Facile encapsulation strategy for uniformly-dispersed catalytic nanoparticles/carbon nanofibers toward advanced Zn–air battery

Abstract

Catalyst-loaded carbon nanofibers (CNFs) have been a promising platform to enhance the efficiency of key electrochemical reactions, crucial for the operation of fuel cells, metal–air batteries, and the reduction of pollutants. A persistent challenge, however, has been the uniform distribution and strong binding of ultrafine active catalytic nanoparticles on the CNF surfaces. To overcome this issue, this study introduces a facile synthesis strategy using electrospun blended nanofibers of polyacrylonitrile and poly(4-vinylpyridine) (P4VP) for the uniform attachment of various metal catalyst ions. Metal complex ion precursors are selectively loaded into P4VP domains and subsequent one-step pyrolysis to readily obtain CNFs densely populated with uniformly dispersed metal nanoparticles enveloped by a P4VP-derived nitrogen-doped carbon layer. The final C@Ir/CNF1000, composed of iridium nanoparticles (average diameter: ∼3.38 ± 1.18 nm) on CNFs with a carbon overlayer and carbonized at 1000 °C, significantly improves the dual-functionality and cycle stability of zinc–air batteries. Our findings present a novel and scalable strategy for improving catalytic efficiency in energy technologies, marking a significant contribution to the field of sustainable energy and environmental applications.

---

Hyeong Min Jin

Prof. Hyeong Min Jin is an Assistant Professor in the Department of Organic Materials Engineering at Chungnam National University. He earned his PhD in Materials Science and Engineering from KAIST in 2017 under the supervision of Prof. Sang Ouk Kim. Following this, he served as a postdoctoral research fellow at the University of Chicago in 2018, working with Prof. Paul F. Nealey. From 2019 to 2022, he was a beamline scientist specializing in small-angle neutron scattering (SANS) at the Korea Atomic Energy Research Institute (KAERI). His current research focuses on the synthesis of mesoporous organic/inorganic hybrid materials through molecular self-assembly for energy device applications, as well as in-depth structural analysis of porous materials using X-ray and neutron scattering techniques.

---

Introduction

The innovative development of catalyst-loaded carbon nanofibers (CNFs) has been one of the important breakthroughs in energy and environmental science applications.^1,2 CNFs, recognized for their high specific surface area, superior electrical conductivity, mechanical durability, and chemical stability, offer an ideal support for the uniform dispersion of catalytic nanoparticles.^3–5 This synergistic integration not only enhances the efficiency of various electrochemical processes, such as the oxygen reduction in proton exchange membrane fuel cells (PEMFCs) and (dis)charging processes in metal–air batteries, but also extends to environmental applications like catalytic converters for pollution control.^6–10

To achieve high catalytic efficiency and durability, retaining ultrafine catalytic nanoparticles in a well-dispersed state on CNFs is essential.^11,12 In electrochemical systems, achieving a uniform dispersion of ultrafine catalytic nanoparticles on CNFs is essential to enhance the accessibility of active sites and improve catalytic efficiency. This dispersion also mitigates adverse effects such as poisoning, Ostwald ripening, metal dissolution, and agglomeration, thereby preserving the electrochemical surface area (ECSA) under severe operational conditions.^13–15 Therefore, recent strategies to address these issues have explored chemical modification of CNFs (e.g. nitrogen or sulfur doping) to improve the binding affinity between catalytic nanoparticles and their supports, alongside the introduction of thin encapsulation layers.^16–23 Notably, the deployment of thin nitrogen-doped (N-doped) carbon or graphitic layers has been actively studied, typically following a three-step process involving (i) the synthesis of the catalyst, (ii) deposition of a carbon precursor, and (iii) subsequent pyrolysis, with various carbon precursors like polydopamine,^24,25 polyaniline,^26,27 dicyandiamide,²⁸ and so on.²⁹ Although these strategies improve activity and durability, the production process exhibits significantly low efficiency due to the involvement of a multi-step and complicated thermal process.^16,19,30

Herein, a facile method for the synthesis of minuscule iridium (Ir) nanoparticle catalysts encapsulated in a thin N-doped carbon layer on CNF supports is presented. Introducing polyacrylonitrile (PAN)/poly(4-vinylpyridine) (P4VP) blended precursor nanofibers, Ir complex ions selectively coordinate with the pyridine group in the P4VP domains. Subsequent one-step pyrolysis process of these Ir complex ion loaded precursor nanofibers allows the synthesis of sub-3 nm metal catalysts uniformly dispersed on CNF (C@Ir/CNF) with strong metal–support interaction (SMSI).^31,32 This strategy is not limited to Ir, but can be universally applied to a variety of metal complex ions capable of coordinating with the pyridine group, including Pt and Au. Furthermore, the versatility of this strategy is demonstrated in its application as a cathodic material in Zn–air batteries, showing significant improvements in electrochemical catalytic activity and long-term stability.^33–35 The aqueous Zn–air battery with the air-breathing electrode of Ir loaded CNFs from the aforementioned strategy exhibited an open-circuit voltage (OCV) of 1.420 V and an energy density of 69.326 mW cm⁻², with an efficiency of 62.34% even after 70 cycles (lasting over 30 hours), thereby indicating stable operational characteristics.

Experimental

Chemicals

Polyacrylonitrile (PAN: M_v ∼150 000), poly(4-vinylpyridine) (P4VP: M_w ∼160 000), N,N-dimethylformamide (DMF, Sigma-Aldrich), and hydrochloric acid (HCl, ACS reagent, 37%) were purchased from Sigma-Aldrich. Dihydrogen hexachloroiridate(IV) hydrate (99.9%-Ir), hydrogen tetrachloroaurate(III) hydrate (99.8%-Au), potassium tetrachloroplatinate(II) (99.9%-Pt) were purchased from Strem chemicals, Inc. All the chemicals were used without additional purification.

Preparation of C@Ir/CNF

The 11 wt% PAN/P4VP (mass ratio of PAN : P4VP = 9 : 1) solution in DMF was electrospun at an applied voltage of 15 kV. As-spun PAN/P4VP precursor nanofibers (diameter: ∼250 nm) were stabilized at 250 °C under vacuum for 2 hours. The thermally stabilized PAN/P4VP nanofibers were then immersed into 2 mM dihydrogen hexachloroiridate in 1 wt% HCl aqueous solution for 30 min, forming a P4VP-Ir complex. Then the PAN/P4VP-Ir nanofibers were washed with deionized water several times and then dried in a vacuum oven at 60 °C for 12 h. The PAN/P4VP-Ir nanofibers were then pyrolyzed using a tube furnace at 700, 800, 900, and 1000 °C under Ar atmosphere, which produced N-doped carbon layers encapsulated catalytic Ir nanoparticles on CNFs (denoted as C@Ir/CNF). The synthesized catalysts were named as C@Ir/CNF700, C@Ir/CNF800, C@Ir/CNF900, and C@Ir/CNF1000 based on their pyrolyzed temperature.

Materials characterizations

The structure analysis and elemental mapping for the catalysts were characterized using a field-emission scanning electron microscope (FE-SEM, S-4800, HITACHI), 200 kV field-emission transmission electron microscope (FE-TEM) (JEM-2100F HR, JEOL). Scanning-TEM (STEM) and energy dispersive X-ray spectroscopy (EDS) were carried out using the same FE-TEM. Mass loading of metallic species in the catalyst was determined through thermogravimetric analysis (TGA). (MaxRes TGA2, Mettler toledo) at a ramp rate of 10 °C min⁻¹, up to 1000 °C in the air. X-ray photoelectron spectrometer (XPS) (K-Alpha+, Thermo Fisher Scientific) was used to analyze the composition and valence state of the catalyst surface. Atomic structure of the catalysts was characterized by X-ray diffraction (XRD) (MiniFlex 600, Rigaku Corporation). In order to ascertain the presence and dimensions of the pores within the samples, Barrett–Joyner–Halenda (BJH) analysis was conducted utilizing N₂ as the adsorptive gas. The equilibration interval was set between 10 and 30 seconds, thus ensuring the accuracy of the adsorption measurements. Prior to the adsorption analysis, the samples were degassed at 300 °C for over 12 hours to remove any moisture or contaminants from the sample surfaces. The graphitization of the samples as a function of pyrolysis temperature was verified by I_D/I_G using Raman microscopy spectroscopy (LabRam Soleil, Horiba France) with a 514 nm laser, 1800 g mm⁻¹, ND 10% filter, acquisition time of 5 seconds, accumulations number of 5, and a spectral range of 600–2200 cm⁻¹.

Electrochemical characterization for OER and ORR

All electrochemical measurements were conducted using a typical three-electrode cell system with a rotating disk electrode (RDE) employing a common glassy carbon (GC) electrode at room temperature. In the three-electrode setup, a GC electrode served as the working electrode, Hg/HgO electrode and Pt electrode served as the reference electrode and the counter electrode. Potentials measured in this setup were all converted into reversible hydrogen electrode (RHE) scale. For the preparation of the working electrode, ink was prepared by DI water/IPA solvent (v/v = 3/1) with 2.5 mg of the sample and 2.5 mg of Ketjen black, after that, an ultrasonication for 30 minutes to obtain a homogeneous ink solution. Then, 8 μl of the catalyst ink was dropped onto the GC electrode and dried for 1 hour at 50 °C to complete the preparation of the working electrode. Before conducting electrochemical measurements, the electrolyte was purged for 20 minutes to ensure N₂ or O₂ saturation. Cyclic voltammetry (CV) curves were performed at a scan rate of 20 mV s⁻¹ from 0 to 1.1 V vs. RHE in both N₂-saturated and O₂-saturated 1 M KOH solution. Linear sweep voltammetry (LSV) was conducted at a rotation speed of 1600 rpm and a scan rate of 5 mV s⁻¹ from 1.2 to 1.7 V vs. RHE in O₂-saturated electrolyte. Electrochemical double layer capacitance (EDLC) was performed at the non-faradic region of the CV curves in the range of 0.85–0.95 V, with a scan rate of 20 mV s⁻¹.

Fabrication of homemade aqueous Zn–air batteries

The zinc–air battery (ZAB) was assembled as follows. First, ink was prepared for the cathode with a uniformly coated catalyst. The ink was made by dispersing 2.5 mg of the sample and a conductive additive (Ketjen black) in 0.95 g of ethanol with 0.05 g of a 10 wt% Nafion binder through 30 minutes of ultrasonication. The manufactured ink was drop-coated onto the GDL substrate, with a loading of 1 mg cm⁻². Subsequently, the assembly of the ZAB utilized a Zn-metal anode, a GDL with a catalyst cathode, and 6 M KOH + 0.2 M zinc acetate as the electrolyte. Both the cathode and the anode had dimensions of 1 × 3 cm. The characterization of the manufactured ZAB was conducted as follows. After stabilization, cyclic tests were performed at a current density of 5 mA cm⁻², with each cycle lasting 24 minutes (12 minutes for each charge and discharge process). The specific capacity during this time was 0.833 mA h cm⁻². Polarization curves were obtained by sweeping the current up to 200 mA at a current density of 10 mA s⁻¹.

Results and discussion

The fabrication procedure of C@Ir/CNF mainly consists of (i) electrospinning of PAN/P4VP precursor nanofibers, (ii) subsequent P4VP–metal complex coordination, and (iii) a one-step pyrolysis process (Fig. 1a). Initially, a blended solution of PAN and P4VP dissolved in DMF solvent is electrospun into PAN/P4VP precursor nanofibers. Next, the PAN/P4VP precursor nanofibers are stabilized at 250 °C under vacuum. Mesoscale phase separation between PAN and P4VP domains occurs, producing PAN nanofibers with evenly distributed mesoscale P4VP domains. The metal precursor is selectively loaded into the P4VP domains by immersing the PAN/P4VP nanofiber into a precursor solution in which a metal complex (Ir, Pt, or Au) is dissolved in an acidic aqueous solution (1 wt% HCl). During this process, the pyridinic-N of P4VP is protonated under acidic conditions, forming a coordination bond with the negatively charged metal complex (Fig. 1a, second step). Subsequently, the pyrolysis process under an Ar atmosphere is carried out to carbonize the polymers, and simultaneously to calcine unnecessary organic species and reduce metal ions to form tiny metal nanoparticles. This step synthesizes the C@M/CNF (M: Ir, Pt, or Au), composed of a thin N-doped carbon layer covering the metal catalyst and CNF support. Interestingly, the P4VP domain is simultaneously converted into a thin N-containing carbon overlayer, encapsulating the catalytic nanoparticle to mitigate particle overgrowth and Ostwald ripening (Fig. 1b). This carbon overlayer acts as an armor to withstand electrochemical degradation through the synergistic effects of SMSI.

Fig. 1 (a) Schematic illustration of C@Ir/CNF synthesis. (b) SEM image of C@Ir/CNF700. TEM images of (c and d) C@Ir/CNF700 and (e and f) C@Ir/CNF1000. STEM and EDS mapping images of (g) C@Ir/CNF700, (h) C@Ir/CNF1000, (i) C@Pt/CNF700, and (j) C@Au/CNF700.

To investigate how the minuscule metallic nanoparticle catalyst – here, Ir is a representative – could be formed, we conducted SEM analysis (Fig. 1b). The SEM image of C@Ir/CNF700 shows the nanofibers with a diameter of about 250 nm, and well-developed mesopores were clearly observed on the surface, while any nanoparticles were not detectable. Compared to neat CNFs from PAN nanofibers (Fig. S1a†), the CNFs from PAN/P4VP nanofibers possess mesopores on their surface. This is because mesoscale phase separation occurs between PAN and P4VP. During the pyrolysis process, the low carbon yield of the P4VP domains lead to their removal, forming mesopores in the CNFs.^36–38 HR-TEM images of C@Ir/CNF700 showed uniformly distributed Ir nanoparticles on the CNF surface without aggregation (Fig. 1c). As shown in the magnified HR-TEM image in Fig. 1d, sub-2 nm Ir nanoparticles were uniformly dispersed and encapsulated by a thin carbon layer, which is beneficial for enhancing catalytic activity. Even C@Ir/CNF1000, which was pyrolyzed at 1000 °C, showed uniformly distributed Ir nanoparticles on the CNF surface (Fig. 1e). Due to the higher pyrolysis temperature, the average nanoparticle size was somewhat coarsened and crystallized, but it remained below 5 nm. Such ultrafine and uniformly distributed Ir nanoparticles are mainly attributed to the encapsulation of the N-doped carbon thin layer derived from the P4VP domain (Fig. 1f). In addition, no Ir nanoparticles were observed in the pyrolyzed sample originating from neat PAN nanofibers without the P4VP domain, confirming that the P4VP domain selectively forms complexes with metal ions (Fig. S1b†). BJH analysis further confirmed the development of mesopores in C@Ir/CNF700 (Fig. S2†). EDS elemental mapping was performed to determine the distribution of macroscopic elements within the nanofibers, confirming that Ir species were uniformly distributed throughout the nanofibers (Fig. 1g and h). The mapping images also confirmed that nitrogen species, derived from both P4VP and PAN, were uniformly distributed in the nanofibers. The presence of these nitrogen species can have a beneficial impact on the catalyst by facilitating electron transfer and providing active sites to enhance catalytic performance,^39–42 while notably enhancing the binding affinity between the catalytic nanoparticles and the CNFs.^22,35,43 Meanwhile, C@M/CNFs with other metal species (M: Pt or Au) were also synthesized to demonstrate the universality of our synthesis strategy for a variety of metal species. As shown in Fig. 1i, C@Pt/CNF800, synthesized via the same route by forming a P4VP-Pt complex, exhibited uniformly distributed sub-10 nm Pt nanoparticles on the CNF surface. Similarly, C@Au/CNF800, synthesized by the same route, also showed well-dispersed sub-10 nm Au nanoparticles on the CNF surface (Fig. 1j). These results highlight the universality of our approach, combining two distinctive polymers that serve as electrically conductive carbon (PAN) and catalyst-generating sites (P4VP).

To confirm the effectiveness of controlling the particle size of the metallic catalyst through encapsulation by the P4VP-derived carbon layer, quantitative particle size analysis according to pyrolysis temperature was performed (Fig. 2). All the TEM images showed homogeneously dispersed Ir catalysts on the one-dimensional (1D) CNFs regardless of the pyrolysis temperatures (Fig. 2a–d). The particle size distributions measured from the TEM images at each pyrolysis temperature (700, 800, 900, and 1000 °C) are shown in Fig. 2e–h. For the sample pyrolyzed at the lowest temperature of 700 °C, the mean diameter of the nanoparticles was only 1.40 nm, with a very narrow size distribution (standard deviation of 0.25 nm). It was confirmed that as the pyrolysis temperature increased, the mean particle size and size distribution gradually increased. Nevertheless, for the C@Ir/CNF1000, the mean particle size remained very small at 3.36 nm, confirming that encapsulation with a thin carbon layer effectively mitigates particle growth and coarsening of metal species during the pyrolysis process (Fig. 2h).

Fig. 2 (a–d) TEM images of (a) C@Ir/CNF700, (b) C@Ir/CNF800, (c) C@Ir/CNF900, and (d) C@Ir/CNF1000. Particle size histograms of (e) C@Ir/CNF700, (f) C@Ir/CNF800, (g) C@Ir/CNF900, and (h) C@Ir/CNF1000.

To evaluate the mass loading of Ir species in the samples, thermogravimetric analysis (TGA) was performed on C@Ir/CNF700 and C@Ir/CNF1000 in air (Fig. 3a). The mass loss observed up to 100 °C was attributed to water evaporation and was excluded from the mass loading calculations (wt%).⁴⁴ The Ir mass loading was determined to be 9.12 wt% for C@Ir/CNF700 and 14 wt% for C@Ir/CNF1000, based on the residual mass after carbon decomposition around 300–400 °C. This increase in Ir content for C@Ir/CNF1000 can be attributed to the reduced nitrogen content, as shown in Fig. 1g and h, resulting in a relatively higher fraction of Ir. Additionally, C@Ir/CNF1000 exhibited a higher carbon decomposition temperature than C@Ir/CNF700, which can be attributed to increased graphitization at higher temperatures.

Fig. 3 (a) TGA data of C@Ir/CNF700 and C@Ir/CNF1000. (b) XRD data of C@Ir/CNF700 and C@Ir/CNF1000. (c) Raman spectra and (d) I_D/I_G ratio of C@Ir/CNFs according to pyrolysis temperature. XPS data of (e) C@Ir/CNF700 and (f) C@Ir/CNF1000.

XRD analysis was performed to verify the crystal structure of C@Ir/CNF700 and C@Ir/CNF1000 (Fig. 3b). The XRD pattern of C@Ir/CNF700 showed two broad diffraction peaks located at 25.7° and 43.9°, corresponding to the (002) and (100) planes of carbon. The absence of diffraction peaks other than the carbon peaks indicates that the Ir nanoparticles were too small (mean diameter of 1.40 nm) and close to an amorphous state to produce detectable diffraction peaks when pyrolyzed at 700 °C, which aligns well with the TEM results in Fig. 1d. In contrast, the XRD pattern of C@Ir/CNF1000 showed a sharper Ir (111) peak at 41.02°, along with additional diffraction peaks at 46.98°, 69.20°, and 83.72°, corresponding to Ir (200), Ir (220), and Ir (311), respectively. This indicates that the Ir nanoparticles in C@Ir/CNF1000 were well-crystallized, with a mean size of ∼2.46 nm, calculated from the Ir (111) peak using the Scherrer equation (Table S1†), consistent with the TEM results shown in Fig. 2. Additional XRD analysis of C@Ir/CNF samples (Fig. S3†) confirms that no iridium diffraction peaks are observed at 700 °C, while clear Ir crystallization peaks emerge and develop at temperatures around 800 °C and higher (900 and 1000 °C). This observation suggests that the crystallization of Ir nanoparticles begins near 800 °C.^44–46 To investigate carbon-related traits, Raman spectroscopy analysis was further conducted, providing insights into the graphitic structure of the CNFs (Fig. 3c). Two distinct Raman peaks at 1350 cm⁻¹ and 1580 cm⁻¹ correspond to the D band and G band, respectively. As the pyrolysis temperature increased, the ratio of D- and G-band intensities (I_D/I_G), which is related to the degree of disorder in carbon, decreased, indicating improved carbon crystallinity at higher pyrolysis temperatures (Fig. 3d).⁴⁷ Therefore, the CNFs produced at high pyrolysis temperatures exhibit high crystallinity, which is beneficial for their electrical conductivity and chemical stability. However, this also slightly increases the size of the Ir nanoparticles, reducing the electrochemical specific surface area (ECSA) available for catalytic reactions.

XPS analysis of C@Ir/CNF700 and C@Ir/CNF1000 was performed to confirm the surface composition and specific chemical and bonding states (Fig. 3e and f). The top panel of each figure presents a full scan of the XPS spectra, confirming that both samples consist of C, N, O, and Ir. Given the nitrogen-rich nature of both PAN and P4VP, a substantial amount of nitrogen (10.79 at%) was doped into the carbon matrix after pyrolysis at 700 °C. N-doping of CNFs significantly enhances their electrical conductivity by increasing charge carrier density. Furthermore, the incorporation of electron-rich and electron-withdrawing nitrogen species, such as pyridinic and graphitic nitrogen, generates highly active sites that synergistically enhance reaction kinetics and catalytic durability.^48–50 Prior to powdering, XPS analysis of C@Ir/CNF700 and C@Ir/CNF1000 revealed Ir atomic fractions of 3.69 at% and 4.27 at%, respectively, corresponding to surface mass fractions of 36.64 wt% and 41.20 wt% (Table S2†). These values are notably higher than the bulk Ir content measured by TGA (∼9 wt%) in Fig. 3a. This discrepancy arises from the surface-sensitive nature of XPS, which probes only the top ∼10 nm of the material and effectively captures the elemental distribution on the CNF surface.⁵¹ The higher Ir surface fraction observed by XPS highlights that Ir nanoparticles are predominantly localized on the CNF surface, as intended in this study, where they maximize the number of catalytically active sites. This result underscores the successful design of our synthesis strategy, which focuses on the surface distribution of Ir species to optimize the active surface area. In the case of C@Ir/CNF1000, the atomic fractions of N and O are diminished.^47,52 The three panels below show XPS spectra of C 1s, N 1s, and Ir 4f for each sample. From the deconvolution of the C 1s spectra of each sample, the C–N peak for C@Ir/CNF700 shows abundant N content in the carbon matrix. However, the high C–C peak intensity indicates increased graphitization of the CNFs as the pyrolysis temperature increased. For the N 1s spectra, a more distinct difference appears between the two samples. The core-level N 1s spectra of C@Ir/CNF700 show strong Ir–N peaks, indicating that ultra-small Ir nanoparticles are stabilized by N atoms in the thin carbon layer. Meanwhile, the peak intensity of graphitic N increased as the pyrolysis temperature increased to 1000 °C. For the Ir 4f spectra, Ir 4f and IrO₂ 4f peaks are primarily present, indicating that the surfaces of the Ir nanoparticles were inevitably oxidized to native IrO_x. C@Ir/CNF1000 exhibited stronger Ir 4f peaks due to the more reductive conditions at elevated pyrolysis temperatures. Notably, the native IrO_x species also serve as excellent catalysts.

The electrochemical catalytic activity of the as-prepared C@Ir/CNF was evaluated using the rotating disk electrode (RDE) method in a 1 M KOH solution. To assess the catalytic activity for oxygen, cyclic voltammetry (CV) curves of C@Ir/CNFs and commercial Ir/C are presented in Fig. 4a. While no significant peaks were observed in the N₂-saturated solution, cathodic oxygen reduction reaction (ORR) peaks were detected in the range of 0.55–0.75 V vs. the reversible hydrogen electrode (RHE) in the O₂-saturated solution. The starting point of the ORR peaks represents the onset potential (E₀) of the Ir catalyst.^53–55 This confirms the catalytic activity of the C@Ir/CNF samples for oxygen reduction.^56–58

Fig. 4 (a) CV curves of Ir/C and C@Ir/CNFs in N₂-saturated electrolyte (dotted line) and O₂-saturated electrolyte (solid line) with 1 M KOH. (b) OER-LSV curves of Ir/C and C@Ir/CNFs in 1 M KOH. (c) Tafel plots of Ir/C and C@Ir/CNFs derived from OER-LSV curves in (b). (d) Electrical double-layer capacitance (EDLC) based on scan rate dependence of Ir/C and C@Ir/CNFs. (e) OER-LSV curves at the 100th cycle of Ir/C and C@Ir/CNFs in 1 M KOH. (f) Mass activity and specific activity derived from OER-LSV curves at the 100th cycle.

Subsequently, the oxygen evolution reaction (OER) activity of C@Ir/CNFs was assessed through OER linear sweep voltammetry (LSV) curves to evaluate their applicability in Zn–air batteries (Fig. 4b and S5†).^59–61 The overpotential (η) required to reach a current density (j) of 1 mA cm⁻² (due to the low Ir content (∼10 wt%) in the C@Ir/CNFs) ranged from 183.17 to 228.07 mV for the C@Ir/CNF samples, showing an increasing trend with higher pyrolysis temperatures. The results demonstrate a dual effect of increasing pyrolysis temperature on the OER activity, as described below. The amorphous and small-sized Ir nanoparticles in C@Ir/CNF700 exhibit higher surface reactivity (i.e. higher surface energy)^62,63 and a greater density of active sites, contributing to its superior initial OER activity. Elevated pyrolysis temperatures enhance the electrical conductivity of CNFs by increasing their crystallinity. However, the crystallization and growth of Ir nanoparticles at higher temperatures reduces the surface reactivity and density of active sites. Although the improved conductivity helps mitigate these effect, it still results in a slight reduction in the initial OER activity compared to C@Ir/CNF700.^64,65 A similar trend can be observed in the OER-LSV curves without conductive material (Fig. S6†).

To assess the OER kinetics, Tafel plots derived from the OER-LSV curves are presented in Fig. 4c and S7,† with Tafel slope and overpotential values summarized in Fig. S8.† Among the samples, C@Ir/CNF700 exhibited the lowest Tafel slope of 250.49 mV dec⁻¹, indicating superior initial OER kinetics. Except for C@Ir/CNF1000, higher pyrolysis temperatures resulted in increased Tafel slopes, suggesting a decline in catalytic performance. For C@Ir/CNF1000, however, the enhanced Ir crystallinity at the highest pyrolysis temperature improved OER kinetics, resulting in a Tafel slope of 293.83 mV dec⁻¹. Similar to OER activity, OER kinetics also showed a decreasing trend with increasing pyrolysis temperatures. This reduction is attributed to the loss of catalytically active sites due to Ir particle growth.

To further analyze this behavior, the electrochemical active surface area (ECSA), which reflects the number of active catalytic sites for electrochemical reactions, was calculated by measuring the current density at various scan rates in the non-faradaic region (0.85–0.95 V) (Fig. S9†). The double-layer capacitance (C_dl) derived from these measurements is presented in Fig. 4d and S10.† The C_dl, which correlates directly with the ECSA, confirms that higher pyrolysis temperatures result in a reduction of active sites due to Ir particle growth, consistent with the morphological trends observed in Fig. 2.

Building upon these findings, OER-LSV tests were conducted over 100 cycles to evaluate the impact of pyrolysis temperature-induced structural changes on the long-term stability of the catalyst. The results, presented in Fig. 4e, demonstrate how the increased crystallinity of the carbon overlayer and Ir nanoparticles contributes to catalytic stability.^66–68 Despite the reduction in active catalytic sites due to Ir particle growth at higher pyrolysis temperatures, the improved crystallinity of the carbon overlayer enhances its encapsulation effect, while the enhanced structural stability of the CNF support enables C@Ir/CNF1000 to achieve superior catalytic stability and exceptional long-term performance in OER reactions. The mass activity (MA) and specific activity (SA) values, calculated from the loaded Ir at η = 300 mV after 100 cycles, are shown in Fig. 4f. Unlike during the initial cycle (Fig. S11†), the MA and SA values of C@Ir/CNF1000 after 100 cycles confirm its highest catalytic stability, followed by C@Ir/CNF900, C@Ir/CNF800, and C@Ir/CNF700.

In addition, to confirm the ORR properties, ORR-LSV evaluation of C@Ir/CNFs was conducted (Fig. S12†).^69,70 The onset potential (E_oc) and half-wave potential (E_1/2) values of the commercial Ir/C and C@Ir/CNF catalysts are similar, which is attributed to the intrinsic ORR catalytic properties of Ir in each sample (Fig. S13†). Furthermore, the limiting current density (j_L) and Tafel slope values from the ORR-LSV curves (Fig. S14 and S15†) show a trend similar to the OER, where increasing pyrolysis temperatures reduce catalytic activity and slow kinetics due to Ir particle growth and a decrease in active surface area.

Finally, an aqueous zinc–air battery (ZAB) cell was assembled to evaluate the reversible catalytic characteristics of C@Ir/CNFs. The ZAB cell consisted of a zinc metal anode, a C@Ir/CNF cathode, and an aqueous electrolyte (6 M KOH + 0.2 M ZnAc), as depicted in Fig. 5a and S16.† To confirm the cell stability before cycling, the open-circuit voltage (OCV) was measured for 1 hour, revealing that the values for C@Ir/CNF700 (1.452 V) and C@Ir/CNF1000 (1.420 V) were higher than that of commercial Ir/C (1.405 V) (Fig. 5b). Furthermore, as shown in the polarization curves in Fig. 5c, the ZAB cell with C@Ir/CNF700 exhibited a higher power density of 69.326 mW cm⁻² compared to Ir/C (58.327 mW cm⁻²), suggesting the excellent ability of C@Ir/CNF700 to recharge and discharge rapidly. In addition, the specific capacity normalized to the mass of consumed Zn at a constant discharge current density (25 mA cm⁻²) was 515.03, 617.65, and 647.62 mAh g_Zn⁻¹ for Ir/C, C@Ir/CNF700, and C@Ir/CNF1000, respectively (Fig. S17†). The higher specific capacity of C@Ir/CNF1000 compared to C@Ir/CNF700 can be attributed to the increased crystallinity and enhanced electrical conductivity of the CNFs at higher carbonization temperatures, which facilitated more efficient electron transfer. Additionally, the crystallized Ir nanoparticles in C@Ir/CNF1000 provided greater structural stability, minimizing degradation during the reaction, whereas the amorphous Ir nanoparticles in C@Ir/CNF700 were less stable under operating conditions. To examine the stability of the ZAB associated with catalyst degradation and oxidation of Ir nanoparticles, a cyclic test was carried out with each charge–discharge period lasting 12 minutes at a current density of 5 mA cm⁻² (Fig. 5d). The charging voltage during the 1st cycle, reflecting the initial OER activity, was lower for C@Ir/CNF700 (2.006 V) and Ir/C (1.999 V) than for C@Ir/CNF1000 (2.085 V). Similarly, the voltage gap in the cycling curve followed the trend Ir/C (0.773 V) < C@Ir/CNF700 (0.821 V) < C@Ir/CNF1000 (0.899 V). In contrast, the cycling stability of the catalysts indicated that C@Ir/CNF1000, despite having the highest initial voltage gap, demonstrated the best cycling performance compared to C@Ir/CNF700 and Ir/C. After 30 hours, the ZAB cell with C@Ir/CNF1000 recorded the lowest charging voltage (2.096 V) compared to C@Ir/CNF700 (2.394 V) and Ir/C (2.495 V), with voltage gaps of 0.932, 1.259, and 1.386 V, respectively. The battery cell performance, along with detailed experimental conditions, is compared with data from other literature sources in Table S3.†

Fig. 5 (a) Schematic illustration of a home-made Zn–air battery. (b) Stable retention of open-circuit voltage (OCV) for ZAB with Ir/C, C@Ir/CNF700, and C@Ir/CNF1000. (c) Discharge polarization curve and power density of ZAB with Ir/C, C@Ir/CNF700, and C@Ir/CNF1000. (d) Charge–discharge cycling curves of ZAB with Ir/C, C@Ir/CNF700, and C@Ir/CNF1000. (e) Energy efficiency of ZAB with Ir/C, C@Ir/CNF700, and C@Ir/CNF1000.

Moreover, to confirm the oxidation of the metal catalyst during the operation of ZAB cells, ex situ XPS and XRD analysis was conducted (Fig. S18 and S19†). The XPS spectrum after the 10th cycle confirmed a high level of IrO₂, indicating that the oxidation of the Ir catalyst occurs during cycling in aqueous media. However, in the XRD spectrum, due to the high intensity of peaks from the underlying gas diffusion layer (GDL), peak of IrO₂ were not clearly confirmed. Regarding discharge–charge efficiency, the ZAB cells with Ir/C and C@Ir/CNF700 exhibited the highest initial efficiency (Fig. 5e). However, as the cycle number increased, C@Ir/CNF1000 maintained a remarkable efficiency of 62.34%, outperforming the other two samples even after 70 cycles.

Taking all of the above into consideration, we observed a trade-off between catalytic activity and durability. The schematic illustration in Fig. 6 compares the structural and electrochemical characteristics of C@Ir/CNF700 and C@Ir/CNF1000, highlighting the effects of crystallinity, conductivity, and strong metal–support interaction (SMSI) on the catalyst's performance and stability. While C@Ir/CNF700 initially exhibits high catalytic activity, it suffers from low crystallinity and conductivity, resulting in rapid degradation. In contrast, C@Ir/CNF1000 achieves a balance of high crystallinity, conductivity, and SMSI, ensuring superior long-term stability and catalytic performance in ZAB cells.⁷¹

Fig. 6 (a) Schematic illustration of the pyrolysis temperature dependence of C@Ir/CNFs. (b) C@Ir/CNFs in a Zn–air battery at the initial cycle. (c) C@Ir/CNFs in a Zn–air battery after long-term cycling.

Conclusion

In conclusion, a facile synthesis method for catalyst-loaded CNFs encapsulated in N-doped carbon layers was successfully demonstrated. The simple one-step thermal process for synthesizing metal catalysts on CNFs ensured the uniform dispersion of various types of ultrafine catalytic nanoparticles on the CNF surface. These catalyst-loaded CNFs produced through the facile process exhibited enhanced electrochemical performance and durability. They showed significant improvements in catalytic activity and stability, particularly as cathodes in Zn–air batteries. Future research should focus on expanding the types of metal catalysts and further refining the synthesis process to enhance the efficiency and application range of these nanostructured materials. The potential of this technology to be scaled up and integrated into existing industrial processes could lead to significant advancements in pollution control and energy conversion systems, thereby supporting global efforts toward a cleaner and more energy-efficient future.

Data availability

Data for this article are available at Open Science Framework at URL: https://doi.org/10.17605/OSF.IO/ASGKC.

Author contributions

Conceptualization: Hyeong Min Jin, Ji-Won Jung, Seong-Woon Yoon, Dae-Kwon Boo; data curation: Seong-Woon Yoon, Dae-Kwon Boo; formal analysis: Hyeong Min Jin, Ji-Won Jung; funding acquisition: Seong-Woon Yoon, Dae-Kwon Boo; investigation: Hyeong Min Jin, Ji-Won Jung, Seong-Woon Yoon, Dae-Kwon Boo; methodology: Hyeong Min Jin, Ji-Won Jung; project administration: Hyeong Min Jin, Ji-Won Jung; resources: Hyunmin Na, Su-Ho Cho; software: Hyeong Min Jin, Ji-Won Jung; supervision: Hyunmin Na, Su-Ho Cho, Hyun-Soo Chang, Tae-Yeon Kim, Ji Sung Park; validation: Hyun-Soo Chang, Tae-Yeon Kim, Ji Sung Park; visualization: Seong-Woon Yoon, Dae-Kwon Boo; writing – original draft preparation: Seong-Woon Yoon, Dae-Kwon Boo, Hyeong Min Jin, Ji-Won Jung; writing – review & editing: Hyeong Min Jin, Ji-Won Jung.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research was supported by the National Research Foundation of Korea (NRF) grant funded by the Ministry of Science (RS-2024-00345469, RS-2023-00213749, RS-2023-00243617) and the National Research Council of Science & Technology (NST) grant by the Korea government (MSIT) (CRC23031-220). This research was supported by Korea Institute for Advancement of Technology (KIAT) grant funded by the Korea Government (MOTIE) (P0023727). This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. RS-2023-00217778).

References

1. J.-W. Jung, J. S. Nam, K. Klyukin, D.-Y. Youn and I.-D. Kim, Nano Energy, 2021, 83, 105821.
2. J.-W. Seo, J.-H. Park, J.-W. Jung and S.-J. Choi, Sens. Actuators, B, 2024, 412, 135750.
3. J. S. Bonso, G. D. Kalaw and J. P. Ferraris, J. Mater. Chem. A, 2014, 2, 418–424.
4. G. Wang, B. Qian, Y. Wang, Q. Dong, F. Zhan and J. Qiu, New J. Chem., 2016, 40, 3786–3792.
5. S. N. Arshad, M. Naraghi and I. Chasiotis, Carbon, 2011, 49, 1710–1719.
6. Y. Li, B. Zou, C. Hu and M. Cao, Carbon, 2016, 99, 79–89.
7. Y. Shen, L. Li, K. Xiao and J. Xi, ACS Sustainable Chem. Eng., 2016, 4, 2351–2358.
8. S. Ma, Y. Wang, Z. Liu, M. Huang, H. Yang and Z.-L. Xu, Chem. Eng. Sci., 2019, 205, 181–189.
9. X. Zhou, B. Liu, Y. Chen, L. Guo and G. Wei, Mater. Adv., 2020, 1, 2163–2181.
10. X. Yan, X. Xiao, C. Au, S. Mathur, L. Huang, Y. Wang, Z. Zhang, Z. Zhu, M. J. Kipper, J. Tang and J. Chen, J. Mater. Chem. A, 2021, 9, 21659–21684.
11. G. Xia, L. Zhang, F. Fang, D. Sun, Z. Guo, H. Liu and X. Yu, Adv. Funct. Mater., 2016, 26, 6188–6196.
12. X. Lu, R. Chen, S. Shen, Y. Li, H. Zhao, H. Wang, T. Wu, Y. Su, J. Luo, X. Hu, S. Ding and W. Li, Nano Lett., 2024, 24, 5490–5497.
13. B. Zhang, Y. Yu, Z. Huang, Y.-B. He, D. Jang, W.-S. Yoon, Y.-W. Mai, F. Kang and J.-K. Kim, Energy Environ. Sci., 2012, 5, 9895–9902.
14. S. Sakurai, T. Tsuji, J. He, M. Yamada and D. N. Futaba, ACS Appl. Nano Mater., 2024, 7(11), 12745–12751.
15. D. Xue and J.-N. Zhang, Ind. Chem. Mater., 2024, 2, 173–190.
16. M. Karuppannan, Y. Kim, S. Gok, E. Lee, J. Y. Hwang, J.-H. Jang, Y.-H. Cho, T. Lim, Y.-E. Sung and O. J. Kwon, Energy Environ. Sci., 2019, 12, 2820–2829.
17. Y. Hu, T. Shen, X. Zhao, J. Zhang, Y. Lu, J. Shen, S. Lu, Z. Tu, H. L. Xin and D. Wang, Appl. Catal., B, 2020, 279, 119370.
18. Y. Nie, S. Chen, W. Ding, X. Xie, Y. Zhang and Z. Wei, Chem. Commun., 2014, 50, 15431–15434.
19. S. G. Ji, H. C. Kwon, T.-H. Kim, U. Sim and C. H. Choi, ACS Catal., 2022, 12, 7317–7325.
20. X. Tong, J. Zhang, G. Zhang, Q. Wei, R. Chenitz, J. P. Claverie and S. Sun, Chem. Mater., 2017, 29, 9579–9587.
21. D. Lee, Y. Kim, J. Song, H. J. Choi, M. Karuppannan, Y.-H. Cho and O. J. Kwon, ACS Appl. Energy Mater., 2022, 5, 596–603.
22. H. Lee, Y.-E. Sung, I. Choi, T. Lim and O. J. Kwon, J. Power Sources, 2017, 362, 228–235.
23. K. Ham, S. Chung and J. Lee, J. Power Sources, 2020, 450, 227650.
24. M. Wei, L. Zhang, H. Zhang, G. Zhang, Q. Zhang and B. Zhang, Carbon, 2023, 213, 118254.
25. C.-L. Zhang, H. Li, Q. Zhang, F.-H. Cao, Y. Xie, B.-R. Lu, W.-D. Zhang, H.-P. Cong and H. Li, Chem. Eng. J., 2021, 420, 127705.
26. M. Dirican, M. Yanilmaz, A. M. Asiri and X. Zhang, J. Electroanal. Chem., 2020, 861, 113995.
27. X. Xu, F. Yao, O. A. A. Ali, W. Xie, S. F. Mahmoud, P. Xie, S. M. El-Bahy, M. Huang, C. Liu, R. Fan, Z. Guo, A. Du, D. Estevez, F. Qin, H. Peng, D. P. Young and H. Gu, Adv. Compos. Hybrid Mater., 2022, 5, 2002–2011.
28. H. Liu, W. Fan, H. Lv, W. Zhang, J. Shi, M. Huang, S. Liu and H. Wang, ACS Appl. Energy Mater., 2021, 4, 2327–2338.
29. H.-F. Li, C. Wang, L.-F. Liu, N. Xie, M. Pan, P. Wu, X.-D. Wang, Z. Zeng, S. Deng and G.-P. Dai, J. Nanopart. Res., 2020, 22, 10.
30. J.-H. Park, K. Kim, X. Wang, M. Huda, Y. Sawada, Y. Matsuo, N. Saito and M. Kawasumi, J. Power Sources, 2023, 580, 233419.
31. P. Liu, A. Klyushin, P. Chandramathy Surendran, A. Fedorov, W. Xie, C. Zeng and X. Huang, ACS Nano, 2023, 17, 24395–24403.
32. P. Yin, S.-C. Shen, L.-L. Zhang, X.-S. Zheng, M. Zuo, Y.-W. Ding and H.-W. Liang, Cell Rep. Phys. Sci., 2022, 3, 100984.
33. Z. Guo, C. Li, W. Li, H. Guo, X. Su, P. He, Y. Wang and Y. Xia, J. Mater. Chem. A, 2016, 4, 6282–6289.
34. S. Li, W. Chen, H. Pan, Y. Cao, Z. Jiang, X. Tian, X. Hao, T. Maiyalagan and Z.-J. Jiang, ACS Sustainable Chem. Eng., 2019, 7, 8530–8541.
35. R. Akmalia, F. Balqis, M. F. Andriani, Y. Irmawati and A. Sumboja, J. Energy Storage, 2023, 72, 108743.
36. Y. Yang, Z. Wang, Z. Liang, L. Shen, C. Guo, Y. Shi, H. Tan, Z. Lu and C. Yan, ACS Appl. Mater. Interfaces, 2022, 14, 43690–43700.
37. Y. H. Jang, S. T. Kochuveedu, Y. J. Jang, H.-Y. Shin, S. Yoon, M. Steinhart and D. H. Kim, Carbon, 2011, 49, 2120–2126.
38. A. T. Rodriguez, X. Li, J. Wang, W. A. Steen and H. Fan, Adv. Funct. Mater., 2007, 17, 2710–2716.
39. M. Wu, X. Yang, X. Cui, N. Chen, L. Du, M. Cherif, F. K. Chiang, Y. Wen, A. Hassanpour, F. Vidal, S. Omanovic, Y. Yang, S. Sun and G. Zhang, Nano-Micro Lett., 2023, 15, 232.
40. L. Bai, D. Wang, H. Shen, W. Wang, S. Li and W. Yan, J. Mater. Chem. A, 2024, 12, 3997–4007.
41. A. Muthurasu, I. Pathak, D. Acharya, Y. R. Rosyara and H. Y. Kim, J. Mater. Chem. A, 2024, 12, 1826–1839.
42. M. B. Poudel, S. Vijayapradeep, K. Sekar, J. S. Kim and D. J. Yoo, J. Mater. Chem. A, 2024, 12, 10185–10195.
43. X. Chen, S. Yan, S. Wen, J. Chen, J. Xu, C. Wang and X. Lu, J. Colloid Interface Sci., 2023, 641, 782–790.
44. N. Saadatkhah, A. Carillo Garcia, S. Ackermann, P. Leclerc, M. Latifi, S. Samih, G. S. Patience and J. Chaouki, Can. J. Chem. Eng., 2019, 98, 34–43.
45. J. Hämäläinen, T. Hatanpää, E. Puukilainen, T. Sajavaara, M. Ritala and M. Leskelä, J. Mater. Chem., 2011, 21, 16488–16493.
46. N. Bhuvanendran, S. Ravichandran, W. Zhang, Q. Ma, Q. Xu, L. Khotseng and H. Su, Int. J. Hydrogen Energy, 2020, 45, 6447–6460.
47. Z. Zhang, Y. Zhang, X. Mu, J. Du, H. Wang, B. Huang, J. Zhou, X. Pan and E. Xie, Electrochim. Acta, 2017, 242, 100–106.
48. S. Wang, C. Dai, J. Li, L. Zhao, Z. Ren, Y. Ren, Y. Qiu and J. Yu, Int. J. Hydrogen Energy, 2015, 40, 4673–4682.
49. L. Lu, X. Cao, Z. Shen, L. Li, J. Huo, W. Chen, C. Liu and H. Liu, Sustainable Mater. Technol., 2020, 26, e00221.
50. Z. R. Ismagilov, A. E. Shalagina, O. Y. Podyacheva, A. V. Ischenko, L. S. Kibis, A. I. Boronin, Y. A. Chesalov, D. I. Kochubey, A. I. Romanenko, O. B. Anikeeva, T. I. Buryakov and E. N. Tkachev, Carbon, 2009, 47, 1922–1929.
51. J. Lefebvre, F. Galli, C. L. Bianchi, G. S. Patience and D. C. Boffito, Can. J. Chem. Eng., 2019, 97, 2588–2593.
52. L. Wan, J. Wang, L. Xie, Y. Sun and K. Li, ACS Appl. Mater. Interfaces, 2014, 6, 15583–15596.
53. Y. Zhang, H. Zhang, Y. Zhang, Y. Ma, H. Zhong and H. Ma, Chem. Commun., 2009, 43, 6589–6591.
54. K. Chakrapani and S. Sampath, Phys. Chem. Chem. Phys., 2014, 16, 16815–16823.
55. S. G. Peera, A. K. Sahu, S. D. Bhat and S. C. Lee, RSC Adv., 2014, 4, 11080–11088.
56. B. Chen, X. He, F. Yin, H. Wang, D. J. Liu, R. Shi, J. Chen and H. Yin, Adv. Funct. Mater., 2017, 27, 1700795.
57. Y. Ying, J. F. Godinez Salomon, L. Lartundo-Rojas, A. Moreno, R. Meyer, C. A. Damin and C. P. Rhodes, Nanoscale Adv., 2021, 3, 1976–1996.
58. Z. Guo, F. Wang, Y. Xia, J. Li, A. G. Tamirat, Y. Liu, L. Wang, Y. Wang and Y. Xia, J. Mater. Chem. A, 2018, 6, 1443–1453.
59. J. Li, J. Song, B.-Y. Huang, G. Liang, W. Liang, G. Huang, Y. Qi Jin, H. Zhang, F. Xie, J. Chen, N. Wang, Y. Jin, X.-B. Li and H. Meng, J. Catal., 2020, 389, 375–381.
60. A. Sumboja, J. Chen, Y. Zong, P. S. Lee and Z. Liu, Nanoscale, 2017, 9, 774–780.
61. L.-N. Han, L.-B. Lv, Q.-C. Zhu, X. Wei, X.-H. Li and J.-S. Chen, J. Mater. Chem. A, 2016, 4, 7841–7847.
62. M. Elmaalouf, et al. , Nat. Commun., 2021, 12, 3935.
63. E. Willinger, C. Massué, R. Schlögl and M. G. Willinger, J. Am. Chem. Soc., 2017, 139, 12093–12101.
64. L. Luo, S. Wang, Z. Wu, Z. Qin, H. Zhu, W. Fan, M. Dong and J. Wang, Appl. Catal., A, 2022, 641, 118699.
65. Q. Zhang, Z. Duan, Y. Wang, L. Li, B. Nan and J. Guan, J. Mater. Chem. A, 2020, 8, 19665–19673.
66. L. Zhang, H. Jiang, M. Tang, Y. Jiang, B. Tang, H. Tan, Y. Kong and H. Hu, Appl. Catal., B, 2024, 354, 124151.
67. D. Deng, K. S. Novoselov, Q. Fu, N. Zheng, Z. Tian and X. Bao, Nat. Nanotechnol., 2016, 11, 218–230.
68. D. Deng, L. Yu, X. Chen, G. Wang, L. Jin, X. Pan, J. Deng, G. Sun and X. Bao, Angew. Chem., Int. Ed., 2013, 52, 371–375.
69. Y. Ma, Y. Xiao, Y. Ge, D. Gao, Y. Zhang, Z. Li and Y. Han, J. Mater. Chem. A, 2024, 12, 2004–2010.
70. Y. Zhang, H. Wu, J. Ma, J. Luo, Z. Lu, S. Feng, Y. Deng, H. Chen, Q. Wang, Z. Miao, P. Rao, N. Yu, Y. Yuan, J. Li and X. Tian, J. Mater. Chem. A, 2024, 12, 26484–26492.
71. Y. Zhang, J. Wang, M. Alfred, P. Lv, F. Huang, Y. Cai, H. Qiao and Q. Wei, Energy Storage Mater., 2022, 51, 181–211.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ta08388a

‡ Contributed equally to this work.

---