Regulating surface electron structure of PtNi nanoalloy via boron doping for high‐current‐density Li‐O2 batteries with low overpotential and long‐life cyclability

The realization of high‐efficiency, reversible, stable, and safe Li‐O2 batteries is severely hindered by the large overpotential and side reactions, especially at high rate conditions. Therefore, rational design of cathode catalysts with high activity and stability is crucial to overcome the terrible issues at high current density. Herein, we report a surface engineering strategy to adjust the surface electron structure of boron (B)‐doped PtNi nanoalloy on carbon nanotubes (PtNiB@CNTs) as an efficient bifunctional cathodic catalyst for high‐rate and long‐life Li‐O2 batteries. Notably, the Li‐O2 batteries assembled with as‐prepared PtNiB@CNT catalyst exhibit ultrahigh discharge capacity of 20510 mA·h/g and extremely low overpotential of 0.48 V at a high current density of 1000 mA/g, both of which outperform the most reported Pt‐based catalysts recently. Meanwhile, our Li‐O2 batteries offer excellent rate capability and ultra‐long cycling life of up to 210 cycles at 1000 mA/g under a fixed capacity of 1000 mA·h/g, which is two times longer than those of Pt@CNTs and PtNi@CNTs. Furthermore, it is revealed that surface engineering of PtNi nanoalloy via B doping can efficiently tailor the electron structure of nanoalloy and optimize the adsorption of oxygen species, consequently delivering excellent Li‐O2 battery performance. Therefore, this strategy of regulating the nanoalloy by doping nonmetallic elements will pave an avenue for the design of high‐performance catalysts for metal‐oxygen batteries.


| INTRODUCTION
5][6] Especially, Li-O 2 batteries with reactants from the external environment have attracted intensive attention because of their ultrahigh theoretical energy density (3505 W•h/kg). 7,8Typically, Li-O 2 batteries work by forming and decomposing Li 2 O 2 products at the cathode, which run by oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) during discharge and charge processes, respectively. 9However, there is a huge gap between actual and theoretical performances, which is resulted from the vulnerability and dendrites issues of Li metal at anode, the sluggish reaction kinetics of ORR/ OER, the electronic insulation, and insolubility of discharge products (e.g., Li 2 O 2 ), which severely increase the overpotential and decrease the energy efficiency of Li-O 2 batteries. 10,11Additionally, the excessive overpotential readily leads to the decomposition of organic liquid electrolytes and the corrosion of carbon-based materials at the cathode, especially under high current density. 12,13Subsequently, the formation of side products results in higher overpotential in return; this vicious circle degrades and disables the Li-O 2 battery performance, especially at high current density. 14,150][21] Typically, Bruce and colleagues 22 constructed a porous gold electrode to improve the reversibility and rate performance of Li-O 2 batteries.Moreover, the OER kinetics on the Au electrode is approximately 10 times faster than that on the carbon electrode.Besides, other noble metal-based nanomaterials, such as Pt, 23 Pd, 24 and Ru 25 were also applied in the Li-O 2 batteries to boost the ORR/OER performance.
However, due to the high cost and scarcity of noble metals, commercial development of Li-O 2 batteries could not be afforded.Alternatively, alloying is an innovative approach to simultaneously improve catalytic activity, stability, and utilization of precious metals.To lower the cost of catalysts and maintain the activity undegraded simultaneously, the noble metals are generally combined with secondary transition metals to disperse the active phase (noble metals) and module their electronic structure. 21,26,27For instance, Amine and colleagues [28][29][30] proposed bimetallic Pt-based cathode catalysts of PtCo and PtCu for Li-O 2 batteries, exhibiting dramatically lower charging overpotential than those of pure Pt catalysts due to structural regulation of transition metals.Further, Wu and colleague 31 developed Pt/NiO catalyst with very limited Pt loading for Li-O 2 cells.Owing to the optimized electrochemical conductivity and synergistic effect of Pt and Ni, the resultant Li-O 2 battery showed much reduced electrochemical polarization and excellent cycling stability.3][34][35][36][37][38][39] Given the similarly large OER overpotential (usually over 1.0 V) of nonaqueous Li-O 2 batteries during charge, it would be efficient to boost the OER if the boron element is introduced into the designed electrocatalysts for nonaqueous Li-O 2 batteries.Despite great efforts on bimetallic noble metal-based catalysts, it is full of challenges to module the structure of bimetallic alloy as bifunctional ORR/OER electrocatalysts for Li-O 2 battery with enhanced cycling stability, capacity, and rate capability simultaneously.
Herein, we proposed a versatile surface-engineering strategy to prepare PtNiB nanoalloy on carbon nanotube (PtNiB@CNT) electrocatalysts by B doping, which can efficiently regulate the electronic structure of PtNi.The Li-O 2 battery assembled with as-prepared PtNiB@CNT electrocatalyst exhibited ultrahigh discharge capacity of 20510 A•h/g, narrow discharge/charge overpotentials of 0.48 V at 1000 mA/g and 1000 mA•h/g, excellent cycling stability, and high rate capability.Further, the analysis of Li 2 O 2 discharge products with ex situ characterization techniques and chromogenic reaction confirmed the good reversibility of the Li-O 2 reaction during discharge and charge processes.Importantly, both experimental and theoretical results revealed that the outstanding Li-O 2 battery performance was attributed to the precise surface engineering of the PtNi nanoalloy electronic structure via B doping, simultaneously regulating the adsorption of related intermediates on the PtNiB surface.

| Preparation of PtNiB@CNTs
Twenty milligrams of CNTs (Aladdin Co., Ltd.) were dispersed in 50 mL deionized (DI) water via ultrasonic processing for 5 h to form a uniform suspension, and then 5 mg H 2 PtCl 6 •6H 2 O and 20 mg NiCl 2 •6H 2 O were dissolved in the as-prepared suspension of CNTs through vigorous stirring for 0.5 h.Subsequently, 10 mL NaBH 4 solution (30 mg) was added into the solution dropwise under vigorous stirring.After stirring for another 2 h, PtNiB@CNTs were filtered and washed with DI water, followed by drying at 40 °C under a vacuum for 24 h.Pt@CNTs and NiB 12 @CNTs were prepared similarly to the PtNiB@CNTs but only with H 2 PtCl 6 •6H 2 O and NiCl 2 •6H 2 O, respectively.PtNi@CNTs were prepared similarly to PtNiB@CNTs, except that reductant NaBH 4 was replaced with H 2 .Specifically, 5% (V/V) H 2 /Ar was inflated into the precursor suspension at a flow rate of 50 mL/min under vigorous stirring for 2 h.

| Electrochemical measurements
Ten milligrams of PtNiB@CNT catalysts (PtNi@CNTs, Pt@CNTs or NiB 12 @CNTs) that were obtained with polyvinylidene fluoride at a mass ratio of 9:1 were ground with two to three drops of N-methylpyrrolidone for 30 min to prepare the slurry.Then, the slurry was brush coated on carbon paper with a diameter of 16 mm and dried at 120 °C under vacuum for 12 h to fabricate cathodes with a catalyst loading of ~0.15 mg.The Li-O 2 battery was assembled with the as-prepared cathode, a glass fiber separator (Whatman GF/D), and 1.0 mol/L lithium bis(trifluoromethanesulfonyl)imide in tetraethylene glycol dimethyl ether as an electrolyte and a Li foil anode in an Ar-filled glove box (H 2 O < 0.01 ppm, O 2 < 0.01 ppm).The Li-O 2 batteries were tested as CR2032 button cell batteries with holes on the cathode side.Then, the cells were measured in 1 atm O 2 atmosphere (H 2 O < 0.1 ppm) with a CT3001A-type battery testing system (LAND Electronics Co. Ltd.).The current density and specific capacity were calculated based on the mass of cathode catalysts.Cyclic voltammetry (CV) curves were conducted on a CHI760E electrochemical workstation at a scan rate of 0.1 mV/s between 2.0 and 4.5 V.The electrochemical impedance spectroscopy (EIS) tests were also conducted on CHI760E in a frequency range of 10 −2 −10 5 Hz.

| Theoretical simulation
The first-principles calculations were performed using the projector-augmented wave method implanted in the Vienna Ab-initio Simulation Package.The Perdew-Burke-Ernzerhof functional with the generalized gradient approximation exchange correlation was used to describe the exchange-correlation interactions.The wave functions were expanded by using the plane-wave energy cutoff of 500 eV.Brillouin-zone integrations were approximated by using special k-point sampling of Monkhorst-Pack scheme with a k-point mesh of 2 × 2 × 1.The supercell lattice and atomic coordinates were fully relaxed until the total energy per cell and the force on each atom are less than 1.0 × 10 −6 eV and 0.01 eV/Å, respectively.For all surfaces, a 4 × 4 supercell was used and a vacuum region of 20 Å was set along the z-direction to eliminate the interaction between periodic images.The surfaces of PtNi and PtNiB (111) were modeled using four atomic layers, with the top two layers and adsorbate permitted to completely relax.

| RESULTS AND DISCUSSION
As schematically illustrated in Figure 1A, the PtNiB@CNT catalyst was prepared by reducing the Pt 4+ and Ni 2+ with NaBH 4 as a reducing agent and doping source to form PtNiB nanoalloy on CNTs.The formation mechanism of boride is depicted in Equation (1) as follows 40 : where M represents metal elements.
The morphology and microstructure of as-prepared PtNiB@CNTs are shown in Figure 1 and Supporting Information: Figure S1.SEM image of PtNiB@CNTs displays similar nanotube-like morphology to those of PtNi@CNTs, Pt@CNTs, and NiB 12 @CNTs (Supporting Information: Figure S1).TEM image of PtNiB@CNTs exhibits the ultrasmall nanoparticles loaded on the CNTs matrix (Figure 1B).As depicted in high-angle annular darkfield scanning transmission electron microscopy (HAADF-STEM) image, the relatively bright spots are uniformly distributed on the CNTs substrate at a mean diameter of 2.25 nm without obvious agglomeration (Figure 1C).The high-resolution TEM (HRTEM) image (Figure 1D) confirms the lattice spacing of 0.225 and 0.195 nm, corresponding to the (111) and (200) planes of PtNiB nanoalloy, respectively. 39EDX displays the homogeneous dispersion of Pt, Ni, and B in the nanoparticle with a molar ratio of 1:4:19 (Supporting Information: Table S1), verifying the successful B doping into PtNi nanoalloy (Figure 1F-I and Supporting Information: Figure S2B).
XRD patterns are used to analyze the crystal structures of as-prepared catalysts.Besides the same peaks at 25.8°a ttributed to the (002) plane of CNTs (JCPDS No. 41-1487; Figure 2A), both PtNiB@CNTs and PtNi@CNTs exhibit similar diffraction peaks to Pt@CNTs at about 39.8°, 46.3°, and 67.5°, which are indexed to (111), (200), and (220) planes of Pt (JCPDS No. 04-0802), respectively.Specifically, these three diffraction peaks of PtNi@CNTs and PtNiB@CNTs shift to a lower angle in sequence compared with those of Pt@CNTs (Figure 2B), which validates the formation of PtNi alloy in PtNi@CNTs and PtNiB alloy in PtNiB@CNTs, respectively. 21,26In addition, the diffraction peaks located at 34.4°and 59.9°are associated with NiB 12 (JCPDS No. 24-0788).Therefore, the XRD patterns together with TEM and EDS results clearly confirm the formation of the well-dispersed PtNiB nanoalloy on CNTs.
Figure 2C shows the XPS of PtNiB@CNTs, proving the existence of Pt, Ni, B, C, and O elements.Highresolution XPS of Pt 4f peaks is deconvolved into four peaks including Pt 0 and Pt 2+ (Figure 2D).Obviously, the binding energy of Pt 4f peak in PtNiB@CNTs slightly upshifts toward higher binding energy, compared to those of PtNi@CNTs and Pt@CNTs.Moreover, high-resolution XPS of Ni 2p peaks is also deconvolved into four new peaks, two of which are satellite peaks.On the contrary, the binding energy of Ni 2p peak in PtNiB@CNTs negatively shifts toward a lower energy, relative to those in PtNi@CNTs and NiB 12 @CNTs (Figure 2E).Apparently, the observed opposite binding energy shifting of Pt 4f and Ni 2p peaks demonstrates that electrons are partially transferred from Pt to Ni in PtNiB nanoalloy. 39,41 demonstrate the superiority of PtNiB@CNTs as ORR/OER catalysts, the electrocatalytic performance of Li-O 2 battery based on PtNiB@CNTs as cathode was examined within a voltage window of 2.0-4.5 V at a scan rate of 0.1 mV/s (Figure 3A).As shown in the CV curves (Supporting Information: Figure S3A), there is an obvious reduction peak at ~2.42 V and a well-marked oxidation peak at ~3.22 V in the O 2 atmosphere.In contrast, the CV curve displays barely any reduction or oxidation peaks in the Ar atmosphere, which clearly demonstrates the high ORR and OER activities of PtNiB@CNT catalyst.Meanwhile, the representative voltage profiles of PtNiB@CNT, PtNi@CNT, Pt@CNT, and NiB 12 @CNT cathodes at a current density of 1000 mA/g and a cutoff capacity of 1000 mA•h/g are given to compare the overpotentials at a mid-capacity of 500 mA•h/g (Figure 3B).Obviously, the overpotential of PtNiB@CNT cathode is only 0.48 V, which is much lower than those of PtNi@CNTs (1.28 V), Pt@CNTs (1.36 V), and NiB 12 @CNTs (1.42 V).The full discharge/charge capacities of these cathodes were tested at a current density of 100 mA/g (Figure 3C), showing that the values of discharge specific capacity follow the order of PtNiB@CNTs (20510 mA•h/g) > PtNi@CNTs (17833 mA•h/g) > Pt@CNTs (17692 mA•h/g) > NiB 12 @CNTs (10201 mA•h/g).Besides, F I G U R E 2 Phase and electron structure of PtNiB@CNTs.(A) XRD patterns and (B) selected magnification from (A) of PtNiB@CNTs, PtNi@CNTs, Pt@CNTs, and NiB 12 @CNTs.(C) XPS spectrum of PtNiB@CNTs.(D) High-resolution Pt 4f XPS spectra of PtNiB@CNTs, PtNi@CNTs, and Pt@CNTs.(E) High-resolution Ni 2p XPS spectra of PtNiB@CNTs, PtNi@CNTs, and NiB 12 @CNTs.CNT, carbon nanotube; XRD, X-ray diffraction; XPS, X-ray photoelectron spectroscopy.
the PtNiB@CNT cathode also exhibits the highest coulombic efficiency of 93.34% compared to the PtNi@CNTs (70.40%),Pt@CNTs (72.11%), and NiB 12 @CNTs (58.87%), thus demonstrating superior reversibility of Li-O 2 battery.The EIS curves show that the Nyquist semicircle of PtNiB@CNTs is smaller than other counterpart catalysts (Figure 3D), which indicates lower charge-transfer resistance.Furthermore, the overpotentials slightly increase with the current density increasing from 200 to 500 and 1000 mA/g (Figure 3E), thus suggesting excellent rate performance.
More importantly, the cycling stability of these as-fabricated cathodes at 200, 500, and 1000 mA/g and a cutoff capacity of 1000 mA•h/g were performed (Figure 3F,G and Supporting Information: Figure S3B,C).It is disclosed that the PtNiB@CNT cathode steadily runs up to 210 cycles even at a large current density of 1000 mA/g.While the PtNi@CNT, Pt@CNT, and NiB 12 @CNT cathodes gradually fade away after 123, 84, and 79 cycles, respectively.It is indicated that the PtNi nanoalloy via B doping undoubtedly contributes to lowering the overpotentials and enhancing cycling stability.Additionally, the Li-O 2 batteries assembled with PtNiB@CNT cathode can also achieve stable cycling over 60 cycles (600 h) and 118 cycles (472 h) at smaller current densities of 200 and 500 mA/g with a fixed capacity of 1000 mA•h/g, respectively (Supporting Information: Figure S3B,C).The initial high charge overpotentials are possibly attributed to the unstable surfaces of the PtNiB@CNT catalysts, which would gradually stabilize after several cycles. 23,31,4231][43][44][45][46][47][48][49] To further elucidate the electrocatalytic mechanism of PtNiB@CNTs for Li-O 2 battery, the ex situ SEM, EIS, ex situ XRD patterns, and chromogenic reaction with Ce 4+ of the cathode at different stages were carried out.As illustrated in Figure 4A-C, the obvious film-like products are formed on the PtNiB@CNT cathode after full discharge (Figure 4B), which is different from the smooth surface of the cathode in the initial stage (Figure 4A).It is worth noting that the film-like discharge products disappear after fully recharging to 4.5 V (Figure 4C).It is well known that the membraneous discharge products are more easily decomposed and thus the corresponding Li-O 2 battery can run more reversibly due to the larger contacted area with the cathode. 50Meanwhile, EIS curves of PtNiB@CNT cathode at various stages are presented in Figure 4D.It is clearly seen that the charge-transfer resistance (R ct ) of PtNiB@CNTs increases from 37.8 Ω at the initial stage to 42.4 Ω after the first discharge because the insulating discharge products cover the surface of the cathode. 44As expected, the R ct almost returns to the original value (38.5 Ω) after the first recharging, thus suggesting outstanding reversibility for the Li-O 2 battery.Significantly, the crystal structure of discharge products was analyzed by ex situ XRD patterns (Figure 4E).It is disclosed that three additional diffraction peaks appear at 32.9°, 35.0°, and 58.8°after deeply discharging to 2.0 V at 200 mA/g, which are indexed to (100), (101), and (110) crystal planes of Li 2 O 2 (JCPDS No. 09-0355).Subsequently, these three peaks of Li 2 O 2 disappear without the formation of other phases after recharging to 4.5 V at 200 mA/g, thus indicating excellent reversibility.
Further, the chromogenic reaction is widely applied in qualitative/quantitative analyses of Li 2 O 2 products. 51,52pecifically, Ce 4+ in H 2 SO 4 aqueous solution is usually F I G U R E 4 Analysis of discharge/charge mechanism of PtNiB@CNT cathodes for Li-O 2 batteries.(A-C) SEM images of PtNiB@CNTbased cathode (A) at the initial stage, (B) after full discharge to 2.0 V, and (C) after full recharge to 4.5 V. (D) Nyquist plots and (E) XRD patterns of PtNiB@CNT cathode at different stages.(F) Photographs of PtNiB@CNT cathode at different stages soaked in 0.5 mmol/L Ce 4+ in 0.5 mol/L H 2 SO 4 (i: pristine Ce 4+ aqueous; ii: initial stage; iii: after first discharge; iv: after first recharge; v: after full discharge to 2.0 V).CNT, carbon nanotube; XPS, X-ray photoelectron spectroscopy; XRD, X-ray diffraction.used as a calibration agent for peroxide, described in Equation ( 2) as follows 53 : Because the light yellow Ce 4+ can become colorless Ce 3+ after reacting with peroxide, Ce 4+ in an H 2 SO 4 aqueous solution is also used to estimate whether Li 2 O 2 forms or does not form after discharging.As presented in Figure 4F, the colors of the blank contrast sample and the Ce 4+ aqueous solution soaked with the pristine electrode are the same light yellow, while the Ce 4+ aqueous solution with PtNiB@CNT electrodes after the first discharge, first recharge, and full discharge became colorless, unchanged, and colorless, respectively, thus indicating reversible formation/decomposition during discharge/charge processes.Taken together, it is clearly demonstrated that the PtNiB nanoalloy dramatically improves the reversible formation/decomposition of film-like Li 2 O 2 without other side products during the discharge/charge processes in the Li-O 2 batteries.
To deeply understand the role of catalyst in the discharge/charge processes and the mechanism of ORR/ OER, DFT calculations were conducted to uncover the electronic structure of the catalyst and free energy diagrams during discharge/charge processes.As shown in the electron localization function plots (Figure 5A-D), the localized electrons near the Pt and Ni atoms lightly offset to B atoms, which is beneficial for lowering the activation energy barrier of oxygen species on PtNiB nanoalloy (Figure 5A,B). 39Moreover, it can be clearly seen from Figure 5C,D that the localization of electrons around the adsorbed LiO 2 on PtNiB nanoalloy is stronger than that on PtNi nanoalloy, suggesting more easy charge transfer between catalyst and oxygen species. 44Further, the specific processes of ORR and OER are investigated by calculating the Gibbs free energy of every assumed step on PtNi (Figure 5E) and PtNiB nanoalloy (Figure 5F).Typically, the discharge procedure (ORR) proceeds through three possible elemental pathways shown in Equations (3-5) as follows: (4) where the * represents different adsorbed species on the surface of catalysts. 44Certainly, the opposite direction of these three equations means the charge procedure (OER).As shown in Figure 5, the values in vertical coordinate at U = 0 V, U DC , U EQ , and U C represent free energy diagrams of ORR/OER at zero potential, maximum discharge potential, equilibrium potential, and minimum charge potential, respectively.The vital descriptor in theory for evaluating the catalytic activity of different catalysts, overpotential, is calculated as 54 It is clearly underscored that the overpotential of Li-O 2 battery in theory with PtNiB@CNTs as a cathode catalyst (1.29 V) is much smaller than that with PtNi@CNTs as a cathode catalyst (3.72 V), which is consistent with the experimental results of excellent ORR/OER activity of PtNiB@CNTs.Therefore, it is concluded that the B doping into PtNi alloy plays a key role in tailoring the electron structure of the catalyst and optimizing the adsorption of intermediates for enhanced Li-O 2 battery performances.

| CONCLUSIONS
In summary, we demonstrate a general surface engineering strategy to fabricate PtNiB nanoalloy supported on CNTs electrocatalyst for high-performance Li-O 2 batteries, delivering a low overpotential of 0.48 V at a high current density of 1000 mA/g, large specific capacity of 20510 mA•h/g, excellent rate performance, and ultra-long cycling life up to 210 cycles at 1000 mA/g under the fixed capacity of 1000 mA•h/g.Moreover, the analysis with ex situ SEM, EIS, ex situ XRD, and chromogenic reaction on PtNiB@CNT-based cathode elucidates the almost reversible formation and decomposition of film-like Li 2 O 2 on the cathode during discharge and charge processes.Critically, the results from the perspective of experiments and DFT calculations reveal that the excellent performance of PtNiB@CNT catalyst for Li-O 2 battery is attributed to the surface engineering of PtNi nanoalloy via doping B, subsequently tailoring the electron structure of nanoalloy and optimizing the adsorption of oxygen species.This study will provide insights into regulating the nanoalloy by doping nonmetallic elements for enhancing the performance of metal-air batteries.

F
I G U R E 3 Electrochemical performance of PtNiB@CNT electrocatalyst for Li-O 2 batteries.(A) Schematic of PtNiB@CNT-based Li-O 2 battery.(B) The representative discharge-charge curves at a current density of 1000 mA/g and a fixed capacity of 1000 mA•h/g.(C) The full discharge/charge curves between 2.0 and 4.5 V were obtained at 100 mA/g.(D) Nyquist plots at 0.1 mV/s from 4.5 to 2.0 V. (E) Rate performance at a current density from 200 to 500 to 1000 mA/g under a fixed capacity of 1000 mA•h/g and (F) discharge-charge profiles at different cycles of PtNiB@CNT cathode.(G) Discharge/charge terminal voltages at 1000 mA/g and fixed capacity of 1000 mA•h/g of PtNiB@ CNT, PtNi@CNT, Pt@CNT, and NiB 12 @CNT electrocatalysts.(H) The comparison of discharge/charge overpotentials and initial full discharge capacity of PtNiB@CNT-based Li-O 2 batteries with previously reported works.CNT, carbon nanotube.

F
I G U R E 5 Simulation calculation of electronic structure and free-energy landscape of PtNi and PtNiB.(A-D) ELF plots of (A) PtNi, (B) PtNiB, and adsorbed LiO 2 on (C) PtNi and (D) PtNiB.(E, F) Calculated free energy diagrams for the ORR/OER reactions of Li-O 2 batteries on (E) PtNi@CNTs and (F) PtNiB@CNTs electrocatalysts.CNT, carbon nanotube; ELF, electron localization function; OER, oxygen evolution reaction; ORR, oxygen reduction reaction.