Introduction

Zn-air batteries (ZABs) are considered as one of the most viable options to next-generation energy storage system because of their high theoretical energy density (1086 W·h/Kg) and low cost [1,2,3]. To achieve the high performance of ZABs, the noble metal catalysts (such as Pt-based catalysts) are often used for accelerating the slow oxygen reduction reaction (ORR) at air electrode [4, 5]. Recently, heteroatom-doped (N, S, B, P, etc.) metal-free carbon catalysts have been generally considered as the most promising alternatives to replace the air electrode of noble metal Pt-based catalysts, which resolves the scarcity and expensive bottlenecks [6, 7].

Heteroatom doping, such as N, S, etc., can manipulate the absorption of O2 and improve ORR kinetics through the redistribution of charge density of adjacent C atoms. Nevertheless, the current study of heteroatom-doped carbon suffers from low doping contents and unstable configuration of heteroatoms obtained by the post-doping. It is of great significance to select low-cost, natural treasure-waste biomass as carbon precursor and N, S sources, to achieve heteroatom self-doping [8,9,10,11,12]. Additionally, the discrete distribution of active sites in the N, S co-doped carbon catalysts brings the requirement of hierarchically porous structure, which ensures an efficient mass transfer [13, 14]. Templates are commonly employed to regulate the porosity, including acid oxidation (H3PO4, HNO3, H2O2, etc.) [15], alkali oxidation (KOH, NaOH, etc.) [16], and metal salt activation (ZnCl2, KCH3COO, NaNH2, KHCO3, etc.) [3, 17,18,19,20,21]. However, the active sites in the micropore channels formed by a single template to prepare biomass-derived carbon materials would be covered by other reaction intermediates and eventually lead to rapid deactivation [5, 22, 23]. Therefore, the hard-soft template (SiO2-ZnCl2) strategy is employed to satisfy accessible catalytically active sites, shorten the diffusion pathways of reactants and products, and increase the specific surface area [20, 24,25,26,27,28,29,30]. On the basis of the covered active sites and the unstable configuration of heteroatoms, the waste biomass is applied as carbon precursors and heteroatoms sources to realize heteroatoms self-doping; meanwhile, the hard-soft template strategy is employed to optimize pore structure features.

Hence, N, S co-doped porous carbon materials have been synthesized by annealing the mixtures of Pointed tail leaf (carbon precursor and N, S sources) and templates (SiO2-ZnCl2) through the hard-soft template approach. The as-prepared material has the unique mesopore-dominated hierarchical porous structure, high specific surface area, and high N, S doping content. The optimal catalyst (NSC-PT-0.1) exhibits a comparable ORR electrocatalytic activity with acid resistance (E1/2 = 0.69 V) and basic resistance (E1/2 = 0.82 V), while the stability and methanol resistance are superior to Pt/C. Furthermore, the assembled ZAB using NSC-PT-0.1 as air electrode reveals a higher peak power density of 155.55 mW cm−2 and better cyclic stability (216 h) than Pt/C-based ZAB. Inspired by above performance, we have prepared many other mesopore-dominated hierarchical heteroatom-doped carbon materials by pyrolysis of the mixture of waste biomasses (Carrot, Whorled stonecrop, and Ginkgo leaf) and the optimal template dosages. This study not only develops a promising hard-soft template strategy for preparing economical and efficient catalysts for ZABs but also provides attractive solutions tackling green technology challenges of severe environmental issues and energy crises.

Experimental section

Materials

Silica (SiO2, ~ 30 nm), zinc chloride (ZnCl2, 98%), potassium hydroxide (KOH, 85%), zinc acetate (Zn(Ac)2, AR), sulfuric acid (H2SO4, 98%), hydrofluoric acid (HF, 40%), isopropyl alcohol ((CH3)2CHOH, AR), and polytetrafluoroethylene (PTFE, 60 wt%) were purchased from Aladdin Reagent Co., Ltd. Commercial Pt/C (20 wt%) electrocatalyst and Nafion solution (5 wt%) were provided by Shanghai Hesen Electric Co., Ltd. Four kinds of natural waste biomass (the leaf of Pointed tail, Carrot, Whorled stonecrop, and Ginkgo) were purchased from a vegetable market in Xinxiang and washed for further application. All the chemicals were used directly without further purification.

Preparation of N, S co-doped hierarchically porous carbon materials

The dried power of Pointed tail leaf (1 g) and SiO2 (x g, x = 0.04, 0.07, 0.1, 0.5, and 1 g, respectively) was firstly mixed and thoroughly ground with ZnCl2 (0.682 g) in an agate mortar for 30 min. The precursors were pyrolyzed in a tube furnace in N2 gas flow at 300 °C for 2 h and continuously heated to 900 °C for 2 h. Subsequently, the pyrolyzed product was etched with 15% HF solution for 24 h to remove the hard templates (SiO2) and other impurities from the pyrolysis of soft templates (ZnCl2). Finally, the product was washed with a large amount of deionized water, filtered, and placed in a blast drying oven at 80 °C for 12 h, to obtain N, S co-doped porous carbon materials named NSC-PT-x. Two control samples, NSC-PT-Z (ZnCl2) and NSC-PT–S (SiO2), were synthesized following the same procedure.

To exploit the universality of soft-hard templates, the Carrot (i.e., CL), Whorled stonecrop (i.e., WS), and Ginkgo leaf (i.e., GL) were employed as carbon precursors and N, S sources to achieve the N, S co-doped hierarchically porous carbon catalysts. The optimal hard-soft templates were employed to synthesize the catalysts, which are donated as NSC-CL-0.1, NSC-WS-0.1, and NSC-GL-0.1. Meanwhile, the control samples without templates were also obtained, which were named as NSC-CL-0, NSC-WS-0, and NSC-GL-0, respectively.

Characterization

Microscopic surface morphology and crystal structure of the synthesized materials were characterized by the field emission scanning electron microscope (FE-SEM) and transmission electron microscope (TEM) on Hitachi High-Technologies Corporation SU8010 and JEOL JEM-2100. The crystal phases of materials were identified by X-ray diffraction (XRD) patterns on the Bruker-D8 instrument with Cu-Kα radiation. Raman spectra were measured directly on LabRAM HR Evolution with 532-nm excitation laser to study the degree of defects. A thermogravimetric analysis/differential thermal analysis (TG/DTA) thermal analyzer (NETZSCH STA449F3) was used for testing weight losses in N2 atmosphere up to 1000 °C. The Brunauer–Emmett–Teller (BET) surface area and pore size distribution were analyzed by N2 adsorption–desorption isotherms on the Quantachrome Autosorb Station iQ2 instrument. X-ray photoelectron spectroscopy (XPS) was used to characterize the surface chemical state analysis of samples on Thermo Scientific ESCALAB250. The contact wettability was measured by the water contact angle test on KRüSS DSA25.

Electrochemical testing of ORR

All electrochemical experiments were carried out on an electrochemical workstation (CHI660E) with a three-electrode system. Hg/HgO electrode and saturated calomel electrode were regarded as the reference electrodes in 0.1 M KOH and 0.5 M H2SO4, Pt foil (1 cm2) electrode as the counter electrode, and catalyst-coated glassy carbon electrode (GCE, d = 3 mm or 5.61 mm) as the working electrode, respectively. The electrolytes were N2/O2-saturated 0.1 M KOH or 0.5 M H2SO4.

Preparation of working electrode: 100 μL Nafion solution (5 wt.%) and 4.9 mL ultrapure water were dispersed in a 5-mL sample tube to obtain 0.1 wt% Nafion solution. 1 mg catalyst and 60 μL 0.1 wt% Nafion solution were ultrasonically dispersed for 30 min to obtain the uniform catalyst ink. Subsequently, 6/18 μL catalyst ink was loaded on a clean glassy carbon electrode (GCE, d = 3/5.61 mm) to acquire the working electrode for cyclic voltammetry (CV) and linear sweep voltammetry (LSV) technologies. Besides, the Pt/C electrode was also prepared using the same method.

ORR tests: A certain amount of 0.1 M KOH or 0.5 M H2SO4 electrolyte was added into a five-port electrolytic cell, and a constant flow of N2 or O2 was introduced to the solution to saturation. Electrode activation was performed by CV technique and tested at scan rates of 10 mV s−1 with potential intervals of 0.07 ~ 1.17 V (0.1 M KOH) or 0.17–1.27 V (0.5 M H2SO4), respectively. The electrochemical active surface area (ECSA) of the catalysts was evaluated by measuring the double-layer capacitance (Cdl) with different scan rates (5, 10, 15, 20, 30, 40, and 50 mV s−1). The accelerated stability tests and methanol tolerance experiments were carried out by the CV curves. The scanning speed of the LSV curve was 10 mV s−1, the potential interval was the same as that of CV test, and the rotation speed was 1600 rpm. All potentials were recorded and converted to the potential versus RHE based on the equations of E(RHE) = E(Hg/HgO) + 0.0591 pH + 0.098 V (pH = 13 for 0.1 M KOH) and E(RHE) = E(SCE) + 0.0591 pH + 0.24 V (pH = 0 for 0.5 M H2SO4).

The hydrogen peroxide yield (H2O2%) and the number of electron transfers (n) of catalysts were calculated according to the following equations:

$$n = \frac{{{\text{4I}}_{{\text{d}}} }}{{{{{\text{I}}_{{\text{r}}} } \mathord{\left/ {\vphantom {{{\text{I}}_{{\text{r}}} } {{\text{N + I}}_{{\text{d}}} }}} \right. \kern-\nulldelimiterspace} {{\text{N + I}}_{{\text{d}}} }}}}$$
(1)
$${\text{H}}_{2} {\text{O}}_{2} = \frac{{{\raise0.7ex\hbox{${2I_{{\text{r}}} }$} \!\mathord{\left/ {\vphantom {{2I_{{\text{r}}} } N}}\right.\kern-\nulldelimiterspace} \!\lower0.7ex\hbox{$N$}} }}{{{\raise0.7ex\hbox{${I_{{\text{r}}} }$} \!\mathord{\left/ {\vphantom {{I_{{\text{r}}} } {N + I_{d} }}}\right.\kern-\nulldelimiterspace} \!\lower0.7ex\hbox{${N + I_{d} }$}} }} \times 100\%$$
(2)

where Id is the disk current, Ir refers to the ring current, and N is the current collection efficiency of the Pt ring (N = 0.37).

Assembly and performance test of Zn-air batteries

Assembly of Zn-air batteries: 3 mg catalyst, 200 μL (CH3)2CHOH solution, 16 μL 5 wt% Nafion solution, and 8 μL 5 wt% PTFE were mixed and ultrasonically treated for 30 min to obtain uniformly dispersed catalyst ink. The catalyst ink was loaded on carbon paper with the loading of 2 mg cm−2 and dried naturally. The air electrode was also prepared by the same process using the commercial Pt/C catalyst. The ZAB was assembled with the catalyst-coated carbon paper as the air cathode, a polished Zn foil (0.2 mm) as the anode, and 6.0 M KOH containing 0.2 M Zn(Ac)2 solution as the electrolyte.

Performance test of Zn-air battery: The charge–discharge polarization curves were performed by LSV at a scan rate of 5 mV s−1. The galvanostatic discharge curves were tested by chronopotentiometry (CP) technology at different galvanostatic current densities (1, 2, 5, 10, 20, 50, and 100 mA cm−2). In addition, the cycling stability was detected by using the CP technique at 2 mA cm−2 with a cycling interval of 10 min (5 min for charging and 5 min for discharging).

Results and discussion

Morphology and composition characterization

The mesopore-dominated N, S co-doped hierarchical porous carbon materials have been synthesized by the hard-soft template strategy, as shown in Scheme 1. The morphologies of NSC-PT series of catalysts were characterized by SEM. Figure 1a reveals a disorderly stacked fragmentary structure of NSC-PT-0. With the increased amounts of hard template SiO2 from 0.04 to 0.1 g, porous structures are gradually generated (Fig. S1a–b and Fig. 2b). Such condition of 0.1 g SiO2 under the presence of ZnCl2 soft template can provide hierarchically porous structure (Fig. 1b). Subsequently, with the hard template dosage continuously increasing from 0.5 to 1 g, more macropores are formed (Fig. S1c, d). When a single soft template or hard template is introduced, the surface morphology of NSC-PT-Z (Fig. 1c) and NSC-PT–S (Fig. 1d) becomes fluffy and porous [31, 32]. The as-prepared NSC-PT-0.1 catalyst with two templating agents, possesses the rich mesoporous morphology of the SiO2 template. More detailed porosity information will be discussed in following BET data. TEM characterization analysis is applied to further study the morphology and structure of the catalysts. The mesopore-dominated porous structure can be found of the NSC-PT-0.1 catalyst in Fig. 1e, f. The HR-TEM image of NSC-PT-0.1 is further observed (Fig. 1g), and the edge of the material shows a large number of discrete lattice fringes with lattice spacings of about 0.34 nm, which correspond to the (002) crystal planes of graphitized carbon [33, 34]. The result indicates that the NSC-PT-0.1 undergoes partial graphitization after high-temperature pyrolysis, which can facilitate the transport of electrons [35]. According to the HR-TEM and element surface scan images (Fig. 1i–l), the elements of C, O, N, and S are uniformly distributed along with the structure of mesopore-dominated carbon, implying the successful preparation of hierarchically nanostructured carbon materials.

Scheme 1
scheme 1

Schematic illustration of synthetic process for NSC-PT, NSC-CL, NSC-WS, and NSC-GL series of catalysts

Full size image
Figure 1
figure 1

SEM images of NSC-PT-0 (a), NSC-PT-0.1 (b), NSC-PT-Z (c), and NSC-PT–S (d); TEM (e–f), HR-TEM (g), HR-TEM and element scanning images of NSC-PT-0.1 (h–l)

Full size image
Figure 2
figure 2

N2 adsorption/desorption isotherms (a), pore size distribution curves (b), XRD patterns (c), and Raman spectra (d) images of NSC-PT-x, NSC-PT-Z, and NSC-PT-S; TGA curves (e), high-resolution C 1s (f), N 1s (g), S 2p (h), and O 1s XPS spectra of NSC-PT-0.1 (i)

Full size image

The influence of template dosages on pore structure and specific surface area of NSC-PT-x has been investigated. Although all carbon materials show type-IV N2 adsorption/desorption isotherms, the porosity changes with the amount of template agent (Fig. 2a, b, Fig. S2a, b, and Table S1). NSC-PT-0 is obtained by carboning the pure Pointed tail leaf without the hard-soft template, whose specific surface area (SBET) is 247.35 m2 g−1, corresponding to the rough topography of the SEM image. As the hard template dosage increases, the determined SBET values firstly increase and then decrease, and the NSC-PT-0.1 shows the largest SBET of 2357.55 m2 g−1. Meanwhile, the dAverage values are mainly distributed in the range from 3.0 to 8.5 nm, proving the successful synthesis of mesopore-dominated porous carbon materials. In addition, without the introduction of the template, the specific surface area (SBET) of NSC-PT-0 is just 247.35 m2 g−1. As just a single soft template or hard template is added, the SBET (1095.55 and 1128.51 m2 g−1) of the catalyst increases significantly. Notably, when hard-soft templating agents are both added, the SBET (2357.55 m2 g−1) of the catalyst is almost double that of adding only one single templating agent. Besides, two broad diffraction peaks around 23° and 44° in the XRD patterns of all catalysts represent the (002) and (101) planes of graphitized carbon in Fig. S2c and Fig. 2c, respectively [36]. The wider diffraction peak (002) of NSC-PT-0.1 indicates that the (002) crystal plane peak is weakened more obviously, meaning higher defectivity [37]. Besides, the D-band corresponds to disorder and defects in the graphite crystal lattice, whereas the G-band reflects crystalline ordering of the graphitic basal-plane. The ID/IG value of NSC-PT-0.1 (1.15) is larger than other catalysts (Fig. 2d, Fig. S2c, d, and Table S1), indicating more defects are generated, corresponding to XRD results.

Thermogravimetric analysis (TGA, Fig. 2e) of NSC-PT-0.1 in the N2 atmosphere was performed to further explore the pore formation mechanism of carbon materials. The high-temperature pyrolysis reaction is mainly divided into three stages: (i) 25–267 °C, the evaporation of free and bound water on the surface occurs; meanwhile, the soft templates of ZnCl2 molecules enter the carbon frameworks and react with water molecules to produce ZnO (formula (1)) [31]. (ii) 267–538 °C, the weight loss at this stage is mainly attributed to the decomposition of cellulose, hemicellulose, and lignin in the waste biomass [38]. (iii) 538–900 °C, the reduction of ZnO in carbon frameworks happens to obtain metallic Zn and CO/CO2 (formula (2)) [39]. After reaching the boiling point of Zn (~ 900 °C), the Zn elements turn into a gaseous state and leave the carbon skeletons [40]. Based on the above reactions, the formation mechanism of pore structure of NSC-PT-0.1 was speculated during the pyrolysis. The evaporation of water and the escape of Zn vapor from the surface mainly provide the microporous structure, whereas the mesoporous structure is fabricated by ZnO in the carbon frameworks and the hard template of SiO2. The gas molecules (CO/CO2) in the molten carbon frameworks also easily rupture to form a meso/microporous structure [41]. The mesopore-dominated hierarchically porous structure not only facilitates the diffusion of oxygen molecules and exposes more catalytically active sites, but also effectively promotes the permeation process of electrolytes in active materials.

$${\text{ZnCl}}_{{2}} {\text{ + H}}_{{2}} {\text{O }} \to {\text{ ZnO + HCl}}$$
(3)
$${\text{ZnO + C }} \to {\text{ CO/CO}}_{{2}} {\text{ + Zn}}$$
(4)

The full XPS spectra contain C, N, O, and S elements are detected (Fig. S3a), matching well with the uniform elemental distribution in the HR-TEM image. Four component peaks of C 1 s spectrum with binding energies at ~ 284.5 eV, ~ 285.1 eV, ~ 286.2 eV, and ~ 288.7 eV, correspond to C=C, C−N/C−C, C−O, and O−C=O species (Fig. 2f). Figure 2g shows that the N 1 s high-resolution spectrum can be fitted to four positions at ~ 398.1 eV (pyridinic-N), ~ 399.8 eV (pyrrolic-N), ~ 401.2 eV (graphitic-N), and ~ 402.7 eV (oxidized-N). Among them, pyridinic-N and graphitic-N with suitable electronegativity and electron configuration have been confirmed to provide catalytically active centers for ORR [42]. The high-resolution spectrum of S 2p can be divided into three peaks (Fig. 2h), C−S−C (~ 163.3 eV), C=S (~ 165.2 eV), and C−SOx−C (~ 168.2 eV). Doping of N, S heteroatoms changes the valence states of adjacent carbon atoms in the carbon frameworks, thereby improving ORR electrocatalytic activity [43]. Finally, the high-resolution spectrum of the O element is fitted and analyzed (Fig. 2i). It can be seen that the peaks with different intensities appear at ~ 531.8 eV, ~ 532.4 eV, ~ 533.2 eV, and ~ 534.1 eV, corresponding to C=O, C−OH, C−O−C, and O−C=O groups [44]. Due to the electronegativity difference between carbon and heteroatoms (δC = 2.55, δN = 3.04, and δS = 2.58), the carbon-heteroatom bond is easily polarized and induces positive/negative charges on adjacent carbon atoms, which is conducive to oxygen adsorption and reduction [45,46,47]. In detail, the surrounding carbon atoms are positively charged due to the appearance of the above oxygen-containing functional groups, which can adsorb OH− or H2O through electrostatic attraction. It can effectively improve the electron transfer between the material surface and reaction intermediates, thereby enhancing the ORR catalytic activity [48]. In addition, the hydrophilic groups can improve the hydrophilicity of carbon materials by utilizing the interaction between hydrogen bonds and water molecules. The water contact angle of NSC-PT-0.1 is determined to be 3° (Fig. S3b), indicating that the material has super hydrophilicity. The hydrophilicity promotes the efficient transport of mass/ions and improves the electrocatalytic activity of the materials.

Electrocatalytic performance

The electrocatalytic performance of the as-prepared catalysts was evaluated in alkaline and acidic media using a three-electrode system. The ORR performance at alkaline condition was firstly investigated. The CV curves, of NSC-PT-x with different ratios of the hard-soft templates, are tested in 0.1 M KOH saturated with N2, which reveals a rectangular-like appearance without cathodic oxygen reduction peaks in Fig. S4a. However, in the O2-saturated electrolyte, the CV curves immediately exhibit an obvious cathodic peak, suggesting that all the catalysts have electrochemical reduction ability to oxygen. It is observed that the optimal ORR activity is achieved for NSC-PT-0.1 (Es = 0.84 V). In order to evaluate the kinetics mechanism of electrocatalysts toward ORR, the LSV plots were obtained by the specific rotating ring disk electrode (RRDE) from Pine company. As depicted in Fig. S4b, c and Table S2, NSC-PT-0.1 possesses an exceptional half-wave potential of 0.83 V with an onset potential of 0.93 V, better than those of NSC-PT-0.04 (E1/2 = 0.77 V, Eonset = 0.89 V), NSC-PT-0.07 (E1/2 = 0.80 V, Eonset = 0.90 V), NSC-PT-0.5 (E1/2 = 0.79 V, Eonset = 0.90 V), and NSC-PT-1 (E1/2 = 0.78 V, Eonset = 0.89 V). To further determine the ORR path, a RRDE technology is performed to evaluate the corresponding electron transfer number n and generation of H2O2%. The H2O2% and the n for NSC-PT-0.1 are 9.38% and 3.81, indicating high four-electron path selectivity (Fig. S4d). Moreover, the excellent ORR kinetics of the NSC-PT-0.1 catalyst is confirmed by a lower Tafel slope of 123 mV·dec−1 compared to of the other control samples (Fig. S4e and Table S2). Figure S4f and Table S2 show that NSC-PT-0.1 has a higher Cdl value (68.69 mF cm−2) than other control catalysts, indicating that this catalyst has larger ECSA. It is beneficial for the exposure of active sites, promoting the transfer rate of oxygen-containing intermediates, thereby improving ORR reaction activity and kinetics [33, 44]. It can be seen that NSC-PT-0.1 possesses better ORR performance than other samples in 0.1 M KOH.

Based on the exploration of biomass with different ratios of hard-soft templates, we further study the effect of hard and soft templates. The ORR activity of catalysts is firstly explored by measuring CV in O2-saturated 0.1 M KOH media (Fig. 3a). The ORR peak potential of NSC-PT-0.1 appears at 0.85 V, which is more positive than those of NSC-PT-0 (0.80 V), NSC-PT-Z (0.82 V), NSC-PT–S (0.83 V), and commercial Pt/C (0.82 V), indicating the higher ORR activity for NSC-PT-0.1. The ORR performances are further examined by the LSV curves at a rotation rate of 1600 rpm through using RRDE in O2-saturated 0.1 M KOH condition (Fig. 3b and Table S2). Figure 3c shows that the Eonset is 0.93 V and the E1/2 is 0.83 V for NSC-PT-0.1, which are more positive than those of NSC-PT-0 (Eonset = 0.88 V, E1/2 = 0.74 V), NSC-PT-Z (Eonset = 0.90 V, E1/2 = 0.80 V), NSC-PT–S (Eonset = 0.91 V, E1/2 = 0.81 V), and commercial Pt/C (Eonset = 0.98 V, E1/2 = 0.86 V). The fast electron transfer rate of NSC-PT-0.1 is certified by the RRDE tests (Fig. 3d). NSC-PT-0.1 reveals a high n value of 3.8–4.0 within the potential window of 0.2–0.8 V, corresponding to a low H2O2 yield below 9.38%, which are even superior to those of Pt/C, suggesting a four-electron ORR pathway, the Tafel slope for NSC-PT-0.1 is 123 mV dec−1, lower than those of NSC-PT-0 (253 mV dec−1), NSC-PT-Z (153 mV dec−1), NSC-PT–S (149 mV dec−1), and Pt/C (144 mV dec−1), demonstrating the fastest ORR kinetics (Fig. 3e). The Cdl derived from electrochemical active area (ECSA) is also applied to evaluate the ORR catalytic activity by determining CV curves with different scan rates from 5 to 50 mV s−1 (Fig. 3f–g). The Cdl of NSC-PT-0.1 is 68.69 mF cm−2, higher than those of NSC-PT-0 (17.95 mF cm−2), NSC-PT-Z (65.42 mF cm−2), NSC-PT–S (29.85 mF cm−2), and Pt/C (38.31 mF cm−2), confirming higher electrochemical activity area. Furthermore, the catalytic cycling stability of the ORR is crucial for evaluating the durability of catalysts. After 10,000 CV cycles, the peak potential and peak current of NSC-PT-0.1 are almost identical to the initial one, while the peak potential of Pt/C shifts negatively by 30 mV, revealing the better cycle stability for NSC-PT-0.1 (Fig. 3h). The tolerance of NSC-PT-0.1 and Pt/C toward CH3OH was tested by injecting 3 mL 3 M methanol solution into the 30 mL 0.1 M KOH. The CV curves of NSC-PT-0.1 show an ignorable change, whereas Pt/C exhibits a typical methanol oxidation peak instead of the ORR peak (Fig. 3i). Therefore, NSC-PT-0.1 reveals high ORR activity, superior methanol tolerance, and stability in alkaline medium. It is related to the hierarchically pore structure of NSC-PT-0.1, which facilitates the exposure of active sites and mass transport [49].

Figure 3
figure 3

ORR performances in 0.1 M KOH: a CV curves, b LSV curves at 1600 rpm, c E1/2 and Eonset, d H2O2% and n, e Tafel plots, f current density at different scan rates (5–50 mV s−1) in the potentiostatic interval (1.07 ~ 1.17 V), g Cdl curves, h stability tests, i methanol tolerance tests images of NSC-PT-0, NSC-PT-0.1, NSC-PT-Z, and NSC-PT–S

Full size image

Excellent ORR property of the catalysts in 0.1 M KOH media stimulates us to study the activity in 0.5 M H2SO4 condition. The electrocatalytic performances of samples with different ratios of the hard-soft template toward ORR are also assessed in O2-saturated 0.5 M H2SO4 electrolyte. RRDE curves of NSC-PT-x at the rotation rate of 1600 rpm are measured and have demonstrated that NSC-PT-0.1 has a more positive half-wave potential (Fig. S5a). Compared with the NSC-PT-0.1 catalyst, the NSC-PT-0.04, NSC-PT-0.07, NSC-PT-0.5, and NSC-PT-1 controls exhibit apparently worse ORR activity, such as the lower E1/2 (0.52 V, 0.64 V, 0.60 V, and 0.61 V, respectively) (Fig. S5b). The n and H2O2% of these catalysts are determined at potentials from 0.2 to 0.8 V by RRDE measurement (Fig. S5c). The average H2O2% yield and n value of NSC-PT-0.1 are calculated to be 2.16% and 3.96, which are better than those of NSC-PT-0.04 (H2O2% = 7.57%, n = 3.85), NSC-PT-0.07 (H2O2% = 3.56%, n = 3.92), NSC-PT-0.5 (H2O2% = 2.75%, n = 3.94), and NSC-PT-1 (H2O2% = 4.98%, n = 3.90), suggesting an effective 4-electron process. It can be seen that NSC-PT-0.1 possesses prominent ORR performance in 0.5 M H2SO4.

Next, in order to study the effect of hard and soft templates, the electrocatalytic performances of the catalysts toward ORR are also assessed in O2-saturated 0.5 M H2SO4 electrolyte. As shown in Fig. 4a and Table S3, the obvious reduction peaks, observed in CV curves under O2-saturated media, elucidate the ORR catalytic activity for prepared samples and Pt/C catalysts. The cathodic peak potential of NSC-PT-0.1 (0.84 V) is much more positive than those of NSC-PT-0 (0.52 V), NSC-PT-Z (0.68 V), NSC-PT–S (0.60 V), and Pt/C (0.81 V), indicating the excellent ORR activity. Figure 4b and Table S3 show the LSV polarization plots of the as-prepared catalysts and 20 wt% Pt/C. As shown in Fig. 4c, NSC-PT-0.1 delivers a more positive onset potential of 0.80 V and half-wave potential of 0.69 V than those of NSC-PT-0 (Eonset = 0.74 V, E1/2 = 0.43 V), NSC-PT-Z (Eonset = 0.79 V, E1/2 = 0.60 V), NSC-PT–S (Eonset = 0.80 V, E1/2 = 0.50 V), and comparable to Pt/C (Eonset = 0.91 V, E1/2 = 0.76 V). The average H2O2% yield and n value of NSC-PT-0.1 are calculated to be 2.16% and 3.96, which is better than those of NSC-PT-0 (H2O2% = 17.30%, n = 3.65), NSC-PT-Z (H2O2% = 3.66%, n = 3.93), NSC-PT–S (H2O2% = 4.48%, n = 3.91), and Pt/C (H2O2% = 7.98%, n = 3.84), suggesting an effective 4-electron process (Fig. 4d). Excellent durability and methanol tolerance are also crucial for ORR electrocatalysts. As depicted in Fig. 4e, after 5000 continuous CV cycles, the peak potential of NSC-PT-0.1 is almost identical to the original one, much better than that of Pt/C. Additionally, the resistance of methanol crossover has been tested by injecting 3 mL 3 M methanol into the electrolyte during the CV test. As shown in Fig. 4f, the obvious methanol poisoning effect containing a distinct methanol redox peak and a significant current drop is observed for Pt/C, whereas the current response of NSC-PT-0.1 barely decays upon the addition of CH3OH. The consequence demonstrates that NSC-PT-0.1 exhibits more excellent tolerance to methanol than Pt/C catalyst. By making a preliminary summary, NSC-PT-0.1 possesses superior stability and selectivity toward ORR in 0.1 M KOH electrolyte, which can be regarded as a promising electrocatalyst for practical application [50, 51].

Figure 4
figure 4

ORR performances in 0.5 M H2SO4: a CV curves, b LSV curves at 1600 rpm, c E1/2 and Eonset, d H2O2% and n, e stability tests, f methanol tolerance tests images of NSC-PT-0, NSC-PT-0.1, NSC-PT-Z, and NSC-PT–S

Full size image

Universality study of other biomass-derived catalysts

To prove the versatility of the hard-soft template strategy, the optimal dosages of hard and soft templates have been selected to prepare other biomass-derived carbon materials (NSC-CL-0/0.1, NSC-WS-0/0.1, and NSC-GL-0/0.1). Figure 5a–c displays the SEM images of three different hierarchical N, S co-doped carbon materials (NSC-CL-0.1, NSC-WS-0.1, and NSC-GL-0.1) derived from the Carrot, Whorled stonecrop, and Ginkgo leaf. It is worth noting that they all possess the similar hierarchical structures to the NSC-PT-0.1. Moreover, structural characteristics were further confirmed by XRD and Raman patterns (Fig. S6). XRD patterns of these catalysts show the diffraction peaks at around 23° and 44° assigned to the (002) and (101) crystal planes of graphitic carbon, similar to the patterns of NSC-CL-0, NSC-WS-0, and NSC-GL-0 [52]. Additionally, the Raman spectra confirm that the higher ID/IG values are achieved for NSC-CL-0.1 (1.14), NSC-WS-0.1 (1.13), and NSC-GL-0.1 (1.13) than those of NSC-CL-0 (1.03), NSC-WS-0 (1.07), and NSC-GL-0 (1.10), implying the existence of richer defects (Table S4). To further confirm the effect of templating agents on pore structure, N2 adsorption/desorption isotherms and pore size distribution tests were also performed for the samples, as shown in Fig. 5d, e. NSC-CL-0/0.1, NSC-WS-0/0.1, and NSC-GL-0/0.1 have type-IV isotherms. It is calculated that the SBET, VTotal, and VAverage have a significant increase with the addition of the hard-soft template agent. The pore structures are similar to those of NSC-PT-0.1, revealing the universality of the hard-soft template strategy.

Figure 5
figure 5

SEM images of NSC-CL-0.1 (a, the insert is Carrot), NSC-WS-0.1 (b, the insert is Whorled stonecrop), NSC-GL-0.1 (c, the insert is Ginkgo leaf); N2 adsorption/desorption isotherms d, and pore size distribution curves e, E1/2 and Eonset, H2O2% and n in 0.1 M KOH (f, g), E1/2 and Eonset, H2O2% and n in 0.5 M H2SO4 h, i images of NSC-CL, NSC-WS, NSC-GL series of catalysts

Full size image

As the electrochemical performance of catalysts relies on mesopore-dominated hierarchical porous structure and the amount of catalytically active sites, we further investigate the ORR activity of various carbon-based materials derived from different biomasses, including the NSC-CL-0.1, NSC-WS-0.1, and NSC-GL-0.1. In terms of E1/2 in 0.1 M KOH media and 0.5 M H2SO4 media, the ORR activity of NSC-CL-0.1 (0.80 V, 0.65 V), NSC-WS-0.1 (0.80 V, 0.62 V), and NSC-GL-0.1 (0.81 V, 0.68 V) is much higher than those of NSC-CL-0 (0.71 V, 0.42 V), NSC-WS-0 (0.75 V, 0.42 V), and NSC-GL-0 (0.72 V, 0.55 V). To further determine the ORR path, a RRDE technology is also performed to evaluate the n and generation of H2O2%. A low HO2− (in 0.1 M KOH) or H2O2 (in 0.5 M H2SO4) yield (< 10%), and an electron transfer number n for NSC-CL-0.1, NSC-WS-0.1 and NSC-GL-0.1 approaching four are determined at potentials from 0.2 to 0.8 V, indicating high four-electron path selectivity (Fig. S4d, 5c). Figure S4f and Table S2 show NSC-CL-0.1, NSC-WS-0.1, and NSC-GL-0.1 have a higher Cdl value (47.84, 65.40, and 56.33 mF cm−2) than other NSC-CL-0 (18.51 mF cm−2), NSC-WS-0 (13.92 mF cm−2), and NSC-GL-0 (22.07 mF cm−2) catalysts, indicating the larger ECSA in 0.1 M KOH. The hard-soft templates are beneficial for providing more active sites and promoting the transfer rate of oxygen-containing intermediates, thereby improving ORR reaction activity and kinetics [53]. It can be seen that NSC-CL-0.1, NSC-WS-0.1, and NSC-GL-0.1 possess better ORR activity than other control samples in 0.1 M KOH and 0.5 M H2SO4 media. In summary, the hard-soft template method can not only promote the exposion of active sites but also generate more edge and defect sites, enabling the synthesized carbon-based catalysts with excellent ORR activity [54, 55].

Zn-air battery performance

To evaluate the practical application of NSC-PT-0.1 catalyst, a ZAB was assembled through adopting carbon paper-supported NSC-PT-0.1 catalyst (2 mg cm−2) as the air cathode, Zn foil as the anode, and 6.0 M KOH with 0.2 M Zn(Ac)2 as the electrolyte (Fig. 6a). The ZAB with commercial Pt/C as cathode catalyst was also assembled. The NSC-PT-0.1-based ZAB exhibits a peak power density of 155.55 mW cm−2, which is significantly higher than that of the Pt/C-based ZAB (137.49 mW cm−2) (Fig. 6b). In the charge–discharge polarization curves, the NSC-PT-0.1-based ZAB has a lower charge–discharge gap of 0.87 V at a current density of 20 mA cm−2 than that of the Pt/C-based ZAB (1.09 V) (Fig. 6c). As displayed in Fig. 6d, the discharge curves of NSC-PT-0.1-based ZAB possess higher values than that of Pt/C-based ZAB at a series of current densities (1 ~ 100 mA cm−2). It is attributed to the mesopore-dominated hierarchically porous structure and heteroatoms-doped of NSC-PT-0.1, which provides effective mass transfer channels for the diffusion of O2 and electrolytes [56,57,58]. As a result, the discharge voltage resumes reversibly once the current density reduces to 1 mA cm−2, indicating the superior rate performance. The battery driven by NSC-PT-0.1 catalyst exhibits a stable open-circuit voltage (1.46 V) superior to Pt/C-based ZAB (1.44 V) in Fig. 6e. Remarkably, three series-connected NSC-PT-0.1-based ZABs can successfully power a 3.7 V light-emitting diode (LED) screen (Fig. 6f) or a 3.0 V electronic toy (Movie SI), demonstrating the feasible application of NSC-PT-0.1 catalyst. Besides, the cycle stability of ZAB with NSC-PT-0.1 catalyst was also evaluated by continuous galvanostatic discharge–charge test at 2 mA cm−2. The NSC-PT-0.1-based ZAB can repeatedly cycle at 2 mA cm−2 for a total of 216 h in Fig. 6g. Compared to a Pt/C-based ZAB, a slightly lower initial charge–discharge voltage gap is observed for the NSC-PT-0.1-based ZAB, indicating a better rechargeability (Fig. 6h). After about 102 h, the NSC-PT-0.1-based ZAB shows a negligible voltage change, whereas Pt/C demonstrates a significantly increase in the voltage gap under the same condition, demonstrating superior stability of the NSC-PT-0.1-based ZAB. The above results verify the promising potential of NSC-PT-0.1 as a high-efficiency and durable catalyst for ZABs [59].

Figure 6
figure 6

a Schematic diagram of ZAB with NSC-PT-0.1 as the air cathode. b Discharge polarization curves and power density curves of NSC-PT-0.1-based and Pt/C-based ZABs. c Charge and discharge polarization curves. d Discharge curves at different current densities from 1 to 100 mA cm−2. e Open-circuit voltage. f Photographs of the LED screen powered by three series of NSC-PT-0.1-based ZABs. g–h Long-term discharge–charge cycling stability curves at 2 mA cm.−2

Full size image

Conclusions

In summary, a feasible hard-soft (SiO2–ZnCl2) template strategy has been developed to construct mesopore-dominated porous heteroatom-doped carbon materials from various biomass precursors (Pointed tail, Carrot, Whorled stonecrop, and Ginkgo leaf). The employed SiO2 and ZnCl2 templates result in the optimized mesopore-dominated hierarchical porous carbon material (NSC-PT-0.1) with large SBET of 2357.55 m2 g−1 and high N, S content of 2.32 at.%. The unique structure not only promotes the exposion of assessable active sites, but also increases the contact wettability of catalyst and electrolyte, improving the transmission efficiency of O2 and ions. Therefore, NSC-PT-0.1 exhibits excellent alkaline (Eonset = 0.93 V, E1/2 = 0.82 V, and jL = 7.84 mA cm−2) and acidic (Eonset = 0.80 V, E1/2 = 0.69 V, and jL = 5.08 mA cm−2) ORR catalytic activity, long-term stability (alkaline 10,000 CV cycles and acidic 5000 CV cycles), and methanol tolerance. Furthermore, when NSC-PT-0.1 is employed as a catalytic layer of air electrode, the assembled ZAB exhibits a better discharge power output capability (155.55 mW cm−2) and long-term cycle stability (216 h) than Pt/C-based ZAB. Besides, the Carrot, Whorled stonecrop, and Ginkgo leaf have been simultaneously chosen to confirm the versatility of the hard-soft template method. This work provides a promising approach for synthesizing low-cost, efficient, and stable ORR electrocatalyst for ZABs through the high-quality utilization of biomass resources.