3.1. Morphological and structural characterization
The Co2P/Co3Fe7@N-C heterostructure was synthesized by a one-step sintering method, which adjusted the mass ratios of cobalt source to iron source. As shown in Fig. 2a and Fig. S1, the X-ray diffraction (XRD) results of Co2P/Co3Fe7@N-C show that the diffraction peaks are distinguished as Co2P (PDF # 32–0306) and Co3Fe7 (PDF # 48-1817), which demonstrates that the Co2P/Co3Fe7@N-C nano-heterostructure was successfully synthesized using the one-step sintering method [28, 29] reveal that the morphology of Co2P/Co3Fe7@N-C assembled by the N-doped carbon nanotubes encapsulated Co2P/Co3Fe7 alloy nanoparticles. Furthermore, the high-resolution TEM (HRTEM) shows a typical interface with typical lattice spacing of 0.20 nm and 0.22 nm, which corresponds to the (121) lattice plane of Co3Fe7 and the (110) of lattice plane of Co2P, respectively (Fig. 2c) [30]. Notably, the distinct nanointerfaces and subtle lattice distortions between Co3Fe7 and Co2P species may have promoted the synergistic effect of Co2P/Co3Fe7@N-C, which provides more active sites, enhances transferred efficiency of interfacial charge, and decreases the reaction-free energy of the adsorption intermediates, thus enhancing the catalytic performance. We further verified these results by X-ray photoelectron spectroscopy (XPS) analysis and first principles calculations with DFT. In addition, we observe rough N-doped carbon layers with diameters of 5.75 nm by the HRTEM, which further ensures that the catalysts are uncorroded by alkaline environments [31, 32]. As shown in Fig. 2d, the high-angular dark-field scanning transmission electron microscope (HAADF-STEM) image and the corresponding energy dispersive spectroscopy (EDS) elemental mappings indicates that the elements of Co, Fe, and P are uniformly distributed in the N-doped carbon nanotubes. Moreover, compared with the intensity of Co, P, and Fe counts, the EDS line-scan elementals (Fig. 2e) further indicate a distinct interfacial structure between Co2P and Co3Fe7, which is consistent with the HRTEM and XRD results.
We further explored the surface electronic state and electronic interactions of Co2P/Co3Fe7@N-C using XPS. Figure 3a displays the full XPS spectra of Co2P/Co3Fe7@N-C, in which Fe, Co, O, N, and P peaks are detected in Co2P/Co3Fe7@N-C, which corresponds to the EDS element mappings and line-scan elementals. The high-resolution Fe 2p XPS spectra of Co2P/Co3Fe7@N-C is shown in Fig. 3b, the peaks at 707.3/718.4, 709.9/720.6, 712.3/725.2, and 714.3/727.1 eV can be recognized properly as Fe0, Fe2+, Fe3+, and satellite species, respectively [33]. As shown in Fig. 3c, the Co 2p XPS spectrum of Co2P/Co3Fe7@N-C can be divided four species, whereas the two peaks locates at 778.1 and 793.6 eV can be recognized as Co0, the peaks locates at 780.5 and 796.5 eV corresponded to Co3+, the peaks of 782.5 and 798.9 eV are assigned to Co2+ because of the form of Co-P and Co-Nx, and another two peaks locates at 785.7 and 803.1 eV corresponded to satellite peaks [34, 35]. Significantly, the positive shift of binding energy for Co2P/Co3Fe7@N-C compares with the Co 2p of Co/Co2P@N-C, indicating the strong electronic interaction between the interface of Co2P and Co3Fe7 [36, 37]. The strong electronic interaction of Co2P/Co3Fe7@N-C may have stimulated the charge transfer at the interface and the adsorption or desorption of intermediate products during oxygen electrolysis activity [26, 27]. These results are beneficial for bifunctional OER/ORR activity, which are further verified by the DFT results. Additionally, the presence of Fe0 and Co0 further indicates the formation of Co3Fe7 alloy, which was in agreement with the XRD and TEM results. Furthermore, the N 1s XPS spectrum can be divided into three subpeaks locate at 398.6 eV of pyridinic-N, 399.9 eV of pyrrolic-N, and 401.1 eV of graphitic-N (Fig. 3d and Table S1) [29, 38, 39]. The P 2p XPS spectrums of Co2P/Co3Fe7@N-C and Co/Co2P@N-C are shown in Fig. 3e. One peak locates at 129.6 eV corresponds to the Co-P bond, which suggests the formation of Co2P. The other peak locates at 133.2 eV corresponds to the P-O bond, which is derived from atmosphere O adsorption. The higher ratio of pyridinic-N and graphitic-N in Co2P/Co3Fe7@N-C compare with Co/Co2P@N-C may have accelerated oxygen adsorption and cut down the ORR overpotential because of their excellent electron-accepting ability [40, 41]. Moreover, compare with the Raman spectra of Co/Co2P@N-C (ID/IG: 1.97), Co2P/Co3Fe7@N-C shows a higher intensity ratio ID/IG of 2.47, demonstrates its comparatively low graphitization degree, which can provide more active sites for electrocatalytic performance, as shown in Fig. 3f [42, 43]. In addition, as shown in Fig. S2, Co2P/Co3Fe7@N-C possesses a high Brunauer-Emmett-Teller (BET) specific surface area of 146.9 m2 g-1 and pore volume of 0.206 cm3 g-1, respectively.
3.2. Electrocatalytic ORR and OER performance of Co2P/Co3Fe7@N-C
According to the above results, we evaluated the bifunctional oxygen reaction of Co2P/Co3Fe7@N-C and Co/Co2P@N-C in a standard three-electrode configuration with a scan rate of 2 mV s− 1. According to Fig. 4a and Fig. S3, the Co2P/Co3Fe7@N-C electrocatalyst shows significant ORR activity with a high half-wave potential (E1/2 = 0.86 eV), limited current density (jL = 4.2 mA cm− 2), which is superior to other amounts of Co:Fe for Co2P/Co3Fe7@N-C catalysts. The Co2P/Co3Fe7@N-C catalyst can compete with the Pt/C catalyst (E1/2 = 0.84 eV; jL = 4.4 mA cm− 2), which demonstrates its superior ORR performance. The corresponding Tafel slope (Fig. 4b) of Co2P/Co3Fe7@N-C is 98 mV dec− 1, which is smaller than the Co/Co2P@N-C (125 mV dec− 1), further revealed the excellent ORR kinetics. To further explore kinetic activity, we calculate the number of electron transfers (n) for the electrocatalysts according to the Koutecky-Levich (K-L) equation, as shown in Fig. S4, S5, S6, S7, and S8 [44]. The calculated average number of n values for Co2P/Co3Fe7@N-C are about 3.8 during ORR activity, suggesting a four-electron pathway in ORR activity. Furthermore, we observe that the average n values are also 3.8 and the yield of HO is below 7 % (Fig. 4c) for Co2PCo3Fe7@N-C in the potential region of 0.3–0.7 V, which is close to that of the Pt/C catalyst (n: 3.90; HO: 6 %), which indicates that the electrocatalst could efficiently motivate the generation of oxygen at a relatively low overpotential [45, 46].
To further prove the bifunctional electrocatalytic performance, we tested the OER activities of Co2P/Co3Fe7@N-C and the contrasted catalysts in 1 M KOH electrolyte with a scan rate of 2 mV s− 1. As shown in Fig. 4d, Fig. 4e and Fig. S9, Co2P/Co3Fe7@N-C exhibits excellent OER activity with a smaller overpotential of 1.47 V and a Tafel slope of 144 mV dec− 1, comparing with Co/Co2P@N-C (1.58 V; 169 mV dec− 1), Ir/C (1.53 V; 153 mV dec− 1), and other amounts of Co:Fe for Co2P/Co3Fe7@N-C catalysts. To further understand the OER reaction kinetics and effective active sites of the catalysts, we adopted the electrochemical impedance spectrums (EIS) and electrochemically active surface areas (ECSA). As shown in Fig. S10 and S11, the Co2P/Co3Fe7@N-C shows the smallest charge transfer resistance (Rct) of 70 Ω and a double layer capacitance (Cdl) of 72.0 mF cm− 2, comparing with Co/Co2P@N-C (Rct: 160 Ω; Cdl: 35.8 mF cm− 2), and other amounts of Co:Fe for Co2P/Co3Fe7@N-C catalysts. These results imply that the Co2P/Co3Fe7@N-C catalyst with amounts of exposed active sites and rapid electron transfer process during the OER activity. Significantly, the electrocatalytic stability of Co2P/Co3Fe7@N-C for both OER and ORR were also evaluated. As shown in Fig. S12, the Co2P/Co3Fe7@N-C maintains 93.3% of the initial ORR currents even after 10 h of testing, comparing with Pt/C catalyst (69.1%). In addition, nearly no overpotential changes were observed for Co2P/Co3Fe7@N-C in the subsequent 5000 cycles (Fig. S13), indicating it’s superior stability for ORR performance. Meanwhile, as shown in Fig. S14 and S15, the Co2P/Co3Fe7@N-C maintains 97.1% of the initial OER currents and its overpotential (at 10 mA cm− 2) values do not further change in the subsequent 5000 cycles, also demonstrating its long-term stability in OER. More importantly, all the XRD diffraction traces corresponded to Co2P (PDF # 32–0306) and Co3Fe7 (PDF # 48-1817), and no additional diffraction peaks are detected, indicating an excellent stability of Co2P/Co3Fe7@N-C during in ORR or OER processes (Fig. S16). More significantly, the Co2P/Co3Fe7@N-C shows a relatively small overall OER/ORR (ΔE = Ej=10 - E1/2) of 0.61 V compares with Co/Co2P@N-C (0.76 V), commercial Pt/C + Ir/C catalysts (0.69 V), and other recently reported bifunctional OER/ORR electrocatalysts (Fig. 4f and Table S2).
In order to thoroughly understand the key role of synergistic effect at the heterointerface between Co2P and Co3Fe7 for bifunctional OER/ORR performance, the DFT calculations were adopted. As shown in Fig. 5a, the oxygen intermediates (O*, OH*, and OOH*) adsorbs on optimized Co2P/Co3Fe7 interfacial structure, where occurs the typical four steps ORR/OER reaction [47]. Based on the reaction pathway, we calculate the free energy of OER and ORR process for the Co2P/Co3Fe7 and Co2P (Fig. 5b and Fig. 5c). It can be obviously seen from Fig. 5b, the rate-determining step (RDS) for Co2P/Co3Fe7 and Co2P are O* to OOH*. The Co2P/Co3Fe7 displays smaller ΔG values of 2.10 eV than the Co2P (3.79 eV), which indicates interfacial and synergistic effect between Co2P and Co3Fe7 as a key role for efficient OER activity [48]. For the Co2P, the formation of OH* step is endothermic reaction, indicating the step is RDS at U = 0 eV (Fig. 5c). Comparing with the Co2P, the Co2P/Co3Fe7 shows that all the elementary steps are exothermic dramatically, which is more beneficial for the ORR catalytic reaction. Moreover, Fig. 5d further shows the optimized atomic structure of the interface formed between Co2P slab and Co3Fe7, where the charge density difference illustrate that electrons accumulate at the interface and transfer from Co2P to Co3Fe7, corresponding with the XPS results. Importantly, as shown in Fig. 5e, the d-band center of Co2P/Co3Fe7 is smaller than that of Co2P, which can effectively regulate adsorption capacity of OER/ORR intermediates [49]. Thus, both theoretical and experimental results prove that the synergistic effect of Co2P/Co3Fe@N-C heterointerface is beneficial to the bifunctional OER/ORR performance and further application in the ZABs.
3.3. Liquid Zn-air battery test
To confirm the practical application of bifunctional Co2P/Co3Fe7@N-C catalyst as an air-cathode in ZABs, we assembled liquid ZABs using Co2P/Co3Fe7@N-C catalysts loaded on carbon cloth as the air-cathode (insert in Fig. 6a). The OCP of the Co2P/Co3Fe7@N-C-based ZAB is about 1.48 V, which is slightly larger than the Pt/C + Ir/C-based ZAB (1.46 V) (Fig. 6a), and competes with the other recently reported bifunctional electrocatalysts-based ZABs (Table S3). As shown in Fig. S17 and Fig. 6b, Co2P/Co3Fe7@N-C-based ZAB shows a smaller charge/discharge gaps and superior maximum power density of 152.3 mW cm− 2 compares with that of Pt/C + Ir/C-based ZAB (105 mW cm− 2). In addition, as shown in Fig. 6c, the Co2P/Co3Fe7@N-C-based ZAB not only can be discharged from 1 mA cm− 2 to a high current density of 50 mA cm− 2, but also be superior to Pt/C + Ir/C-based ZAB, indicating it can be operated at a wide range of current densities. In Fig. S18, the specific capacity is calculated based on the mass of consumed Zn and is about 858 mAh g-1 Zn for Co2P/Co3Fe7@N-C ZAB at a current density of 10 mA cm− 2. Significantly, as shown in Fig. 6d, the Co2P/Co3Fe7@N-C-based ZAB shows 266 h (1596 cycles) of excellent stability at a current density of 2 mA cm− 2, whereas the Pt/C + Ir/C-based ZAB lasts only about 147 hours (882 cycles). As shown in Fig. 6e, the Co2P/Co3Fe7@N-C-based ZAB shows a smaller range of voltage gap 0.02 V (from 0.79 V to 0.81 V) than that of Pt/C + Ir/C-based ZAB (0.41 V: from 0.7 V to 1.11 V). Furthermore, even at 10 mA cm− 2 for charging/discharging cycles, the Co2P/Co3Fe7@N-C-based ZAB also exhibits long-term stability of about 60 h (Fig. S19). Importantly, the red light-emitting diodes (LEDs) and blue LEDs display panel were driven by two-liquid ZABs connected in series with Co2P/Co3Fe7@N-C-based ZABs, respectively (Fig. S20).
3.4. All-solid-state Zn-air battery test
All-solid-state ZAB was also assembled with Co2P/Co3Fe7@N-C, zinc plate, and alkaline polyvinyl alcohol (PVA) gel as air cathode, anode, and solid electrolyte, respectively, as shown in inside of Fig. 7a. The Co2P/Co3Fe7@N-C-based all-solid-state ZAB displays an OCP of 1.4 V and power density of 62.5 mW cm-2, which is better than the Pt/C + Ir/C-based all-solid-state ZAB (OCP: 1.38 V; 38.1 mW cm-2) and competes with the recently reported papers in Fig. 7a, Fig. 7b, Fig. S21, and Table S4. Notably, the homemade all-solid-state ZAB still exhibits an excellent stability after 60 h and no significant voltage decay during charging and discharging (Fig. 7c). As shown in Fig. 7d and Fig. S22, the smart button ZAB was assembled with Co2P/Co3Fe7@N-C as an air cathode and the can provide an OCP about 1.4 V, slightly higher than Pt/C + Ir/C-based ZAB (1.38 V). As shown in Fig. 7e, the red LED are driven by two-button cells connected in series with Co2P/Co3Fe7@N-C-based ZABs. The Co2P/Co3Fe7@N-C-based button ZAB also shows long charge-discharge cycles performance (Fig. 7f). These results further confirmed the potential of applying Co2P/Co3Fe7@N-C to portable devices.