The scheme of preparation procedure of NBC@MWCNTs; (b,c) SEM morphology of NBC@MWCNTs; (d) Elemental distribution mapping images of elements N, B, O, and C in NBC@MWCNTs products.

The nature of different elements shown in the materials was further determined by using X-ray photoelectron spectroscopy (XPS). As illustrated in Fig. 2a, no other peaks besides those for N, B, C and O can be seen in the wide survey spectrum, providing further evidence of the purity of this material. In Fig. 2b, the high-resolution of C 1 s, it can be observed that an intensive peak is exhibited at around 284.6 eV, suggesting high quantities carbon in the sample37. To compare with the main carbon peak, the high-resolution peaks located at 532.1 eV (Fig. 2c), 191.6 eV (Fig. 2d), 398.6 eV and 400.4 eV (Fig. 2e) are ascribed to O, B, and N at different microenvironments41,42,43. The quantitative estimation of different elemental species from the XPS peak intensity suggests the atomic percentage ratios for the NBC@MWCNTs are 73.73% for C 1 s, 9.79% for N 1 s, 8.60% for O 1 s and 7.88% for B 1 s, respectively. The high contents of N, B, and O containing species might serve as electrochemically active sites for making big contribution to the enhancement of lithium storage properties44,45,46,47.

Figure 2

XPS and Raman spectra of NBC@MWCNTs. (a) survey spectrum; (b) C 1 s spectrum; (c) O 1 s spectrum; (d) B 1 s spectrum; (e) N1s spectrum, and (f) Raman spectrum of NBC@MWCNTs.

Figure 2f displays the room temperature Raman spectrum of the sample for further investigation of the NBC@MWCNTs. Two characteristic carbon peaks are found at around 1351 and 1584 cm−1, which demonstrates the presence of carbon in the form of disordered structure. It is well known that the band around 1351 cm−1 (D band) with a lower intensity is related to the disordered carbon, and at 1584 cm−1 (G band) with a high intensity is assigned to graphitic carbon35. The latter peak corresponds to an E2g mode of graphite, which is due to the sp2-bonded carbon atoms in a two-dimensional hexagonal graphitic layer37. In addition, it is interesting to note that the peak intensity ratio between the D and G bands (ID/IG) is ~ 0.9, demonstrating the partial crystallization of carbon species (CNT/carbon layer) during the coating process and high temperature thermal treatment.

Furthermore, transmission electron microscope (TEM) and high resolution TEM were utilized to carefully characterize the morphology and nanostructure of the obtained NBC@MWCNTs, and the results are displayed in Fig. 3. It can be clearly seen that the NBC@MWCNTs consist of homogeneously distributed CNTs wrapped by amorphous thin carbon layer marked with arrows in Fig. 3a. Besides, the corresponding selected area electron diffraction (SAED) pattern (Fig. 3b) shows a typical diffraction ring pattern that could be exactly identical to the coiled MWCNTs. The designed unique structure can be clearly revealed by a high resolution TEM image at the edge of the material (Fig. 3c). A MWCNT is coherently coated by a thin coating layer with some nanopores with a size of about 1 nm. The unique embedded structure is expected to be more beneficial for enhancing the lithium storage properties, and ensure the fast transportation of electrons inside the CNTs, when used as anodic materials in LIBs.

Figure 3

(a,b) low magnification TEM image and the corresponding SAED pattern of NBC@MWCNTs; (c) High resolution TEM image of a single CNT wrapped by nanoporous materials in NBC@MWCNTs; (d) High resolution TEM image and FFT pattern of the typical structure in NBC@MWCNTs; (e) HAADF image and corresponding elemental distribution mappings of N, B, O, and C in the NBC@MWCNTs.

The typical high resolution TEM images in Fig. 3d (lower left and right) show common interplanar spacing of 0.34 nm for MWCNTs. Besides, the corresponding fast Fourier transform (FFT) pattern illustrated in the upper right also demonstrates the diffraction features of coiled MWCNTs (Fig. 3c). In order to further understand the nature of the product, high-angle annular dark-field scanning TEM (HAADF-STEM) was carried out. A typical HAADF-STEM image taken from the MWCNTs is displayed in Fig. 3e, where a few MWCNTs can be detected at the edge of the materials. The corresponding EDX elemental mappings performed on the same region indicated by the square (Fig. 3e) reveal that the designed materials are rich with carbon, nitrogen, boron, and oxygen. Taken together, it can be concluded that all four elements are coexisted and distributed homogenously in the matrix material, and nitrogen and boron are well doped onto the surface of MWCNTs.

In order to systematically evaluate the lithium storage properties of the MWCNTs, Swagelok-type cells with two-electrode configuration were assembled in an argon-filled glovebox with lithium foil serves as counter/reference electrode. Figure 4a shows the initial three consecutive cyclic voltammetry (CV) cycles at a scan rate of 0.2 mV s−1 in the voltage range of 0.01–3.0 V versus Li/Li+. Compared with the subsequent two cycles, there exist two additional reduction peaks at approximately ~ 1.2 and ~ 0.7 V in the first cycle. These two peaks could be assigned to preliminary decomposition of the used electrolyte35. Furthermore, the reduction current in the first cycle below 1.3 V is significantly higher than that of the following two cycles. This extra current could be ascribed to the electrolyte being irreversibly reduced at low voltage to form a protective solid electrolyte interphase (SEI) layer48. In the subsequent cycles, these two peaks become weaker and slight shift to higher potential at around ~ 1.1 and ~ 1.7 V. There are two broad oxidation peaks in the following anodic processes at ~ 1.3 and ~ 2.3 V. The former peak is related to the partial decomposition of SEI components such as LiF, Li2CO3, and RCO2Li, and the latter peak is attributed to the reversible oxidation of N by other functional groups37,49,50,51. Accordingly, it can be seen that the following CV curves are quite similar after the first cycle, suggesting that both the formed SEI layer and the electrode material are highly stable.

Figure 4

CV profiles of NBC@MWCNTs electrode in a potential range of 0.01 and 3 V: (a) at a scan rate of 0.2 mV s−1 of the initial three cycles; (b) at different scan rates from 1 to 1000 mV s−1; (c) Galvanostatic discharge/charge voltage profiles of different cycles at a current density of 0.1 A g−1 between 0.01 and 3 V for the NBC@MWCNTs electrode; (d) EIS of NBC@MWCNTs anode in a test window range from 100 kHz to 0.01 Hz before and after 10 electrochemical cycles for the cycling performance test at 0.1 A g−1. The inset is the equivalent circuit model of the studied system.

Next, to study the electrochemical stability of the NBC@MWCNTs electrode, the same battery was further investigated with various scan rates increasing from 1 to 1000 mV s−1. As plotted in Fig. 4b, it can be observed that the LIB based on the NBC@MWCNTs anode has a very large discharge/charge current range. With the scan rate being increased step by step, the data still present good electrochemical stability. These positive results would indicate again that the fabricated NBC@MWCNTs electrode could supple a high-power output and fast charge, which is an ideal high performance anode for next-generation electric vehicles.

Figure 4c presents galvanostatic discharge/charge voltage curves of the as-obtained NBC@MWCNTs electrode for the first ten cycles at a current density of 0.1 A g−1 between 0.01 and 3 V. Surprisingly, at this current density, in the first cycle, an extreme high discharge capacity as high as 2701 mAh g−1 can be observed, which is calculated based on the total weight of the NBC@MWCNTs. In the following charge process, it still can maintain a high reversible capacity of about 1227 mAh g−1, which is more than 3.3 times higher than that of the theoretical maximum capacity for graphitic anode (LiC6, 372 mAh g−1). Accordingly, the irreversible capacity leads to an initial Coulombic efficiency of 45.4% (CE: defined by the charge capacity divided by the discharge capacity). In fact, the large irreversible capacity loss is an expected phenomenon for carbonaceous and other metal-oxides anodes in LIBs37,52. It is generally believed that the low efficiency value is probably because of the interfacial reactions, and the irreversible lithium ions insertion into some uncertain positions in the NBC@MWCNTs during the first discharge process53. However, this irreversible capacity loss could be much improved by prelithiation and/or some additional modifications in future experiments. In the following cycles, the capacity values of each discharge or charge are similar, indicating a stable lithiation and delithiation processes. Even after the tenth cycle, the charge capacity is still up to 1008 mAh g−1, corresponding to a high Coulombic efficiency of 95.8%. These similar features are consistent with the results that observed in CV curves.

To evaluate the effects of the NBC@MWCNTs electrochemical properties, subsequent electrochemical impedance spectroscopy (EIS) measurement were conducted on the NBC@MWCNTs electrode before cycling, and after tenth cycle, as plotted in Fig. 4d. The measured impedance data were analyzed by fitting to an equivalent circuit model (inset of Fig. 4d). The symbols denote the experimental data, while the continuous lines represent the fitted spectra. It is obvious to see that the charge transfer resistance (R2) changes from ~ 400 Ω of the fresh battery to ~ 120 Ω of the cycled battery based on the fitting results. This is commonly occurring in activity process, which is related to the electrochemical reactions of NBC@MWCNTs electrode with lithium ions, and other additional irreversible reactions37. This behaviour should be favourable for the transportation of lithium ions and electrons during repeated lithiation/delithiation processes, which agrees well with the results obtained from galvanostatic discharge/charge tests (Fig. 4c).

The rate capability is an important parameter of LIBs for many applications that require fast discharge/charge rates. Herein, in order to confirm the high rate performance of the NBC@MWCNTs, the NBC@MWCNTs LIB was tested under different discharge/charge current densities from 0.3 to 67.5 A g−1 over 120 cycles. As shown in Fig. 5a, the battery was first tested from a small current density of 0.3 A g−1 to an extreme high current density of 30 A g−1 for each of 10 cycles. As a result, the battery could exhibit high capacity retention from ~ 870 to ~ 90 mAh g−1 and present a constant charge capacity in 1, 3, 10, and 30 A g−1 in the first test series. Moreover, the capacity could recover to a high capacity of 870 mAh g−1 when the current density is set back to 0.3 A g−1 after 60 cycles, showing highly flexible current test ability. This remarkable rate capability could be repeated again in the following high-rate experiment. It can be seen from the second test series, the battery still possess high lithium storage performance. It is noteworthy that, even after testing at an ultra-high current density of 67.5 A g−1, the electrode can still return to the original value when the current density suddenly decreases to 0.3 A g−1 after 120 cycles. This result demonstrates that the NBC@MWCNTs electrode has excellent high current or power output ability, which means it is a potential candidate to reduce charging time of rechargeable batteries to make it possible to promote the popularity of electric vehicles.

Figure 5

(a) Rate capability at various current densities from 0.3 to 67.5 A g−1, and (b) Cycling stability and Coulombic efficiency versus cycle number over a long life of 5000 cycles at a high current density of 5 A g−1.

Furthermore, the ultra long term cycling stability of the NBC@MWCNTs electrode was also evaluated and the discharge/charge capacities, as well as Coulombic efficiency versus cycle number over 5000 cycles in the voltage range of 0.01–3 V at a high current density of 5 A g−1 are displayed in Fig. 5b. It should be mentioned that the measurement protocol was started with a small current density of 0.1 A g−1 for ten cycles to activate the anodic electrode thoroughly. As anticipated, the NBC@MWCNTs electrode also preforms unprecedented cycling stability, retaining high specific capacity of ~ 400 mAh g−1 after 5000 cycles with a superior average Coulombic efficiency of more than 99.2% at a high current density of 5 A g−1.

To gain more insight into post-morphology of NBC@MWCNTs electrode, the cycled cell was manually disassembled and obtained cycled electrode was rinsed with propylene carbonate for further SEM observations. As evidenced in Fig. 6, it is interesting that the surface morphology of the NBC@MWCNTs electrode involves negligible changes even after 5000 discharge/charge cycles. More specifically, the as-prepared NBC@MWCNTs electrode still remains integrity and contains no large cracks, and contacts with the current collector tightly. These findings would account for the enhanced reaction kinetics and improved performance. In other words, the excellent electrochemical properties are therefore related to the good conductive paths for electrons, more lithium ions accessible sites introduced by N, B, and O doping, and good structural stability of the as-prepared electrodes.

Figure 6

(a,b) SEM images at different magnifications of NBC@MWCNTs electrode at fully charged state after 5000 repeated electrochemical cycles under a constant current density of 5 A g−1.

In conclusion, highly porous N, B-doped carbon@MWCNTs were successfully prepared by pyrolysis of a thin layer of metal–organic framework complex coated on MWCNTs. When used as anode materials for LIBs, the fabricated NBC@MWCNTs electrode exhibits high stable capacity over an ultra-long cycling lifespan (5000 cycles) and superior rate capability even at very high current density (67.5 A g−1). Therefore, it is obvious that the NBC@MWCNTs could hold better electrochemical performance than previous reported carbon-based materials. Such impressive lithium storage properties could be ascribed to the synergistic coupling effect of the distinctive structural features, the reduced diffusion length of lithium ions, more active sites generated by doped atoms for lithium storage, as well as the enhancement of the electrode structural integrity. Our ongoing efforts are aligned to ascertain and mitigate the irreversible capacity losses in the first few cycles. Taken together, these results indicate that the N, B-doped carbon@MWCNTs materials may have great potential in future-generation LIBs for energy storage applications.