3.1. Structure and morphology
The structure of CoFe2O4@NCNF was studied with X-ray diffraction (XRD) at first. As shown in Fig. S1, the broad peak centered at 26.1° is attributed to (002) plane of carbon in graphitic phase.27 The other diffraction peaks located at 35.4, 43.1 and 74.1° are in agreement with (311), (400) and (533) planes of CoFe2O4 (PDF# 22-1086). The structure of NiFe2O4@NCNF was also confirmed by XRD as shown in Fig. S2 (PDF# 10–0325). Raman spectrum was employed to obtain more structural information from NCNF in CoFe2O4@NCNF (Fig. 1a). Two obvious intensive peaks appear at 1375 and 1583 cm− 1, corresponding to D and G bands of carbon material. D band originates from disordered or defective carbon, while G band is attributed to carbon in sp2-hybridized graphitic phase.28,29 Intensity ratio between D and G bands can reflect the degree of defect in carbon material. In general, high ID/IG value means the existence of a large number of defects.30,31 For CoFe2O4@NCNF, ID/IG value is 1.16, which implies there are a lot of defects in NCNF. This is benefit for electron transportation.
Elements and oxidation states were studied by X-ray photoelectron spectroscopy (XPS). Survey spectrum confirms C, N, O, Fe and Co co-exist in CoFe2O4@NCNF (Fig. S3). Detail information is further provided by high resolution XPS spectrum. High resolution C 1s spectrum can be deconvoluted into four peaks, which correspond to C-C (284.7 eV), C-N (285.7 eV), C-O (286.8 eV) and O = C-O (289.2 eV) (Fig. 1b).32 High resolution N 1s spectrum implies pyridinic-N (398.5 eV), pyrrolic-N (400.5 eV), graphitic-N (401.1 eV) and oxidated-N (403.8 eV) exist (Fig. 1c).33 In high resolution O 1s spectrum, three peaks at 530.4, 531.6 and 533.3 eV correspond to lattice oxygen, vacancy oxygen and surface oxygen, respectively (Fig. 1d).34 In high resolution Fe 2p spectrum, the first pair of doublet peaks at 711.1 and 723.7 eV are assigned to Fe 2p3/2 and Fe 2p1/2 of Fe2+ (Fig. 1e).35,36 The second doublet peaks at 713.5 and 726.3 eV match with Fe 2p3/2 and Fe 2p1/2 of Fe3+. The left two peaks located at 718.5 and 731.6 eV are satellite peaks. For high resolution Co 2p spectrum, Co 2p3/2 and Co 2p1/2 of Co3+ emerge at 780.9 and 796.4 eV (Fig. 1f). Co 2p3/2 and Co 2p1/2 of Co2+ appear at 783.6 and 798.1 eV.37 The XPS results match perfectly with recently reported spinel ferrites.
Morphology of CoFe2O4@NCNF was studied with scanning electron microscopy (SEM) and transmission electron microscopy (TEM). SEM image demonstrates CoFe2O4@NCNF with diameter about 70 to 80 nm staggers and interconnects together, which exhibits similar appearance with PPy nanofiber (Fig. 2a and S4). This illustrates the anchoring of CoFe2O4 does not alter morphology. Furthermore, cross-linked CoFe2O4@NCNF possesses rough surface. TEM image shows CoFe2O4 particle with small size about 8 to 12 nm dispersed homogeneously on NCNF (Fig. 2b and 2c). High resolution TEM (HRTEM) provides more structural information. In HRTEM image, the lattice plane with interlayer distance about 0.26 nm is observed, which corresponds to (311) crystal plane of CoFe2O4 perfectly (Fig. 2d).38 Energy dispersive spectroscopy (EDS) mapping demonstrates Co, Fe and O distribute evenly throughout NCNF (Fig. 2e and S5). This further suggests CoFe2O4 particles have generated successfully on NCNF.
3.2. ORR activity study
ORR activity of CoFe2O4@NCNF, NiFe2O4@NCNF, NCNF and Pt/C were tested by cyclic voltammetry (CV) technique in 0.1 M KOH saturated by O2. The scanning rate is 50 mV·s− 1. CoFe2O4@NCNF, NiFe2O4@NCNF and Pt/C possess obvious reduction peaks (Fig. 3a). In contrast, the reduction peak vanishes in NCNF. For CoFe2O4@NCNF, the reduction peak appears at 0.81 V, which is adjacent to Pt/C (0.82 V) and confirms its ideal ORR activity. This is higher than NiFe2O4@NCNF (0.78 V). Linear sweep voltammetry (LSV) curve at 1600 rpm reveals E1/2 of CoFe2O4@NCNF achieves 0.83 V (Fig. 3b). The value is close to commercial Pt/C (0.84 V) and higher greatly than NiFe2O4@NCNF. CoFe2O4@NCNF also possesses large limiting current density of 6.03 mA·cm− 2. This further implies excellent ORR activity of CoFe2O4@NCNF.
In order to calculate electron transfer number (n) in ORR, LSV curves at different rotational speeds were tested with rotating disk electrode (RDE) technique. The limiting current density adds obviously with increasing of rotational speed, implying ORR follows the first-order kinetic process (Fig. 3c).39 The Koutecky-Levich (K-L) plot of CoFe2O4@NCNF derived from LSV curves suggests J− 1 keeps perfect linear relationship with ω−1/2 (Fig. 3d). At 0.4, 0.5, 0.6, 0.7 V (vs. RHE), all the calculated n values approach 4.0, which indicates ORR follows four-electron process. LSV curves of NiFe2O4@NCNF were also measured at different rotational speeds and n values are also close to 4.0 (Fig. S6a and b). The ORR performance of CoFe2O4@NCNF is comparable with recently reported electrocatalysts as shown in Table S1. The stability is an important factor to evaluate the practicability of an electrocatalyst in ZAB. For CoFe2O4@NCNF, stability was tested by chronoamperometry technique. After 12 h, the current density still remains as high as 97.6% (Fig. 3e). On the contrary, current density of commercial Pt/C decreases to 80.2% in the same time. Furthermore, LSV curve of CoFe2O4@NCNF does not change before and after 5000 cycle test (Fig. 3f). These results imply CoFe2O4@NCNF possesses excellent ORR stability.
To study ORR performance of NCNF, NiFe2O4@NCNF and CoFe2O4@NCNF in detail, rotating ring-disk electrode (RRDE) technique was employed to obtain n value and H2O2 selectivity (H2O2%). Based on disk and ring current, n value of CoFe2O4@NCNF ranges from 3.79 to 3.91 with potential scope 0 to 0.8 V (Fig. 4a and 4b). These values can be compared with Pt/C with n value from 3.80 to 3.94. The n values of NiFe2O4@NCNF and NCNF are 3.67 to 3.91 and 2.68 to 3.64 in the same potential scope, which are lower obviously than CoFe2O4@NCNF. Furthermore, the H2O2 selectivity of CoFe2O4@NCNF is below 10.5%, which is close to commercial Pt/C (Fig. 4c). The H2O2 selectivity is lower significantly than NiFe2O4@NCNF and NCNF. Above facts illustrate CoFe2O4@NCNF is an excellent four-electron ORR electrocatalyst with low H2O2 selectivity.
The ORR activity was also studied in neutral electrolyte and LSV was conducted in 0.1 M phosphate buffer saline (PBS) buffer saturated by O2. At 1600 rpm, E1/2 value of CoFe2O4@NCNF reaches 0.58 V with limiting current density 5.68 mA·cm− 2 (Fig. S7a). For commercial Pt/C, E1/2 and limiting current density are 0.57 V and 4.71 mA·cm− 2. It can be anticipated CoFe2O4@NCNF possesses better ORR performance than commercial Pt/C under neutral condition. LSV curve of CoFe2O4@NCNF was further conducted at different rotating speed (Fig. S7b). The electron transfer number was determined by the K-L equation, which suggests four-electron ORR happens under alkaline condition (Fig. S7c). The ORR activity was further studied by RRDE technique (Fig. S8a). With potential scope from 0 to 0.55 V (vs. RHE), electron transfer number fluctuates between 3.82 and 3.92. The H2O2 selectivity is lower than 9.1% (Fig. S8b and S8c). This performance is better than commercial Pt/C with n value from 3.78 to 3.90 and H2O2 selectivity undulating around 11%. The facts indicate CoFe2O4@NCNF is a reliable ORR electrocatalyst under neutral condition.
3.3. OER activity study
For ZAB, charge performance is decided by OER activity greatly. Here the OER activity of CoFe2O4@NCNF was evaluated by LSV in 1.0 M KOH. Its η10 is 334 mV, which is higher slightly than commercial RuO2 (283 mV) (Fig. 5a). NCNF is inactive in OER. As for NiFe2O4@NCNF, the η10 value is 358 mV. OER kinetics of above electrocatalysts was also evaluated by Tafel slope. Tafel slope of CoFe2O4@NCNF is 72.2 mV·dec− 1 (Fig. 5b), which is lower than NiFe2O4@NCNF (84.7 mV·dec− 1). The value is close to commercial RuO2 (56.9 mV·dec− 1). This indicates CoFe2O4@NCNF exhibits a quick kinetic process in OER. The activity of CoFe2O4@NCNF can be comparable with recently reported OER electrocatalysts (Table S2). The stability of CoFe2O4@NCNF was studied chronoamperometry technique. After 18000 s, current density decay of CoFe2O4@NCNF is very limited (Fig. S9). As for commercial RuO2, only 68.2% current density remains. This implies CoFe2O4@NCNF exhibits more excellent OER stability than RuO2.
To explore the difference OER performance between CoFe2O4@NCNF and NiFe2O4@NCNF, electrochemical impedance spectroscopy (EIS) was conducted. Nyquist diagram shows CoFe2O4@NCNF possesses a smaller semicircle diameter than NiFe2O4@NCNF (Fig. S10). This implies CoFe2O4@NCNF exhibits lower charge transfer resistance than NiFe2O4@NCNF, which facilitates electron movement and improves OER. Moreover, electrochemical double-layer capacitance (Cdl) was calculated from CV curves at different scan rates from 10 to 100 mV·s− 1 (Fig. S11a and 11b). Cdl value of CoFe2O4@NCNF is 24.52 mF·cm− 2, which is higher than NiFe2O4@NCNF (7.69 mF·cm− 2) (Fig. 5c). Since Cdl value is positively correlated with electrochemical active surface area (ECSA), these imply CoFe2O4@NCNF can provide more active points than NiFe2O4@NCNF in OER.40 The magnitude of potential gap, ΔE (ΔE = η10-E1/2), is usually employed to measure bifunctional performance of an electrocatalyst.41 In general, low ΔE means good bifunctional electrocatalytic performance. CoFe2O4@NCNF has a low ΔE of 0.73 V (Fig. 5d). This also implies it is an ideal cathode material for ZABs.
3.4. Mechanism study of ORR/OER
To analyze the different ORR and OER activities between CoFe2O4@NCNF and NiFe2O4@NCNF, density functional theory (DFT) calculation was applied.42 As for ORR, O2 is absorbed on electrocatalyst, which is reduced to OOH*, O*, OH* and OH− in turn (Table S3). For CoFe2O4@NCNF and NiFe2O4@NCNF, the free energy of each reductive step is thermodynamically favorable downhill as U = 0 V (Fig.. 6a and 6b). After U adds to 1.23 V, the first and second steps (O2 + H2O + e = OOH* + OH− and OOH* + e = O* + OH−) become uphill.43 It can be observed clearly, CoFe2O4@NCNF needs lower energy than NiFe2O4@NCNF to complete the barrier. To evaluate ORR activity in theory, limiting potential (the highest potential to make all elementary reactions exothermic) is an important parameter, which is abbreviated as UL(ORR). For CoFe2O4@NCNF and NiFe2O4@NCNF, the UL(ORR) values are 0.95 and 0.74 V, respectively. Higher UL(ORR) of CoFe2O4@NCNF implies it exhibits more excellent ORR activity than NiFe2O4@NCNF.44 For OER, as U = 0 V, free energy of each step is uphill for both CoFe2O4@NCNF and NiFe2O4@NCNF (Fig.. 6c and 6d). At U = 1.23 V, the first two reactions (OH− - e = OH*; OH* + OH− - e = H2O + O*) are still uphill and become rate determining steps.45 It is obviously, CoFe2O4@NCNF needs lower energy to overcome the barrier than NiFe2O4@NCNF. Furthermore, the UL(ORR) (limiting potential in OER) of CoFe2O4@NCNF is 1.51 V, which is lower than NiFe2O4@NCNF with UL(ORR) 1.78 V. This confirms superior OER activity of CoFe2O4@NCNF over NiFe2O4@NCNF.
3.5. Performance of rechargeable ZAB
Considering perfect ORR and OER activities, rechargeable ZAB was assembled with CoFe2O4@NCNF as cathode material. In this ZAB, polished zinc plate acts as anode and 6.0 M KOH containing 0.2 M Zn(OAc)2 serves as electrolyte. The open circuit voltage (OCV) of this ZAB achieves 1.45 V (Fig. 7a). Its peak power density reaches 85.9 mW·cm− 2 (Fig. 7b). This is higher obviously than ZAB with Pt/C + RuO2 as cathode material, which originates from perfect ORR activity of CoFe2O4@NCNF. Furthermore, four red light-emitting diodes (LED) can be light up by two ZABs connected in series manner (Fig. 7c). Multimeter also suggests the OCV of ZAB achieves 1.45 V. As revealed by charge/discharge polarization curves, ZAB shows low charge voltage and high discharge voltage. At 10 and 20 mA·cm− 2, the voltage gaps are merely 0.42 and 0.61 V, which are lower obviously than ZAB built by Pt/C + RuO2 (Fig. S12).
Rate capability of ZAB was evaluated with current densities from 1 to 10 mA·cm− 2. At 1, 2, 4, 8 and 10 mA·cm− 2, discharge voltage plateaus are 1.26, 1.24, 1.21, 1.16 and 1.13 V (Fig. 7d). After 5 h discharge test, the current density recovers to original value and discharge voltage also returns to 1.26 V. The discharge voltage and current attenuation can almost be negligible, which demonstrates its promising discharge rate performance. The specific capacity and energy density can be calculated based on the mass of Zn consumed. For ZAB, its specific capacity and energy density are 807.4 mA·h·g− 1 and 963.6 Wh·kg− 1 at 10 mA·cm− 2 (Fig. 7e). The performance is better than ZAB built by Pt/C + RuO2 with specific capacity and energy density 788.5 mA·h·g− 1 and 930.4 Wh·kg− 1. Cycling stability was studied with galvanostatic charge/discharge tests at 1 mA·cm− 2 with 30 min for each cycle (Fig. 7f). ZAB shows constant charge (2.03 V) and discharge voltages (1.19 V) with the gap of 0.84 V for 120 h. On the contrary, charge and discharge voltage gap increases obviously after 80 h for ZAB built by Pt/C + RuO2. Above results further show the advantage of CoFe2O4@NCNF over Pt/C + RuO2 as cathode material in ZAB.