Low-temperature Liquid Exfoliation of Milligram-scale Single Crystalline Few-layer β12-Borophene Sheets as Ecient Electrocatalysts for Lithium– Sulfur Batteries

Two-dimensional (2D) borophene is predicted as an ideal electrode material for lithium sulfur (Li-S) batteries because of low-density, metallic conductivity, high Li-ion surface mobility and strong interface bonding energy to polysulde. But until now, 2D borophene-based Li-S batteries have not yet been achieved due to the absence of massive synthesis method. Herein, we developed a novel low-temperature liquid exfoliation (LTLE) method for scalable synthesis of single crystalline 2D few-layer β 12 -borophene sheets with a symmetry. The as-synthesized 2D sheets were used as the polysulde immobilizers and electrocatalysts of Li-S batteries for the rst time. The resulting Li-S cells employing borophene sheets delivered a strikingly high areal capacity of 5.2 mAh cm − 2 at a high sulfur loading of 7.8 mg cm − 2 with an ultralow capacity fading rate (0.039 % per cycle) in 1000 cycles, outperforming most of the Li-S batteries employing other 2D materials. Under the help of few-layer β 12 -borophene, their high-activity behaviors should be attributed to the signicant enhancement of both the Li-ion’s surface migration and the adsorption energy for Li 2 S n clusters based on density functional theory (DFT) models. Our research reveals great potential of 2D β 12 -borophene sheets in future high-performance Li-S batteries.


Introduction
The rapid development of electrochemical energy storage devices in the elds of electric vehicles, portable electronic devices and large-scale smart power grids continuously drive the researchers to explore lower cost, higher energy density, and better safety batteries than current lithium-ion batteries [1][2][3][4] . Among many candidates, lithium sulfur (Li-S) batteries have been gaining the global attention due to their overwhelming energy density (2600 Wh kg -1 ), natural abundance and environment-friendly of sulfur feedstock [5][6][7][8][9] . However, the existence of internal polysul de shuttling, large volume expansion of sulfur and sluggish redox kinetics inevitably lead to the sharp deterioration of the electrochemical performances of Li-S batteries [10][11][12] .
Recently, two-dimensional (2D) materials with strong in-plane covalent bonds and weak interlayered van der Waals (vdW) forces have been intensively studied because of their superior advantages over traditional bulk materials for Li-S cell applications 27,28 . 2D materials such as siloxane 29 , black phosphorene 30,31 , BN 32,33 , C 3 N 4 34,35 , and MXene 36,37 , were found to exhibit excellent catalytic activities towards polysul des because of their extraordinary surface properties. However, most of these 2D material-based Li-S cells still have some disadvantages, such as low capacity 38 , slow charge-discharge rate 39,40 , structure instability 38,41 , and poor cyclic stability 32 . Hence, the design and development of novel 2D materials are highly demanded towards high-performance Li-S batteries with large catalytic activity, high-e cient adsorption, fast conversion of polysul des and long-term durability. As a typical 2D Dirac material consisted of the lightest solid element, 2D borophene with unique surface con guration and complex multicenter-two electron bonds has been earlier predicted as an ideal electrode material for Li-S batteries due to its native metallic conductivity 42 , large elastic modulus 43 , heavy anisotropy 44 , high Fermi velocity (6.6×10 5 m/s) 45 , excellent thermal and chemical stability 46 , large Li-ion surface mobility as well as strong bonding energy to polysul de clusters 47 . However, borophenebased Li-S cells have not yet been achieved for practical use so far owing to the absence of a facile route for the scalable production of 2D borophene nanomaterials.
In this work, we developed a low-temperature liquid exfoliation (LTLE) strategy for scalable production of single crystalline borophene sheets as e cient polysul de electrocatalyst for Li-S batteries. Few-layer 2D borophene sheets with β 12 -phase were thus identi ed with an average ake size of ~3 μm and an ultrathin thickness less than 10 atomic layers. The β 12 -borophene sheets exhibited extraordinary performances as e cient immobilizer and electrocatalyst for advanced Li-S batteries, showing excellent rate performance of 721 mAh g -1 at 8 C (1 C = 1675 mAh g -1 ) and an ultralow decay rate of less than 0.039 % in 1000 continuous cycling measurements. More impressively, the areal capacity can arrive as high as 5.2 mAh cm -2 at a large sulfur loading of 7.8 mg cm -1 in lean electrolyte with a ratio of electrolyte to sulfur (E/S) ratio of 6.8 ml g -1 . Our work suggests that single crystalline few-layer borophene sheets hold great potential for high-e ciency Li-S batteries.

Results And Discussions
The LTLE synthesis process of few-layer borophene sheets is schematically illustrated in Fig. 1a. By optimizing the sonication power and solvent concentration, massive production of 2D borophene sheets has been successfully achieved. N-methyl pyrrolidone (NMP) was found to be the most effective among a series of solvents adopted in our experiment, as seen in Supplementary Figs. 1 and 2. The color of the product solution is dark-brown or dark-black in Fig. 1b, varying with the sheet concentration. Moreover, the mass of 2D sheets reaches as high as 10 mg and their yield is over 20 %, evidently increased compared with previous reports (≤ 10 % 48,49 ). The low-temperature approach is thus believed to improve the exfoliation e ciency of non-layered bulk materials, because of which signi cantly enhances the anisotropy discrepancy between the in-plane and out-of-plane covalence bonds 50 . Scanning electron microscope (SEM) and atomic force microscope (AFM) images of the as-grown products are respectively in Fig. 1c, d, where ultrathin 2D borophene sheets are observed to have an edge length of 2 ~ 5 µm and exhibit uniform and smooth appearance. The thickness of 2D sheet is observed to be only 1.32 ~ 2.32 nm, suggesting its ultra-thin nature.
X-ray diffraction (XRD) pattern of 2D borophene sheets ( Supplementary Fig. 3) shows the same characteristic diffraction pattern as that of the theoretically calculated β 12 -borophene using density functional theory (DFT), clearly different from that of the bulk β-rhombohedral boron powder (JCPDS No. 00-031-0207). As seen in Fig. 1e, the X-ray photoelectron spectrum (XPS) of B 1s core level is consisted of two characteristic components, attributed to the B-B species at 187.5 eV 51 and the B-O species at 189.1 eV 52 , respectively. And the molar ratio of the B-B to B-O species is estimated to be more than 94 %, suggesting a majority of pure boron composition in 2D borophene sheets. The minor B-O species are supposed to originate from the edge oxidation of borophene sheets during the short exposure to the air after being taken out for XPS measurements ( Supplementary Fig. 4) 52 .
Raman spectroscopy was employed to better differentiate the 2D borophene sheets from the bulk boron powders 53,54 . Four Raman peaks of 2D borophene sheets are clearly identi ed (Fig. 1f) as the ngerprints of β 12 phase 55 , differing from those of bulk boron with β-rhombohedral phase. Accordingly, the strong peak at ~ 268 cm − 1 is ascribed to the out-of-plane bending vibration mode of β 12 phase 55 .
And the other peaks at ~ 423, ~901 and ~ 1017 cm − 1 are respectively indexed as the , and modes, resulting from the in-plane stretching modes of β 12 phase 55 .
Transmission electron microscopy (TEM) was performed to determine the surface con guration of 2D  Fig. 5a). Therefore, the atomic layer numbers of the as-synthesized borophene sheets should be less than 10, unveiling the ultrathin nature of few-layer borophene. In addition, the 2D β 12 -borophene sheets are thermodynamically stable as evidenced by the absence of any negative frequency in the entire Brillouin zone ( Fig. 2d) according to the density functional perturbation theory (DFPT). Similar calculations are carried out on the (104) plane ( Supplementary Fig. 5), which are also in good agreement with our experimental results. High-angle annular dark-eld scanning transmission electron microscope (HAADF-STEM) and energy dispersive X-ray spectroscopy (EDX) mapping ( Fig. 2e-h) images also reveal a uniform distribution of boron element across the 2D sheet with a pure boron content over 98 %, which is in good consistent with the electron energy loss spectrum (EELS) ( Supplementary Fig. 5b).
Based on all characterizations mentioned above, we can thus conclude that single crystalline few-layer borophene sheets with β 12 -B 5 phase were successfully synthesized using LTLE. In comparison with other  Table 1, our route is thus low-cost, facile and highe cient for scalable production of single crystalline β 12 -borophene sheets towards practical applications such as Li-S battery.
To demonstrate the catalytic activity of few-layer β 12 -borophene sheets for Li-S battery, the potentiostatic experiments were carried out to monitor the liquid-solid conversion in the nucleation and growth of Li 2 S from polysul des. The galvanostatic discharge was respectively performed on CNT/β 12 -borophene (CNT/B) and bare CNT hosts at 2.05 V, in which 0.01 V overpotential was used to induce the generation of Li 2 S. All the cells reached the highest potentiostatic current after about 1000 s, but the nucleation abilities of Li 2 S were found to be completely different, causing a capacity of 193 mAh g − 1 and 72 mAh g − 1 for CNT/B and CNT electrodes, respectively (Fig. 3a). Besides, the dissolution ability of solid Li 2 S was also remarkably promoted by the implantation of few-layer β 12 -borophene. After the fully conversion of sulfurs into Li 2 S, the Li 2 S dissolution was kinetically evaluated by using a potentiostatic charge process.  29 . In contrast with the Li-S battery using non-undecorated CNT electrode, the Li-S battery using CNT/B electrode possessed higher current density (Fig. 3d). As shown in Fig. 3e, the Tafel plots of the rst oxidation process of the cells using CNT and CNT/B were respectively 57 and 29 mV dec − 1 , where the smaller Tafel slope of the CNT/B-based cell suggests that the β 12 -borophene induces higher surface reaction rates. In addition, the simulated interfacial impedance of the Li-S cells sharply decreased from 41.2 Ω to 24.9 Ω when the electrodes changed from CNT to CNT/B (Fig. 3f), re ecting the β 12borophene was more favorable for the interface electrochemical reactions 8 .
Considering the distinguished electrocatalytic reactivity and polysul de interactions of the CNT/B-based Li-S battery in the sulfur redox reactions, their actual working performances were further evaluated by regarding the bare CNT-based Li-S battery as a reference. In our experiments, the same amount of polysul de (Li 2 S 8 ) solution was added as active material ( Supplementary Fig. 6). As seen in Fig. 4a, the CV pro les of CNT/B-based battery overlap each other and exhibit excellent reversibility in the redox process, revealing the high-e ciency utilization of sulfur. The galvanostatic charge/discharge pro les are shown in Fig. 4b. The high reversible speci c capacities of 1329, 1236, 1159, 1057, 993, and 919 mAh g − 1 were obtained at 0.3, 0.5, 1, 2, 3, and 5 C rates (1 C = 1675 mAh g − 1 ), respectively. Even if the current density increased to 8 C, the CNT/B-based Li-S battery still remained an ultrahigh capacity of 721 mAh g − 1 . More signi cantly, after returning current density back to 0.3 C, a reversible capacity of 1216 mAh g − 1 recovered immediately with a columbic e ciency of nearly 100 % (Fig. 4c). By contrast, the battery using bare CNT electrode exhibited inferior rate performances, such as a lower initial capacity of 981 mAh g − 1 at 0.3 C and a rapider degradation into 394 mAh g − 1 with the increase of capacity to 8 C as well as unsatisfactory capacity restoration after high-rate test (Fig. 4c). As shown in Fig. 4d, the CNT/B-based battery possessed a much lower polarization voltage of 188 mV than the CNT-based battery (217 mV), further revealing the outstanding catalytic property of β 12 -borophene sheets for polysul de conversion.
The CNT/B cathode also exhibited excellent cycling stability at current density of 0.5 C, as found in mAh g − 1 after 300 cycles, resulting in a fast-fading rate of 0.2 % per cycle ( Fig. 4e and Supplementary  Fig. 7). In addition, both of the high-and low-plateau capacities of CNT/B electrode were much better than bare CNT electrode, demonstrating few-layer β 12 -borophene sheets can effectively suppress the polysul de diffusion and improve the polysul de immobilization (Supplementary Fig. 9) 56 . Moreover, high areal sulfur loadings of 5.3 mg cm − 2 and 7.8 mg cm − 2 with low E/S ratios of 9.8 and 6.8 ml g − 1 were respectively performed on the Li-S batteries to test the high-energy density behaviors. It was noted that the areal capacities of the CNT/B-based Li-S cells can reach up to 4.6 and 5.2 mAh cm − 2 when the capacities respectively adopted 871 and 661 mAh g − 1 (Fig. 4f), which were much higher than those of 4.0 mAh cm − 2 for commercial Li-ion batteries 57 . Impressively, the CNT/B-based battery could preserve an enough high reversible capacity of 572 mAh g − 1 with an extremely-low capacity decay rate of 0.039 % per cycle after 1000 long-term cycles, re ecting excellent cycling stability ( Fig. 4g and Supplementary Fig. 8).
Finally, we calculated the adsorption energy of soluble polysul des on a monolayer β 12 -borophene using DFT calculation to comprehend the improvement mechanism of β 12 -borophene sheets on Li-S batteries, as observed in Supplementary Fig. 10. Figure 5a gives the optimized con gurations of S 8 and Li 2 S n on monolayer β 12 -borophene sheet. Based on the DFT calculations, S 8 has the weakest adsorption energy on borophene of only 1.23 eV among all con gurations, and the adsorption energy gradually increases with the progression of the polysul des' lithiation and eventually arrives at 3.8 eV for the fully-lithiated Li 2 S (Fig. 5b). The adsorption energy of polysul des on β 12 -borophene is far higher than that on CNT (below 1 eV), unveiling that β 12 -borophene can anchor polysul de and inhibit the shuttle of lithium polysul de more effectively than CNT. Figure 5c shows the typical partial density of states (PDOS) of Li 2 S 4 on monolayer β 12 -borophene, and more details can be seen in Supplementary Fig. 11. The 2p orbital electrons of Li 2 S 4 and β 12 -borophene were found to overlap near the Fermi level, suggesting the formation of a strong chemical bonding between β 12 -borophene and Li 2 S 4 cluster. This is probably originated from a strong charge transfer of 0.22 e from β 12 -borophene to Li 2 S 4 cluster (Fig. 5d) based on the charge density difference and bader charge analysis. The strong chemical interaction between β 12borophene and Li 2 S 4 cluster can be also ascertained because the dark-yellow color of Li 2 S 4 solution will gradually attenuate with the increase of the mixing time with β 12 -borophene ( Supplementary Fig. 12).
Furthermore, the diffusion barrier of Li + on β 12 -borophene was deduced to be only 0.10 eV (Fig. 5e), much lower than that (0.28 eV) on CNT (Supplementary Fig. 13). The enhanced surface migration of Li + on borophene would further accelerate the nucleation and decomposition of Li 2 S n and thus improves the capacity and charge-discharge rate of Li-S battery 60 .
In summary, we have developed a novel, facile and high-yield LTLE strategy to produce single crystalline few-layer β 12 -borophene sheets. As promising 2D electrode materials, the β 12 -borophene sheets were rstly used as e cient polysul de-conversion electrocatalysts for Li-S batteries. Due to the usage of fewlayer β 12 -borophene sheets, the CNT/B-based Li-S batteries exhibited a high areal sulfur loading of 5.2 mgh cm − 2 at 7.8 mg cm − 2 under a low E/S ratio of 6.8 ml g − 1 at 0.3 C. Compared with the CNT-based Li-S cell, the CNT/B-based Li-S cell exhibited a better rate performance of as high as 721 mAh g − 1 at 8 C and a much lower decay rate of only 0.039 % in 1000 cycles. By DFT calculations, β 12 -borophene had a lower surface diffusion barrier of Li ion and a stronger adsorption for Li 2 S n clusters than CNT, which can effectively inhibit the shuttle effect of polysul des and accelerate their decomposition at the same time. These should be responsible for the extraordinary catalytic activity of β 12 -borophene towards polysul des in the CNT/B-based Li-S cell. Therefore, our strategy will pave a new way for the design of high-energy rechargeable batteries through the exploration of 2D boron-based nanomaterials.

Methods
Synthesis of few-layer β 12 -borophene sheets. The low-temperature liquid exfoliation (LTLE) method was rstly developed to synthesize β 12 -borophene sheets at milligram scale by using boron powder (99.8 %, Zhongnuo Incorp., China) as source materials. Firstly, 20 ~ 50 mg boron powers were added into 50 ml N,N-Dimethylformamide (NMP, 99.9 %, Innochem. Incorp., China) to form uniform and well-dispersed solution by several minutes' stirring, as seen in Figure S1. Secondly, the boron-power solution was transferred into the ethanol path and treated at -20~-25 ℃ in the tip-type ultrasonicator equipped with cooling system (SXSONIC Incorp., China), where the ultrasonic power was kept at 800 W and the treatment lasted for 4 ~ 8 h. Thirdly, the product solution was statically settled at room temperature for 48 ~ 72 h to enough precipitate the undissolved boron powder. Finally, the suspension was centrifuged at about 10000 ~ 11000 revolutions per minute (rpm) for 30 minutes to obtain solid products. After the above synthesis process, the mass of the collected 2D sheets was ranging from 4 to 10 mg. Accordingly, the yield of 2D few-layer borophene by LTLE way can reach as high as over 20 %, which is much higher than those by many other methods in previous reports (Supplementary Table 1) 48,49 .
Material characterizations. The morphology of β 12 -borophene sheets was investigated by SEM (Zeiss Supra 60) and AFM (Bruker Dimension Fastscan). XPS (Thermo sher Nexsa), XRD (D-MAX 2200 VPC) and Raman spectroscope (inVia Re ex, 532-nm laser) were respectively used to analyze the chemical compositions of the sample. UV-vis spectroscopy (UV-3600) was applied to determine the energy-band structure and absorption coe cient of β 12 -borophene sheets. TEM and HRTEM (FEI Titan 80-300) were employed to ascertain the lattice structure of the product. The STEM and elemental mapping were performed on a JEM ARM200F thermal-eld emission microscope with a probe Cs-corrector working at 200 kV. For the HAADF imaging, the convergence angle of ~ 23 mrad and collection angle range of 68 ~ 174 mrad were adopted for the incoherent atomic number imaging. Both the elemental composition and distribution were analyzed on the energy dispersive X-ray analyzer (EDS, EX-230 100m 2 detector) equipped with the microscope.
Preparation of Li 2 S 8 catholyte. The sources of sulfur and Li 2 S with a molar ratio of 7:1 were put into an appropriate amount of 1 mol l − 1 lithium bis (tri uoromethanesulfonyl) imide (LiTFSI). Secondly, the LiTFI solution was added into the mixed solvent of 1,3-dioxolane (DOL) and 1,2-dimethoxyethane (DME) (volume ratio (v:v): 1:1). Thirdly, 1 wt.% LiNO 3 was used as additive by vigorous magnetic stirring at 50 ℃ until the sulfur powder were fully dissolved. The concentration of Li 2 S 8 was ranging from 0.15 to 2 mol l − 1 .
Fabrication of CNT electrodes. 5 g commercial multiwalled CNT (length ~ 50 nm, Aladdin, China) powders were dispersed into 50 ml Triton X-100 aqueous solution (0.01 wt.%, Secco Romeo, China) to form uniform and homodisperse solution by ultrasonication for 2 h. Subsequently, the obtained CNT solution was ltered through nylon lm under vacuum. After three-time washing by deionized water and drying for 2 h at 60 ℃ in vacuum oven, the free-standing CNT paper was peeled from the nylon lm. Finally, the obtained CNT paper was cut into desired disks as the free-standing electrode.
Assembly of symmetric cells for kinetic evaluation of polysul de conversion. CNT/B (with a mass loading of about 1 mg β 12 -borophene sheets) or bare CNT electrodes were used as both working and counter electrodes. And 40 µl catholyte (0.5 mol l − 1 Li 2 S 6 and 1.0 mol l − 1 solution of LiTFSI with 1 wt.% LiNO 3 in DOL and DME, v/v = 1:1) was added into each coin cell. The CV behaviors of the symmetric cell were tested at a scan rate of 10 mV s − 1 , in which the voltage window ranged from − 0.8 to 0.8 V.
Measurement on the nucleation and dissolution of Li 2 S. The CNT/B or CNT lm electrodes were used as cathodes and Li foils were employed as the anodes. Also, 20 µl Li 2 S 8 solution (0.15 mol l − 1 ) was applied as catholyte, and 20 µl electrolyte without Li 2 S 8 was used as anolyte. For the nucleation and growth of Li 2 S, the assembled cells were rst discharged galvanostatically to 2.06 V at 0.112 mA, and then discharged potentiostatically to 2.05 V until the current dropped to below 10 − 5 A. The deposition capacities of Li 2 S were calculated according to the Faraday's law. For the Li 2 S dissolution, the assembled cells were rstly galvanostatically discharged to 1.80 V at 0.10 mA, and subsequently galvanostatically discharged to 1.80 V at 0.01 mA for fully transforming sulfur species into solid Li 2  Both lattice parameters and atomic positions were optimized by conjugate gradient method, and the convergence criteria for energy and force were eV and eVÅ -1 , respectively. The kinetic energy cutoff for plane waves was set at 450 eV. The Brillouin zones were sampled with Å -1 spacing in reciprocal space by the Monkhorst-Pack scheme 66 . The high symmetry Kpoints for band structure and phonon dispersion curves were generated by AFLOW package 67

. And
Grimme's DFT-D3 van der Waals corrections with the Becke-Jonson damping 68,69 was employed. The phonon spectrum was calculated by DFPT method implemented in Phonopy program 69 . Also, the crystal structures were visualized by VESTA package 70 .
Computational methods of the adsorption energy of few-layer β 12 -borophene. First-principle calculations were implemented using VASP 61 software package. The PBE 65 functional of generalized gradient approximation (GGA) was used for the exchange-correlation. The basis set utilized PAW pseudopotential method 63,64 , and the energy cutoff was set at 400 eV. The self-consistent eld (SCF) tolerance was eV and the force convergence criterion for atomic relaxation was 0.02 eV Å −1 . A Monkhorst-Pack k-point mesh with different sizes was chosen to meet various requirements, where is for the geometrical relaxation, is for the calculation of electronic structure and is for the calculation of adsorption. The vdW forces between Li 2 S n and β 12 -borophene sheet or CNT were accurately obtained by the DFT-D3 method 68  CNT was used for the adsorption energy and CI-NEB calculations, respectively. The adsorption energy ( ) was derived using the following equation:

Data availability
The authors declare that all the data supporting the ndings of this study are available within the article and its Supplementary

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SIBoropheneLiSbattery20210204.docx