Interfacial Engineering of N‐Doped Carbon Host via Pore and Oxygen Coregulation for Superior Zinc‐Iodine Batteries

ABSTRACT

Aqueous zinc–iodine batteries (AZIBs) are promising for grid‐scale energy storage yet face challenges of polyiodide shuttling and sluggish kinetics in the iodine cathode. Nitrogen‐doped porous carbons are widely used as host materials to mitigate these issues. However, the potential roles of different nitrogen configurations and their synergy with trace oxygen doping in the iodine conversion process remain unclear. Herein, a hierarchical N/O codoped porous carbon (ONC) is synthesized through a simple citrate‐assisted pyrolysis method and utilized as an advanced host. The codoping strategy modulates the electronic structure of the carbon surface, significantly enhancing the chemical adsorption and catalytic conversion of iodine species, while the multi‐scale pores provide effective confinement. This synergy, combined with abundant surface defects, reduces the charge transfer barrier and facilitates ion transport, resulting in rapid surface conversion kinetics. Consequently, the ONC/I₂ cathodes deliver a high capacity of 205.8 mAh g–¹ at 0.1 A g–¹ and excellent stability, retaining 59% of their initial capacity over 10,000 cycles at 1 A g–¹. This work provides deep insights into the synergistic mechanisms of heteroatom codoping in carbon hosts, guiding the rational design of high‐performance iodine cathodes through the integration of pore engineering and interfacial chemistry.

A sodium citrate‐assisted pyrolysis route converts ZIF‐8 into O/N codoped carbon with hierarchical porosity as an advanced iodine host. The regulated interfacial chemistry enhances iodine chemisorption and conversion, while the porous structure confines polyiodides. This strategy significantly boosts reaction kinetics and ion transport, resulting in Zn‐I₂ batteries with excellent capacity and cycling stability.

1. Introduction

Benefiting from the abundance, cost‐effectiveness, and suitable theoretical capacity of iodine (211 mAh g–¹), iodine has made metal iodine batteries (such as lithium‐ and zinc‐iodine systems) promising candidates for next‐generation energy storage [1, 2, 3]. Among these systems, aqueous zinc iodine batteries (AZIBs) employ aqueous electrolytes and highly compatible zinc anodes, demonstrating remarkable advantages such as economic feasibility, environmental benignity, intrinsic safety, and broad power density applicability. Consequently, AZIBs are regarded as highly promising for large‐scale grid energy storage applications [4, 5]. However, this battery system still faces several critical challenges at the cathode‐electrolyte interface, including the notorious shuttling effect and sluggish reaction kinetics. The soluble polyiodides (such as I₃ ^– and I₅ ^–) formed during cycling can leak into the electrolyte and potentially migrate to the zinc anodes, triggering parasitic reactions. This leads to rapid capacity decay, reduced Coulombic efficiency (CE), and irreversible consumption of active materials on both electrodes [6]. Moreover, the sluggish conversion kinetics of iodine, accompanied by its intrinsically low electrical conductivity, restrict the practical utilization of cathode materials and impede the achievement of high power density in AZIBs.

To solve these issues, carbonaceous materials with excellent electric conductivity, high porosity, and large specific surface area have been widely employed as host matrices for iodine in cathode fabrication [7]. The microporous structure (<1 nm) in carbon hosts effectively suppresses the formation and dissolution of I_3– and other intermediates, mitigating the shuttle effects [8]. Pores in the range of 1–3 nm provide large pore volumes for high iodine loading and facilitate electrolyte infiltration, while with relatively weaker confinement capability for polyiodides [9]. It should be noted that numerous “dead pores”, with insufficient contact with electrolytes, limit the full utilization of iodine active materials. Therefore, rational design of pore structures in carbon hosts is crucial for enhancing the electrochemical performance of iodine cathodes. Surface chemistry regulation is another critical factor in optimizing carbon hosts. Introducing non‐metal hetero‐atoms (such as N, P, S, B, O) has been demonstrated as an effective approach to improve their performance [10, 11, 12, 13, 14]. Heteroatom doping induces charge redistribution, generates structural defects, and creates unsaturated coordination sites, which collectively regulate the surface polarity of carbon, thereby strengthening the chemical adsorption of iodine and polyiodides. This compensates for the limitation in physical adsorption of pure carbon matrices relying solely on van der Waals interactions [15]. Furthermore, the hydrophilic functional groups on the surface improve the electrode wettability and promote the Zn²⁺ transport, elevating the utilization of active materials and specific capacity [16, 17]. Additionally, such heteroatoms often exhibit electrocatalytic activities toward iodine conversion reactions, accelerating the interfacial charge transfer and enhancing the rate capability of batteries [10]. Heteroatom‐doped carbon substrates also serve as essential supports for fabricating single‐atom metal catalysts [18, 19, 20]. Constructing suitable pore structures coupled with tailored surface chemical environments remains a key challenge for further advancing the performance of AZIBs.

Among these non‐metal doped carbons, N‐doped carbons are the most widely studied in fields such as electrocatalysis, capacitive deionization, and zinc‐iodine batteries [21, 22]. Gong et al. [23] prepared hierarchical porous N‐doped carbon hosts for iodine cathodes, achieving a high specific capacity of 219.3 mAh g–¹ at 1C and excellent stability over 10,000 cycles at 5C. This enhanced performance is attributed to the abundant pyridine‐N sites and well‐connected micro/meso‐pores. Liu and coworkers [10] utilized carbon hosts with rich graphitic‐N sites to catalyze the rapid conversion of I_3–, achieving a high energy density of 320 Wh kg–¹ and stable performance over 10,000 cycles. They attributed the superior electrocatalytic activity primarily to graphitic‐N, highlighting its crucial role in enhancing the reaction kinetics compared to other nitrogen configurations. Ye et al. [24] developed an optimized preparation strategy for zeolitic imidazole framework‐derived carbons through pre‐activation of the precursors, resulting in defect‐rich carbon hosts anchored with Ni/Zn bimetallic sites. The iodine electrodes based on the host materials delivered a high specific capacity of 219 mAh g–¹ at 5 A g–¹ and long‐term cycling stability over 20,000 times at 2 A g–¹. Notably, the preparation of porous carbon often results in incidental oxygen doping, influenced by precursor structures and environmental conditions. It is reported that O‐doping in carbon hosts enhances their hydrophilicity and adsorption of Zn²⁺, potentially improving the iodine utilization and contributing additional capacity via a pseudocapacitive mechanism [14, 16, 17, 25]. Zhou and coauthors [16] demonstrated that the surface hydroxyl groups (−OH) exhibit electrocatalytic activity toward iodine redox reactions, enhancing reaction kinetics, reversibility, and iodine nucleation. However, research on N and O codoped carbons remains relatively limited. The influence of trace oxygen on the reaction kinetics and underlying mechanism of iodine conversion is unclear. The synergistic effects between N and O doping require further in‐depth investigations. For instance, how the introduction of O modulates the electrocatalytic activity of various N sites and alters iodine conversion pathways still lacks systematic understanding.

In this work, O and N codoped hierarchical porous carbon hosts (ONC) was synthesized through the pyrolysis of zeolitic imidazole framework‐8 (ZIF‐8) pre‐absorbed with trisodium citrate and utilized as iodine host for AZIBs. Compared with the N‐doped microporous carbon (NC) derived from ZIF‐8, the effects of pore structure and surface chemical environment on the electrochemical performance of the iodine cathode were systematically studied. The electronic configuration of the carbon surface regulated by N and trace O, along with their influence on the kinetics of the iodine conversion reaction, was investigated at the atomic scale using DFT calculations. Thanks to the well‐interconnected pore structure and abundant polar surface sites, the ONC/I₂ cathodes exhibit high active material utilization, strong polyiodides confinement capability, and rapid Zn ion transport, delivering a reversible specific capacity of 205.8 mAh g–¹ at 0.1 A g–¹. The batteries based on the ONC/I₂ cathodes demonstrate an excellent rate capability of 178.3 mAh g–¹ at 1 A g–¹ and superior cycling stability over 10,000 cycles. This study highlights that simultaneous optimization of pore structure and surface chemistry serves as an effective interfacial engineering strategy for enhancing conversion kinetics and suppressing the shuttling effect.

2. Experimental

2.1. Materials

Preparation of ZIF‐8: 1.2 g of Zn(NO₃)₂·6H₂O (SCRC, AR) and 2.6 g of 2‐methylimidazole (SAITONG, >99%) were separately dissolved in 40 mL of methanol (Fuchen, >99.5%). The two solutions were then mixed, vigorously stirred for 2 h, and aged for 12 h at room temperature. The resulting white precipitates were collected by centrifugation at 8,000 rpm, washed with methanol three times, and dried at 60°C for 12 h in air.

Preparation of NC: Typically, 1 g of ZIF‐8 sample was placed in a quartz boat and heated to 900 °C with a heating rate of 5 °C min–¹, kept for 2 h, and naturally cooled down to room temperature in a tube furnace under the protection of high‐purity N₂ (99.999%). The obtained products were denoted as NC.

Preparation of ONC: 0.5 g of as‐synthesized ZIF‐8 was immersed in 20 mL of aqueous solution containing trisodium citrate (SCRC, AR, 0.15 mol L–¹), stirred for 3 h, and aged for 12 h at room temperature. The product was collected by filtration, dried at 60°C for 12 h in air, and labeled as SC‐ZIF‐8. The ONC sample was then obtained by pyrolyzing SC‐ZIF‐8 under the same conditions as for NC.

2.2. Characterization

The morphology of the samples was characterized using a field emission scanning electron microscope (SEM, HITACHI, JSM‐7610F) and a transmission electron microscope (TEM, JEOL2100F). The lattice structure was analyzed by X‐ray diffraction (XRD, Rigaku, SmartLab) with Cu Kα radiation. The surface elemental composition was determined by X‐ray photoelectron spectroscopy (XPS, Thermo‐Fisher, ESCALAB 250Xi) using Al Kα radiation. Raman spectra were acquired on a Renishaw InVia Qontor spectrometer equipped with a 532 nm laser as the excitation source. N₂ adsorption‐desorption isotherms of the carbon hosts were measured at 77 K using a Micromeritic ASAP2460 system. The mass loading of iodine was evaluated by a PerkinElmer STA 8000 thermal analyzer. Dissolved iodine species were monitored using an ultraviolet visible spectrometer (UV‐vis, Persee, TU‐1950).

Adsorption tests: A saturated iodine solution was first prepared by adding ​excess​ iodine to 10 mL of water and stirring overnight. The saturated iodine was obtained by collecting the supernatant after the undissolved iodine particles had settled. Subsequently, 20 mg of NC and ONC powders were separately immersed in 9 mL of the saturated iodine solution. The color changes of mixtures were observed and recorded after 20 min, 4 h, 24 h, and 48 h. For I₃ ^– adsorption, ZnSO₄ solution (2 mol L⁻¹) containing 4 mmol L⁻¹ of ZnI₂ and 2 mmol L⁻¹ of I₂ was used to generate the I₃ ^– species, with all other conditions identical to those of I₂ adsorption tests.

2.3. Electrochemical Measurement

Preparation of iodine cathode: The carbon‐iodine composites were prepared by manually grinding porous carbons (NC or ONC) and I₂ powder at a weight ratio of 1:1 in an agate mortar, followed by heating at 120 °C for 4 h in a sealed vessel. The composite was mixed with carboxymethyl cellulose and conductive carbon black in a weight ratio of 7:2:1, and dispersed in deionized water to form a homogeneous slurry. The slurry was coated onto graphite paper (thickness, 25 µm), dried at 30 °C in air overnight, and then punched into 12 mm disks to serve as cathodes. The mass loading of iodine in cathodes was determined as 0.6–0.7 mg cm⁻². The specific capacity was calculated based on the iodine mass.

Battery assembly and measurements: The electrochemical performance of iodine cathodes was evaluated using CR2025 coin cells with a commercial Zn anode (diameter, 12 mm), a glass fiber separator (Olegee, thickness, 290 µm), and 20 µL of 2 M ZnSO₄ electrolyte. Electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV) were conducted on a CHI 660E electrochemical workstation. Galvanostatic charge‐discharge (GCD) tests were carried out using LAND battery test systems at 25 °C under thermostat control. Galvanostatic intermittent titration technique (GITT) measurements were performed with a current density of 100 mA g⁻¹, a pulse width of 120 s, and a relaxation interval of 300 s. The diffusion coefficients were calculated according to the following equation,

$$D=\frac{4}{\pi \tau }{\left(\frac{m_B{V}_M}{M_{B}S}\right)}^2{\left(\frac{\mathrm{\Delta }{E}_S}{\mathrm{\Delta }{E}_t}\right)}^2$$

where τ is the pulse time; m _B, V _m, and M _B denote the mass, molar volume, and molar mass of iodine active materials, respectively; S refers to the electrode area; ΔE _s and ΔE _t represent the voltage changes during the steady state and the pulse phase, respectively.

2.4. DFT Calculation

All calculations were performed using spin‐unrestricted density functional theory (DFT) as implemented in the Vienna Ab initio Simulation Package (VASP) [26, 27, 28]. The exchange correlation functional was treated within the generalized gradient approximation (GGA) in the Perdew‐Burke‐Ernzerhof (PBE) form [29]. The projector augmented wave (PAW) method was used to model the interactions between ionic cores and valence electrons [30, 31]. A 5×5 graphite monolayer was adopted as the pure carbon substrate, with a 10 Å vacuum layer added on both sides along the z‐direction to avoid periodic interactions. NC and ONC models were constructed by partially substituting C atoms with N or O atoms. A plane wave energy cutoff of 500 eV was used. The Brillouin zone was sampled with a 2×2×1 k‐point mesh generated by the Monkhorst‐Pack method. The convergence criteria for energy and forces were set to 10–⁶ eV and 0.01 eV Å–², respectively. Van der Waals interactions were incorporated through Grimme's DFT‐D3 method with Becke‐Johnson damping [32, 33]. The adsorption energy (E_ad) was defined as: E _ad = E _A+S—E _A—E _S, where E _A+S, E _A, and E _S denote the total energy of the adsorbed system, the isolated adsorbates, and the clean substrates, respectively. Free energy (ΔG) was calculated as: ΔG = ΔE + ΔZPE—TΔS, where ΔE is the electronic energy obtained from DFT, ΔZPE refers to zero‐point energy, and ΔS represents the entropy.

3. Results and Discussion

As depicted in Figure 1a, ZIF‐8 was successfully synthesized through a coprecipitation method. The obtained ZIF‐8 was then immersed in the trisodium citrate solution to obtain the SC‐ZIF‐8 precursor through a simple adsorption process. Both ZIF‐8 and SC‐ZIF‐8 were pyrolyzed at 900°C under N₂ atmosphere to yield NC and ONC materials, respectively. Iodine was subsequently loaded into the pores of NC and ONC via a melting diffusion method. SEM imaging reveals that the ZIF‐8 particles exhibit rhombic dodecahedral morphology with an average size of ∼100 nm (Figure 1b). Upon immersion of ZIF‐8 in an aqueous sodium citrate solution, a colloidal mixture formed as the pH changed from an initial 8.67 to a final 9.67 (Figure S1), suggesting successful citrate adsorption and partial ligand exchange. The SC‐ZIF‐8 sample maintains this dodecahedral shape with negligible changes in particle size (Figure 1c). The XRD pattern of ZIF‐8 matches those reported in the literature [34], confirming the successful synthesis of ZIF‐8 (Figure S2). Furthermore, the XRD pattern of SC‐ZIF‐8 retains all the characteristic features of ZIF‐8 without significant peak shifts, additional diffraction peaks, or loss of crystallinity. Meanwhile, the FT‐IR spectra of ZIF‐8 and SC‐ZIF‐8 display nearly identical features (Figure S3a), indicating that only trace amounts of trisodium citrate were incorporated into the framework. XPS analysis further confirms the successful uptake of citrates, as evidenced by the presence of Na in SC‐ZIF‐8, demonstrating (Figure S3b‐f). High‐resolution C 1s and O 1s spectra show signals corresponding to carboxylate groups, which also support the presence of citrates. Notably, the Zn 2p peaks shift slightly by 0.2 eV toward lower binding energy, suggesting possible coordination between citrates and Zn in ZIF‐8. The ligand displacement by citrates leads to the release of 2‐methyle imidazole, whose subsequent hydrolysis results in the observed increase in pH.

FIGURE 1.

(a) Scheme of the NC and ONC preparation process; (b–e) SEM images of (b) ZIF‐8, (c) SC‐ZIF‐8, (d) NC, and (e) ONC samples; (f–g) TEM images of (f) NC and (g) ONC; (h) N₂ adsorption/desorption isotherm and (i) corresponding pore distribution; (j) EDX mapping of the ONC/I₂ composite; (k) Raman spectra of NC and ONC; (l) TG analysis of NC/I₂ and ONC/I₂ composites.

After high‐temperature calcination, NC samples maintain the original dodecahedral shape with a slightly rough surface (Figure 1d). In contrast, ONC samples largely preserve the polyhedral shape but exhibit concave and hollow features (Figure 1e). TEM analysis further confirms the distinct structural difference between the two samples: NC exhibits solid polyhedral characteristics, whereas ONC shows hollow and concave features (Figure 1f, g). The distinct morphology of ONC, compared to NC, is likely attributed to the O‐rich environment generated during the decomposition of trisodium citrate [35], which promotes the oxidation of Zn‐N coordination centers in ZIF‐8. The formed ZnO is reduced by surrounding carbon species at high temperatures, leading to the formation and sublimation of metallic Zn, along with etching of C and N within the framework [36]. The sodium cations are first converted into oxides, then reduced by the carbon matrix to metallic species. Subsequently, these species evaporate and intercalate into the graphene‐like layers, swelling and expanding the carbon microstructure. The decomposition of trace trisodium citrate adsorbed on the surface induces the collapse of the initial rhombic dodecahedral structure of ZIF‐8, resulting in the formation of hollow porous carbon materials. The hollow architecture enhances the porosity and specific surface area of the carbon hosts, which could provide abundant sites for iodine loading and reaction. Moreover, the pyrolysis of citrates contributes an additional carbon source, which can further facilitate the development of the porous architecture.

N₂ adsorption/desorption isotherms further reveal the pore structural changes in NC and ONC (Figure 1h). NC exhibits a typical Type I isotherm with sharp gas adsorption at low relative pressure (P/P₀ ∼0.01), demonstrating a microporous structure. In contrast, ONC displays a clear hysteresis loop in the high relative pressure range (P/P₀ 0.5 – 1), illustrating the hierarchical pore structure from micropores to mesopores [37]. The additional meso‐porosity contributes to a higher specific surface area in ONC (1039.3 m² g–¹) compared to NC (923.5 m² g–¹), thereby exposing more active sites. As shown in Figure 1i, analysis using the Barrett‐Joyner‐Halenda desorption model confirms an increase in average pore size from 2.84 nm for NC to 3.58 nm for ONC. The expanded mesoporous structure improves electrolyte infiltration and Zn²⁺ transport, ultimately leading to higher specific capacity and rate capability [38]. Furthermore, the micro/mesoporous architecture combined with the large specific surface area is favorable for both physical confinement and chemical anchoring of iodine species [39].

Raman spectra of NC and ONC display characteristic D and G bands at 1351.2 and 1587.3 cm–¹, corresponding to the vibration of sp³ defective carbon and the in‐plane stretching vibration of sp² hybridized graphitic carbon, respectively. The intensity ratio of the D to G band (I_D/I_G) reflects the defect features within the carbon materials [40]. ONC displays a higher I_D/I_G ratio of 1.31 than that of NC (1.29), implying the incorporation of trisodium citrate promotes the formation of structural defects, which is consistent with the TEM results. Rich defect sites facilitate enhanced adsorption of polyiodides by the carbon materials, thereby suppressing the shuttling effect [24].

XRD patterns of both NC and ONC show two broad peaks at 25° and 45° without additional peaks, indicating amorphous carbon and the absence of crystalline impurities (Figure S4). After I₂ loading in NC and ONC, the composites retain the initial amorphous carbon features without iodine‐related diffraction peaks, indicating the uniform distribution of molecular iodine within the carbon host (Figure S5). Energy dispersive x‐ray spectroscopy (EDS) elemental mapping further confirms the homogeneous distribution of C, N, O, Zn, and I in the ONC/I₂ composite (Figure 1j). TG analysis verifies the successful incorporation of iodine in both carbon hosts (Figure 1l). Both samples display slight weight loss below 100 °C, which is attributed to the evaporation of surface‐adsorbed iodine or moisture, with ONC/I₂ showing less weight change. A significant mass loss occurs subsequently with increasing temperature, resulting from the sublimation of I₂ confined within the pores. The mass loading of iodine is determined to be ∼44.1% for ONC/I₂ and 42.2% for NC/I₂ at 490 °C. The higher iodine content in ONC/I₂ stems from its expanded mesoporous structure, which not only provides more confinement sites but also facilitates efficient iodine diffusion into the micropores beyond the outer surface [41].

XPS analysis reveals the presence of C, N, O, and Zn in both samples (Figure S6a). Additionally, weak Na signals originating from citrate residues are observed in ONC. These residual Na species can be readily removed by rinsing with diluted hydrochloric acid and water (Figure S7) and exhibit negligible impact on the electrochemical performance. High‐resolution N 1s XPS spectrum of NC is deconvoluted into four subpeaks, which are assigned to the pyridinic‐N at 398.3 eV, pyrrolic‐N at 399.9 eV, graphitic‐N at 401.1 eV, and oxidized‐N at 402.4 eV (Figure 2a). A corresponding C‐N peak is observed at ∼285.9 eV in the C 1s spectrum (Figure S6c). Compared to NC, the N 1s peaks in ONC shift toward higher binding energy, and the total N content decreases from 11.8% to 9.7% in ONC, which is attributed to N loss during the thermal decomposition of trisodium citrate. Moreover, the distribution of N components differs between the two samples. The proportions of pyridine‐N and graphitic‐N contents decrease from 57.9% and 16.5% in NC to 51.3% and 11.6% in ONC, respectively, while the proportion of pyrrolic‐N increases from 25.6% to 37.1% (Figure S6b). High‐resolution O 1s spectra of both samples are fitted with three peaks, corresponding to C‐O, C = O, and ‐COOH (Figure 2b). Similarly, C‐O and C = O contributions are also identified in the C 1s spectra within the range of 287–289 eV (Figure S6c). In ONC, the O 1s peaks shift slightly to lower binding energy, and the O content increases to 6.0% compared to 5.8% in NC, consistent with the incorporation of O‐rich trisodium citrate. The introduction of O‐containing functional groups enhances the hydrophilicity of carbon hosts, improving the electrolyte wettability [42]. Meanwhile, increased O‐doping possibly improves the adsorption capability for polyiodides and accelerates the reaction kinetics of iodine conversion. Weak Zn 2p signals are detected in both samples, originating from incompletely removed zinc during pyrolysis (Figure S6d). Compared to NC (2.3%), ONC contains lower residual zinc content (1.2%). As mentioned above, the O‐rich atmosphere induced by the decomposition of trisodium citrate promotes the oxidation and sublimation of Zn centers, contributing to the development of mesoporous structure and increased specific surface area. These structural regulations facilitate the full exposure of active sites and the rapid iodine conversion reaction, resulting in an improvement in electrochemical performance.

FIGURE 2.

High‐resolution (a) N 1s and (b) O 1s XPS spectra of NC and ONC; (c) Optical images and UV‐vis spectra of I₃ ^– and saturated I₂ solutions after immersed with NC and ONC; (d–f) Adsorption energy and (g) corresponded models of (d) I^–, (e)I₂, and (f) I₃ ^– species on different substrates.

Building on these structural advantages, we next investigated the adsorption of iodine species in the two samples, as shown in Figure 2c. After being treated with ONC, the Zn(I₃)₂ solution shows more rapid decolorization compared to that with NC under identical conditions. The UV‐vis spectrum of the solution after immersion of NC for 48 h displays two peaks of I_3– at around 290 and 360 nm, while is absent in the case of ONC. This indicates the stronger I_3– adsorption capacity of ONC relative to NC, which is attributed to its larger specific surface area and abundant O‐containing functional groups. The hierarchical porous structure facilitates efficient diffusion of I_3– into micropores through the mesopore channels, leading to rapid adsorption. Similarly, ONC exhibits faster color fading than NC when added to the saturated I₂ solution (Figure 2c). UV‐vis analysis confirms the higher I₂ adsorption capability of ONC, as indicated by a lower residue I₂ peak intensity in the solution compared to the case with NC. After immersion in ZnSO₄ electrolyte for 72 h, the ONC/I₂ system shows a weaker I^– peak around 230 nm due to the leakage of I₂ than the NC/I₂ system (Figure S8), indicating stronger confinement of I₂ by ONC. The well‐developed micro/mesoporous architecture, larger specific surface area, and rich surface O‐containing groups enable ONC to strongly suppress the dissolution of iodine species through both physical confinement and chemical anchoring [43].

DFT calculations reveal the distinct interaction behaviors between iodine species (I₂, I–, I_3–) and various carbon substrates (Figures 2d‐f). The N‐5 and N‐6 models with hole defects are utilized to represent pyrrolic‐ and pyridinic‐N in the samples, respectively, while the defect‐free N‐G model corresponds to graphitic‐N. All edge carbon atoms are saturated with hydrogen atoms to meet their coordination requirements. The pure carbon substrate exhibits weak interactions with I₂, I^–, and I₃ ^–. I₂ molecules adsorb parallel to the graphene surface, with each I atom located at the hollow site of hexagon carbon rings. I₃ ^– adopts a similar configuration with all three I atoms occupying hollow sites of carbon rings, whereas I^– tends to adsorb at the top sites of C atoms (Figure S9). The N‐doping enhances the adsorption of iodine species onto the substrates. Among the N‐doped configurations, graphitic‐N (N‐G) exhibits the strongest adsorption, followed by pyridinic‐N (N‐6), and then pyrrolic‐N (N‐5). On the N‐G substrate, I^– and I₃ ^– preferentially adsorb at the top sites of C adjacent to the N atom, while I₂ occupies hollow and CN bridge sites. The strong electronegativity of N induces the positive charge on surrounding C atoms, strengthening electrostatic interaction with the iodine species. On N‐5 and N‐6 type substrates, iodine species tend to adsorb at hole sites, emphasizing the crucial role of defect engineering in optimizing the performance of carbon hosts. Further introduction of O enhances the adsorption affinity for iodine species across all three configurations, suggesting that the trace O content improves the chemical entrapment effect of N‐doped carbon hosts. For I₃ ^– adsorption, the interaction strength of N/O codoped substrates follows the order: N‐G > N‐6> N‐5 (Figure 2f). In the case of I₂ and I^–, the N‐5 model displays significant enhancement, especially for I₂ adsorption where its adsorption energy even exceeds that of N‐G surface, resulting from its out‐of‐plane distortion after adsorption. On N/O codoped substrates, the favorite adsorption sites vary: I– adsorbs at the top site of O atom, I₂ occupies the top sites of both N and O atoms, and I_3– adsorbs at the O top site and hollow sites near the N atom. The iodine species preferentially occupy the defect sites near O atoms for N‐5 and N‐6 models of N/O codoped substrates. Given that ONC has a lower total nitrogen contents but a higher pyrrolic‐N content than NC, we propose that the synergistic sites between pyrrolic‐N and O play a dominant role in iodine adsorption. These results highlight the significant role of O‐doping and defect sites in enhancing the physicochemical adsorption capability of N‐doped carbon hosts toward iodine species.

CV curves of ONC/I₂ and NC/I₂ exhibit a pair of redox peaks at 1.223/1.264 V and 1.209/1.258 V, respectively, corresponding to the I₂/I^– conversion process (Figure 3a). ONC/I₂ displays the smaller peak potential difference of 41 mV than the NC/I₂ of 49 mV, along with higher peak current density, demonstrating its lower polarization and better reversibility. The enlarged specific surface area in ONC provides abundant reactive sites for the charge transfer process of I₂/I^–. The electrochemical active surface area (ECSA) is evaluated through CV scanning at varied rates (1–10 mV s–¹) within the range of 0.6–1.1 V. ECSA values of NC and ONC are determined to be 94.7 and 103.5 F g–¹, respectively, consistent with the BET analysis (Figure S10). This result suggests that ONC possesses more reactive sites and more effective active interface. The surface chemistry of carbon hosts may also affect the kinetics of I₂ reduction reaction (IRR). The Tafel slope of ONC/I₂ (149.0 mV dec–¹), derived from the CV curve, is lower than that of NC/I₂ (185.6 mV dec–¹), demonstrating reduced overpotential and enhanced reaction kinetics [44]. EIS plots of NC/I₂ and ONC/I₂ show depressed semicircles in the high frequency range and slopes in the low frequency region, corresponding to the interfacial charge transfer resistance (R_ct) and Warburg impedance (Z_w), respectively. The ONC/I₂ system displays a smaller R_ct compared to NC/I₂, confirming the more efficient charge transfer process, in agreement with the CV results. The apparent activation energy (E_a), calculated from the EIS analysis at different temperatures according to the Arrhenius equation: 1/R_ct = A·exp[E_a/RT], is found to be 34.2 kJ mol–¹ for ONC/I₂ and 39.9 kJ mol–¹ for NC/I₂ (Figure 3d). Notably, XPS analysis reveals lower N and Zn contents in ONC than in NC, implying that the enhanced IRR kinetics originate from the trace O‐doping.

FIGURE 3.

(a) CV curves, (b) Tafel plots, (c) EIS, and (d) Arrhenius plots of NC and ONC; (e‐f) Gibbs free energy diagram of I₂/I⁻ through (e) consecutive and (f) concerted mechanism; (g) Electron localization function and (h) differential charge density of graphene, N‐G model N‐doped C, and N/O codoped C with I₂ adsorption (up panel: top view, low panel: side view).

DFT calculations are employed to compare the two reaction pathways of IRR on various carbon‐based substrates(Figure 3e–f) [45] The overall reaction is:${\mathrm{I}}_2+2{\mathrm{e}}^{-}\to 2{\mathrm{I}}^{-}$. Two mechanisms are considered:

• consecutive mechanism

  $$\mathrm{Heyrovsky}\mathrm{step}:\ast {\mathrm{I}}_2+{\mathrm{e}}^{-}\to \ast \mathrm{I}+{\mathrm{I}}^{-}$$

  $$\mathrm{Volmer}\mathrm{step}:\ast \mathrm{I}+{\mathrm{e}}^{-}\to {\mathrm{I}}^{-}$$
• concerted mechanism

  $$\mathrm{Tafel}\mathrm{step}:\ast {\mathrm{I}}_2\to 2\ast \mathrm{I}$$

  $$\mathrm{Volmer}\mathrm{step}:\ast \mathrm{I}+{\mathrm{e}}^{-}\to {\mathrm{I}}^{-}$$

The results indicate that the consecutive pathway (Heyrovsky‐Volmer) is preferred across all carbonaceous substrates, whereas the concerted route (Tafel‐Volmer) shows a larger free energy increase for the dissociative adsorption of I₂ molecule, making it less favorable. This finding is consistent with the previous report [46]. N and O heteroatom doping facilitates the IRR process compared to pure carbon substrates in both mechanisms. For the consecutive mechanism, the N‐G configuration shows the lowest calculated overpotential of 0.18 V among the N‐doped carbon models, followed by N‐6 (0.32 V) and N‐5 (0.37 V). Further introduction of O reduces the overpotential across all configurations. Specifically, the N‐6 and N‐G models achieve overpotentials of 0.03 V and 0.14 V, respectively, which are lower than that of N‐5 (0.23 V) (Figure 3e). In contrast, O‐doping slightly increases the dissociative adsorption free energy for N‐5 and N‐6 configurations relative to their N‐doped counterparts (Figure 3f), which is possibly attributed to the distribution of polar sites and defects. Among N/O codoped models, the N‐G configuration demonstrates the lowest dissociative free energy of 0.78 eV. When considering the I_3– intermediate species in IRR, the introduction of N in carbon substrates also reduces the free energy change (Figure S11). N/O codoped configurations further lower this value compared to N‐doped substrates. The N‐G model shows the smallest free energy change among the three configurations. Notably, although the N‐6 model under N/O codoping exhibits the lowest overpotential in the consecutive mechanism, its weak adsorption of iodine species may limit its ability to suppress the shuttling effect and enhance cycling stability. In contrast, N‐G sites demonstrate superior electrocatalytic activity for IRR and chemical anchoring capability, and the introduction of trace O further enhances the performance of N‐doped carbon hosts.

Electron localization function (ELF) analysis indicates that the N site shows higher values than the C sites in the N‐G model of N‐doped carbon substrate (Figure 3g). The incorporation of electron‐rich N increases the overall charge density. Meanwhile, the region near C─N bonds exhibits enhanced electron delocalization. Similarly, in the N/O codoped carbon substrate, the O site displays even higher ELF values than the N site, also resulting from the increased charge density due to O‐doping. The region surrounding the C─O bonds likewise shows stronger electron delocalization. Comparable trends are observed in the N‐5 and N‐6 models, even though the N‐6 structure exhibits significant out‐of‐plane surface distortion (Figure S12). The increased electron delocalization strengthens the interaction with I₂. Differential charge density analysis further confirms enhanced electron transfer from the substrates to the adsorbed I₂after heteroatom doping with N and O (Figures 3h and S13). Bader charge analysis shows negligible charge transfer (0.04e) between I₂ and pure carbon surface, while higher values of 0.18e and 0.33e are observed for the N‐G model under N‐doping and N/O codoping, respectively. Heteroatom doping with electron‐rich N and O enhances the electron density of substrates, promoting the charge transfer to I₂. This effect not only strengthens the adsorption of iodine species but also promotes the IRR. For instance, the distance between adsorbed I₂ and substrates reduces from 3.70 Å on pure carbon to 3.60 Å on N‐G model under N‐doping, and further to 3.29 Å on N‐G model under N/O codoping. Although I₂ tends to occupy the hole sites on N‐5 and N‐6 models, reducing in the adsorption distance from the surface is also consistently observed. The transfer of electron density from the substrates to I₂ molecule weakens the covalent I─I bond. The bond length of adsorbed I₂ increases from 2.684 Å on a pure carbon surface to 2.828 Å on the N‐G model under N‐doping, and further to 2.971 Å on the N‐G model under N/O codoping. This stretching of the I─I bond facilitates IRR and improves the reaction kinetics. The regulation of electronic structure by introducing trace O in N‐doped carbon host enhances the catalytic activity of N‐doped carbon substrates, which helps explain the more rapid reaction kinetics observed on ONC, while with lower N contents compared to NC.

GCD measurements of ONC/I₂ show a higher specific capacity of 205.8 mAh g–¹ at 0.1 A g–¹ than that of NC/I₂ (170.0 mAh g–¹). The value, approaching the theoretical capacity of the iodine reduction reaction, reflects the high utilization of active materials in the ONC system. Meanwhile, the voltage hysteresis of ∼32 mV at the half‐discharge state for ONC/I₂ is lower than that of ∼43 mV for NC/I₂, indicating reduced polarization. Chronoamperometric tests of ONC/I₂ at 0.6 V deliver a higher capacity of 200.9 mAh g–¹ than that of NC/I₂ (159.5 mAh g–¹), which underscores the enhanced I₂ reduction efficiency enabled by the synergistic effect of hierarchical porous structure and optimized surface chemistry. The ONC/I₂ system exhibits excellent rate performance over the wide current density range of 0.1–2.0 A g–¹ (Figure 4c, d). Specifically, the ONC/I₂ electrode delivers specific capacities of 191.8, 183.6, 175.0, and 161.0 mAh g⁻¹ at 0.2, 0.5, 1.0, and 2.0 A g⁻¹, respectively, all higher than those of NC/I₂ at the same current density (156.5, 148.6, 141.2, and 130.2 mAh g⁻¹). When the current density is returned to 0.1 A g⁻¹, ONC/I₂ recovers a reversible capacity of 191.1 mAh g⁻¹, whereas NC/I₂ restores only 159.7 mAh g⁻¹. Moreover, as current density increases, GCD curves of both systems remain stable with a modest increase in voltage hysteresis. The rational design of pore structure and surface chemistry collectively contributes to the outstanding rate capability of the ONC/I₂ system compared to other reported carbon hosts (Figure 4 g) [39, 47, 48, 49, 50, 51, 52, 53].

FIGURE 4.

(a) GCD and (b) chronoamperometric curves of NC and ONC; (c) rate capability and (d) GCD curves at different current density of NC and ONC; (e, f) cycle performance of NC and ONC based Zn‐I₂ batteries; (g, h) Comparison of the rate capability and cycle performance of this work with reported representative Zn‐I₂ batteries; (i) Optical images of Zn‐I₂ pouch cells based on ONC powering light‐emitting diodes with “HFUT” and driving a small electronic fan; (j) Cycle performance and (k) GCD curves of the pouch cell; the inset shows operating voltage recorded during testing.

The ONC/I₂ system also exhibits excellent cycle performance (Figure 4e, f). ONC/I₂ retains 87.0% of the initial capacity after 200 cycles at 0.1 A g⁻¹, outperforming NC/I₂ (82.4%). ONC/I₂ shows a higher initial specific capacity of 178.3 mAh g⁻¹ at a current density of 1.0 A g⁻¹ than the NC/I₂ (153.7 mAh g⁻¹). After 10,000 cycles, ONC/I₂ retains a reversible capacity of 105.5 mAh g⁻¹ with 59% retention of the initial value, significantly exceeding that of NC/I₂ (88.5 mAh g⁻¹). Both systems exhibit nearly 100% Coulombic efficiency (CE) throughout long‐term cycling, with an average CE of 99.7% for NC/I₂ and 99.8% for ONC/I₂, indicating exceptional long‐term stability. The acid‐washed ONC, with residue Na species removed, shows a negligible difference in electrochemical behavior from the untreated ONC, delivering a specific capacity of 186.5 mAh g–¹ and retaining 80.2% of its capacity after 1000 cycles at 1.0 A g⁻¹ (Figure S14). The physical confinement and chemical anchoring effect of ONC on iodine species suppresses the irreversible loss of active materials during cycling, resulting in the excellent electrochemical stability that is competitive with other reported host materials (Figure 4 h) [11, 39, 47, 48, 51, 52, 53, 54]. Soft‐pack cells based on ONC/I₂ electrodes (size: 2.0×2.5 cm²) were assembled and evaluated (Figure 4i, k). The pouch‐cell delivers a maximum specific capacity of 165.4 mAh g–¹ at 0.5 A g–¹. It achieves excellent cycling stability, retaining 98.3% of maximum capacity and an average CE of 99.4% over 400 cycles. Two batteries connected in parallel successfully powered the LEDs and a small electronic fan, demonstrating the practical application potential of the ONC/I₂ electrode.

The electrochemical kinetics of ONC/I₂ and NC/I₂ systems are further investigated through CV tests at different scan rates (0.2 – 1.0 mV s–¹), as shown in Figures 5a and S15a. Both systems exhibit an increased peak potential difference and enhanced peak current density with increasing scan rates. The peak current density (I) is fitted against the scan rates (v) according to the power law equation [55]: log I = log a + b log v, where a and b are the adjustable parameters. The b value approaching 1.0 suggests the capacitive control process, whereas the value close to 0.5 indicates the diffusion limited mechanism. The calculated b values of the anodic peaks are 0.89 and 0.90 for ONC/I₂ and NC/I₂, respectively, while the cathodic peaks yield b values of 0.95 and 0.97. These values indicate the dominant contribution of capacitive behavior, suggesting the rapid surface‐mediated conversion of I₂. The pseudocapacitive contributions are further quantified through the equation [55]: I = k₁ v + k₂ v ^1/2, where k₁ v and k₂ v ^1/2 refer to capacitive and diffusion related contributions. Both systems display higher capacitive contribution with the increase of scan rates (Figures 5c, d, and S15, 16). Notably, the ONC/I₂ displays higher capacitive contribution than the NC/I₂ across all scan rates. For example, at 0.2 mV s–¹, the capacitive contribution reaches 93.09% for ONC/I₂ compared to 88.43% for NC/I₂. The enhanced capacitive contribution in ONC/I₂ suggests more favorable surface conversion kinetics of iodine species, which are attributed to the hierarchical porous structure and reinforced catalytic activity by O‐doping.

FIGURE 5.

(a) CV curves, (b) Log(current density) versus Log(sweep rate) plots, and (c) Capacitive contribution of ONC/I₂ cathodes; (d) Comparison of capacitive contribution; (e) GITT curves and (f) corresponding ΔV_|CCV‐QOCV| versus normalized time plots; (g, h) Adsorption configuration and energy of (g) H₂O and (h) ZnI₂ on graphene, N‐5 model of N doped C and N/O doped C substrates; (i) In‐situ Raman spectra of ONC and NC based iodine cathodes.

GITT is utilized to evaluate the internal polarization of AZIBs based on NC/I₂ and ONC/I₂ (Figure 5e). ONC/I₂ displays smaller voltage hysteresis during both charging and discharging processes than NC/I₂, demonstrating lower charge transfer resistance and ohmic impedance. The analysis of the difference between quasi‐open circuit voltage and the closed‐circuit voltage (ΔV_|CCV‐QOCV|) shows greater internal impedance in the NC/I₂ system compared to the ONC/I₂ (Figure 5f). The micro/mesoporous architecture and surface O‐containing groups improve the hydrophilicity and electrolyte infiltration of the iodine cathode, facilitating the rapid Zn²⁺ transport and diffusion within the pore channels. EIS and GITT fitting analysis both demonstrate the higher diffusion coefficients for ONC/I₂ compared to NC/I₂ (Figure S17).

DFT calculations verify the strengthened adsorption of ZnI₂ and H₂O on N‐doped and N/O codoped carbon substrates (Figure 5g, h). The pure carbon surface exhibits a repulsive interaction with ZnI₂, showing a positive adsorption energy of 2.87 eV due to its nonpolar delocalized electron distribution. Introducing electron‐rich N modulates the electron density of the carbon surface, resulting in enhanced adsorption of ZnI₂. The adsorption energy for the N‐5, N‐6, and N‐G models of N‐doped carbon is ‐1.07, −0.73, and ‐0.40 eV, respectively. Incorporation of O further increases the adhesion while maintaining the same adsorption energy trend as N‐doped substrates. Benefiting from the synergistic coordination of N and O sites, the N‐5 model yields the lowest adsorption energy (−1.60 eV), followed by N‐6 (−0.96 eV), and then defect‐free N‐G of −0.43 eV (Figure S18). The pure carbon surface attracts the H₂O molecule through the conjugated π‐electron and dipole interaction, with an adsorption energy of ‐0.14 eV. For N‐doped models containing hole defects, specifically the N‐5 and N‐6 configurations, water adsorption is enhanced through hydrogen bonding, yielding adsorption energies of −0.48 and −0.31 eV, respectively. The introduction of O further increases H₂O adsorption, with adsorption energy of −0.90 eV for N‐5 and −0.37 eV for N‐6. In contrast, the defect‐free N‐G models of N‐doped and N/O codoped substrates show only slight improvements, with adsorption energies of −0.17 and −0.19 eV, respectively (Figure S19). Overall, the N‐5 model demonstrates the strongest adsorption capability toward both ZnI₂ and H₂O, while the N‐G model shows the weakest. These results suggest the significant roles of pyrrolic‐N sites, hole defects, and trace O‐doping in facilitating the electrolyte wettability and Zn²⁺ transport. The synergistic effect of these features promotes rapid I₂ conversion kinetics, achieving high utilization of active materials and specific capacity, while also suppressing the formation of polyiodide intermediates. In‐situ Raman spectra of NC/I₂ show the I_3– peaks at ∼110 cm–¹ in the high voltage region, with significantly higher intensity than those of ONC/I₂. The strong adsorption of iodine species by ONC further helps to suppress the dissolution and subsequent shuttling of polyiodides.

4. Conclusion

In summary, this work demonstrates a simple yet highly effective trisodium citrate post‐treatment strategy to modulate the pore structure and surface chemistry of ZIF‐8‐derived carbon, obtaining N/O codoped hierarchical porous carbon materials with abundant defects. The micro/mesoporous channels and enlarged specific surface area provide adequate I₂ loading sites. Meanwhile, the electronic configuration of the carbon surface is regulated by N/O codoping, which enhances the chemical anchoring and electrocatalytic activity towards I₂ conversion, lowers the interfacial charge transfer barrier, and thereby promotes rapid reaction kinetics and high iodine utilization. The incorporation of trace O is crucial in enhancing the adsorption capability and catalytic activity of N‐doped carbons. Furthermore, the well‐developed hierarchical porosity and N/O codoping improve the electrolyte wettability and Zn²⁺ ion transport, thus accelerating the kinetics of the iodine conversion reaction. The synergistic effects of physical confinement by multi‐scale pores, chemical entrapment by surface functional groups, and rapid reaction kinetics collectively suppress the irreversible loss and shuttling effect of iodine species during cycling. DFT calculations further highlight the distinct contributions of various N‐active sites to the iodine reduction reaction upon trace O‐doping. Benefiting from these advantages, AZIBs based on ONC/I₂ deliver high iodine utilization, achieving a reversible capacity of 205.8 mAh g⁻¹ at 0.1 A g⁻¹. The batteries exhibit excellent rate capability and long‐term cycling stability, retaining 178.3 mAh g⁻¹ at 1 A g⁻¹ and maintaining 59% capacity retention over 10,000 cycles. This study provides fundamental insights into the synergistic design of pore structure and interface chemistry for optimizing carbon hosts in advanced AZIBs.

Conflicts of Interest

The authors declare no conflicts of interest.

Data availability

The data that support the findings of this study are available from the corresponding author, [Ni, Gang], upon reasonable request.

Supporting information

FT‐IR and XPS spectra of ZIF‐8 and SC‐ZIF‐8; XRD patterns and XPS spectra of ONC and NC; UV‐vis spectra and optical images of ONC/I₂ and NC/I₂ desorption; DFT models, ECSA of ONC and NC; Free energy diagram, ELF, and Differential charge density of N doped C and N/O doped C; Capacitive contribution of ONC/I₂ and NC/I₂ electrodes; Fitted diffusion coefficient; Adsorption energy of H₂O and ZnI₂ on substrates.