High-efficiency oxygen electrocatalyst for Zn-air batteries on CoMn alloy encapsulated in N-doped carbon architectures

As an energy conversion/storage technology, Zn-air batteries (ZABs) recently have received enormous attention on account of their high energy density, cost-effectiveness, and outstanding safety.^1–6 Unfortunately, the commercialization of ZABs is largely limited by the shortage of high active oxygen electrocatalysts during the discharge and charge processes.^7–10 Nowadays, noble metal catalysts (Pt, Ru, and Ir-based) are still recognized as the best-performing ORR and OER electrocatalysts.^11–14 Unfortunately, it suffers from scarce resources, single-function, and instability that limit the commercialization of noble-metal catalysts.^15–19 In this context, exploring a high-efficiency without noble metal electrocatalysts for ORR and OER is imperative for both scientific and industrial fields.

In recent decades, transition metal compounds (which include nitrides, carbides, oxides, alloys, and so on) have been used as alternatives to Pt-based and RuO₂-based catalysts for ORR and OER electrocatalysts owing to their earth abundance, low cost, environmental friendliness, and outstanding ORR/OER.^20,21 In particular, an alloy consisting of two or more metal species has been found with high bifunctional activities for ORR/OER.^22,23 Owing to the difference in electronegativity, the electrons transfer between the different metal species, leading to electron rearrangement.^24,25 The boosting effect of reconfigured electronic structure on electrocatalytic activity is of great significance for ORR and OER.^26–28 Up to date, the overall ORR/OER performances of the alloy catalysts are still plagued by intrinsic poor electrical conductivity. Furthermore, when exposed to air, the alloy will be oxidized easily, resulting in poor catalytic stability.^29,30 Recent research demonstrates that loading alloy on N-doped nanocarbon substrates is an effective strategy for the accomplishment of ORR/OER catalytic activity because of the high conductivity of N-doped nanocarbon substrates.^31,32 Co-based catalysts have been demonstrated with good activity for ORR with superior durability in alkaline medium. The active sites for the ORR in the single-metal Co–N–C catalyst were confirmed owing to the Co–N and Co–Co bonds. The density of states of Co–N across the Fermi level is demonstrated by density functional theory calculation. The metallic structures of Co–N can facilitate efficient electrons for effective electrocatalysis.^33,34 Mn-based catalysts have been demonstrated with highly active sites for OER.^35–41 In the case of transition metal alloy catalysts based on the Co and Mn elements, few studies had been done, involving the synthetic strategies and the mechanism of their synergy. Therefore, new insights into the active centers of CoMn alloy loading on N-doped nanocarbon are needed for the accomplishment of bifunctional catalysts.

Carbon materials [which include carbon nanotubes (CNTs), graphene, carbon nanocages, and so on] hold great potential as candidates to replace Pt-based and RuO₂-based catalysts for ORR/OER because of their favorable conductivity, high specific surface area, controlled microstructures, and excellent durability.^42–44 Until now, carbon materials have been prepared via various approaches. Among them, metal-organic frameworks (MOFs)-derived carbon-based materials attracted extensive studies because of their well-defined morphology, controllable pores, high surface area, and tunable structures, which are beneficial for the exposure of sufficient active sites and providing electron/mass transportation for ORR/OER.^45–48 Motivated by the advantages, carbonaceous material derived from MOFs is expected to be a promising method for the construction of well-defined ORR/OER electrocatalysts.

Herein, facile polydopamine-pyrolytic strategies were developed for in situ immobilizing CoMn alloy encapsulated in N-doped hollow carbon architectures (NHC). CoMn-PDA@ZIF-8 was obtained by using ZIF-8 as a template and associated with the dopamine to incorporate Co and Mn elements. After the simple pyrolysis of a mixture of CoMn-PDA@ZIF-8 and melamine, a well-defined CoMn alloy encapsulated in nitrogen-doped hollow carbon architectures was successfully obtained and denoted as CoMn@NHC. The CoMn alloy and combination with nitrogen-doped hollow carbon architectures dramatically boost the activities toward ORR and OER. The generated CoMn@NHC composite expressed outstanding electrocatalytic performances for ORR (E_1/2 = 0.87 V) and OER (E_i=10 = 1.51 V). The homemade aqueous ZABs equipped with CoMn@NHC show a larger peak power density of 92.3 mW cm⁻² and good long-term stability. This study affords an economical strategy for designing high-efficient bifunctional oxygen catalysts for ZABs.

The typical synthetic route of CoMn@NHC was schematically depicted in Fig. 1. First, CoMn-PDA@ZIF-8 is obtained via a facile polydopamine process by adding an aqueous solution containing dopamine, Co(NO₃)₂, and Mn(NO₃)₂ in an alkaline Tris buffer with ZIF-8. According to Fig. 3(a), CoMn-PDA@ZIF-8 has the same crystalline texture as ZIF-8. Then, CoMn-PDA@ZIF-8 mixed with melamine is pyrolyzed in the N₂ atmosphere at 920 °C. Furthermore, HC, NHC, Mn@NHC, and Co@NHC are also prepared through a similar procedure without melamine, cobalt salt, or manganese salt (Fig. 1, see details in the supporting file). The morphologies and microstructure of the obtained materials are investigated by transmission electron microscopy (TEM) and EDS line scan analysis. The TEM image of HC derived from the PDA@ZIF-8 carbonized without melamine, cobalt salt, and manganese salt exhibits a well-defined hollow carbon architecture, which is inheriting the morphology structure of ZIF-8 (Figs. S1a and S1b). NHC shows similar morphologies to those of HC (Fig. S1c) with thinner graphitization carbon nanosheets. It is indicated that the introduction of N is beneficial for the formation of graphite carbon. After intruding on the cobalt salt or manganese salt during the polydopamine process, the morphology of the obtained sample has a slight change. According to the EDS element mapping [Figs. S2(a, b) and S3(a, b)], Co and Mn have been successfully introduced into the nitrogen-doped hollow carbon architectures. Mn is distributed on the carbon framework homogeneously while Co is of a distinct nanoparticle with an obvious lattice spacing of 0.206 nm, corresponding to the facets of Co (111) encased in hollow carbon architectures [Fig. S3(c, d)].⁴⁹ The TEM image and EDX line-scan element profile over one particle of CoMn@NHC reveal that CoMn alloy is successfully embedded in N-doped hollow carbon. The obvious lattice fringe spacing of 0.340 nm belongs to the (002) lattice plane of C [Fig. 2(e)], while the lattice fringe spacing of 0.203 nm is smaller than that of Co (111) facets of 0.206 nm, owing to some Mn atoms doped into Co phases to form Co–Mn alloy. The formation of Co–Mn alloy is also confirmed by EDX line-scan element analysis over individual nanoparticles. A well-distribution of Co and Mn elements over individual nanoparticles is obvious, as shown in Fig. 2(f). According to the morphologies and microstructure analysis, a well-defined CoMn alloy encapsulated in nitrogen-doped hollow carbon architectures is successfully fabricated.

The crystalline structures of HC, NHC, Mn@NHC, Co@NHC, and CoMn@NHC are ascertained by XRD [Fig. 3(b)]. All the samples display diffraction peaks at 26.1°, which are indexed to the (022) facet of graphitic carbon. Co@NHC displays three diffraction peaks at 44.2°, 51.5°, and 75.6°, representing the crystal faces of (111), (200), and (220) based on the standard card of metal Co (PDF#15-0806).⁵⁰ CoMn@NHC shows three peaks at 43.6°, 50.8°, and 75.0°, which are situated between the characteristic peaks of Co and Mn and further demonstrate the formation of CoMn alloy. It is noteworthy that the XRD patterns of Mn@NHC and CoMn@NHC show no obvious diffraction peaks of Mn-based crystal structures, probably for the low content (0.27%) of Mn. Raman spectrum of HC, NHC, Mn@NHC, Co@NHC, and CoMn@NHC is shown in Fig. 3(c). All curves exhibit two dominant peaks at 1342 cm⁻¹ (D band) and 1588 cm⁻¹ (G band),⁵¹ respectively. The peak intensity ratio (I_D/I_G) value of HC, NHC, Mn@NHC, Co@NHC, and CoMn@NHC is 1.13, 1.08, 0.99, 0.96, and 0.92, respectively. The smallest I_D/I_G of CoMn@NHC indicates the highest graphitization degree of carbon substrates, which is beneficial for electron transport for the electrocatalytic process. XPS measurement is applied to further detect the surface composition and bonding states of CoMn@NHC. The full XPS spectrum (Fig. S4) of CoMn@NHC shows the presence of Co, Mn, O, N, and C with atomic percentages of 0.58%, 0.27%, 1.24%, 4.7%, and 93.21%, respectively. There are no obvious peaks associated with Mn in the XPS survey spectrum may be due to the low content of Mn. The C 1s spectrum [Fig. 3(d)] is best fitted to three peaks including C–C (284.7 eV), C=N (285.3 eV), and C–O (288.6 eV).⁵² According to the result, heteroatoms were introduced into the carbon lattice successfully. The N 1s spectrum confirms four types of N species [Fig. 3(e)], including pyrrolic N (400.7 eV), pyridinic N (398.7 eV), graphitic N (402.2 eV), and oxidized N (405.1 eV),⁵³ which are the key role for ORR/OER electrocatalytic activity. For Co 2p spectrum [Fig. 3(f)], there are four fitted peaks: Co⁰ species (780.2 and 796.1 eV) and satellite peaks (785.3 and 802.5 eV), which are consistent with the results of XRD and TEM.⁵⁴ For Mn 2p spectrum [Fig. 3(g)], it displays five peaks: Mn-N_x species (640.3 eV), Mn²⁺ (642.1 and 653.1 eV), and the satellite peak (645.4 and 655.1 eV).⁵⁵ The deconvolution O1s [Fig. 3(h)] displays two peaks at 531.4 and 532.7 eV, assigning to O=C and O–C bonds, while the peak at 529.5 eV is attributed to O²⁻, referring to the slight oxidation of Mn.⁵⁶ 

Moreover, N₂ isothermal adsorption/desorption experiment is employed to study the surface area and porous feature of CoMn@NHC. As revealed in Fig. 3(i), the isotherm curve of CoMn@NHC presents typical type-IV isotherms with a significant hysteresis loop over a P/P₀ range of 0.43–0.98, indicating the characteristic representation of meso-porous structure.⁵⁷ According to the pore size distribution, both micropores and mesopores exist in CoMn@NHC, which are beneficial for electron transport and mass transportation during the electrocatalytic reactions. Furthermore, CoMn@NHC has a larger surface area of 245.95 m² g⁻¹ and a total pore volume of 0.96 cm³ g⁻¹, benefitting by providing abundant interconnected channels for the diffusion of reaction species into the interior of the as-prepared catalyst and exposing more possible active sites.

The ORR or OER activities of HC, NHC, Mn@NHC, Co@NHC, CoMn@NHC, and Pt/C were systematically studied in a three-electrode system. According to the CV curves under O₂-saturated KOH [Fig. 4(a)], CoMn@NHC shows an obvious peak of ORR signals, higher than that of HC, NHC, Mn@NHC, and Co@NHC, suggesting its outstanding activity for ORR. Notably, no cathodic current responses of all catalysts in the N₂-saturated KOH solution. The remarkable electrocatalytic performances of CoMn@NHC were confirmed by LSV. The LSV curve of CoMn@NHC in Fig. 4(b) possesses E_1/2 at 0.87 V and outperforms HC (0.78 V), HC (0.79 V), Mn@NHC (0.82 V), Co@NHC (0.84 V), and Pt/C (0.86 V). Moreover, the higher E_1/2 value of CoMn@NHC also delivers better ORR performance than that of the other reported alloy catalysts in the literature (listed in Table S2). Figure S5 shows ORR polarization curves of CoMn@NHC under various rotating speeds. It reveals that the limiting current density was positively correlated with rotation speeds, suggesting diffusion-controlled processes of ORR on the CoMn@NHC electrode. Furthermore, the charge transfer resistance of all catalysts was evaluated by electrochemical impedance spectroscopy (EIS). As shown in Fig. S6, the semicircle of the Nyquist diagram at low frequency (high Z) shows that the R_ct value increases sequentially, CoMn@NHC < Co@NHC < Mn@NHC < NHC < HC. The tendency demonstrates that the incorporation of N and metals can significantly enhance electrical conductivity and result in fast electron transfer kinetics. In order to further figure out the ORR mechanism of CoMn@NHC, the electron transfer number (n) and the H₂O₂ production of CoMn@NHC were investigated by RRDE measurement. As illustrated in Fig. 4(c), the n value of CoMn@NHC is almost 4 and the H₂O₂ yield is below 3% in the potential region of 0.20–0.9 V, confirming a 4e⁻-dominated reduction pathway of the ORR process at CoMn@NHC. Durability is an essential criterion for practical applications in ZABs. A continuous CV stability test is employed to explore the durability of CoMn@NHC. As revealed in Fig. 4(d), the E_1/2 of CoMn@NHC shows slight voltage decay about negative shift 2 mV after 10 000 CV cycles, demonstrating the superior durability of CoMn@NHC. The resistance to methanol of electrocatalysts was conducted by the methanol crossover test. As shown in Fig. 4(e), CoMn@NHC displays negligible variation after the injection of methanol, while Pt/C shows a sharp decrease of over 50%, confirming its good selectivity for the ORR than methanol oxidation. The stability of CoMn@NHC is further studied by the chronoamperometry measurement. As demonstrated in Fig. 4(f), the current retention ratio of CoMn@NHC still maintains 95.9% after 54 000 s, while Pt/C suffers a remarkable decay (59.8%), demonstrating its excellent stability for ORR. In summary, the obtained CoMn@NHC electrocatalyst exhibits outstanding activity, strong stability, and superior selectivity for ORR.

Besides ORR, the OER performance of as-prepared materials was also assessed. As shown in Fig. 4(g), HC and NHC show a poor current response of OER. After the introduction of Mn or Co, the OER activity expresses a significant improvement. Notably, CoMn@NHC displays a superior OER activity with E_i=10 = 1.51 V, which is smaller than that of Co@NHC and Mn@NHC. The Tafel slope is a key indicator for evaluating the OER kinetics. If the value of the Tafel slope is smaller, the reaction kinetics during molecular adsorption and activation is more favorable. The Tafel slope of CoMn@NHC is explored by LSV curves. As revealed in Fig. 4(h), CoMn@NHC exhibits a Tafel slope value of 96 mV dec⁻¹, which is comparable to that of RuO₂ (82.3 mV dec⁻¹), suggesting excellent OER reaction kinetics. Furthermore, the stability of CoMn@NHC is further validated via a continuous CV test. As exhibited in Fig. 4(i), it shows a slight variation (about 1 mV) after 1000 cycles, demonstrating good stability of CoMn@NHC. The series of experiments demonstrate that CoMn@NHC has excellent bifunctional catalytic activity, possessing actual feasibility to replace Nobel metal catalysts in ZABs.

Benefiting from the outstanding dual electrocatalytic activities, CoMn@NHC is further assembled into ZABs to assess its practical feasibility in energy transduction devices [Fig. 5(a)]. A homemade ZABs is loaded with carbon paper equipped with CoMn@NHC or Pt/C + RuO₂ as the cathode, Zn flake as the anode, and a mixed solution [6.0M KOH +0.2M Zn (CH₃COO)₂] as the electrolyte. The ZABs driven by CoMn@NHC afford a OCV of 1.45 V [Fig. 5(b)], which is higher than that of the Pt/C + RuO₂- based ZABs (1.42 V), proving the better electrocatalytic activity of CoMn@NHC. In addition, the open-circuit voltage shows no decay over 70 min, giving evidence of the excellent stability of CoMn@NHC. Two ZABs with CoMn@NHC as the cathode in series can power up a 2.2 V LED lamp [Fig. 5(c)], manifesting its attractiveness for application in actual devices. Figure 5(d) reveals that the ZABs with CoMn@NCNC deliver a peak power density of 92.3 mW cm⁻², outperforming the Pt/C + RuO₂-based ZABs (79.8 mW cm⁻²). The cycling stability of CoMn@NHC is further evaluated by a charging–discharging test. The ZABs with CoMn@NHC and Pt/C + RuO₂ are cycled at 5 mA cm⁻² with each cycle being 20 min. The voltage gap of CoMn@NHC for both charge and discharge procedures reveals without obvious decay after 273 h [Fig. 5(e)], while the voltage gap of Pt/C + RuO₂ dropped evidently within 100 h (Fig. S7), confirming its excellent stability. The galvanostatic cycling curves at different current densities have been displayed in Fig. S8. The Zn-air battery based on CoMn@NHC discharges stably at different densities. After the current density reduced from 50 to 2 mA cm⁻², the discharge potential returned to normal. These results indicate that the CoMn@NHC based Zn-air battery has excellent discharge rate performance and good reversibility. For all the results, CoMn@NHC shows outstanding electrocatalytic performance in actual ZABs.

To summarize, CoMn alloy encapsulated in N-doped hollow carbon architectures was developed through facile polydopamine and pyrolytic strategies. The obtained CoMn@NHC nanohybrid shows excellent ORR and OER electrocatalytic activity compared with the commercial Pt or RuO₂. The ZABs assembled with CoMn@NHC exhibit a high OCV of 1.45 V, a large peak power density of 92.3 mW cm⁻², and excellent cycling stability of nearly 300 h, surpassing those of commercially used Pt/C + RuO₂. The achieved high dual oxygen electrocatalytic performance can be owing to the synergistic effect between CoMn alloy and N-doped hollow carbon matrix. This study provides significant knowledge for the construction of high-efficient bifunctional oxygen catalysts for application in actual facilities.

See the supplementary material for the Experimental section, Figs. S1–S8, and Table S1.

Financial support from the National Natural Science Foundation of China (Grant No. 31901272) and the Jiangsu Province Key Laboratory of Biomass Energy and Materials (Grant No. JSBEM-S-201906) is acknowledged.

The authors have no conflicts to disclose.

Wenshu Zhou: Writing – original draft (lead); Writing – review & editing (lead). Yanyan Liu: Writing – review & editing (supporting). Dichao Wu: Software (supporting). Shuling Liu: Writing – review & editing (supporting). Pengxiang Zhang: Software (supporting). Gaoyue Zhang: Methodology (supporting). Kang Sun: Formal analysis (supporting). Jianchun Jiang: Methodology (equal).

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