Atomically ordered non-precious Co3Ta intermetallic nanoparticles as high-performance catalysts for hydrazine electrooxidation

Nano-ordered intermetallic compounds have generated great interest in fuel cell applications. However, the synthesis of non-preciousearly transition metal intermetallic nanoparticles remains a formidable challenge owing to the extremely oxyphilic nature and very negative reduction potentials. Here, we have successfully synthesized non-precious Co3Ta intermetallic nanoparticles, with uniform size of 5 nm. Atomic structural characterizations and X-ray absorption fine structure measurements confirm the atomically ordered intermetallic structure. As electrocatalysts for the hydrazine oxidation reaction, Co3Ta nanoparticles exhibit an onset potential of −0.086 V (vs. reversible hydrogen electrode) and two times higher specific activity relative to commercial Pt/C (+0.06 V), demonstrating the top-level performance among reported electrocatalysts. The Co-Ta bridge sites are identified as the location of the most active sites thanks to density functional theory calculations. The activation energy of the hydrogen dissociation step decreases significantly upon N2H4 adsorption on the Co-Ta bridge active sites, contributing to the significantly enhanced activity.


Results
Synthesis and characterization of Co 3 Ta intermetallic NPs. To prepare early transition metal intermetallic compounds, extremely strong reductants and high temperatures are usually adopted, which leads to significant agglomeration and unwanted particle growth 28,[30][31][32][33] . With this in mind, we instead added surface treated carbon supports to the reaction solution before the co-reduction of metal salts (details can be found in the Methods section). Supplementary Fig. 1 shows the powder X-ray diffraction patterns of the precursors and products. After the precursors were treated at 400°C for 3 h, the diffraction peaks were identified but appeared broad, allowing for assignment of Co 3 Ta as a face-centered cubic (fcc) structure in the Pm-3m space group (JCPDS,(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28). Broad diffraction peaks are indicative of small particles size. A representative HAADF-STEM overview image of the Co 3 Ta NPs is shown in Fig. 1a. The Co 3 Ta NPs are well dispersed on the carbon supports, with an average diameter of 5 nm, which is much smaller than that of other reported Ptbased early transition metal intermetallic compounds (Supplementary Table 1). The high-resolution transmission electron microscopy (HRTEM) image of Co 3 Ta NPs is shown in Fig. 1b, with lattice spacing is 2.10 Å corresponding to the (111) plane of intermetallic Co 3 Ta. The same lattice spacing measured for other NPs ( Supplementary Fig. 2) suggests that Co 3 Ta NPs have (111) basal planes. Energy-dispersive spectroscopy (EDS) mapping and line scanning profiles of a single Co 3 Ta NP are presented in Fig. 1c-e and Supplementary Fig. 3, respectively, revealing that both Co and Ta are homogeneously dispersed. Combined with the result of inductively coupled plasma (ICP) analysis, the molar ratio of these two elements is 75.95:24.05 (Fig. 1f), that is 3:1 ratio of Co 3 Ta.
High-magnification HAADF-STEM imaging was employed to reveal the ordered intermetallic crystal structure of Co 3 Ta. From the ordered arrangement of atoms, we can identify the ordered intermetallic structure of Co 3 Ta. Figure 2a shows a representative atomic resolution image of Co 3 Ta along the [111] zone axis, with lattice spacing of 2.56 Å corresponding to the (110) plane. Because of the 'Z-contrast' of Ta and Co, the Ta columns display a higher intensity than that of the Co columns in HAADF-STEM images. It can be seen that each Ta atom is surrounded by six Co atoms in a periodic hexagonal array. The corresponding fast Fourier transformation (FFT) pattern ( Fig. 2b) is matched with the simulated image in Fig. 2c, further demonstrating the crystallinity and ordered intermetallic phase of Co 3 Ta. This ordered intermetallic structure also presents along the [211] zone axes. In Fig. 2d, a HAADF-STEM image of Co 3 Ta NP is viewed along the [211] zone axis with lattice spacings of 2.09 and 2.56 Å, which are assigned to (111) and (110) lattice fringes of the Co 3 Ta intermetallic structure. Due to the overlap of Ta and Co atoms in some positions along the [211] zone axis, the 'Z-contrast' of Ta and Co is small. The FFT pattern is shown in Fig. 2e. To further confirm the ordered structure, we have simulated the diffraction patterns of an ideal Co 3 Ta crystal along its [211] zone axis (Fig. 2f), which matches the experimental result. This is conclusive evidence that the atomically ordered Co 3 Ta intermetallic nanostructures have been successfully synthesized.
To further determine the entire ordered intermetallic structure of the Co 3 Ta NPs, we carried out X-ray absorption fine structure (XAFS) measurements for the Co 3 Ta NPs and other possible phases. Comparisons between the calculated and experimental absorption patterns of Co and Ta in the ordered intermetallic structure of Co 3 Ta and other potential samples (i.e., CoO, Co 3 O 4 , Co foil, Ta 2 O 5 , and Ta powder) are shown in Fig. 2g, h. The Computational Methods show how to obtain the theoretical spectrum. The experimental spectrum of the sample has been successfully reproduced, both for the peak positions and intensities, by using the theoretical Co 3 Ta configuration. Using the "fingerprint" of the Co K-edge and Ta L 3 -edge XAFS, we can easily determine the dominant existence of the ordered intermetallic Co 3 Ta phase structure. As to the small difference observed in the A peak site (Fig. 2g), this may be attributed to oxidation of the Co 3 Ta surface upon exposure of sample in air for relatively long period during the measurements. X-ray photoelectron spectroscopy (XPS) ( Supplementary Fig. 4) also confirms the presence of a slight oxidation, which is in agreement with previous metal nanocrystal studies 3,[34][35][36] . The XAFS results strongly suggest that the as-prepared product is an ordered intermetallic Co 3 Ta phase structure as a whole.
Electrochemical performance of Co 3 Ta intermetallic NPs. Direct hydrazine fuel cell (DHFC), as a clean energy for future transportation vehicles and portable devices, has drawn increasing attention in recent years due to its remarkably fascinating highlights [36][37][38][39][40][41] . For example, not only do DHFC has a higher energy density and theoretical voltage than those of hydrogen fuel cell and most direct liquid fuel cells, but also a much safer handling system than that of gas state of hydrogen fuel cell, and more environmentally friendly CO 2 -free products than other direct liquid fuel cells. However, in previous studies, the onset potentials (E on ) of the reported electrocatalysts for HzOR remained very high, indicating that these electrocatalysts require much higher potentials (overpotentials) to start the HzOR and realize a certain degree of hydrazine oxidation. As a result, the practical energy density is reduced. Here, we found that the ordered intermetallic Co 3 Ta has an ultrahigh intrinsic activity toward hydrazine electrooxidation in an alkaline medium, notably an ultralow E on of −0.086 V vs RHE.
To characterize the intrinsic activity of the Co 3 Ta catalyst, the kinetic currents shown in this work are all normalized to the electrochemically active surface area (ECSA) for a quantitative and more convincing comparison. According to the previous reports [42][43][44] , the ECSA was determined by measurement of the electrochemical double-layer capacitance (EDLC) (Supplementary Fig. 5). The electrocatalytic HzOR activities of Co 3 Ta/C NPs measured in different concentrations of hydrazine with scan rates of 5 mV s −1 are displayed in Fig. 3a. From the linear sweep voltammetry (LSV) curve for the solution without hydrazine, we can see that no obvious anodic current appears in the potential window. However, when placed in a 0.1 M hydrazine solution, a significantly rising anodic current appears. Furthermore, the current density increases with increasing hydrazine concentration, indicating that the Co 3 Ta/C NPs are highly efficient for the HzOR.
In order to further evaluate the intrinsic superior catalytic performance of Co 3 Ta/C NPs, several samples (including the precursors, Co/C, XC-72, commercial Ir/C, and commercial Pt/C) were employed as control samples in a solution containing 0.2 M hydrazine and 3 M KOH at a scan rate of 5 mV s −1 , as shown in Fig. 3b. The onset potential is usually regarded as a very important evaluation criterion to determine the catalytic performance of catalysts. It can be seen that the E on of Co 3 Ta/ C NPs is −0.086 V (vs RHE; i.e., −1.175 V vs SCE), which is 16, 35, and 141 mV lower than that of Co/C, commercial Pt/C, and commercial Ir/C, respectively (Supplementary Fig. 6). Compared with that of other catalysts reported in the recent literatures (Supplementary Table 2), the onset potential of the ordered intermetallic Co 3 Ta/C catalyst is the lowest, indicating that Co 3 Ta/C NPs have an ultrahigh intrinsic activity toward HzOR. In addition, the Tafel slope of Co 3 Ta/C NPs (56.9 mV dec −1 ) is lower than those of Co/C NPs (60.6 mV dec −1 ) and commercial Pt/C (73.7 mV dec −1 ) (Fig. 3c), demonstrating that Co 3 Ta/C NPs increase the current density more quickly than other samples during the HzOR. Notably, at a potential of +0.06 V (vs RHE) shown in Fig. 3d, Co 3 Ta/C NPs produced a praiseworthy current (25.2 mA cm −2 ), which is 1.67 and 2.02 times higher than that of Co/C NPs (15.1 mA cm −2 ) and commercial Pt/C (12.5 mA cm −2 ), respectively. Furthermore, compared with our previous studies (e.g., ultrathin nickel nanosheet arrays (68 mA mg −1 ) 45 and ultrathin nickel-cobalt alloy nanosheet arrays (92 mA mg −1 ) 46 ), the ordered intermetallic Co 3 Ta shows a considerably high mass activity of 534 mA mg −1 at a potential of +0.05 V (vs RHE) in 0.5 M hydrazine solution. The above results demonstrate that Co 3 Ta/C NPs show ultrahigh electrocatalytic activity for the HzOR. Stability is another important parameter to evaluate the practical performance of a catalyst. As shown in Supplementary  Fig. 7, long-term stability of the Co 3 Ta/C NPs for 12,000 s under a constant potential (+0.1 V vs RHE) in 0.2 M hydrazine solution was investigated. After this harsh 12,000 s test, Co 3 Ta/C NPs show a loss of only 12.8% of the initial current density (Fig. 3e). In contrast, Co/C NPs and commercial Pt/C exhibit a serious current density loss of 40.8% and 38.4%, respectively. Moreover, after 12,000 s, the E on of the Co 3 Ta/C catalyst shows a positive shift of only 10 mV ( Supplementary Fig. 8), electrochemical impedance spectroscopy (EIS) displays negligible changes (Supplementary Fig. 9), and the structure remains intact with no particle agglomerations ( Supplementary Fig. 10). The above results demonstrate that Co 3 Ta/C NPs are remarkably stable. The excellent stability and ultrahigh electrocatalytic activity of Co 3 Ta/C NPs may be ascribed to its atomically ordered structure and electronic effect of Co 3 Ta.
X-ray absorption near-edge structure (XANES) analysis. Figure 4a, b show the K-edge and L 3 -edge X-ray absorption nearedge of Co and Ta elements in the ordered intermetallic structure of Co 3 Ta and other potential phases (i.e., bulk Co foil, CoO, Co 3 O 4 , bulk Ta powder, and Ta 2 O 5 ). It can be seen that the Co Kedge and Ta L 3 -edge X-ray absorption edge energy is arranged in  49,50 , and the Ta absorption edge energy in Co 3 Ta also exhibits a small positive shift with respect to that of bulk metallic Ta. Both are related with the size effect and the tuning of electronic structure of the Ta atom by Co, suggesting that the 5d hole in Ta atom increases 51,52 . An increase in the 5d hole will promote donation of electrons from the reactant orbital to the Ta 5d orbital. Thus, the tuning of electronic structure of Co and Ta likely facilitates the adsorption of N 2 H 4 through Density functional theory (DFT) calculations. Identifying the location of the active sites in an intermetallic is conducive to the design of optimal catalysts. Generally, there are several forms of hydrazine adsorption on Co 3 Ta catalyst surface 53 . As shown in Supplementary Fig. 11, the Ta-Co-cis conformation has the lowest adsorption energy (Supplementary Table 4), indicating the most stable adsorption configuration. It is preferable for the N-N molecular axis projection onto the surface be parallel with the Ta-Co bridge. That means the Co-Ta bridge sites can be identified as the location of the most active sites of HzOR in the ordered Co 3 Ta intermetallic.
DFT calculations were carried out to further reveal the origin of the superior intrinsic activity of ordered intermetallic Co 3 Ta toward hydrazine oxidation. The surface structure plays an important role about the computations of adsorption of hydrazine on metal surface 54 . Based on the HRTEM, XRD results, and the analysis of surface energies (Supplementary  Table 5), the dehydrogenation process of hydrazine was analyzed over the Co 3 Ta (111) surface, with the dehydrogenation process of hydrazine over the (111) surface of Pt used as a control, as shown in Supplementary Fig. 12. Because both the Ta-Co-cis and Pt-anti conformations have the lowest adsorption energies (Supplementary Tables 4, 6), these conformations were selected as the adsorption configurations for DFT calculations. The density of states (DOS) of the Co 3 Ta (111) surface indicated that the electron structure of the surface is similar to bulk metal, and the surface DOS distribution crosses over the Fermi level, which benefits electron transport during the HzOR process (Supplementary Figs. 13 and 14). Because of the magnetic properties of Co, the spin down parts are larger than the spin up parts near the Fermi energy level. Figure 4c, e illustrate the stepwise N 2 H 4 molecular dehydrogenation process on the Pt (111) and the Co 3 Ta (111) surface, respectively, which consist of four intramolecular dehydrogenation steps (N 2 H 4 → N 2 H 3 → N 2 H 2 → N 2 H → N 2 ) 41,55-58 . Figure 4d shows the free energy profiles of the HzOR on the Pt (111) and Co 3 Ta (111) surfaces accordingly. The first dehydrogenation step (N 2 H 4 → N 2 H 3 ) is exothermic on the Co 3 Ta surface (0.65 eV), which is much higher than on the Pt surface (0.37 eV). This demonstrates that the first dehydrogenation step is more readily carried out on the surface of Co 3 Ta than that of Pt. Though the second dehydrogenation step (N 2 H 3 → N 2 H 2 ) is endothermic on both surfaces, the activation energy is much easier to overcome on the Co 3 Ta surface (0.03 eV) than on the Pt surface (0.31 eV). Overall, from the state of adsorbed N 2 H 4 to the state of adsorbed N 2 , the exothermic energy on the Co 3 Ta surface (1.83 eV) is much larger than that on the Pt surface (1.25 eV), indicating a more thermodynamically favorable catalytic process over the ordered intermetallic Co 3 Ta.
Based on the discussion above, ordered intermetallic Co 3 Ta exhibits ultrahigh electrocatalytic activity toward hydrazine electrooxidation, including an ultralow onset potential, low Tafel slope, and high current density. The origin of the ultrahigh hydrazine oxidation activity can be explained as follows. First, the electronic structure of the Ta atom is tuned by Co and there is a synergistic effect between Co and Ta in Co 3 Ta. Second, the activation energy of the hydrogen dissociation step decreases significantly during HzOR. Both will intrinsically contribute to the superior electrocatalytic activity of non-precious ordered nano-intermetallic Co 3 Ta toward hydrazine oxidation.  Fig. 4 X-ray absorption near-edge structure analysis and theoretical calculations. a The normalized X-ray absorption near-edge at the Co K-edge of the Co 3 Ta/C NPs, Co foil, CoO, and Co 3 O 4 . b The normalized X-ray absorption near-edge at the Ta L 3 -edge of the Co 3 Ta/C NPs, Ta powder, and Ta 2 O 5 . c A schematic illustration for the stepwise N 2 H 4 molecular dehydrogenation process on the Pt (111) surface. Navy blue, orange, and yellow balls represent Pt, N, and H atoms, respectively. d Free energy profiles of the HzOR on the Pt (111) and Co 3 Ta (111) surfaces. e A schematic illustration for the stepwise N 2 H 4 molecular dehydrogenation process on the Co 3 Ta (111) surface. Purple and cyan balls represent Co and Ta atoms, respectively

Discussion
In summary, we have demonstrated the excellent performance and stability of novel Co 3 Ta NPs as an intermetallic electrocatalyst that contains non-precious metals and only early transition metals. The use of air-free synthetic conditions and subsequent annealing leads to the formation of atomically ordered Co 3 Ta NPs with a uniform particle size of 5 nm. As an electrocatalyst for hydrazine oxidation reaction (HzOR), Co 3 Ta/C NPs exhibit high stability and a higher electrocatalytic performance than conventional electrocatalysts in terms of their low onset potentials (−0.086 V vs RHE) for fuel oxidation. Theoretical calculations reveal that the activation energy of hydrogen dissociation decreases significantly upon N 2 H 4 adsorption on the Co-Ta bridge active sites, which not only increases the number but also enhances the activity of the active sites, contributing to the considerably enhanced HzOR activity. The extraordinarily high performance of the supported ordered Co 3 Ta intermetallic nanocrystals provides a very promising alternative to the conventional Pt/C catalyst for the HzOR in direct liquid fuel cells.

Methods
Synthesis of Co 3 Ta intermetallic nanoparticles. Co 3 Ta intermetallic NPs were synthesized by co-reduction of tantalum and cobalt salts under anhydrous and anaerobic conditions, followed by annealing. In a typical procedure, 0.18 mmol CoCl 2 and 0.065 mmol TaCl 5 were dissolved in 35 mL of rigorously dried and degassed tetrahydrofuran (THF) and diglyme solution in an argon-filled glovebox (both O 2 and H 2 O concentrations were <0.1 ppm), and stirred to form a clear solution. Then 60 mg of treated XC-72 carbon powders were added to the clear solution, and the mixture was stirred continuously for 8 h to form a uniformly dispersed solution. Next, sodium triethylborohydride (NaEt 3 BH, 1 M in THF, Sigma-Aldrich) was injected into the mixture under vigorous stirring and left to stir overnight. Finally, the sample was separated from the above mixture via centrifugation without contacting air. The sample was then washed with rigorously dried and degassed THF and hexanes, and dried at 60°C for 8 h in a glovebox. The obtained product was quickly transferred to a tube furnace and treated under flowing H 2 /Ar at 300°C for 3 h. The sample was then washed with argon-saturated ultrapure water and dried at 80°C for 8 h in a vacuum oven. The precursor was again transferred to tube furnace and annealed at 400°C for 3 h. The obtained intermetallic Co 3 Ta NPs were finally stored in a glovebox under Ar for further characterization and electrochemical measurements. The synthesis of Co/C NPs followed a similar procedure to that of Co 3 Ta NPs, except the addition of TaCl 5 .
Material characterizations. The morphologies and composition of samples were characterized by high-resolution transmission electron microscopy (HRTEM; JEOL, JEM-2100F, 200 kV) equipped with an energy-dispersive X-ray spectrometry (EDS) instrument. Atomic structural characterization of the samples was measured using a spherical aberration corrected transmission electron microscope (TEM; Titan-G2, 300 kV). During the TEM measurements, electron exposures employed should be very low to minimize irradiation damage. X-ray diffraction (XRD) patterns were collected on an X-ray diffractometer (Bruker D8, Cu K α , λ = 1.5406 Å, 40 kV, and 40 mA) with a counting time of 8 s, recorded with 2θ ranging from 15°to 70°. X-ray photoelectron spectroscopy (XPS) measurements were performed using a Thermo VG Scientific ESCALAB 250 spectrometer (Al K α , 200 W). The XAFS (Ta L 3 -edge and Co K-edge) spectra were collected at beamline BL14W1 of the Shanghai Synchrotron Radiation Facility. It should be noted that the samples were kept under an argon atmosphere before all of the above characterizations to avoid oxidation.
Electrochemical measurements. All the electrochemical measurements of the HzOR were carried out in a standard three-electrode electrochemical cell at room temperature using a BioLogic SP 240 electrochemical workstation. The glassy carbon (GC) film, saturated calomel electrode (SCE), and Co 3 Ta/C-coated GC rotating disk electrode (geometric area 0.1257 cm 2 ) were used as the counter electrode, reference electrode, and working electrode, respectively. The working electrode was prepared as follows: 3 mg of Co 3 Ta/C sample was dispersed in 1 mL of isopropyl alcohol and 5% Nafion solution, then the mixture was transferred to an ultrasonic bath to form a homogeneous catalyst ink. A 15 μL aliquot of the obtained suspension was pipetted onto the GC electrode and dried naturally for electrochemical measurements. The working electrode fabrication procedures of (20 wt%) and Ir/C (10 wt%) followed the same procedure with that of Co 3 Ta/C sample. Both Pt/C and Ir/C loadings are 0.36 mg cm −2 . The HzOR tests were conducted in 3 M KOH solution containing 0.5 M hydrazine at a rotation rate of 2000 rpm. In this work, the potentials measured (SCE) were converted to reversible hydrogen electrode (RHE) using the conversion method: Computational methods. All spin unrestricted DFT calculations were performed by using the Vienna Ab-initio Simulation Package (VASP) 59,60 with Perdew-Burk-Ernzerhof (PBE) 61 functional. The planewave basis (kinetic energy cutoff values ECUT = 520 eV) and projector-augmented wave (PAW) pseudopotential 62,63 were employed. The atomic positions were fully optimized until the Hellmann-Feynman force was <0.05 eV/Å and total energy convergence criterion was set to 1 × 10 −4 eV. The two Pt (111) and Co 3 Ta (111) surfaces were employed in our calculation using the 441 super cell, which includes four atomic layers, with the atomic position of the two top layers optimized, while the two bottom layers was fixed. The adsorption energies (E ads ) which one N 2 H 4 molecule adsorbed on Co 3 Ta (111) surface with different layers were shown in Supplementary Table 7. The results indicated that four layers surface model in our calculation was workable. To avoid the interaction of neighboring images,~15 Å vacuum layer was set to the direction of the c axis. The Monkhorst-Pack 64 k-point grid 3 × 3 × 1 was set in all calculations. The calculation of the Gibbs free energy of the intermediates followed the Nørskov method 65 .
The calculation of the theoretical XAFs spectra of Co 3 Ta: the calculation of the theoretical spectra was performed by the software Artemis based on a standard model of Co 3 Ta with a Pm-3m space group (JCPDS, No. 15-0028). The core of the calculation was based on the framework of FEFF's multiple scattering path expansion, where the simulated spectra is the summation of one or more scattering paths computed by FEFF. To obtain the theoretical spectra, the ATOM module was ran with a cluster size of 5.5 Å and longest scattering path of 5.0 Å. The calculation was done in the R space within an R range of 1.05-3.88 Å during which the Fourier transformation was conducted. Finally, the theoretical XAFS spectra were acquired by inverse Fourier transformation from R space and compared with experimental results.
The Gibbs free energy ΔG is defined as follows: ΔG = ΔE + ΔZPE − TΔS + ΔG U + ΔG pH , where ΔE is the change of electronic energy obtained from DFT calculations, ΔZPE is the change of zero-point energy, and ΔS is the entropy difference (see the values in Supplementary Table 8). ΔG U is the free energy contributed by the electrode potential (ΔG U = −n e U (n e represents the number of electrons transferred in the corresponding elementary steps and U is the electrode potential, respectively)). ΔG pH is the free energy related to the H + concentration. The ΔG U and ΔG pH are set to zero in our calculation. The contribution of vibration of all adsorbed species were considered in our calculation.
The calculated Gibbs free energy change (ΔG) is related to the onset potential in experiment. The theoretical E onset value could be obtained by the formula E onset = Max[ΔG 1 , ΔG 2 , ΔG 3 , ΔG 4 ]/e. It should be noted that the theoretical E onset value obtained by the above formula are somewhat different from experimental value. This difference is caused by many factors involved in the real reaction conditions, such as solvent effect, coverage degree of catalytic species, experimental temperature, etc.

Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.