1 Introduction

The ever-increasing energy consumption and related environmental concerns are driving the consistent development of electrochemical energy storage and conversion technologies [1]. Various advanced renewable electricity devices, such as primary batteries, rechargeable batteries, fuel cells, and supercapacitors, have been proposed and applied in the commercial market. Among these diverse energy devices, metal-air batteries are a kind of electrochemical device in which the electricity is produced by metal oxidation and oxygen (from air) reduction. Owing to the facile fabrication, natural abundance, favorable safety, and eco-friendliness of Zn metal, Zn-air batteries with a high theoretical energy density of 1 086 Wh kg−1 (fivefold larger than that of traditional lithium-ion batteries [2, 3]) have attracted increasing attention in recent years (Fig. 1a, b) [4, 5].

Fig. 1
figure 1

Current progress in Zn-air battery research. a The number of publications about Zn-air batteries in the past five years. b Comparison of practical and theoretical energy densities for a range of batteries (the cell level refers to all anode (Zn) materials without any packaging materials or leads) [11,12,13,14,15,16]. c Reported electrocatalysts in Zn-air batteries to date. The different species of metal compounds are marked with different colors; more intense colors indicate more reported literature, whereas lighter colors indicate fewer reports. Updated on March 15, 2022

However, despite the advantages of Zn-air batteries, their commercialization is still hindered by several drawbacks, such as the insufficient mass transport of oxygen electrodes and the intrinsically sluggish reaction kinetics of the oxygen evolution reaction (OER) and the oxygen reduction reaction (ORR), resulting in low cyclability of Zn-air batteries. To enhance the intrinsically sluggish kinetics, oxygen electrodes loaded with efficient oxygen catalysts are required to accelerate the ORR and/or the OER [6]. Currently, both ORR and OER catalysts are mostly based on noble metals (Pt, Ru, Ir, and so on), which usually exhibit excellent activities [7, 8]. However, their relatively low stability and high cost still hinder their large-scale application in the air cathodes of Zn-air batteries. Therefore, most reports on oxygen catalysts for Zn-air batteries have focused on designing non-noble metal–based catalytic sites with efficient and intrinsically stable catalytic activities. Furthermore, unlike the rotating disk electrode (RDE) technique used in laboratories for ORR testing, practical Zn-air batteries are susceptible to the influence of mass and air transport. Since the oxygen-catalytic reactions in Zn-air batteries generally occur at the solid (active catalysts)–liquid (electrolytes)–gas (O2) three-phase interface, accelerating O2 transport and enlarging their contact areas play key roles in boosting their performance [9]. Overall, the design of air cathodes with well-defined oxygen-catalytic centers and nanostructured supports is of great importance in providing a wide range of possibilities for designing advanced Zn-air batteries [10].

In recent decades, nanostructured carbon-based electrodes with enormous potentials, such as larger surface areas, efficient oxygen-catalytic centers, hierarchically porous matrices, superior durability, and low cost, were regarded as the most promising oxygen electrodes for Zn-air batteries systems [17,18,19,20]. Stimulated by the above-mentioned superiority and the necessary pursuit of structural design for carbon electrodes, several previous reviews have summarized this research field in recent years. As early as 2017, the design of a hierarchical and freestanding structured air cathode was discussed by the Shen group, with a specific emphasis on the reaction mechanism and optimization strategies of air cathodes for Zn-air batteries [9]. Later, the Zhang group reviewed fundamental studies on the structural design of oxygen catalysts in Zn-air batteries [21]. In another review, they presented the advanced progress in the atom-scale modulation and nanostructure design of carbon-based bifunctional oxygen-catalytic materials for metal-air batteries [22]. Recently, the advances of flexible Zn-air batteries were reviewed by the Huang group, who mainly focused on 3D self-supporting oxygen catalysts, types of solid-state electrolytes, and the rational design of Zn anodes [23]. However, a comprehensive review of all carbon-based oxygen catalysts, including catalytic sites, nanostructures, synthesis methods, precursor design, and developments for advanced Zn-air batteries, is still lacking. Thus, the advanced applications of flexible Zn-air batteries and high-power Zn-air batteries are summarized herein. More importantly, a comprehensive perspective with insight into the advantages, key problems, and further promising solutions of carbon-based oxygen catalysts for Zn-air batteries is provided in this review.

Herein, we present a timely and comprehensive review that summarizes the reported nanostructured carbon-based electrodes for advanced Zn-air batteries in terms of tailoring oxygen-catalytic sites and designing carbon supports. First, the working mechanisms and principles of oxygen electrodes for primary and rechargeable Zn-air batteries are briefly introduced. Subsequently, we systematically summarize the methods applied for tailoring different types of oxygen-catalytic sites, including (i) metal-free heteroatoms, (ii) metal single-atom sites, and (iii) metal alloys and compounds. Then, we outline the strategies used for designing carbon supports for advanced oxygen electrodes, such as conventional carbon nanomaterials, porous carbon, and organic precursors-derived carbon. In this review, we pay particular attention to the versatile synthetic strategies, characterization methods, and in-depth understanding of the relationships between oxygen-catalytic sites/nanostructures and electrode performance. Moreover, we also briefly comment on recent progress in the development of flexible and high-power Zn-air batteries for practical applications. Finally, we thoroughly discuss future development directions, challenges, and perspectives on the rational design of nanostructured carbon-based electrodes in Zn-air batteries. It is believed that this review can inspire future developments of fast-kinetic and efficient Zn-air batteries for use in a broad range of energy fields.

2 Principles and Mechanisms of Oxygen Electrodes in Zn-Air Batteries

2.1 Battery Components and Working Principles

The main structure of a primary Zn-air battery, which is composed of a Zn anode, an electrolyte, an air cathode with active materials to promote the ORR [24], and a gas diffusion layer that allows air to enter into the electrolyte, is illustrated in Fig. 2a. The basic working principle of a primary Zn-air battery during the discharge process in an alkaline electrolyte is illustrated in Fig. 2b.[25, 26] Specifically, the Zn anode first liberates two e and is oxidized to zinc ions (Zn2+). The released e migrates via an external circuit to the air cathode, where it contacts O2 on the surface of the catalysts. Then, the O2 is reduced through the ORR to form hydroxide ions (OH) at the solid (active catalysts)–liquid (electrolytes)–gas (O2) three-phase interface [25], and the full battery reaction is Zn + O2 to ZnO.

Fig. 2
figure 2

a Schematic working principle and b discharge reaction equations of a primary Zn-air battery in an alkaline medium. c Schematic working principle, d discharge reaction equations, and e charge reaction equations of a rechargeable Zn-air battery in an alkaline medium

Theoretically, the ORR in a Zn-air battery is fully reversible. Therefore, a bifunctional (ORR and OER) oxygen-catalytic electrode will allow realization of a rechargeable Zn-air battery, where the charging process is promoted by the OER on the cathode, and ZnO is reduced back to Zn and deposited on the anode [27]. Generally, the properties of rechargeable Zn-air batteries are particularly dependent on the kinetics of both the ORR and the OER, which are closely related to the natural activities of the bifunctional catalysts, as well as the reactant diffusion efficiency at the reaction interface, as mentioned above [12]. The main structure of a rechargeable Zn-air battery, which is similar to that of a primary battery but with different electrolytes and catalyst layers, is illustrated in Fig. 2c. The working principle of a rechargeable Zn-air battery during the discharging and charging processes in an alkaline electrolyte is illustrated in Fig. 2d, e [6].

2.2 The ORR and OER in Zn-Air Batteries

The ORR, a crucial reaction on the oxygen cathode in Zn-air batteries, is a multielectron (2e or 4e) transfer reaction. As summarized in Table 1, in alkaline electrolytes, the ORR can proceed either by a two-step 2e pathway with the formation of H2O2 or by a more efficient 4e process to directly reduce O2 to OH. Therefore, the 4e process possesses the highest O2 utilization efficiency and is remarkably favorable for practical Zn-air battery applications [28,29,30,31,32].

Table 1 The fundamental reaction equations of the ORR in alkaline aqueous solution

To better understand the mechanism of the ORR, the four elementary reaction steps of the 4e pathway in an alkaline electrolyte are listed below and in Fig. 3a:[33]

$${* + \text{O}}_{{2}} {\text{(g) + }}\,{\text{H}}_{2} \text{O(l)}\,{ + }\,{\text{e}}^{ - } { } \to \text{OOH*}\,{ + }\,\text{OH}^{ - } \text{(aq)}$$
$$\text{OOH*}\,{ + }\,\text{e}^{ - } { } \to \text{ O*}\,{ + }\,\text{OH}^{ - } \text{(aq)}$$
$$\text{O*}\,{ + }\,\text{H}_{2} \text{O(l)}\,{ + }\,\text{e}^{ - } { } \to \text{ OH* }\,{ + }\,\text{OH}^{ - } \text{(aq)}$$
$$\text{OH*}\,{ + }\,\text{e}^{ - } { } \to \text{ OH}^{ - } \text{(aq)}\,{ + *}$$

where * is the catalytic site of the catalyst and OOH*, O*, and OH* are the intermediates adsorbed on the catalytic sites. It has been reported that the free energies of all intermediates can be used to evaluate the selectivity and activity of ORR electrocatalysts. Therefore, researchers have evaluated the binding energies of O*, OH*, and OOH* on different metal surfaces by density functional theory (DFT) calculations, and a volcano plot correlating the theoretical ORR activity and the binding energies (ΔGOH) was constructed, as shown in Fig. 3b [34, 35], optimizing ΔGOH is the key to effectively improving the ORR activity [36, 37].

Fig. 3
figure 3

Reproduced with permission from Ref. [40]. Copyright 2016, National Academy of Sciences. b ORR volcano relationship based on metal alloys in the N-doped four-atom vacancy of graphene (denoted as N8V4 (blue) and N6V4 (red)). Reproduced with permission from Ref. [41]. Copyright 2019, American Chemical Society. c The OER mechanism on a layered double hydroxide (LDH) surface in an alkaline electrolyte. Reproduced with permission from Ref. [42]. Copyright 2018, Springer Nature. d OER volcano relationship based on metal oxides. Reproduced with permission from Ref. [34]. Copyright 2017, National Academy of Sciences

a The ORR mechanism of N-doped carbon (NC) catalysts in an alkaline electrolyte.

The OER is the reverse of the ORR and also involves different reactive species in an alkaline medium (OH) [38]. The well-accepted reaction mechanism of the OER under alkaline conditions involves the formation and transformation of OH*, O*, and OOH* intermediates at a catalytic site (M) as well as electron and proton transfer, as shown in Fig. 3c. During the OER, the M–O bonding interactions play a key role in each step. Thus, researchers have related the OER performance of catalysts with a general descriptor, namely, the binding energy difference between O* and OH* (ΔGO − ΔGOH), and a volcano plot between the theoretical overpotential of the OER and the free energy ΔGO − ΔGOH has been constructed (Fig. 3d) [34, 39]. Generally, noble metals and metal oxides of Ru/Ir are the best for the acidic OER, while catalysts derived from transition metals are more favorable for catalyzing the OER in an alkaline medium.

3 Regulation of Oxygen-Catalytic Sites

3.1 Metal-Free Heteroatoms

3.1.1 Heteroatoms for the ORR

In designing oxygen-catalytic sites, metal-free heteroatom (i.e., N, B, P, O, or S)-doping of carbon is often regarded as a promising method to optimize electrocatalytic activity by perturbing the charge/spin distribution of the sp2 carbon plane [43,44,45,46]. In particular, N-doped carbon catalysts are widely studied for use in the ORR because their doped sites have greater electronegativity than the C sites, which can help change the electronic state of adjoining carbon atoms and the chemical adsorption mode of intermediates during the catalytic process. Representative NC-based oxygen electrocatalysts can date back to 2009; the Dai group developed a N-doped carbon nanotube (CNT) as an efficient catalyst for the ORR [47]. Since then, different NC catalysts have been designed by controlling the different types of N species, i.e., graphitic N, pyridinic N, pyrrolic N, oxidic N, triazinic N [48], and the previously reported sp-hybridized N [49], as summarized in Fig. 4a. It has been reported that the ORR catalytic activity of these N species decreases in the order pyridinic N > pyrrolic N > graphitic N > oxidized N > C (carbon) [50]. However, several N species coexist in N-doped carbon catalysts, and the individual contribution to the ORR is difficult to define. To date, various N-doped carbon catalysts, such as N-doped CNT, N-doped graphene (graphene) [51, 52], N-doped graphitic arrays [53], N-doped carbon [54, 55], and carbon nitride (C3N4)-based carbon materials [56, 57], have been extensively investigated for ORR applications. For example, a N-doped carbon nanofiber (N-CNF) with an abundantly N-doped active site and a N atom content of 5.8% showed an onset potential (Eonset) of 0.85 vs. the reversible hydrogen electrode (RHE), a high electron-transfer number (3.97 at 0.8 V vs. RHE), and good durability in an alkaline electrolyte [58].

Fig. 4
figure 4

Copyright 2018, Elsevier B.V. e Linear sweep voltammetry (LSV) plots of NPMC, RuO2, and Pt/C catalysts in 0.1 M (1 M = 1 mol L−1) KOH. f Discharge/charge cycling stability of Zn-air batteries assembled with NPMC-1000 cathodes at 2 mA cm−2. ef Reproduced with permission from Ref. [74]. Copyright 2015, Springer Nature

a The typical doping type of N catalytic centers in N-doped carbon materials [48, 49]. b Comparison of the electronegativity and relative atomic radius of heteroatoms doped in carbon materials [71]. c Schematic diagram of NPS-G. d Calculated adsorption energies of ORR intermediates on NPS-G [69]. cd Reproduced with permission from Ref. [69].

In addition, carbon catalysts codoped with other heteroatoms (B, S, O, P, etc.) with different electronegativities (Fig. 4b), such as N/S- [59], B/N- [60], N/P- [61, 62], N/F- [63], and P/C3N4-codoped carbon catalysts [64], have been further explored and generally exhibited better ORR performance owing to their synergetic effects [65, 66]. In particular, boron carbon nitride (BCN) nanosheets with a large surface area of 817 m2 g−1 displayed a higher Eonset of 0.94 V and better durability than a commercial Pt/C catalyst in an alkaline solution [67]. This outstanding ORR activity may be ascribed to the complementary effects of B/N-codoping and the tunable electronic structure of carbon. In another effort, the S,N-codoped bamboo carbons (SNBCs) displayed a high half-wave potential (E1/2) of 0.850 V vs. RHE and superior stability. A primary Zn-air battery equipped with an SNBC12 cathode presented a high cell performance of 156 mW cm−2 [68]. Furthermore, triply doped carbon catalysts have also been fabricated. A typical example is a N/P/S triply doped graphene-like carbon (NPS-G, Fig. 4c) electrocatalyst with appropriate N, P, and S percentages, which displayed better ORR performance than any singly doped counterparts in an alkaline medium. Such improved ORR performance was derived from the facilitated chemisorption of oxygen-containing intermediates on the heteroatoms of the NPS-G electrocatalysts that decreased the charge transfer resistance (Fig. 4d). Furthermore, the primary Zn-air battery prepared by using an NPS-G cathode presented a large power density and specific capacity [69].

3.1.2 Heteroatom Sites for Bifunctional Oxygen Catalysis

Despite their favorable ORR activity, the application of NC catalysts in rechargeable Zn-air batteries is still hindered by their inactivity for the OER [70, 71]. To date, very few B-doped and P-doped carbon–oxygen catalysts have been reported for rechargeable Zn-air batteries [72, 73]. Moreover, multi-doped carbon materials are promising as bifunctional oxygen catalysts for rechargeable Zn-air batteries. As a representative example, the Dai group fabricated N/P-codoped mesoporous carbon (NPMC-1000) foam with a high surface area (∼1 663 m2 g−1), which exhibited high activities for both the ORR and OER owing to the N/P-codoping, edge effects, and highly porous network (Fig. 4e). The rechargeable Zn-air battery assembled with metal-free NPMC-1000 cathodes cycled steadily for 600 cycles at 2 mA cm−2 (Fig. 4f) [74].

3.2 Metal Single-Atom Sites

3.2.1 Single-Atom Sites for the ORR

In addition to the metal-free heteroatom-containing catalytic centers, carbon-supported single-atom catalysts have emerged as promising ORR catalysts, which are great for Zn-air batteries owing to their high intrinsic activity and maximum atom efficiency [75,76,77,78]. Currently, the most widely investigated and effective single-atom catalysts for Zn-air batteries are Fe [77, 79,80,81,82,83], Co [84,85,86,87], Cu [88, 89], Mn [90, 91], supported on an N-doped carbon substrate. Among them, Fe-N4 [79, 81, 92] and Co-N4 [85, 87], have been found to be the most active ORR centers under alkaline conditions, with ORR activity and Zn-air battery performance comparable or even superior to those of Pt/C. For example, the 3D hierarchically ordered porous NC with abundant FeN4 active sites displayed a more positive E1/2 of 0.875 V than the state-of-the-art Pt/C catalyst (0.845 V) [81]. The primary Zn-air battery equipped with this catalyst as the air cathode showed a higher power density (235 mW cm−2) and specific capacity (~ 768 mAh g−1) at 20 mA cm−2 than those of Pt/C (192 mW cm−2, 737 mAh g−1). Another example is that a N-doped graphitic nanosheet distributed with CoN4 species (CoN4/NG) showed outstanding ORR performance, with an E1/2 of 0.870 V [87]. Simultaneously, CoN4 species can generally achieve an improved OER activity for rechargeable Zn-air batteries. Thus, the CoN4/NG catalyst as an air cathode can be incorporated into a rechargeable Zn-air battery, yielding a specific capacity and an energy density (730 mAh gZn−1 and 671 Wh kg−1) better than those of Pt/C + IrO2-based batteries (711 mAh gZn−1 and 618 Wh kg−1).

Furthermore, the doping of other heteroatoms (i.e., P, S, or O) on M-Nx/C can further improve their ORR activity [75]. As shown in Fig. 5A a–b, a Cu-S1N3 active site stabilized in MOF-derived CNs (S-Cu-ISA/SNC) was identified by aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (AC HAADF-STEM) and Fourier transform extended X-ray absorption fine structure (FT-EXAFS) spectroscopy [89]. The S-Cu-ISA/SNC with low-valent Cu-S1N3 exhibited outstanding ORR activity, with an E1/2 of 0.918 V vs. RHE, outperforming Cu-ISA/SNC (in which S was separated from Cu, 0.87 V) and Cu-ISA/NC (S free, 0.86 V). Thus, the primary Zn-air battery assembled with the S-Cu-ISA/SNC cathode catalyst exhibited a higher power density of 225 mW cm−2 than that of a Pt/C-based catalyst (155 mW cm−2) and a specific capacity as high as 735 mAh g−1 at a discharge of 10 mA cm−2 (Fig. 5A c). DFT calculations and a volcano diagram of the ORR potential and free energy ∆GO* revealed that the Cu-N4 site with weakened bonding for ORR intermediates showed inferior ORR activity, which could be optimized by the substitution of a S atom for coordinated N (Cu-S1N3) because S atoms have larger atomic radii and lower electronegativity than N atoms, causing the Cu in S-Cu-ISA/SNC to lose fewer valence electrons than the Cu in Cu-ISA/NC (Fig. 5A d–f). Recently, a ZnCo with two single atoms bonded to a NC matrix (named Zn/CoN-C) was fabricated and demonstrated a higher E1/2 of 0.861 V than CoN-C (0.793 V) and ZnN-C (0.706 V), as the dual single-atom sites more effectively accelerated the dissociation of O–O bonds to achieve improved ORR activity [93].

Fig. 5
figure 5

Copyright 2020, Springer Nature. B. a SEM and b AC HAADF-STEM images of the Ni-N4/GHSs/Fe-N4 catalyst. EXAFS fitting curves of c Ni-N4/GHSs and d GHSs/Fe-N4 in R space; the insets in (c) and (d) show atomic structure models of Ni-N4/GHSs and GHSs/Fe-N4. e The optimal adsorption configurations of intermediates on Fe-N4 and Ni-N4. f Reaction-free energies of Fe-N4 and Ni-N4 for the ORR and OER, g PDOS of Fe-N4 and Ni-N4 sites. ag Reproduced with permission from Ref. [97]. Copyright 2020, Wiley–VCH GmbH

A. a XANES spectra of S-Cu-ISA/SNC and reference materials at the Cu K-edge. b FT-EXAFS spectra of S-Cu-ISA/SNC and reference materials at the k3-weighted Cu K-edge; the inset shows a schematic of the structure of S-Cu-ISA/SNC. c Discharge polarization plots and power density curves of Zn-air batteries assembled with S-Cu-ISA/SNC and Pt/C cathodes. d The ORR volcano plot and e the free energy diagram of distinct Cu-centered active sites. f The PDOS of Cu and O* after O* adsorption onto the Cu-S1N3 active center. af Reproduced with permission from Ref. [89].

3.2.2 Single-Atom Sites for Bifunctional Oxygen Catalysis

Moreover, the doping of other heteroatoms is expected to improve the OER activity of M-Nx/C, which enables achieving bifunctional application in rechargeable Zn-air batteries [94,95,96]. This conjecture was first confirmed by the introduction of Fe-Nx moieties onto N and S co-decorated CNTs (S, N-Fe/N/C-CNTs), as reported by the Xie group [95]. Compared with N-Fe/N/C-CNT catalysts, S, N-Fe/N/C-CNTs exhibited decreased free energy for the ORR pathway and also greatly improved OER activity. The rechargeable Zn-air battery integrated with the S,N-Fe/N/C-CNT cathode showed high cycling stability over 100 h without obvious voltage gap fading. Recently, bimetallic single-atom site-doped carbon catalysts also showed great potential in enhancing the OER activity to realize rechargeable Zn-air batteries [97,98,99]. To further enhance the OER performance of one catalyst for rechargeable Zn-air batteries, composites of two different single-atom compounds were developed. Chen et al. designed nonadjacent dual single-atom sites (Ni-N4 and Fe-N4) doped graphene hollow nanospheres (Ni-N4/GHSs/Fe-N4) by integrating Fe(II) phthalocyanine (FePc) onto the surface of a Ni-N4 precursor (SiO2@[Ni(CN)4]2−@GO) through π-π stacking, followed by calcination at 700 °C in a N2 atmosphere and subsequent etching (Fig. 5B a–d) [97]. The fabricated Ni-N4/GHSs/Fe-N4 catalyst exhibited excellent bifunctional oxygen-catalytic activity, with a lower Egap of 0.790 V (Egap indicates the potential between the E1/2 of the ORR and the Ej = 10 of the OER). The outer Fe-N4 active centers were mainly conducive to the ORR, and the inner Ni-N4 active centers facilitated the OER, which was further verified by DFT calculations, as shown in Fig. 5B e–g. Therefore, the rechargeable Zn-air battery driven by the Ni-N4/GHSs/Fe-N4 catalyst displayed an exceptional specific capacity of ~ 778 mAh gZn−1, even better than that of a noble metal-based battery (~ 660 mAh gZn−1).

3.3 Metal Alloys and Compounds

3.3.1 Metal Alloys for Bifunctional Oxygen Catalysis

As metal-free and metal single-atom catalysts still show insufficient activity toward the OER, carbon-supported metal compounds have been employed as catalysts to achieve sufficient bifunctional ORR and OER catalytic activity and rechargeable Zn-air batteries [33, 100, 101]. Integrating carbon materials and metal alloys/compounds may offer more active sites, good conductivity, and strong tolerance. Thus, carbon-supported metal alloys [102,103,104] and metal compounds, including oxides [105,106,107,108], sulfides [109,110,111], carbide [112,113,114], nitrides [115, 116], and phosphide [117,118,119], have been widely investigated and displayed good rechargeable Zn-air battery performance. For example, Fe0.5Co0.5 and Fe0.5Co0.4Ni0.1 alloys confined in N-doped CNTs-grafted porous hybrid fibers (FeCo/FeCoNi@NCNTs-HF) showed a low Egap value of 0.747 V. When incorporated into a self-supported flexible rechargeable Zn-air battery, the initial voltage gap between the charge and discharge potentials was 0.76 V. Impressively, the voltage increased by only 0.07 V after a 670 h (1 005 cycles) long-term cycling test [120]. Furthermore, carbides and nitrides as favorable ORR electrocatalysts have also been extensively explored owing to their satisfactory conductivity and corrosion resistance [121]. As shown in Fig. 6A a–d, Mn0.9Fe2.1C embedded in N-doped graphitic carbon (Mn0.9Fe2.1C/NC) displayed encouraging ORR (E1/2 = 0.780 V) and OER (Ej = 10 = 1.644 V vs. RHE) activities. A rechargeable Zn-air battery assembled with a Mn0.9Fe2.1C/NC cathode displayed a high power density of up to 160 mW cm−2 and functioned steadily for more than 100 h, as well as being able to power an electronic clock (rated voltage: 1.5 V) for over 12 h [114]. Recently, the bimetal nitride (Co, Fe)3N_2D exhibited a potential of 0.34 V at –2 mA cm–2 for the ORR and a lower Ej = 10 of 1.540 V vs. RHE for the OER. When incorporated into a rechargeable Zn-air battery, it showed a maximum power density of 234 mW cm−2 and functioned successively for over 300 h at 30 mA cm−2 without obvious discharge–charge voltage gap fading [122].

Fig. 6
figure 6

Copyright 2019, American Chemical Society. B. a Schematic of the synthetic route and b, c TEM images of the CoSx@Cu2MoS4-MoS2/NSG catalyst. d LSV plots of the CoSx@Cu2MoS4-MoS2/NSG catalyst for the ORR and OER in 0.1 M KOH. e Cell voltage of a single Zn-air battery. f Stability testing of Zn-air batteries driven by the CoSx@Cu2MoS4-MoS2/NSG and Pt/C cathode; the inset shows a photograph of a red light-emitting diode (LED, 5 mm, ≈1.8 V) powered by two tandem Zn-air batteries. af Reproduced with permission from Ref. [109]. Copyright 2020, Wiley–VCH GmbH

A. a Schematic of the sample preparation and b TEM image of the 24MnxFe3−xNC catalyst. c ORR and d OER LSV curves of different catalysts in 0.1 M KOH. e Polarization curves and power densities of Zn-air batteries equipped with different catalysts. ae Reproduced with permission from Ref. [114].

3.3.2 Metal Oxides for Bifunctional Oxygen Catalysis

Moreover, an oxide/hydroxide is also a promising alternative to precious-metal-based ORR catalysts, and a series of NC (e.g., graphene and CNT)-based metal oxide (e.g., Co3O4 and CoO) catalysts were reported by the Dai group, all of which displayed ORR activities comparable to that of commercial Pt/C in an alkaline medium [123, 124]. Furthermore, most of these metal oxides/hydroxides demonstrated excellent OER activity; thus, carbon-based oxides have achieved an excellent bifunctional oxygen-catalytic process and rechargeable Zn-air battery performance [125, 126]. Representatively, Xie and coworkers fabricated a N-doped graphene-coated ultrathin cobalt oxide (CoOx) catalyst that showed outstanding bifunctional oxygen-catalytic activity with a low Egap value of 0.704 V. The maximum specific power of the rechargeable Zn-air batteries equipped with this catalyst was as high as 300 W gcat−1 [125]. Compared with metal oxides, metal sulfides normally show improved OER activity due to unsaturated metal centers that can accelerate the adsorption of oxygen-containing intermediates onto catalytic sites [127, 128]. Furthermore, metal phosphides have also been found to exert a favorable effect on oxygen-catalytic activities and durability by influencing the d-band structure of the metal [129]. For example, the CoPx@CNS catalyst exhibited a high E1/2 of 0.760 V for the ORR and a low overpotential of 1.516 V vs. RHE at 50 mA cm−2 for the OER compared to Co@CNS (1.572 V vs. RHE). The rechargeable Zn-air battery equipped with the CoPx@CNS cathode demonstrated a high peak power density of 110 mW cm−2. and good work stability for more than 130 h [130].

3.3.3 Heterojunctions for Bifunctional Oxygen Catalysis

Furthermore, carbon-supported multiple-metal compounds also showed an advantage in synergistically increasing the bifunctional ORR/OER activity and battery performance by combining the advantages of each material [33, 131, 132]. As depicted in Fig. 6B a–d, a molybdenum dichalcogenide/N,S-codoped graphene homogeneously distributed with core@shell nanostructured CoSx@Cu2MoS4 catalysts (CoSx@Cu2MoS4-MoS2/NSG) exhibited an impressive bifunctional oxygen-catalytic activity with an ultra-low Egap of 0.694 V in an alkaline medium. The rechargeable Zn-air battery constructed by using the CoSx@Cu2MoS4-MoS2/NSG cathode displayed an open-circuit voltage of 1.44 V and excellent cyclic stability (Fig. 6B e–f) [109]. In another reported work, the di-cobalt phosphide (Co2P)-cobalt nitride (CoN) core–shell nanostructures with double active sites coated on N-doped CNTs (Co2P/CoN-in-NCNTs) were prepared via straightforward in situ self-template methods and displayed effective Zn-air battery performance, with a prominent working life of more than 96 h [133]. More recently, N,P-co-doped carbon nanofiber (CNF)-coated FeCo alloys and Co2P (FeCo/Co2P@NPCF) compounds displayed simultaneously improved ORR and OER activity (Egap = 0.770 V) relative to their counterparts of CNF-coated FeCo alloys (FeCo@CF, 0.86 V) and N,P-co-doped CNF-coated Co2P (Co2P@NPCF, 0.94 V). When used as an air cathode in a rechargeable Zn-air battery, these catalysts exhibited outstanding cycling stability, and the discharge–charge voltage gap only decreased from 0.83 to 0.80 V after 642 continuous cycles of testing [134]. Consequently, considerable advancements have been made in carbon-based metal compounds for Zn-air batteries, as listed in Table 2. It is concluded that Co is the best choice and displays better activity for Zn-air batteries among all the transition metal elements (such as Fe, Co, Ni, Cu, and Mn) [135,136,137].

Fig. 7
figure 7

Reproduced with permission from Ref. [150]. Copyright 2015, Wiley–VCH GmbH

a Dimensionally controlled synthesis of BODIPY-based conjugated CPs and the corresponding BN-X catalysts. b TEM images of CP-0D, CP-1D, and CP-2D precursors. c LSV curves of various catalysts in 0.1 M KOH. d Polarization curve and the corresponding power density plot of assembled Zn-air batteries based on different catalysts. e Discharge/charge cycling curve of 2DBN-800.

Table 2 Comparison of the ORR/OER and Zn-air battery performance of carbon-based oxygen catalysts with metal-free heteroatoms, single-metal atoms, and metal alloy and compound sites

4 Designing Carbon Supports for Advanced Oxygen Electrodes

4.1 Conventional Carbon Nanomaterials

To date, several commercial carbon substrates, such as carbon black (CB), CNTs, and graphene, have shown great prospects for use in oxygen-catalytic processes owing to their low costs, high surface areas, and good electroconductivity. CNTs feature slender structures that can facilitate the e and ion transfer between an electrode and an electrolyte. Graphene is the most popular nanocatalyst because of its fast charge-transport mobility, unique physicochemical performance, flexible modified surface, and remarkable electrochemical stability [125]. Therefore, these materials are suitable for constructing carbon-based air cathodes in Zn-air batteries as a result of heteroatom doping or providing support for other active compounds [69, 102, 115, 125, 138,139,140,141].

4.1.1 CB-, CNT-, and Graphene-Based Metal-Free Catalysts for the ORR

Among diverse modification methods, the N doping of carbon is the commonest strategy for improving the ORR activity and Zn-air battery performance of carbon nanomaterials [141,142,143,144,145]. For example, by carbonizing freeze-dried gelatin-KB, a metal-free ORR catalyst of Ketjenblack (KB)-integrated gelatin-derived carbon nanosheets (named GK) was fabricated and exhibited enhanced ORR stability because the incorporated pyridinic N and graphitic N species located near the exposed edge positions markedly facilitated the ORR [141]. The primary Zn-air battery assembled with this catalyst displayed a larger peak power density of 193 mW cm−2 than that of a Pt/C-based battery (188 mW cm−2). N-doped graphene and CNTs have also been widely investigated as high-performance ORR catalysts for Zn-air batteries [51, 146,147,148,149]. For example, a defect-rich N-doped graphene nanoribbon was fabricated by chemical oxidation and unzipping of CNTs, followed by NH3 activation at high temperatures. This oxygen catalyst displayed excellent ORR activities, with an E1/2 of 0.808 V in alkaline conditions. A primary Zn-air battery driven by this cathode delivered a maximum power density of 151 mW cm−2 and outstanding cycling stability over 300 cycles, performing much better than a commercial Pt/C + RuO2-based battery (126 mW cm−2) [147].

Furthermore, increasing the heteroatom density further improved the oxygen-catalytic performance of carbon-based oxygen catalysts. As shown in Fig. 7 a–b, the Feng group realized a series of high-density B,N-codoped carbon, including 0D carbon spheres, 1D CNTs, and 2D carbon nanosheets, by pyrolysis of conjugated polymer-functionalized carbons. Among them, the prepared 2D BN-doped carbon nanosheets exhibit the best ORR catalytic performance (E1/2 = 0.810 V, Fig. 7 c). The as-prepared air electrode based on these catalysts showed very stable discharging/charging performance in primary Zn-air batteries (Fig. 7 d–e) [150]. In addition to the codoping of B and N into a carbon matrix, N and S are widely used in bi-heteroatom-doping systems. The Feng group also prepared N,S-codoped porous carbon sheets (N/S-2DPCs) by pyrolysis of S-enriched 2D polymer-functionalized graphene by NH3 treatment. The N/S ratio of N/S-2DPCs could be well tuned by altering the NH3 activation time, and N/S-2DPCs-60 manifested the best ORR activity (E1/2 = 0.750 V) closest to that of the Pt/C benchmark (0.777 V). When employed as an air cathode in a primary Zn-air battery, outstanding cycling stability for more than 12 h at a current density of 20 mA cm−2 was achieved [151].

4.1.2 CB-, CNT-, and Graphene-Based Single-Atom Catalysts for the ORR

Moreover, the decoration of a single-metal atom or metal compounds onto CB, CNT, and graphene catalysts can further improve the ORR activity [80, 152, 153]. For example, a π-conjugated FePc/graphene composite with abundant Fe–N-C centers (pfSAC-Fe–X) was designed by means of a pyrolysis-free synthetic method (Fig. 8A a–b) [80]. In an AC HAADF-STEM image, the coordinated single Fe atoms were observed as bright dots and were uniformly distributed on the conductive graphene matrix (Fig. 8A c–d). The results of DFT calculations suggested that the electrons of graphene tended to attach to the FePc sites, leading to a Fe–C electron path. The oxygen adsorbed only on the Fe atomic sites and provided superior catalytic sites during the ORR. Therefore, the prepared pfSAC-Fe exhibited highly efficient ORR properties, with E1/2 of 0.910 V and a tolerance for methanol (Fig. 8A e–f). The primary Zn-air battery incorporating pfSAC-Fe exhibited a larger specific capacity of 732 mAh gZn−1 and outstanding stability over 300 h (Fig. 8A g). Recently, a hexagonal Fe2N compound loaded on an NC (Fe2N/NC) catalyst was obtained through carbonization of a CB, FeCl3, and melamine mixture at 250 °C, followed by NH3 activation at 700 °C (Fig. 8B a–c). The as-fabricated Fe2N/NC catalyst, composed of a major Fe2N phase and Fe-N4 sites, displayed better ORR activity (E1/2 = 0.910 V) than NC (~ 0.800 V, Fig. 8B d). The peak power density of a primary Zn-air battery equipped with the Fe2N/NC catalyst as the air cathode reached as high as 155 mW cm−2 . (Fig. 8B e). The results of corresponding DFT calculations indicated that the Fe2N(1̅1̅1) structure had a free energy closer to that of an ideal ORR route than the Fe-N4 sites at 0.85 V vs. RHE (Fig. 8B f–g), revealing that the excellent ORR activity of Fe2N/NC could mainly be attributed to the Fe2N structure rather than to the Fe-N4 [115].

Fig. 8
figure 8

Copyright 2020, National Academy of Sciences. B. a Schematic preparation of the Fe2N/NC hybrid and corresponding crystal cell model of hexagonal Fe2N. b SEM image and c Raman spectra of the Fe2N/NC hybrid. d ORR LSV profiles of electrocatalysts in a 1.0 M NaOH solution. e The power density curves of assembled Zn-air batteries. f Free energy diagrams and g the corresponding calculated Uonset. ag Reproduced with permission from Ref. [115]. Copyright 2020, American Chemical Society

A. a Schematic of the preparation process and b HAADF-STEM image of the pfSAC-Fe electrocatalyst. c FT-XANES spectra of pfSAC-Fe at the Fe K-edge and d simulated structures. e Comparison of the E1/2 and kinetic current density at 0.850 V vs. RHE of the catalysts. f CV curves of pfSAC-Fe-0.2 without and with methanol solutions. g Comparison of the specific capacities of the Zn-air batteries. ag Reproduced with permission from Ref. [80].

4.1.3 CB-, CNT-, and Graphene-Based Catalysts for the OER

To date, most of the reported CB-, CNT-, and graphene-based metal-free and single-atom catalysts have only displayed enhanced ORR activities, and good OER performance has rarely been observed. Recently, one study revealed that incorporating electron-withdrawing oxygen atoms into CNTs results in a large positive charge density on adjacent C atoms, resulting in remarkable OOH* adsorption and enhanced OER activity [154]. Therefore, the O/N-codoped CB, CNTs, and graphene substrates have been considered as promising candidates in the literature [154, 155, 156]. For example, bifunctional metal-free electrocatalysts of O-functionalized CNTs (OCNTs) were reported by Lu and coworkers, and they displayed good oxygen-catalytic performance, with an E1/2 of 0.850 V for the ORR and a low Ej = 10 of 1.546 V for the OER. These OCNTs served as efficient cathode catalysts in a rechargeable Zn-air battery and demonstrated a high peak power density of 130 mW cm−2. Impressively, the discharge–charge voltage of an assembled Zn//OCNT battery manifested negligible fading after 340 h (about 2 040 cycles) of continuous work [154]. Similar results were also realized by Co/N/O triply doped graphene (NGM-Co) with abundant single-atom Co-Nx motifs prepared by dispersing single sites through the structural defects formed in situ during carbonization (Fig. 9A a–b). The as-obtained NGM-Co displayed enhanced bifunctional oxygen-catalytic activity and excellent rechargeable Zn-air battery performance. The equipped rechargeable liquid Zn-air battery showed a high specific capacity of up to 749 mAh g−1 and a steady discharge voltage of 1.12 V at 20.0 mA cm−2, which is sufficient to drive a toy vehicle (Fig. 9A c–d). In particular, when this material was incorporated into a flexible rechargeable Zn-air battery, the discharge–charge voltage was stable for over 18 cycles, even under bending (Fig. 9A e–g) [157].

Integrating graphene carbon substrates with other efficient OER species is a highly effective strategy to realize bifunctional oxygen catalysts [108, 124, 125, 157,158,159,160,161,162,163,164,165,166,167]. For example, the Dai group developed a CoO/N-doped CNT (CoO/N-CNT) hybrid for the ORR and NiFe-LDH/CNT for the OER [124]. The rechargeable Zn-air batteries assembled with mixed CoO/N-CNT + NiFe-LDH/CNT catalysts as cathodes displayed excellent charge–discharge stability over 200 cycles. Recently, a rechargeable Zn-air battery driven by N-doped CNTs coated with NiFe particles (NiFe/N-CNT) showed an ultrahigh peak power density of approximately 300 mW cm−2 and long-life rechargeability for more than 300 cycles [161]. To further prolong the working stability of rechargeable Zn-air batteries, a CNT-based bifunctional catalyst composed of atomically dispersed Co–N-C species with ORR activity integrated with NiFe LDHs with OER activity was fabricated and showed an ultra-low Egap of 0.630 V, far exceeding the Pt/C + Ir/C catalyst (0.77 V) and most of the currently reported catalysts (Fig. 9B a–b) [165, 168, 169]. This type of prepared rechargeable Zn-air batteries exhibited not only a large power density of 173 mW cm−2 but also excellent stability (more than 3 600 cycles at 10 mA cm−2 without obvious voltage gap fading) (Fig. 9B c) [162]. Afterward, a N-doped graphene substrate coating with an ultrathin CoOx layer (1 nm-CoOx/N-RGO) was constructed through a facial ligand-assisted pyrolysis strategy. The 1 nm-CoOx/N-RGO achieved excellent bifunctional oxygen-catalytic activities with a low Egap of 0.704 V, owing to its improved e transfer and abundant active sites. When this material was incorporated into flexible rechargeable Zn-air batteries, good stability for more than 10 h was achieved, even upon bending from 0° to 180° [125].

Fig. 9
figure 9

Copyright 2017, Wiley–VCH GmbH. B. a TEM and b AC HAADF-STEM images of CoNC@LDH. c LSV profiles of distinct catalysts in 0.1 M KOH. The inset shows a comparison of Tafel slopes for ORR/OER. d Long-term stability evaluations of rechargeable Zn-air batteries. ad Reproduced with permission from Ref. [162]. Copyright 2021, Wiley–VCH GmbH

A. a HRTEM image and b N 1 s XPS spectra of the NGM-Co catalyst. c Discharge plots of Zn-air batteries assembled with NGM-Co and Pt/C + Ir/C catalysts at 20.0 mA cm−2. d Picture of a toy car powered by one Zn-air battery based on an NGM-Co catalyst. e The open-circuit voltage of a solid Zn-air battery. f Charge and discharge polarization plots. g Galvanostatic discharge–charge cycling profiles at 1.0 mA cm−2 under bending. ag Reproduced with permission from Ref. [157].

4.2 Porous Carbon Supports

As discussed above, the ORR process of the Zn-air battery usually presents sluggish kinetics because of poor mass loading and insufficient air/ion diffusion, and it is reported that the catalytic performance of an air cathode relies not only on the intrinsic catalytic active sites but also on their assembled nanostructures, especially the porous structures [71, 170]. Therefore, in addition to the inherent activities of catalysts, the porous structure design should be considered in the fabrication of high-performance carbon-based oxygen catalysts. Recently, 3D porous carbon-based oxygen catalysts exhibited excellent advantages in Zn-air batteries due to accelerated reactant diffusion in their pores, their enlarged electrode/electrolyte contact area, and their promising potential in reliable, portable, and flexible devices [100]. Generally, the 3D porous structures of carbon-based catalysts incorporate a multilevel porous configuration over one length scale from micropores to mesopores and macropores. The micropores are favored for the exposure of catalytic centers, while the meso- and macropores benefit fluent mass transportation [171]. Therefore, ideal carbon-based oxygen catalysts should possess both mesoporous and macroporous architectures to achieve high oxygen-catalytic efficiency.

4.2.1 3D Graphene Aerogels

To date, most of the reported porous carbon-based oxygen catalysts used in Zn-air batteries were derived from graphene-based aerogels [72, 100, 172], the assembly of low-dimensional carbon materials [173,174,175,176,177,178,179], or the direct synthesis of carbon materials by chemical vapor deposition (CVD) [51, 146, 180, 181]. For instance, a 3D porous graphene aerogel-loaded Ni/MnO (Ni-MnO/rGO) catalyst was prepared by the sol–gel process followed by pyrolysis and displayed bifunctional oxygen-catalytic activity (Egap = 0.820 V) similar to that of Pt/C + RuO2 catalysts (0.770 V). The rechargeable Zn-air battery assembled with this porous catalyst exhibited better cycling stability for more than 100 cycles than a Pt/C + RuO2-assembled Zn-air battery [100]. Recent work showed that the interconnected ordered macropores of carbon matrices further enhanced the mass transport efficiency and resulted in more exposed active sites than micro-/mesoporous carbon. Therefore, an air cathode of hierarchical porous N,P-doped carbon aerogel with ordered macropores and coated FeP/Fe2O3 compounds was fabricated by freeze-drying and subsequent pyrolysis (Fig. 10A a–d) and displayed an E1/2 (0.838 V) very similar to that of commercial Pt/C (0.819 V) for the ORR and a lower onset potential than benchmark RuO2 for the OER (1.430 V) in an alkaline medium [172]. When incorporated into an aqueous rechargeable Zn-air battery, this material exhibited a high peak power density of 130 mW cm−2 and excellent charge–discharge cycle stability for more than 200 h. Encouragingly, based on the alkaline poly(vinyl alcohol) (PVA) gel electrolyte, the peak power density of a freestanding flexible Zn-air battery can also reach up to 40 mW cm−2.

4.2.2 Low-Dimensional Carbon Assembled 3D Structures

A typical example of integrating 1D CNTs with other low-dimensional materials to prepare 3D catalysts is the fabrication of a 3D freestanding film from a Ni-centered hollow MOF/GO precursor (Fig. 10B a–c) by the Wang group [175]. During the pyrolysis process, MOFs were prepared by wrapping a 1D N-CNT carbon shell around Ni nanoparticles cores, followed by connection to 2D rGO to from a heteronanostructure. Their synergistic effect endows the resulting 3D catalyst with superior oxygen-catalytic activities, with an E1/2 of 0.875 V for the ORR and Ej = 10 of 1.490 V for the OER. When this material was applied as the cathode in a rechargeable Zn-air battery, a high peak power density (~ 117 mW cm−2), specific capacity (706 mAh g−1), and energy density (841 Wh kg−1) were obtained.

4.2.3 3D Carbon Synthesized by the CVD Method

Moreover, the CVD strategy has been extensively applied to the controllable preparation of various carbon nanocatalysts, such as CNTs and graphene, by the transport of gaseous precursors followed by deposition on substrates or templates. As shown in Fig. 10C a–b, 3D bicontinuous mesoporous N-doped graphene (N-MG-800) catalysts were prepared by CVD of triethylene glycol in the presence of NH3 and served as metal-free ORR catalysts [146]. N-MG-800 showed a mesopore size of approximately 25 nm (Fig. 10C c) and a high specific surface area of up to 1 015 m2 g−1, endowing N-MG-800 with a high N-doping content of up to 5.83 At% (At% means the atomic percentage), with 3.86 At% of the desired pyridinic N, and excellent ORR activity comparable to that of a Pt/C catalyst. Thus, the assembled primary Zn-air battery displayed a maximum power density of 270 mW cm−2 and good operational durability over 10 h in an alkaline medium.

Fig. 10
figure 10

Copyright 2020, Wiley–VCH GmbH. B. a Synthesis process and b, c cross-sectional SEM images of Ni@N-HCGHF. ac Reproduced with permission from Ref. [175]. Copyright 2020, Wiley–VCH GmbH. C. a Schematic fabrication of N-MG-800 by CVD. b TEM image of N-MG-800. c Pore size distributions of MG and N-MG-600. ad Reproduced with permission from Ref. [146]. Copyright 2018, Wiley–VCH GmbH

A. a Diagram illustrating the preparation of FeP/Fe2O3@NPCA. b, c SEM image, d HAADF-STEM images and energy-dispersive X-ray spectroscopy mapping of FeP/Fe2O3@NPCA. ad Reproduced with permission from Ref. [172].

4.3 Organic Precursor-Derived Carbon

Compared with the conventional commercial carbon support method, the direct carbonization of organic precursors, including metal-free polymers, metal-coordinated hybrids and polymers, and metal–organic frameworks (MOFs), has the advantages of (a) tunable meso-/microporous structures and (b) a wide selection of active metallic sites and doped heteroatoms. In the following sections, these three types of precursor-derived oxygen catalysts for Zn-air batteries are discussed in detail.

4.3.1 Polymer-Derived Metal-Free Carbon

Polymers containing only heteroatoms in their backbones are usually investigated as precursors for heteroatom-doped carbon catalysts. Polymer-derived metal-free carbon–oxygen catalysts with low cost and various nanostructures have attracted increasing attention and are considered promising candidates in air cathodes of Zn-air batteries for the ORR [182,183,184]. For example, chitosan, a linear polysaccharide with a N content of approximately 7%, was used to fabricate NC oxygen catalysts. A representative porous NC catalyst was obtained by molten salt pyrolysis using chitosan as a precursor and exhibited superior ORR activity, with E1/2 of 0.860 V, compared to commercial Pt/C (0.84 V) in an alkaline medium. When this material was applied as an air cathode in a primary Zn-air battery, a maximum power density of 178 mW cm−2 and high charge–discharge cycling stability over 50 h were achieved [182]. Recently, a porous N-doped carbon (NPC) catalyst with a high specific surface area of up to 1 900 m2 g−1 was obtained by carbonizing glucose in the presence of cyanuric acid and ammonia chloride. The resultant NPC-1000 displayed excellent bifunctional ORR and OER activities, with an Egap of 0.720 V, similar to that of Pt/C + RuO2 catalysts (~ 0.80 V), which enabled the assembly of a rechargeable Zn-air battery with NPC-1000. After 300 cycles for 50 h, the potential gap of this battery remained near 0.85 V without obvious change [183].

Furthermore, conductive polymers, such as polyaniline (PANI), polypyrrole (PPy), and polythiophene (PTh), with desired optoelectronic features, have also been regarded as ideal metal-free carbon–oxygen catalyst precursors for the ORR and Zn-air batteries due to their abundant carbon/heteroatom sources, augmented asymmetrical atomic spin density, and possibly higher conductivity than that of traditional polymer-derived carbon materials [185]. Additionally, it is easy to construct porous nanostructures of these polymers, thus facilitating mass transport and air/electrolyte diffusion. The hard-templating method has been widely used to construct porous structures of polymer-derived carbons for the ORR and Zn-air batteries [186,187,188]. As shown in Fig. 11A a–d, through pyrolysis of o-phenylenediamine with SiO2 colloids, followed by etching and NH3 activation, a meso-/micro-PoPD catalyst with a hierarchically porous structure and N doping was fabricated. Impressively, the primary Zn-air battery driven by the meso-/micro-PoPD displayed superior stability for 60 h, making it superior to meso-PoPD and Pt/C [187]. Instead of the hard-templating method, a self-templated aniline-phytic acid composite was used to fabricate a N/P-co-doped mesoporous carbon foam (NPMC)(Fig. 11B a–b). The resultant NPMC-1000 with a large surface area (1 663 m2 g−1) and abundant N/P doping facilitated mass diffusion during catalysis, resulting in excellent ORR activity (E1/2 = 0.850 V) and enhanced OER activity. The rechargeable Zn-air battery based on the NPMC-1000 catalyst achieved encouraging cycling stability for 30 h at 5 mA cm−2 and 180 discharge/charge cycles over a period of 30 h [74].

Recently developed covalent-organic frameworks (COFs) with p-conjugated and precisely customized architectures and adjustable pore sizes have provided a novel perspective to fabricating the porous carbon–oxygen catalysts [189,190,191,192]. For example, a naphthalene-based COF-derived N/S-co-doped porous carbon (CPN-NS) catalyst with a large surface area of 1 116 m2 g−1 was designed (Fig. 11C a–b) and displayed excellent ORR activity, with an E1/2 of 0.868 V [189]. In addition, a N/P-codoped porous carbon catalyst was fabricated by pyrolysis of N/P-containing COF-coated CNTs. The N/P-codoped catalyst showed excellent primary Zn-air battery performance in terms of a high peak power density (255 mW cm−2) and discharge stability over 20 h (Fig. 11C c) [192].

Fig. 11
figure 11

Copyright 2015, Nature Publishing Group. C. a Schematic of the fabrication process and b TEM image of N/P-co-doped porous carbon obtained at 800 °C. c Discharge curves at various current densities. ac Reproduced with permission from Ref. [192]. Copyright 2017, American Chemical Society

A. a Schematic of the fabrication process and b TEM image of meso-/micro-PoPD. c Pore size distribution and d ORR polarization curves of different catalysts. e The stability of meso-/micro-PoPD and Pt/C. ae Reproduced with permission from Ref. [187]. Copyright 2014. Nature Publishing Group. B. a Preparation of the NPMC electrocatalysts. b SEM images of NPMC-1000; the inset shows pictures of the PANI aerogel gel before (left) and after (right) carbonization. ab Reproduced with permission from Ref. [74].

4.3.2 Metal-Coordinated Hybrids and Polymer-Derived Carbon

Although much effort has been made to explore metal-free carbon-based oxygen catalysts, the sluggish kinetics of the oxygen-involving electrochemistry still significantly limit their performance in Zn-air batteries, especially in the OER process. The introduction of metal ions is regarded as an effective strategy to dramatically decrease the kinetic barrier and facilitate electrocatalytic kinetics [193]. In recent years, metal-coordinated hybrids and polymers, which contain well-distributed coordination metal species in polymeric hybrids or frameworks, have been widely used to fabricate M-NC and carbon-supported metal compounds.

In particular, metal–phenolic hybrid-derived catalysts are receiving dramatically increasing attention owing to their strong binding ability to various metal species.[110, 194,195,196,197,198] Tannic acid, a natural organic ligand with abundant phenolic hydroxyl groups, can self-assemble into a cross-linked porous polyphenolic framework and coordinate with different metal ions [199]. For example, a N/S-co-doped carbon aerogel loaded with a Fe–Ni alloy (Fe–Ni ANC@NSCA) was realized by pyrolysis of a chelate of tannic acid, PANI, Fe3+ and Ni2+ ions (Fig. 12A a–c). Benefiting from the synergistic effects of the N4-Fe–O-Ni-N4 sites, the prepared Fe–Ni ANC@NSCA displayed a balance between the adsorption of reactants and the desorption of products. Thus, the Fe–Ni ANC@NSCA showed excellent bifunctional oxygen-catalytic activity, with an Egap of 0.599 V (Fig. 12A d), far exceeding that of the commercial Pt/C + RuO2 catalyst. Accordingly, the rechargeable Zn-air battery based on this catalyst delivered a peak power density of 140 mW cm−2 and extremely good stability over 500 h at 5 mA cm−2 (Fig. 12A e) [195]. Through a one-pot thermolysis reaction of Fe(II) salts, TA, and disodium hydrogen phosphate, a N/P-codoped carbon with well-dispersed metal phosphide nanoparticles (FePx) and Fe–N-C sites (FePx/Fe–N-C/NPC) was constructed. The synergistic effect again boosted the bifunctional oxygen-catalytic performance and resulted in a smaller Egap of 0.695 V. The prepared rechargeable Zn-air battery exhibited cycling stability for 33 h and less voltage gap fading [197].

Fig. 12
figure 12

Reproduced with permission from Ref. [195]. Copyright 2021, Wiley–VCH GmbH. B. a Synthesis of 2D microporous carbon nanosheets. b SEM image of porous carbon sheets from DA-Fe. c SEM and d HAADF-STEM images of meso-/micro-FeCo-Nx-CN. e ORR LSV curves of NFe/CNs-700–800-NH3 and other counterparts. f Comparison of Egap values of different catalysts. g Discharge plots and corresponding power density of Zn-air batteries assembled by using meso-/micro-FeCo-Nx-CN and Pt/C. h Long-time charge and discharge curves of meso-/micro-FeCo-Nx-CN. a, b, e Reproduced with permission from Ref. [200]. Copyright 2017, Wiley–VCH GmbH. c, d, fh Reproduced with permission from Ref. [204]. Copyright 2018, Wiley–VCH GmbH

A. a Schematic of the synthesis route of the Fe–Ni ANC@NSCA catalyst. b LSV curves of Fe–Ni ANC@NSCA and Pt/C + RuO2. c Discharge polarization and power density curves. d Specific capacity plots. e Stability testing curves of Zn-air batteries assembled with Fe–Ni ANC@NSCA and Pt/C + RuO2.

Dopamine (DA), which contains both a rigid catechol group and a flexible methylene amino group, is a more promising candidate for constructing M-NC catalysts [153, 200,201,202,203]. As an example, a layered metal–organic precursor of DA-intercalated iron(III) chloride hexahydrate, which shows a well-defined layered structure and good flexibility/processability, was reported by our group. This novel metal–organic precursor not only can be used to construct M-Nx-doped 2D microporous carbon nanosheets (NFe/CNs-700–800-NH3) but also can be incorporated into other templates to construct hierarchically porous M-Nx-doped carbon catalysts (meso-/micro-FeCo-Nx-CN) (Fig. 12B a–d) [200, 204]. The prepared NFe/CNs-700–800-NH3 catalyst showed excellent ORR properties, with an E1/2 of 0.863 V (Fig. 12B e). Therefore, the assembled primary Zn-air battery exhibited a higher specific capacity of 750 mAh g−1 than a Pt/C battery at a large discharge current density of 100 mA cm−2. Furthermore, the 2D hierarchically porous meso-/micro-FeCo-Nx-CN catalyst showed excellent ORR/OER activity (Egap = 0.780 V) and stability; thus, the assembled rechargeable Zn-air battery presented superior cycle stability during a long-time charge/discharge process at a current density of 20 mA cm−2 (Fig. 12B f–h).

COFs with transition metal-binding groups, such as bipyridyl [19], porphyrin [86, 205, 206], and phthalocyanine [79, 207,208,209,210,211], were also investigated as metal-coordinated polymeric precursors for effective ORR and OER catalysts in Zn-air batteries [212, 213]. For example, a Fe–N-doped mesoporous carbon catalyst was fabricated by carbonizing the SiO2-templated bipyridine-containing COF (mC-TpBpy-Fe) (Fig. 13A a–d). The obtained mC-TpBpy-Fe displayed markedly improved mass transfer efficiency and enhanced accessibility of the active centers, resulting in better ORR activity (E1/2 = 0.845 V) than C-TpBpy-Fe (0.804 V) without any mesoporous structure (Fig. 13A e). A primary Zn-air battery driven by this catalyst exhibited a peak power density of ∼81 mW cm.−2 and discharging stability over 70 h (Fig. 13A f) [19]. A triazine covalent-organic framework (PTCOF) incorporated with Co nanoparticles (CoNP-PTCOF), which showed well-defined active sites and pores, was readily prepared under mild conditions, and such materials with modulated electronic structures have been investigated as bifunctional electrocatalysts for the ORR and OER. A Zn-air battery assembled with bifunctional CoNP-PTCOF exhibited a smaller Egap of 0.830 V than PTCOF (1.12 V) and better durability for 720 cycles than a battery containing commercial Pt/C and RuO2 [214]. A metal (Fe and Ni)-phthalocyanine-based COF has also emerged as a promising precursor for preparing N-doped carbon-coated FeNi catalysts for bifunctional oxygen catalysts (Fig. 13B a, c). The resultant FeNi-COP-800 displayed a lower Egap of 0.827 V than Fe-COP-800 and Ni-COP-800 (Fig. 13B b). Impressively, the rechargeable Zn-air battery prepared with a FeNi-COP-800 cathode showed the best cycling performance with a low voltage change (decreased from the initial 55.8% to 52.9%) after 175 h, where the loss of voltaic efficiency was only 2.9% (Fig. 13B e) [210].

Fig. 13
figure 13

Copyright 2019, American Chemical Society. B. a Schematic diagram of the synthesis of FeNi-COP-800. b LSV curves of different catalysts for the ORR and the OER; the inset shows the corresponding Tafel plots. c TEM image of FeNi-COP-800. d Discharge polarization and power density plots of FeNi-COP-800- and Pt/C + IrO2/C-based Zn-air batteries. e Charge–discharge cycle plots of assembled Zn-air batteries with FeNi-COP-800 and Pt/C + IrO2/C. ae Reproduced with permission from Ref. [210]. Copyright 2019, Elsevier B.V

A. a Schematic illustration of the preparation of mC-TpBpy-Fe. b SEM image of mC-TpBpy-Fe. c Structural diagram of TpBpy exhibiting the bipyridine coordination sites available in the COF backbone. d N2 adsorption–desorption isotherms of catalysts. e Comparison of the E1/2 and the kinetic current of the ORR catalysts at 0.800 V. f Discharge curve of a primary Zn-air battery driven by mC-TpBpy-Fe at a current density of 20 mA cm−2. af Reproduced with permission from Ref. [19].

4.3.3 MOF-Derived Carbon

MOFs are crystalline porous materials consisting of inorganic moieties (metal ions or metal oxo clusters) and organic linkers (rigid, semi-rigid, or flexible ligands) with well-defined structures and large specific surface areas. This unique coordination structure of MOF precursors can not only inhibit the aggregation or formation of large metal nanoparticles, resulting in the uniform distribution of metal doping throughout the carbon-based matrix, but also produce abundant hierarchical pores with the decomposition of organic ligands during pyrolysis. Furthermore, the diversity of metal ions and organic linkers with different organic groups and structures can lead to a variety of MOF-derived nanostructured carbon materials [215]. Because of these advantages, MOF-derived carbon materials as emerging ORR or bifunctional oxygen catalysts have attracted widespread interest for use in Zn-air batteries [130, 136, 216,217,218,219,220,221,222,223]. MOF-Derived Single-Atom-Decorated Carbon

Zeolitic imidazole frameworks (ZIFs) as a representative MOF series were synthesized by strong bonding between imidazole derivatives and transition metals, especially Zn(II) and Co(II) ions [85, 224,225,226,227,228,229,230,231]. Metal-doped ZIF-8 (Zn) precursors are promising candidates for the fabrication of porous NC and single-atom-doped porous NC catalysts since the melting and escape of Zn at high temperatures prevent the agglomeration of other nanoparticles [89, 232,233,234,235,236]. For example, a hierarchical porous (micro-meso-macropore) NC catalyst (N-HPC) with trace Fe was fabricated by pyrolysis of the assembled ZIF-8 and rGO by means of foam copper impregnated with a ferrous acetate solution [234]. Benefitting from the N-doped, hierarchically porous, and noncovalently joined N-deficient/N-rich heterostructures, the N-HPCs manifested improved adsorption for oxygen intermediates and facilitated electron transfer; thus, a high E1/2 of 0.920 V for the ORR in a 0.1 M KOH solution was achieved. A primary Zn-air battery assembled using a N-HPC cathode catalyst showed a high power density of 158 mW cm−2, far exceeding that of the benchmark Pt/C-based battery (108 mW cm−2). Furthermore, a MOF-derived defect-rich NC can be directly applied as a carbon support for loading single atoms. For instance, a single Cu atom-doped defect-rich NC was also realized by pyrolysis of commercial copper(I) oxide (Cu2O) powder with a ZIF-8-derived NC near the melting temperature of Cu2O (1 500 K). As shown in Fig. 14A a–c, the as-synthesized Cu ISAS/NC electrocatalyst presented outstanding ORR performance, with an E1/2 of 0.920 V (Fig. 14A d). When applied as an air cathode in a primary Zn-air battery, it demonstrated an outstanding power density of 280 mW cm−2, greater than that of a Pt/C-based Zn-air battery (200 mW cm−2, Fig. 14A e) [236]. Recently, the introduction of dual-atom sites has proved to be an effective way to further improve oxygen-catalytic activities [237, 238]. For example, 2D ultrathin porous NC nanosheets with adjacent Fe-N4 and Mn-N4 sites (FeMn-DSAC) were converted by molten salt-assisted pyrolysis of Fe/ZIF-8@Mn/GM composites. Such Fe–Mn dual-sites and 2D porous structures displayed outstanding bifunctional ORR and OER performance, with a low Egap of 0.713 V, compared to other structures such as Fe-SAC, Mn-SAC, and even noble metal-based Pt/C + Ir/C (0.724 V). A flexible rechargeable Zn-air battery assembled with this catalyst exhibited a high specific capacity of 734 mAh g−1 at room temperature; surprisingly, the 86% specific capacity could be maintained even at the ultra-low temperature of − 40 °C [91].

Fig. 14
figure 14

Copyright 2021, Wiley–VCH GmbH

A. a The synthetic procedure of isolated Cu active centers, b AC HAADF-STEM image of the Cu ISAS/N–C catalyst, the single Cu atoms are marked with yellow circles. c FT-EXAFS spectra in k space of the Cu ISAS/N–C catalyst. d E1/2 and kinetic current density (jk) of different catalysts. e Polarization and power density profiles. f Free energy diagram of three desired models (inset) for the ORR. af Reproduced with permission from Ref. [236]. Copyright 2019, Nature Publishing Group. B. a Schematic diagram of the fabrication process and b TEM image of FeCo-NPs/NC. c Durability of FeCo-NPs/NC and Pt/C for ORR. d Comparisons of E1/2 and Ej = 10 of various catalysts. e Illustration of the adsorption of oxygen intermediates on FeCo-NPs/NC. f Free energy plots of the ORR at U = 0 and 1.230 V. af Reproduced with permission from Ref. [243]. MOF-Derived Metal Compound-Decorated Carbon

Moreover, MOFs can also be used to fabricate carbon-based compounds with bifunctional oxygen-catalytic activity for use in rechargeable Zn-air batteries [136, 217, 239,240,241,242,243]. For example, a kind of double-shell nanocages (NC@Co-NGC DSNCs) with NC as the inner shell and Co–N-doped graphitic carbon as the outer shell was prepared through the pyrolysis of core–shell ZIF-8@ZIF-67. The as-prepared NC@Co-NGC DSNCs showed a lower Egap of 0.820 V than single-shell nanocages (Co-NGC DSNCs) in alkaline electrolytes. A rechargeable Zn-air battery driven by these catalysts exhibited an excellent charge and discharge stability over 56 h at a current density of 10 mA cm−2 [240]. Later, the Chen group prepared NC nanorods loaded with a Fe–Ni-Co metal/metal phosphide catalyst (FeNiCo@NC-P) by pyrolysis of the dual-MOF complex Fe2Ni_MIL-88@ZnCo-ZIF, followed by phosphatization. The hierarchically porous FeNiCo@NC-P exhibited a lower Egap of 0.700 V than its counterparts FeNi@C-P and Co@NC-P; therefore, the assembled rechargeable Zn-air battery realized long-time stability for 130 h without obvious voltage gap fading at a current density of 10 mA cm–2 [217]. Recently, a series of bimetallic compounds (e.g., FeCo, FeNi, and CoNi) encapsulated in hollow NC nanocubes were fabricated by pyrolysis of DA-coated MOFs in an NH3 atmosphere (Fig. 14B a–c). In particular, the FeCo-NPs/NC composite displayed excellent bifunctional ORR and OER activities, with an ultra-low Egap of 0.670 V (Fig. 14B d). When incorporated into an all-solid-state rechargeable Zn-air battery, this composite exhibited a large power density (~ 77 mW cm−2) and excellent cycling stability for 32 h, with little potential gap fading at a current density of 1 mA cm−2, even when bent from 0° to 180° [243]. MOF-Derived 3D Porous Carbon

To further enhance the mass diffusion properties of MOF-derived carbon catalysts, a 3D macroporous nanostructure was designed by pyrolysis of SiO2- or PS-templated MOFs [81, 225, 244, 245]. For example, ordered macroporous carbon (ZOMC) with abundant Co–N-C sites prepared from a ZIF-67-coated silica assembly displayed better bifunctional oxygen-catalytic activity than the benchmark Pt/C catalyst [225]. Thus, a rechargeable Zn-air battery equipped with the ZOMC cathode achieved a high power density of up to 221 mW cm−2 and operated for more than 160 h. A trimodal-porous carbon dispersed with Fe and Ni sites (Fe/Ni-Nx/OC) was synthesized by combining PS templates with Fe,Ni-doped-ZIF-8 ((Fe, Ni)-ZiF-8@PS) [245]. This unique porous configuration and the coexisting Fe-N4 and Ni-N4 active centers endowed Fe/Ni-Nx/OC with exceptional ORR activity (E1/2 = 0.938 V, Fig. 15A. When applied as an air cathode in a primary Zn-air battery, Fe/Ni-Nx/OC displayed a large peak power density of 148 mW cm−2. Furthermore, the self-templating method is another important way to construct 3D micro-/mesoporous carbon-based oxygen catalysts via the in situ pyrolysis of organics and the release of small molecules [239, 246]. As an example, the Xu group obtained 3D ordered open carbon cages with micro-/mesopores via direct pyrolysis of the morphology-controlled Fe(III)-doped core@shell Zn@Co-MOFs precursor (Fig. 15B a–c). Such open carbon cages endowed the resultant CoFe20@CC with superior bifunctional catalytic activities for the ORR and OER, comparable to those of noble metal-based RuO2 and Pt/C catalysts. When this material was applied as a cathode in rechargeable Zn-air batteries, a substantially elongated lifetime for 400 cycles (130 h) was achieved, and a negligible voltage loss was observed (Fig. 15B d) [239].

Fig. 15
figure 15

Copyright 2020, Wiley–VCH GmbH. B. a Schematic diagram of the fabrication of porous CN-based oxygen catalysts through the self-template pyrolysis strategy. b Pore size distributions of different precursors. c SEM image of CoFe20@CC. d Discharge–charge cycling test of CoFe20@CC and Pt/C-IrO2 at 5 mA cm−2. ad Reproduced with permission from Ref. [239]. Copyright 2019, Wiley–VCH GmbH

A. a Schematic diagram of the synthesis process, b pore size distributions (top) and SEM images (bottom), c HAADF-STEM image of Fe/Ni-Nx/OC. FT-EXAFS spectra of Fe/Ni-Nx/OC and reference materials at d the Fe K-edge and e the Ni K-edge. f ORR LSV curves of different catalysts in 0.1 M KOH. af Reproduced with permission from Ref. [245].

Based on the above discussion, considerable progress has been made in the design of carbon supports for advanced oxygen electrodes, as listed in Table 3. Conventional carbon nanomaterials exhibit excellent ORR activities and stabilities when incorporated into primary Zn-air batteries. Surface functionalization, heteroatom doping, and defect engineering are commonly used strategies to augment the oxygen-catalytic activities and realize rechargeable performance. Although most of the porous carbon supports and organic precursor-derived carbon electrodes have realized satisfactory bifunctional oxygen activities and rechargeable performance, there are still challenges in achieving a high energy density and durability for practical applications. Therefore, the development of high-power Zn-air batteries for practical applications will be discussed in the following section.

Table 3 ORR/OER and Zn-air battery performance of electrocatalysts with different carbon supports

5 Representative Applications of Nanostructured Carbon-Based Electrodes for Zn-Air Batteries

5.1 Flexible Zn-Air Batteries

Over the years, flexible Zn-air batteries have received increased attention in wearable electronics due to their large theoretical energy density, aesthetic versatility, and easy incorporation into irregular geometric surfaces [247, 248]. Although the preparation of diverse advanced flexible Zn-air batteries has been attempted, along with consecutive promotion in design strategies and assembly technologies [249], the development of specifically functionalized electrodes that exhibit a combination of superior flexibility, safety, comfort, and excellent battery performance remains a considerable challenge [249]. The first obstacle hindering the preparation of flexible Zn-air batteries is achieving air cathode flexibility. Accordingly, carbon-based cathodes with good conductivity, stability, and easily constructed porous structures, such as CNT paper, CNF cloth, graphene or porous graphene film, and carbon foam, have shown promise in applications in flexible Zn-air batteries [72, 180, 250, 251]. Currently, the available flexible carbon-based air cathodes have been mainly constructed by two methods: (i) spraying technique (spraying electrocatalyst ink onto flexible porous current collectors) and (ii) in situ fabrication (in situ preparation of electrocatalysts/current collectors incorporated into free-supporting electrodes).

Spraying is a common method to design flexible electrodes coated with the desired electrocatalysts as long as the electrocatalyst powder can be prepared. Various impressive works related to this method have been reported. Recently, a fabricated N-doped CNT/rGO composite prepared by pyrolyzing MOFs grown on rGO nanosheets exhibited excellent ORR propertied with an E1/2 of 0.850 V, which is very similar to that of noble metal-based catalysts [252]. After being sprayed onto a carbon cloth, the N-CNT/rGO catalyst was applied to construct an excellent air cathode for use in flexible primary Zn-air batteries. Impressively, when the resulting battery was studied in various bending states, the loaded battery still displayed a small voltage gap and steady charge/discharge potentials. In another representative work, a 2D multilayer mesoporous Co3O4/N-rGO catalyst was fabricated via facile solvothermal and oxidation treatment and displayed excellent electrochemical performance with a lower Egap value of 0.930 V than that of a commercial Pt/C catalyst (1.16 V) [126]. As shown in Fig. 16A, after being sprayed onto Toray carbon fibers (the red circle), this Co3O4/N-rGO was knitted into a fiber-shaped rechargeable Zn-air battery with a desirable length and woven into clothing to power various portable products, such as smart glass, wearable sensors, and electronic devices. Nevertheless, the flexibility of oxygen electrodes obtained via spraying mainly depends on the flexibility of the conductive substrates, of which CNTs and graphene showed more favorable deformability than CNF cloth. Furthermore, certifying the strong interaction between the electrocatalyst and current collector is a considerable challenge during the spraying technique. Moreover, the usage of conductive substrates improves the total weight. All of these factors are critical to improving the cycling stability, rate capability, and energy density of the battery. Therefore, developing another synthetic method to maintain both satisfactory interactions and flexibility is necessary.

Fig. 16
figure 16

Copyright 2018, Wiley–VCH GmbH. B. a Preparation, b assembly, and c discharge plots of a flexible Zn-air battery. ac Reproduced with permission from Ref. [253]. Copyright 2020, Wiley–VCH GmbH. C. a Schematic of the material configuration and b charge/discharge cycling curve of a flexible Zn-air battery. ab Reproduced with permission from Ref. [172]. Copyright 2020, Wiley–VCH GmbH. D. a Polarization and power density plots and b the corresponding cycling profile of stretchable Zn-air batteries. ab Reproduced with permission from Ref. [254]. Copyright 2019, Wiley–VCH GmbH. E. a Preparation and b power density retention (%) of a squeezable Zn-air battery. ab Reproduced with permission from Ref. [251]. Copyright 2019, Elsevier B.V

A. a Preparation of a flexible fiber-shaped Zn-air battery. b Photographs of an LED watch and c iPhone 4 s powered by this battery. ac Reproduced with permission from Ref. [126].

The in situ construction strategy of free-supporting electrocatalysts aims to solve the above-mentioned issues inherent to the spraying technique. Generally, the construction of free-supporting electrocatalyst involves an elaborate configuration and comparatively complicated post-processing. According to the literature to date, freestanding air cathodes with high catalytic sites and unique microstructures can be synthesized via three general methods or their combinations: the hydrothermal growth, solvothermal method, electrospinning, or CVD [12]. The resulting composite free-supporting cathode electrocatalyst was fairly flexible and exhibited satisfactory catalytic activities even under various deformation conditions. A typical example is that a flexible freestanding air cathode of oxygenated cobalt vanadium selenide (O-CoxV1−xSe2) grew directly on carbon fiber cloth (CFC) via hydrothermal-assisted selenization (Fig. 16B a). In addition, a CFC-based anode composed of porous 3D ZnO@C-Zn nanorods was also obtained. When poly(vinyl alcohol) (PVA)/KOH was utilized as the gel electrolyte, the fabricated flexible rechargeable Zn-air battery (Fig. 16B b) displayed an excellent specific capacity of 775 mAh gZn−1, far exceeding that of a Pt/C-RuO2 assembled Zn-air battery (Fig. 16B c) [253]. Furthermore, CNFs covered with single Co (or Ni) atoms and NC flakes (named M SA@NCF/CNF) were fabricated by self-assembled growth of ZIF on polyacrylonitrile (PAN) nanofibers and subsequent pyrolysis and etching. The assembled flexible rechargeable Zn-air battery displayed outstanding stability under different deformation states could supply power for three light-emitting diodes (LEDs) simultaneously [250]. However, the extremely compact fibrous configuration led to decreased air/electrolyte penetrability and incompetent air diffusion for ideal battery applications. Simultaneously, the loading contents of the electrocatalysts were also restricted by the small specific surface area of the carbon cloth. Currently, these restrictions can be alleviated by constructing 3D porous carbon aerogels through directional freeze-casting and subsequent pyrolysis, promoting the catalysis reactions at the interface and resulting higher electroconductibility by the absence of polymer additives [172]. For example, a flexible rechargeable Zn-air battery driven by FeP/Fe2O3 nanoparticle-decorated N,P-doped carbon aerogel catalyst cathodes showed a remarkable specific capacity and cyclability under various deformation conditions (Fig. 16C) [172].

In addition, stretchability is another intrinsic feature of flexible Zn-air batteries that can be achieved by designing both stretchable electrodes and gel polymer electrolytes (GPEs). Based on a metal Zn spring, a CNT paper-loaded catalyst, and 1 200% stretchable GPEs, a stretchable fiber Zn-air battery was constructed and showed excellent waterproofness, ambient adaptive capacity, and catalytic performance, even under 500% stretching (Fig. 16D) [254]. On the other hand, compressible performance was also achieved in a carbon-based rechargeable Zn-air battery, and displayed a tiny alteration in power density which was observed after 500 deformation cycles (Fig. 16E) [251]. The cathode and anode were, respectively, fabricated by electroplating Fe-Co3O4 nanowires and Zn nanosheets on 3D porous N-doped carbon foam (NCF).

5.2 High-Power Zn-Air Batteries

Zn-air batteries have shown promising potential as alternatives to energy storage devices; some basic parameters are used to estimate the properties of energy storage systems, such as power density (mW cm−2), specific capacity (mAh gZn−1), cycle life, safety, and cost. There are two major weaknesses of Zn-air batteries: the short life span when recharged electrically and Zn anode dendrite generation resulting in short circuits and shedding of Zn. However, these factors do not affect primary Zn-air batteries because their Zn anodes can be mechanically replaced by new Zn metal. In 1999, the Electric Fuel Ltd. Co. developed a mechanically rechargeable Zn-air battery for military use and a fleet of electric vehicles [255]. Thereafter, various studies have focused on how to reduce the high overpotential in the cathode reaction and offer a high power density by constructing novel catalysts and optimizing air electrode structures, and the previous sections have discussed many carbon-based oxygen electrodes for Zn-air batteries. In particular, 2D Ni-doped CoO nanosheets with abundant micro-/nanostructures exhibit favorable O2 diffusion, enhanced catalytic activation, and improved native ORR performance at the catalytic center by Ni doping. Integration of CB substrates as air cathodes in primary Zn-air batteries resulted in an ultrahigh power density (377 mW cm−2, Fig. 17A a), and integrated all-solid-state coin cells can be used to power an iPhone 7 mobile phone (Fig. 17A b–d) [256].

Fig. 17
figure 17

Copyright 2018, Wiley–VCH GmbH. B. a Top: schematic of the structure of the cable and sandwich forms of Zn-air batteries using PVAA-GO GPE. Bottom: corresponding photographs of connected cells used to power a commercial smartwatch. b Comparison of the cycle time and charge potential of the sandwich form of KI-PVAAGO-based Zn-air batteries with other batteries. ab Reproduced with permission from Ref. [257]. Copyright 2020, Wiley–VCH GmbH. C. a Schematic diagram of the reaction procedures in the IHL and the outer Helmholtz layer (OHL). b Galvanostatic discharge plots of Zn-air batteries, different electrolytes, and the corresponding Zn utilization ratios (ZURs). The insets show photographs of a pristine Zn anode and after discharge in distinct electrolytes. c Storage of Zn-air batteries using distinct electrolytes. ac Reproduced with permission from Ref. [4]. Copyright 2021, National Academy of Sciences

A. a Polarization curves and power density plots of integrated batteries using different catalysts as air electrodes. b Schematic of the structure and photographs of c a button-like Zn-air battery and d an iPhone 7 powered by 4 series-connected cells. ad Reproduced with permission from Ref. [256].

Although considerable effort has been made in developing cathode catalysts, the electrolytes are also bottleneck factors that highly restrict the power density and working life of Zn-air batteries. Optimizing Zn-air batteries as a whole is essential for further performance improvement, and some researchers have always been devoted to developing novel electrolytes. For example, Song and coworkers produced a kind of GPE gel electrolyte via multistep cross-linking reactions involving PVA, GO, and poly(acrylic acid) [257]. The resultant PVAA-GO GPE displayed much greater performance than previously reported PVA electrolytes in terms of ionic conductivity, mechanical strength, and water-retaining properties, which endows rechargeable Zn-air batteries with a minimal charge potential of 1.69 V, a superb energy efficiency of up to 73%, and a life-span working stability over 200 h, even under various severe working conditions. For example, a series of cable and sandwich forms of Zn-air batteries have been, respectively, incorporated into clothing for charging of smartwatches (Fig. 17B), and a sandwich from of rechargeable Zn-air batteries has been incorporated into commercial electroluminescence devices.

The above-mentioned endeavors to promote the performance of Zn-air batteries have concentrated on designing bifunctional cathode catalysts or increasing the Zn anode lifetime via the development of improved electrode structures or electrolyte additives. However, formidable challenges remain in the preparation of secondary Zn-air batteries. Recent studies have reported that near-neutral electrolytes could inhibit the generation of Zn dendrite and carbonates [258, 259]. Thus, based on a CB cathode, a Zn metal anode, and a nonalkaline zinc trifluoromethanesulfonate (Zn(OTf)2) electrolyte, Sun and coworkers developed a Zn-O2/ZnO2 complex that underwent a 2e/O2 pathway as a result of its water-poor and Zn2+-rich inner Helmholtz layer (IHL) on the air cathode induced by the hydrophobic trifluoromethanesulfonate anions (OTf) (Fig. 17C a). This endowed the assembled rechargeable Zn-air full battery with an efficiently reversible redox reaction process, and the battery exhibited outstanding cycling properties in ambient air despite its simple configuration (Fig. 17C b–c). This different interface structure regulated by electrolyte features blazed a new path for electrochemical irreversibility and allowed the realization of secondary Zn-air batteries [4].

6 Conclusions and Perspectives

Nanostructured carbon-based electrodes with different oxygen-catalytic sites and carbon supports appear to be promising catalysts for efficient Zn-air batteries because of their many advantages compared to traditional metal/metal oxide and noble metal-based catalysts, such as larger surface areas, efficient oxygen-catalytic centers, hierarchically porous matrices, superior durability, and low cost. In this timely review, we have provided a comprehensive summary of the recent advances in synthesizing nanostructured carbon-based electrodes to prepare efficient Zn-air batteries, especially in tailoring the oxygen-catalytic sites and designing carbon supports. In this review, we focus on versatile synthetic strategies, characterization methods, and an in-depth understanding of the relationships between oxygen-catalytic sites/nanostructures and the oxygen electrode performance. In addition, we have highlighted the current development of flexible and high-power Zn-air batteries to illustrate their application potential in exploiting scalable, bendable, and wearable power supplies. Finally, a thorough discussion of current primary challenges and future perspectives on the rational design of nanostructured carbon-based oxygen electrodes has been given; we believe that the investigation of oxygen-catalytic carbon electrodes for Zn-air batteries is still in an early stage, and enormous challenges must be solved.

  1. (a)

    As for Carbon-Based Oxygen Electrodes

The synthetic cost of carbon electrodes is a critical issue for commercial use. Based on the calculation in an earlier review, carbon nanomaterials (i.e., CBs, CNTs, graphene) and related catalysts exhibit higher conductivity and a lower cost per mass and volume than many catalysts [260]. However, further optimization of the catalytic performance by introducing heteroatoms, porosity, or porous structures would increase the cost of such carbon catalysts, which may increase the final price to even more than that of commercial Pt/C. Therefore, more effort should be devoted to decreasing the cost of carbon nanomaterial-based electrocatalysts. As for the organic precursor (i.e., metal-free polymer, metal-coordinated hybrid and polymer, and MOF)-derived carbon catalysts, the high cost of the organic molecules and the synthetic processes may also be substantial barriers preventing their commercialization. On the other hand, pyrolysis of precursors to afford large-scale and low-cost carbon-based catalysts is also necessary for practical applications of carbon electrodes.

In addition to considering the scale-up production, the well-defined oxygen-catalytic sites, controllable morphology, and optimized porosities should also be taken into account to balance the performance of electrical conductivity, cycling stability, and mass/air transport. Generally, carbon supports with high-density functional groups or edge sites can endow the anchored catalytic sites with better dispersion and smaller sizes. Furthermore, carbon with different porosities can increase the number of catalytic centers to improve the electrocatalytic activities of carbon electrodes. Further tuning of the configuration of carbon, the heteroatom-doping level, and the degree of graphitization of carbon electrodes is essential for increasing the intrinsic activity of catalytic sites and improving the transport issues. However, the controlled engineering of catalytic sites in carbon electrodes does not seem to be an easy task to achieve. Therefore, organic precursors with well-defined and tailorable structures, particularly those with N/P/S-doping, constitute a powerful toolbox to design tailor-made carbon electrodes for Zn-air batteries. Furthermore, an in-depth understanding of structure-performance relationships involved in Zn-air batteries has become feasible thanks to the elaborate structural control achieved in recent sophisticated studies.

For flexible electrodes, many of the currently synthesized carbon electrodes show promising potential for the fabrication of flexible Zn-air batteries in the form of cable or sandwich structures. However, more effort should be committed to improving the Zn-air batteries’ properties in terms of energy density, cycling life, and fast charging, as well as the ionic conductivity of flexible carbon electrodes. In addition, enormous challenges related to the low-cost, lightweight, and flexible carbon electrodes that can be applied in flexible and wearable commercial devices on a large scale remain to be addressed.

  1. (b)

    Concerning the Development of Electrolytes

The oxygen catalysis process occurs at the solid (active catalysts)–liquid (electrolytes)–gas (O2) three-phase interface; thus, the mutual contraction of electrolytes with O2 dissolved on the surface of the electrode is closely related to the efficiency of the reaction. Furthermore, the wettability of the electrode can affect the contact between the electrolyte and the electrode. Therefore, in addition to regulating oxygen-catalytic sites and designing nanostructured carbon-based electrodes, the properties of electrolytes should also be recognized as key points for future Zn-air battery fabrication. Generally, the widely researched alkaline liquid electrolytes possess the advantages of low toxicity, high ionic conductivity, nonflammability, and low cost. However, the problems associated with corrosion, evaporation, leakage, a narrow electrochemical stability window, and restrictions in miniaturization cannot be ignored. Moreover, carbonate precipitation (e.g., K2CO3 or KHCO3) and the generation of Zn dendrites are expected to decrease battery capacity. An increasing number of studies have reported that near-neutral electrolytes can inhibit the generation of Zn dendrites and that carbonates show better stability and reversibility in ambient environments [4, 258, 259]. Therefore, near-neutral electrolytes could be one of the promising future research directions.

Polymer electrolytes, including gel polymer, solid polymer, and composite polymer electrolytes, exhibit good film-forming features, excellent mechanical strength, and desired flexibility and are favorable for future devices [261, 262]. Gel polymers are widely used in flexible Zn-air batteries because of their high conductivity and simple preparation. Nevertheless, compared to liquid Zn-air batteries, flexible Zn-air batteries still exhibit a relatively low battery performance due to the high ionic resistance of the gel electrolytes. Furthermore, dehydration is a serious problem for gel electrolytes, reducing their ability to stay wet for a long time and thus decreasing their chemical durability, ion conductivity, and performance [23, 263]. Therefore, promising gel electrolytes would require good water retention, high ionic conductivity, and good stability. Currently, much effort is being devoted to these topics, and we believe that developing gel electrolytes could stimulate the commercialization of flexible Zn-air batteries.

  1. (c)

    In Terms of Battery Configuration

The irreversible deposition and accumulation of ZnO, the formation of Zn dendrites, and the corrosion of high-concentration electrolytes to electrodes are key problems affecting practical applications by reducing cycling stability. To solve these problems, modifications of the composition and surface of Zn electrodes have been suggested as feasible strategies. In addition to the above problems, electrolyte leakage and inconvenient transportation are also critical problems hindering the practical application of liquid Zn-air batteries. Therefore, effort should be devoted to researching flexible Zn-air batteries, especially flexible collectors, flexible electrolytes, and packaging methods. Furthermore, Zn-air batteries currently show excellent application prospects in the grid energy storage, automotive and aerospace, and miniaturized equipment fields; thus, developing low-temperature-resistant batteries that work under extreme conditions should be another research direction of Zn-air batteries in the future.

In conclusion, we believe that this review provides new insights into the design of oxygen-catalytic carbon electrodes and paves the way to engineer efficient, robust, and flexible Zn-air batteries to further the development and prosperity of this emerging field. This review also provides perspectives on a powerful toolbox to design tailor-made carbon electrodes by engineering organic precursors with well-defined structures on the atomic, molecular, and nanoscale levels. More importantly, this review could help readers gain better insight into the advantages, key problems, and further promising solutions of carbon-based oxygen catalysts for Zn-air batteries. Looking into the future, along with the fundamental knowledge breakthroughs in the fabrication of carbon-based catalysts and the technological evolutions of Zn-air batteries devices, we believe that the large-scale production and commercial use of oxygen-catalytic carbon electrodes in Zn-air battery devices will not be far from realization.