High capacity graphite-like calcium boridecarbides as a novel anode active material for lithium-ion batteries

Abstract

Graphite-like Ca_0.6B_1.2C_4.8 is reported as a novel anode active material for lithium-ion batteries. Its original specific capacity was 552 mA h g⁻¹, and it was increased to 648 mA h g⁻¹, which is 1.7 times higher than that of graphite, by preliminary extraction of a portion of its calcium atoms. The average lithium deintercalation potential was 1.16 V vs. Li/Li⁺, which is one of the lowest-ever potential values among high capacity carbon-based materials studied so far. We also reveal its unique Li intercalation/deintercalation behavior coincident with a 7% change in the boron-doped graphene interlayer distance while keeping its ordered layer structure. This is the first development of a high capacity and low potential carbon based material which is capable of being prepared by one-pot heat treatment from graphite showing graphite-like reversible structural changes upon Li intercalation/deintercalation.

---

Graphite is the most widely used anode active material for lithium-ion batteries (LIBs) because of its excellent Li intercalation/deintercalation cycle performance and low potential of lithium intercalation.¹ Also the low cost and abundance of graphite make it advantageous for industrial use. Thus, developing a new graphite-based material with a higher capacity while keeping the advantages of graphite may critically accelerate the development of next-generation high capacity LIBs. The specific capacity of graphite is 372 mA h g⁻¹, which corresponds to an intercalation of up to 1 lithium atom per 6 carbon atoms to form Li–graphite intercalation compounds (Li–GIC), LiC₆. From density functional theory (DFT) calculations, the reaction potential of forming LiC₆ from lithium metal and graphite is 0.1 V vs. Li/Li⁺.^2,3 The positive value of reaction potential means that forming LiC₆ is thermodynamically more favorable than lithium metal deposition. On the other hand, intercalation of more than 1 lithium atom per 6 carbon atoms results in a negative reaction potential, which limits the number of lithium atoms capable of being intercalated into graphite.³ As a possible way to overcome the limit, boron doped graphite has been theoretically predicted to be capable of intercalating/deintercalating more than 1 lithium atom per 6 atoms of boron or carbon.^3,4 As graphite becomes electron-deficient upon boron doping, the lithium intercalated state of boron-doped graphite becomes thermodynamically more stable, and more than 1 lithium atom can be intercalated into boron-doped graphite (i.e. 3Li + BC₅ → Li₃BC₅).³ Yamada et al. prepared graphite-like layered B/C and B/C/N materials by a CVD method and showed that they can realize a higher capacity than carbon as the anode material for sodium ion batteries.⁵ However, this hypothesis has not been sufficiently verified experimentally for lithium ion batteries and boron-doped graphite prepared by methods other than CVD because it is difficult to prepare such high-density boron-doped graphite. Usually, unfavorable products such as boron carbide (B₄C) or boron oxides are formed instead of a boron-doped graphene layer during the graphitization of amorphous carbon with boron sources.^6–9

In this study, we report a novel high capacity metal-boridecarbide as an anode of lithium-ion batteries. It is Ca-GIC, Ca_0.6B_1.2C_4.8, with an ordered graphite-like layer structure. This is a newly designed and synthesized material with the lowest boron and calcium densities in composition among previously reported calcium boridecarbides, such as Ca_1.5B₃C₃,¹⁰ CaB₂C₄,^11,12 and Ca_0.75B_1.5C_4.5.¹² By a simple one pot heat treatment of graphite with boron and calcium sources, carbon atoms of graphene-planar six-ring nets of graphite are displaced by boron atoms coincident with the intercalation of calcium atoms into the interlayer of graphite. It occurs without forming unfavorable products such as boron carbide (B₄C) or boron oxides. The specific capacity of Ca_0.6B_1.2C_4.8 is 552 mA h g⁻¹, which is 1.5 times higher than that of graphite, corresponding to an intercalation of 1.9 lithium atoms per 6 atoms of boron or carbon. The average lithium deintercalation potential was 1.16 V vs. Li/Li⁺, which is one of the lowest-ever potential values among high capacity carbon-based materials studied so far. Moreover, by electrochemically removing a portion of the calcium atoms from Ca_0.6B_1.2C_4.8 as a pretreatment, its specific capacity increased to 648 mA h g⁻¹. We also reveal its unique graphite-like Li intercalation/deintercalation behavior coincident with a 7% change in the boron-doped graphene interlayer distance, which corresponds to 0.3 Å, while still keeping its layer structure. This is characterized by the shift of the (001) peak of Ca_0.6B_1.2C_4.8 in XRD patterns upon lithium intercalation/deintercalation. Only a few high capacity carbon-based materials show such clear reversible structural changes in layer distances while keeping the layer structure other than graphite. As such a material, Matsuo et al. reported graphene-like graphite with a high capacity, low potential and reversible structural changes during Li intercalation/deintercalation cycles.^13,14 Ca_0.6B_1.2C_4.8 is another example of such a carbon-based material and the first as a material capable of being prepared by one-pot heat treatment from graphite without any chemical pre-treatment applied to graphite for oxidation or exfoliation to the best of our knowledge. We note that in the most recent report, Li et al. introduced Li₃B₃C₃ as a graphite-like high capacity anode material for Li-ion batteries,¹⁵ for which a reversible specific capacity of 450 mA h g⁻¹ was observed. However, from XRD patterns in the report, the interlayer distance of the Li intercalated state, Li₃B₃C₃, was almost unchanged or slightly decreased (about 0.04 Å) by oxidation current with its quantity of electricity corresponding to an extraction of 1.5 lithium atoms per B₃C₃. It is a very different behavior in terms of the change of interlayer distance because of lithium deficiency compared with previous reports. Fogg et al.¹⁶ reported that the interlayer distance of Li₃B₃C₃, 3.524 Å, increased with the loss of lithium atoms. For example, it increased to 3.806 Å for Li_1.23BC and 3.721 Å for Li_1.68B₃C₃, where these compositions were directly synthesized by a solid phase process and lattice parameters were determined by refinement of the room temperature neutron diffraction data. The increase of the interlayer distance of Li₃B₃C₃ with the loss of lithium atoms is also reproduced by DFT calculations.^4,16 We concluded that the major part of the observed specific capacity in the report by Li et al.¹⁵ did not originate from lithium intercalation/deintercalation, which was the reason why the interlayer distance did not show a plausible change upon lithium intercalation/deintercalation. Thus, it is not appropriate to conclude Li₃C₃B₃ as a graphite-like high capacity anode material for Li-ion batteries according to reports so far.

Fig. 1a shows the unit cell and schematic of Ca_0.6B_1.2C_4.8. This was visualized using the VESTA program.¹⁷ Up to now, three types of graphite-like calcium boridecarbides have been structurally characterized: Ca_1.5B₃C₃,¹⁰ CaB₂C₄,^11,12 and Ca_0.75B_1.5C_4.5.¹² We also investigated the capacity of Ca_0.75B_1.5C_4.5 in the potential range from 0 to 3 V vs. Li/Li⁺, although it was lower than that of graphite as we show below. We assumed that the low capacity of Ca_0.75B_1.5C_4.5 was due to its high content of calcium atoms. As we discuss below, calcium atoms stay in boron-doped graphene interlayers without being extracted during the Li intercalation/deintercalation process from 0 to 3 V vs. Li/Li⁺. This means that they share the interlayer sites with intercalated lithium atoms. Because of the Coulomb repulsion among intercalated cations or the limitation of the number of electrons that boron doped-graphene can accept from cations, fewer lithium atoms may be capable of being intercalated if more calcium atoms exist in advance. This is the reason that we designed a novel composition, Ca_0.6B_1.2C_4.8, whose boron and calcium contents are lower than those in any previously reported calcium boridecarbide. In the synthesis of Ca_0.6B_1.2C_4.8, CaB₆ and CaC₂ were chosen as boron and calcium sources, as in a previous report.¹² Typically, graphite (ca. 20 μm, Kojundo Chemical Lab. Co., Ltd.), CaB₆ (Sigma-Aldrich Co. LLC) and CaC₂ (FUJIFILM Wako Pure Chemical Industries Corp.) are mixed in a stoichiometric ratio and heat-treated at 2073 K in a graphite furnace for 5 hours. Here, we emphasize that a simple one-pot heat-treatment synthesis method is advantageous for mass production. This heat treatment temperature is 200 K and 400 K higher than those for synthesizing Ca_0.75B_1.5C_4.5 and CaB₂C₄, respectively.¹² We also synthesized Ca_0.75B_1.5C_4.5 for comparative experiments. Fig. 1b shows the XRD pattern of the synthesized Ca_0.6B_1.2C_4.8, where the sharp diffraction patterns indicate its highly growing layer structure. The XRD pattern could be well-fitted by Rietveld analysis with the RIETAN-FP program.¹⁸ The length of a side of a six-membered ring consisting of B and C atoms in the layer, (B,C)–(B,C), and the interlayer distance obtained by the Rietveld refinement are summarized in Table 1. For comparison, those of Ca_0.75B_1.5C_4.5 (ref. 12) and graphite¹⁹ are also shown. Ca_0.6B_1.2C_4.8 has a 7% shorter (B,C)–(B,C) length than Ca_0.75B_1.5C_4.5, although no significant change occurred in the interlayer distance. For details on the Rietveld refinement and a comparison with Ca_0.75B_1.5C_4.5, see Fig. S1, S2 and Table S1.†Fig. 1c shows the results of the B 1s spectrum from X-ray photoelectron spectroscopy (XPS), where the intense B–C peak suggests the existence of boron atoms in heterographene planar six-ring nets instead of the formation of boron oxides (BC₂O or BCO₂). The attribution of peaks reflected the former reports on boron-doped graphene^20,21 using the XPSPEAK program.²² We note that the exact compositions of synthesized calcium boridecarbides were slightly different from the ideal ones. From the results of inductively coupled plasma atomic emission spectroscopy (ICP), the compositions were Ca_0.6−xB_1.2C_4.8 and Ca_0.75−yB_1.4C_4.6, where x = 0.08 and y = 0.02 (Table S2†). As pointed out by Wörle et al., deficiencies of intercalated cations in hetero-graphites occur easily.¹² In the present report, we denote these synthesized compounds as “Ca_0.6B_1.2C_4.8” or “Ca_0.75B_1.5C_4.5” for simplicity. See the ESI† for detailed synthesis methods and characterization.

Fig. 1 (a) A unit cell (upper) and schematic (lower) of Ca_0.6B_1.2C_4.8. Cyan: calcium, white: calcium defects, green: boron, and brown: carbon. The ratio of the colors at a site reflects occupancy in the unit cell. (b) XRD pattern of Ca_0.6B_1.2C_4.8 with Miller indices. (c) XPS spectrum of the B 1s orbital in Ca_0.6B_1.2C_4.8.

Table 1 Lattice parameters of Ca_0.6B_1.2C_4.8, Ca_0.75B_1.5C_4.5, and graphite

	(B,C)–(B,C) or C–C distance	Interlayer distance
Ca0.6B1.2C4.8	1.475 Å	4.436 Å
Ca0.75B1.5C4.5 (ref. 12)	1.492 Å	4.429 Å
Graphite^19	1.423 Å	3.355 Å

---

For electrochemical tests, laminate-type cells were set up using Ca_0.6B_1.2C_4.8 as the cathode material and lithium metal as the anode. The electrolyte was 1 M LiPF₆ in ethylene carbonate (EC)–ethyl methyl carbonate (EMC) electrolyte (Kishida Chemical Co., Ltd.). For preparing active material composites, acetylene black (AB, Denka Company Ltd.) and two types of binders, polyvinylidene difluoride (PVDF, Kureha Corp.) and polyimide (PI, I.S.T. Corp.), are used. Charge/discharge tests were performed using a battery charge/discharge system (TOSCAT-3100, TOYO SYSTEM CO., LTD.) at a rate of 3.72 mA g⁻¹. As a comparison, Ca_0.75B_1.5C_4.5 was also tested in the same way, although only a PVDF binder was used. See the ESI† for detailed cell preparation and electrochemical experiments. We denote these test conditions as Ca_0.6B_1.2C_4.8–PVDF, Ca_0.6B_1.2C_4.8–PI, and Ca_0.75B_1.2C_4.8–PVDF. The results are shown in Fig. 2a. Starting from Li intercalation, its 1st reversible capacity was 368 mA h g⁻¹ for Ca_0.6B_1.2C_4.8–PVdF and 552 mA h g⁻¹ for Ca_0.6B_1.2C_4.8–PI in the potential range from 0 to 3 V vs. Li/Li⁺. We point out that the calcium atoms in Ca_0.6B_1.2C_4.8 do not contribute to the cell capacity in the potential range from 0 to 3 V. The calcium atoms were not deintercalated by simply dipping the Ca_0.6B_1.2C_4.8 electrode in the electrolyte (Table S3†), and the capacity corresponding to calcium extraction appeared only in the potential range of over 3 V vs. Li/Li⁺ (Fig. S3†). By sweeping the potential beyond 3.2 V vs. Li/Li⁺, calcium atoms were extracted, and this improved the cell capacity. We return back to this point later. Ca_0.6B_1.2C_4.8–PI showed a higher capacity than Ca_0.6B_1.2C_4.8–PVDF, which suggests that the volumetric change in the active material may affect the performance. The PI binder is usually effective for suppressing a volumetric change of the active material during the charge/discharge process, which prevents the peeling off of active material composites from the Cu current collector or losing of electric contact among the active material composite grains.²³ The specific capacity of Ca_0.6B_1.2C_4.8–PI, 552 mA h g⁻¹, is much higher than that of graphite, 372 mA h g⁻¹. It is also higher than previously reported values for high capacity carbon-based materials, such as boron carbon nitride nanosheets (420 mA h g⁻¹),²⁴ lithium borocarbide (450 mA h g⁻¹),¹⁵ and graphene nanosheets (540 mA h g⁻¹).²⁵ Moreover, the average Li deintercalation potential of Ca_0.6B_1.2C_4.8–PI, 1.16 V vs. Li/Li⁺, is the lowest potential value among these carbon-based materials, which is an important characteristic of anode materials for obtaining high cell voltage. There are reports of carbon-based materials with even higher capacities such as graphene-like graphite (608 mA h g⁻¹),¹⁴ nitrogen-doped porous carbon nanofiber webs (1280 mA h g⁻¹)²⁶ and boron- or nitrogen-doped graphene sheets (1550 mA h g⁻¹).²⁷ However, the simple one-pot heat treatment synthesis of Ca_0.6B_1.2C_4.8 is advantageous in terms of industrial use and mass production compared with these former reported materials, which require chemical processes such as chemical exfoliation of graphite or polymerization of pyrrole followed by heat treatment during the synthesis processes. Also, the true density of Ca_0.6B_1.2C_4.8, 2.01 g cm³, is almost as high as that of graphite, 2.26 g cm³. It is advantageous in terms of volumetric capacity, which is an important characteristic for use in mobile devices or electric vehicles.

Fig. 2 (a) First Li intercalation–deintercalation curve of Ca_0.6B_1.2C_4.8–PI (blue), Ca_0.6B_1.2C_4.8–PVDF (red), Ca_0.75B_1.5C_4.5–PVDF (green), and graphite (black). (b) Replot of (a) with the x-axis as the number of intercalated or deintercalated lithium atoms. (c) Capacity cycle performance of Ca_0.6B_1.2C_4.8–PI (blue) and Ca_0.6B_1.2C_4.8–PVDF (red).

From DFT calculations, a higher boron density results in a higher reaction potential for boron-doped graphite.³ From this point of view, new compositions with even lower boron densities, and thus lower lithium intercalation potentials, may be possible to create considering the flexibility of calcium boridecarbide series in forming multiple compositions with different boron densities. So far, Ca_1.5B₃C₃, CaB₂C₄, and Ca_0.75B_1.5C_4.5 (ref. 10–12) have been reported, and Ca_0.6B_1.2C_4.8 was synthesized in this work.

Fig. 2b is a replot of Fig. 1a, plotting the x-axis as the number of lithium atoms intercalated/deintercalated to/from Ca_0.6B_1.2C_4.8 or Ca_0.75B_1.5C_4.5 calculated from the quantity of applied electricity. For Ca_0.75B_1.5C_4.5–PVDF, 0.88 lithium atoms are deintercalated from each B_1.5C_4.5 during the first Li deintercalation. For Ca_0.6B_1.2C_4.8–PVDF and Ca_0.6B_1.2C_4.8–PI, 1.3 and 1.9 lithium atoms are deintercalated for each B_1.2C_4.8, respectively, during the first Li deintercalation. This demonstrates the hypothesis of intercalation/deintercalation of more than one lithium atom per 6 atoms of boron or carbon by boron doping in graphite, and it is preserved for 7 cycles for Ca_0.6B_1.2C_4.8–PVDF and at least 10 cycles for Ca_0.6B_1.2C_4.8–PI as shown in Fig. 2c. Comparing Ca_0.6B_1.2C_4.8 to Ca_0.75B_1.5C_4.5, 0.15 Ca atoms decreased per 6 B or C atoms whereas 0.38 Li atoms increased as being intercalated per 6 B or C atoms (from 0.88 Li for Ca_0.75B_1.5C_4.5 to 1.26 Li for Ca_0.6B_1.2C_4.8). The difference in these two values can be explained by the difference in the area occupied by Ca and Li atoms on the plane parallel to boron-doped graphene layers. In Ca_0.6B_1.2C_4.8 and Ca_0.75B_1.5C_4.5, the B–C bonding in an sp² configuration contains a π-bonding orbital occupied by a C 2p_z electron and a Ca 4s electron transferred to the B 2p_z orbital.^12,15 After donating two 4s electrons to B 2p_z orbitals, Ca atoms become divalent cation states, which are highly repulsive to other cation atoms on the planes parallel to the boron-doped graphene layer. On the other hand, lithium atoms, which form monovalent cation states and are less ionized in the absence of donation of their electrons to B 2p_z orbitals, should be less repulsive to other cations and occupy smaller areas than Ca atoms. As the result, removing one Ca atom can cause an increase of more than one Li atom capable of being intercalated. For verifying this hypothesis, we discussed the difference in configurations of Ca and Li atoms in the interlayer with DTF calculations which are given in the ESI.†

Next, we discuss the characteristics of lithium intercalation into Ca_0.6B_1.2C_4.8 on the basis of the results of XRD measurements during the Li intercalation/deintercalation process of Ca_0.6B_1.2C_4.8–PVDF and DFT calculations. Fig. 3a shows the XRD patterns of the active material composite on a Cu foil at four different points in the Li intercalation/deintercalation process, namely, the as-prepared active material composite, the point at 0 V vs. Li/Li⁺ after the 1st Li intercalation, 3 V vs. Li/Li⁺ after the 1st Li deintercalation, and 0 V vs. Li/Li⁺ after the 2nd Li intercalation. The (001) peak at around 20° corresponds to the diffraction from boron-doped graphene layers, and it reversibly shifts from 20.1° (as-prepared) or 20.3° (1st Li deintercalated state) to 21.5° (1st Li intercalated states) or 21.6° (2nd Li intercalated states). This corresponds to the decrease in the boron-doped graphene interlayer distance from Li deintercalated states (4.418 Å or 4.375 Å) to Li intercalated states (4.133 Å or 4.114 Å) upon lithium intercalation. The wavelength of Cu-Kα was set to 1.5418 Å. It corresponds to 7% change in the boron-doped graphene interlayer distance. This is the first development of a high capacity carbon-based material other than graphite, which is capable of being prepared from graphite without any chemical processes and clearly shows such reversible structural changes in layer distances observable by the shift of XRD peaks upon lithium intercalation/deintercalation. The value of 4.418 Å for Ca_0.6B_1.2C_4.8 is much larger than graphite (3.356 Å) and the already-known Li-richest boridecarbide, Li₃B₃C₃ (3.529 Å).²⁸ Because of such a wide interlayer distance created by Ca atoms, 1.3 lithium atoms can be intercalated into Ca_0.6B_1.2C_4.8 without further increase of its interlayer distance. From another point of view, considering the interlayer distances of the already-known Li- and Ca-GICs, LiC₆ (3.700 Å)²⁹ and CaC₆ (4.524 Å),³⁰ the intermediate value of 4.133 Å for Li_1.3Ca_0.6B_1.2C_4.8 can be understood as the result of co-existence of two types of intercalants in its interlayer. While Ca atoms expand the interlayer distance, Li atoms contract it. The decrease in the interlayer distance upon lithium intercalation is also supported by DFT calculations. Geometric relaxations were performed for supercells containing 10 formula units of Ca_0.6B_1.2C_4.8 and Li_xCa_0.6B_1.2C_4.8 for x = 0.7, 1.0 and 1.2 using the VASP code.^31,32Fig. 3b shows the calculated interlayer distances. They linearly decrease from x = 0 to x = 1 for Li_xCa_0.6B_1.2C_4.8 though the decrease rate is suppressed at x = 1.2. See the ESI† for detailed calculation methods.

Fig. 3 (a) XRD patterns of the active material composite at four different points during the Li intercalation/deintercalation process for Ca_0.6B_1.2C_4.8–PVDF. Cu-Kα radiation was used. (b) Calculated interlayer distances for Li_xCa_0.6B_1.2C_4.8 where x = 0, 0.7, 1.0 and 1.2.

Next, we verify the effect of calcium content on the capacity. As we discussed above, calcium atoms are not extracted during the Li intercalation/deintercalation process between 0 V and 3 V vs. Li/Li⁺, and they share the interlayer sites with intercalated lithium atoms. Due to the Coulomb repulsion among intercalated cations or the limitation of the number of electrons that boron-doped graphene can accept from cations, more lithium atoms might be intercalated if fewer calcium atoms existed in the interlayer. To prepare calcium deficient compositions, a positive constant current was applied to Ca_0.6B_1.2C_4.8 (Fig. S3†). The potential was over 3 V vs. Li/Li⁺ during calcium extraction. After extraction, the composition of the active material was estimated to be Ca_0.36B_1.2C_4.8 by ICP (Table S3†). To characterize the Ca_0.36B_1.2C_4.8 composite, we compared the results of XRD and cross-sectional SEM images of Ca_0.36B_1.2C_4.8 with those of the as-prepared Ca_0.6B_1.2C_4.8 composite (Fig. 4a and c). In XRD, the peak at 20° of Ca_0.36B_1.2C_4.8, which corresponds to the 001 peak of Ca_0.6B_1.2C_4.8, is broadened toward a higher angle. This broad peak can be attributed to a group of states with shorter interlayer distances than Ca_0.6B_1.2C_4.8 formed by the electrochemical Ca extraction. The interlayer distances of these states range from 4.171 Å (21.3° in the XRD pattern) to 4.436 Å (20.0° in the XRD pattern, same as Ca_0.6B_1.2C_4.8). On the other hand, a new peak at around 23° appeared in Ca_0.36B_1.2C_4.8. This peak can be attributed to the 002 peak of the newly formed 2nd stage of Ca intercalated borocarbide as explained below. Recently, Ishikawa et al. reported the first example of a 2nd stage structure of Ca intercalated borocarbide. By the vapour-phase reaction of Ca with graphite-like layered borocarbide, BC₈, they obtained a 2nd stage structure of Ca intercalated BC₈.³³ The interlayer distance of boron-doped graphene layers was 3.41 Å for the layers without Ca atoms and 4.41 Å for the Ca intercalated layers. Using these interlayer distances, we assumed a 2nd stage model, which is made by alternately removing Ca atom layers from Ca_0.6B_1.2C_4.8 to form the 2nd stage calcium boridecarbide, Ca_0.3B_1.2C_4.8 (Fig. 4b). Then, its 002 (22.7°) and 003 (34.4°) peaks match well with the newly appeared peaks in the XRD pattern of Ca_0.36B_1.2C_4.8. Thus, we concluded that Ca extraction from Ca_0.6B_1.2C_4.8 causes phase separation to two different phases, Ca_0.6B_1.2C_4.8 with Ca defects keeping the 1st stage structures though with shorter interlayer distances, and a newly appeared 2nd stage structure. Ca_0.36B_1.2C_4.8 is a mixture of these two phases. As in cross-sectional SEM images, the aligned contrast variation of the Ca_0.36B_1.2C_4.8 composite indicates the preservation of the layer structures of Ca_0.36B_1.2C_4.8. Also, its coarser contrast variation compared with that of Ca_0.6B_1.2C_4.8 reflects the changes in crystalline structures upon Ca extraction as we explained above. A charge/discharge test was applied to the prepared Ca_0.36B_1.2C_4.8 starting from Li intercalation in the potential range from 0 to 3 V vs. Li/Li⁺. We denote the measurements as Ca_0.36B_1.2C_4.8–PVDF. The first reversible capacity was 648 mA h g⁻¹ (Fig. 4d), which was higher than that of Ca_0.6B_1.2C_4.8–PVDF (368 mA h g⁻¹). This demonstrates that extracting a portion of the calcium atoms is effective for achieving higher capacities. Until now, it is not possible to distinguish the contributions from the 1st stage with Ca defects and the newly appeared 2nd stage to the increase of capacity during the 1st cycle of Ca_0.36B_1.2C_4.8. However, the 002 and 003 peaks of the newly appeared 2nd stage remarkably diminished after the 1st cycle whereas Ca_0.36B_1.2C_4.8 keeps its high capacity at least for 10 cycles (Fig. 4e). Thus, we expect that the increase of capacity in Ca_0.36B_1.2C_4.8 is mainly due to the 1st stage with Ca defects. The capacity retention rate after 10 cycles decreased from 76% for Ca_0.6B_1.2C_4.8–PVDF to 68% for Ca_0.36B_1.2C_4.8–PVDF (Fig. 4e). This seems to be due to the lack of interlayer calcium atoms, which may induce instability of the calcium boridecarbide lattice. The instability of metal-boridecarbides induced by the lack of interlayer cations has also been pointed out in previous reports.^4,13

Fig. 4 (a) XRD patterns of the Ca_0.36B_1.2C_4.8 composite and Ca_0.6B_1.2C_4.8 composite, calculation results for a model of the 2nd stage as shown in (b) and Ca_0.36B_1.2C_4.8 after 1 cycle of lithium intercalation and deintercalation. The XRD pattern for the model of the 2nd stage was calculated and visualized using the VESTA program¹⁷ and (b) model of the 2nd stage, Ca_0.3B_1.2C_4.8, which is made by alternatively removing Ca atom layers from Ca_0.6B_1.2C_4.8 (unit cell in Fig. 1a) though the interlayer distances of Ca intercalated layers and layers without Ca atoms were set to those reported by Ishikawa et al.³³ (c) Cross-sectional images of Ca_0.36B_1.2C_4.8 and Ca_0.6B_1.2C_4.8 composites and (d) first Li intercalation/deintercalation profile of Ca_0.36B_1.2C_4.8–PVDF. (e) Capacity cycle performance of Ca_0.36B_1.2C_4.8.

Finally, we discuss the influence of B-doping on the electrochemical capacity. Fig. 5 shows the relation between boron density and the total number of cations capable of being intercalated into Ca_0.75B_1.5C_4.5, Ca_0.6B_1.2C_4.8, Ca_0.36B_1.2C_4.8 and the already-known Li–GIC, Li₃B₃C₃ and LiC₆. Basically, more cations are capable of being intercalated into the graphite-like structures as the boron density becomes higher. It is because boron-doped graphite has a thermodynamically lower energy state than graphite and more cation atoms can be intercalated. From the DFT calculation, Doh et al. predicted that if the density of boron doping exceeds the composition of BC₅ in graphite-like borocarbide, namely 16% boron doping, the borocarbide is thermodynamically capable of storing 3 Li atoms per 6 B or C atoms. It corresponds to Li₃C₆ or Li₃B₃C₃ configuration, where Li atoms occupy all the centers of hexagons in the honeycomb lattice of graphene or boron-doped graphene layers. Now, both Ca_0.6B_1.2C_4.8 (20%) and Ca_0.75B_1.5C_4.5 (25%) exceed 16% boron doping, and are potentially capable of storing 3 Li atoms per 6 B or C atoms. However, as we pointed above, the Ca content is another factor regulating the intercalation of Li atoms into the calcium boridecarbide. This is why calcium boridecarbides can only store fewer Li atoms than Li₃B₃C₃, and Ca_0.75B_1.5C_4.5 can store fewer Li atoms than Ca_0.6B_1.2C_4.8 even if it has a higher boron content. Here, we can summarize that higher B doping can help to store more cations in the graphite-like structure, though the content of pre-existing Ca atoms is also an important factor for the electrochemical capacity due to intercalation/deintercalation of Li atoms. A higher boron density and fewer Ca atoms are advantageous for high capacity. Also, removing a part of the pre-existing Ca atoms can increase the Li capacity. However, we point out that removing all the pre-existing Ca atoms should induce instability in the calcium boridecarbide lattice analogous to Li₃B₃C₃, where extracting about a half of the pre-existing Li atoms causes the instability of the lattice.^4,15 Also, from the view point of potential in operation, a higher boron density causes higher intercalation/deintercalation potential for Li atoms,³ which is inappropriate for use as the anode material.

Fig. 5 Relation between the boron density in the boron doped-graphene layers and the total number of cations capable of being intercalated into Ca_0.75B_1.5C_4.5, Ca_0.6B_1.2C_4.8, Ca_0.36B_1.2C_4.8 and the already-known Li–GIC, Li₃B₃C₃ and LiC₆.

Conclusions

In this study, we reported a high capacity graphite-like calcium boridecarbide, Ca_0.6B_1.2C_4.8, as a novel anode active material for lithium-ion batteries. It showed a high specific capacity (552 mA h g⁻¹) and one of the lowest lithium deintercalation potentials (1.16 V vs. Li/Li⁺) among previously reported high capacity carbon-based materials. By electrochemically removing a part of the calcium atoms from Ca_0.6B_1.2C_4.8 as pretreatment, its specific capacity increased to 648 mA h g⁻¹. Furthermore, we unveiled its unique Li intercalation/deintercalation behavior coincident with a 7% change in the boron-doped graphene layer distance while keeping its ordered layer structure. Considering the flexibility of preparing different compositions of calcium boridecarbide, a still higher-performance anode may be achieved by designing and tuning the boron and calcium densities. Choosing other dopants for the graphite template may also be effective for changing performance. Furthermore, the material's simple one-pot heat treatment synthesis method is advantageous for industrial use and mass production. We expect this study to shed light on the possibility of achieving an even higher-capacity anode material using a graphite template while keeping the advantages of graphite, such as low potential, excellent cycle performance, and economic efficiency.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors are grateful to Dr Mitsuhiro Hibino for helpful discussions and technical support for Rietveld analysis.

Notes and references

1. N. Nitta, F. Wu, J. Lee and G. Yushin, Mater. Today, 2015, 18, 252–264.
2. K. R. Kganyagoa, P. E. Ngoepea and C. R. A. Catlow, Solid State Ionics, 2003, 159, 21–23.
3. C. Doh, B. Han, B. Jin and H. Gu, Bull. Korean Chem. Soc., 2011, 32, 2045–2050.
4. Q. Xu, C. Ban, A. C. Dillon, S. Wei and Y. Zhao, J. Phys. Chem. Lett., 2011, 2, 1129–1132.
5. K. Yamada, H. Ishikawa, C. Kamiwaki and M. Kawaguchi, Electrochemistry, 2015, 83(6), 452–458.
6. M. Sahoo, K. P. Sreena, B. P. Vinayan and S. Ramaprabhu, Mater. Res. Bull., 2015, 61, 383–390.
7. L. Wang, P. Yu, L. Zhao, C. Tian, D. Zhao, W. Zhou, J. Yin, R. Wang and H. Fu, Sci. Rep., 2014, 4, 5184.
8. X. Zhou, L. Ma, J. Yang, B. Huang, Y. Zou, J. Tang, J. Xie, S. Wang and G. Chen, J. Electroanal. Chem., 2013, 698, 39–44.
9. H. Fujimoto, A. Mabuchi, C. Natarajan and T. Kasuh, Carbon, 2002, 40, 567–574.
10. T. Breant, D. Pensec, J. Bauer and J. Debuigne, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1980, 36, 1540.
11. L. Y. Markovskii and N. V. Vekshina, Zh. Prikl. Khim., 1964, 37, 2126.
12. M. Wörle, D. Widmer, F. von Rohr, F. Krumeich, C. Mensing and R. Nesper, Z. Anorg. Allg. Chem., 2010, 636, 1447–1453.
13. Y. Matsuo and Y. Sugie, J. Electrochem. Soc., 1999, 146, 2011–2014.
14. Q. Cheng, Y. Okamoto, N. Tamura, M. Tsuji, S. Maruyama and Y. Matsuo, Sci. Rep., 2017, 7, 14782.
15. D. Li, P. Dai, Y. Chen, R. Peng, Y. Sun and H. Zhou, J. Phys. Chem. C, 2018, 122, 18231–18236.
16. A. M. Fogg, J. Meldrum, G. R. Darling, J. B. Claridge and M. J. Rosseinsky, J. Am. Chem. Soc., 2006, 128, 10043–10053.
17. K. Momma and F. Izumi, J. Appl. Crystallogr., 2011, 44, 1272–1276.
18. F. Izumi and K. Momma, Solid State Phenom., 2007, 130, 15–20.
19. P. Trucano and R. Chen, Nature, 1975, 258, 136–137.
20. X. Hai, Q. Mao, W. Wang, X. Wang, X. Chen and J. Wang, J. Mater. Chem. B, 2015, 3, 9109–9114.
21. N. Sreekanth, M. A. Nazrulla, T. V. Vineesh, K. Sailaja and K. L. Phani, Chem. Commun., 2015, 51, 16061–16064.
22. R. W. M. Kwok, XPS Peak Fitting Program: XPSPEAK Version 4.1, Department of Chemistry, The Chinese University of Hong Kong.
23. J. Kim, W. Choi, K. Cho, D. Byun, J. Lim and J. Lee, J. Power Sources, 2013, 244, 521–526.
24. W. Lei, S. Qin, D. Liu, D. Portehault, Z. Liue and Y. Chen, Chem. Commun., 2013, 49, 352–354.
25. E. Yoo, J. Kim, E. Hosono, H. Zhou, T. Kudo and I. Honma, Nano Lett., 2008, 8, 2277–2282.
26. L. Qie, W. Chen, Z. Wang, Q. Shao, X. Li, L. Yuan, X. Hu, W. Zhang and Y. Huang, Adv. Mater., 2012, 24, 2047–2050.
27. Z. Wu, W. Ren, L. Xu, F. Li and H. Cheng, ACS Nano, 2011, 5, 463–5471.
28. M. Wörle and R. Nesper, Z. Anorg. Allg. Chem., 1995, 621, 1153–1159.
29. O. Dolotko, A. Senyshyn, M. J. Mühlbauer, K. Nikolowski and H. Ehrenberg, J. Power Sources, 2014, 255, 197–203.
30. N. Emery, C. Hérold, M. d'Astuto, V. Garcia, C. Bellin, J. F. Marêché, P. Lagrange and G. Loupias, J. Power Sources, 2014, 255, 197–203.
31. G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169–11186.
32. G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 6, 15–50.
33. H. Ishikawa, H. Higuchi and M. Kawaguchi, Chem. Lett., 2018, 47, 891–893.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9se00061e

---