Framework-to-carbon conversion strategy for nitrogen–silicon co-doped microporous carbons with superior energy storage performance

Abstract

Developing heteroatom-doped porous carbons with tailored architectures is a promising strategy for enhancing the energy storage performance of next-generation supercapacitors. Herein, we report the design and synthesis of two nitrogen–silicon co-doped microporous carbons (NSi@C-1 and NSi@C-2) prepared from silicon-containing covalent triazine frameworks (Si-CTFs). These Si-CTFs were constructed by the condensation of 4,4′,4″,4‴-silanetetrayltetrabenzaldehyde (Si-4CHO) with two different carboximidamide linkers, namely, terephthalimidamide (TP-2NHNH₂) and [1′-biphenyl]-4,4′-dicarboximidamide (BP-2NHNH₂), resulting in tunable porosity and distinct structural architectures. The direct carbonization route enables the uniform incorporation of N and Si heteroatoms within a hierarchical porous network, simultaneously improving the electrical conductivity and redox activity. NSi@C-2, featuring an extended biphenyl linker, exhibited a higher surface area (732 m² g⁻¹), larger pore volume (0.59 cm³ g⁻¹), and higher electrical conductivity (4.23 S cm⁻¹) than NSi@C-1. Owing to these structural features, the material exhibits exceptional electrochemical properties, achieving a specific capacitance of 403 F g⁻¹ at 0.5 A g⁻¹ and retaining 95.25% of its initial capacitance after 10 000 charge–discharge cycles in a three-electrode system. A symmetric NSi@C-2-based supercapacitor device exhibited a high capacitance of 211 F g⁻¹, an impressive energy density of 152 W h kg⁻¹ at a power density of 1037 W kg⁻¹, and exceptional cycling stability (91.50% retention after 10 000 cycles). The synergistic effects of hierarchical porosity, extended π–π conjugation, and N–Si dual-doping collectively endow NSi@C-2 with superior ion diffusion, efficient electron transport, and abundant active sites. This study establishes a facile framework-to-carbon conversion strategy for constructing multifunctional heteroatom-doped carbons with tunable porosity and electronic properties for advanced energy-storage applications.

---

1. Introduction

Given the growing global energy demand and escalating environmental challenges posed by climate change and pollution, there is an urgent need for sustainable and efficient energy-storage technologies to support the utilization of renewable energy.¹ Among electrochemical storage devices, supercapacitors (SCs) have attracted considerable research interest because they combine high power density, fast charge–discharge capability, excellent cycling life, and robust mechanical/electrochemical stability, making them promising for applications such as regenerative braking, electric-vehicle acceleration, and renewable-energy buffering.^2,3 Depending on the charge-storage mechanism, SCs are generally classified into electric double-layer capacitors (EDLCs), pseudocapacitors, and hybrid capacitors.^4–6 EDLCs store charge through the electrostatic adsorption of electrolyte ions at the electrode/electrolyte interface, while pseudocapacitance arises from fast and reversible faradaic reactions occurring at or near the electrode surface; hybrid capacitors integrate both contributions.^7–9 Carbon materials (e.g., activated carbon, carbon nanotubes, and graphene) are widely used as EDLC electrodes due to their high specific surface area, good conductivity, and outstanding durability.^10–12 However, their energy density and specific capacitance are often limited by the lack of abundant intrinsic redox-active sites. To overcome this issue, incorporating redox-active units that facilitate fast and reversible faradaic reactions is essential to enhance the energy storage capability of the electrodes.¹³ Although metal oxides and conjugated polymers can introduce pseudocapacitive charge storage, their faradaic reactions are frequently accompanied by structural/volumetric changes that can compromise cycling stability and restrict access to active sites.^14–16 Accordingly, rational design strategies—such as engineering porous nanostructures for efficient ion transport and tailoring the surface chemistry through heteroatom incorporation—have become effective pathways to enhance capacitance while maintaining high rate capability and long-term stability.¹⁷

Covalent triazine frameworks (CTFs) are a class of porous organic polymers constructed from robust triazine linkages, offering high chemical/thermal stability and excellent structural designability.^18,19 Their fully conjugated triazine units and nitrogen-rich backbones can modulate charge distribution and provide abundant heteroatom sites, while the framework architecture can be tuned at the molecular level by selecting appropriate building blocks, making CTFs attractive for electrochemical energy-storage applications, including supercapacitors.²⁰ In addition, the intrinsic porosity of CTFs enables electrolyte permeation and exposes a large number of accessible sites, which is beneficial for interfacial charge storage. However, pristine CTFs typically exhibit relatively low electrical conductivity and limited charge-transfer kinetics, which restrict electron transport through the framework and ultimately reduce their capacitive performance.²¹ A practical and widely adopted strategy to address this limitation is to convert CTFs into porous carbons via controlled carbonization. This framework-to-carbon transformation can enhance graphitization and electronic conductivity, generate abundant micropores (and, when properly engineered, hierarchical pores) for efficient ion diffusion, and partially retain heteroatom-derived functionalities from the parent framework.³¹ Consequently, CTF-derived carbons provide a versatile platform for integrating tailored porosity, heteroatom chemistry, and improved charge transport, offering an effective route toward high-performance carbon electrodes for next-generation supercapacitors.

Porous carbon (PC) materials possess interconnected pore networks and large specific surface areas, making them widely attractive for gas adsorption, energy storage, and catalysis.^22–24 In particular, PCs are among the most practical electrode candidates for supercapacitors because of their well-developed porosity, tunable surface chemistry, good electrical conductivity, and low cost.^25,26 Their charge storage is primarily governed by EDLC; however, purely EDLC-type carbon electrodes typically deliver limited energy density, especially under high-power operation, where polarization and ion-transport resistance become more pronounced.^22,25 Therefore, introducing pseudocapacitive contributions through fast, reversible surface redox processes is an effective strategy to enhance energy density while maintaining high power.^25,27,28 Common approaches include incorporating redox-active components (e.g., metal oxides or conductive polymers) or heteroatom doping, which can increase charge-carrier density, improve wettability, and create additional electrochemically active sites.^27–29 Notably, nitrogen-doped porous carbons have received significant attention because N functionalities can enhance ion accessibility, promote surface polarity/wettability, and improve electrical conductivity, thereby improving both capacitance and rate performance. Similarly, silicon incorporation can modulate the electronic structure and surface chemistry of carbon frameworks and introduce polar Si-containing sites (e.g., Si–C/Si–O–C), thereby facilitating electrolyte wetting and charge/ion transport.³⁰ Despite these benefits, the preparation of nitrogen–silicon co-doped porous carbons remains relatively limited compared with the extensive literature on single-heteroatom (e.g., N- or O-doped) carbons. Most reported N/Si-containing carbons are produced via (i) organosilane/sol–gel additive routes (e.g., TEOS/APTES or polysiloxane precursors), followed by pyrolysis or (ii) biomass/self-doping approaches, where the N/Si contents and speciation are largely dictated by the feedstock and activation conditions.^31–38 In these strategies, the Si distribution and bonding configuration (Si–O vs. Si–C) are often strongly process-dependent and not inherently programmed at the molecular level, which can limit the dopant distribution and reproducibility. Therefore, developing a precursor-defined, reproducible strategy to construct N/Si co-doped microporous carbons with favorable conductivity, accessible active sites, and improved electrode–electrolyte interactions is highly desirable for advanced supercapacitor electrodes.

In this study, we employ a precursor-defined framework-to-carbon conversion strategy by first constructing silicon-containing covalent triazine frameworks (Si-CTFs) from a tetrafunctional Si-4CHO node and amidine-based linkers, thereby intrinsically encoding both Si and N within the framework backbone prior to carbonization. Subsequent carbonization at 700 °C directly yields N/Si co-doped microporous carbons without additional dopant impregnation, post-silylation, or template-removal steps, providing a chemically defined and reproducible dual-doping route. Importantly, by changing the linker geometry (terphenyl vs. biphenyl), we demonstrate a clear precursor-structure tunability of pore accessibility and electrochemical kinetics, strengthening the structure–property relationship beyond incremental performance differences. This design allows molecular-level encoding of dual-heteroatom chemistry and enables linker-guided tuning of pore accessibility and electrochemical kinetics (NSi@C-1 vs. NSi@C-2), providing a clearer structure–property relationship and improved reproducibility without additional dopant impregnation or template-removal steps. In contrast to conventional multi-step doping or post-synthetic modification approaches, this method embeds both nitrogen and silicon atoms uniformly within the polymer backbone during framework formation, enabling strong structural integration. Herein, two nitrogen- and silicon-co-doped microporous carbons (NSi@C-1 and NSi@C-2) were synthesized via direct carbonization of silicon-containing covalent triazine frameworks (Si-CTFs), which were obtained from the condensation of 4,4′,4″,4‴-silanetetrayltetrabenzaldehyde (Si-4CHO) with terephthalimidamide (TP-2NHNH₂) and [1′-biphenyl]-4,4′-dicarboximidamide (BP-2NHNH₂), respectively (Fig. 1). Structural and textural analyses reveal that biphenyl-derived NSi@C-2 possesses a higher surface area (732 m² g⁻¹) and larger pore volume (0.59 cm³ g⁻¹) than NSi@C-1, facilitating improved electrolyte penetration and ion accessibility. Consequently, NSi@C-2 exhibits a remarkable specific capacitance of 403 F g⁻¹ at 0.5 A g⁻¹, excellent rate capability, and outstanding cycling stability (95.25% retention after 10 000 cycles). Furthermore, in a symmetric NSi@C-2//NSi@C-2 device, the material exhibited an impressive specific capacitance of 211 F g⁻¹ at 0.2 A g⁻¹, along with a high energy density of 152 W h kg⁻¹ and a power density of 1037 W kg⁻¹, maintaining 91.50% capacitance retention after 10 000 cycles. Overall, this framework-to-carbon conversion route enables simultaneous textural control and heteroatom engineering, offering a practical platform for designing multifunctional heteroatom-doped carbons from structurally defined Si-CTF precursors.

Fig. 1 Schematic of the synthesis and carbonization process of the N- and Si-doped porous carbons (NSi@C-1 and NSi@C-2) derived from silicon-containing covalent triazine frameworks (Si-CTFs).

2. Results and discussion

The silicon-functionalized aldehyde monomer, tetrakis(4-formylphenyl)silane (Si-4CHO), was synthesized via a two-step reaction route. Initially, 4-bromobenzaldehyde dimethyl acetal was dissolved in a dry solvent under a nitrogen atmosphere and treated dropwise with n-butyllithium at −78 °C to generate the corresponding lithiated intermediate. Silicon tetrachloride (SiCl₄) was then added slowly, and the reaction mixture was allowed to warm to room temperature. The reaction was quenched with 2 M HCl and extracted using ethyl acetate. The crude yellow solid, tetrakis(4-(dimethoxymethyl)phenyl)silane [Si(4-C₆H₄CH(OCH₃)₂)₄], was directly hydrolyzed in 2 M HCl/THF (1 : 1) overnight to afford the target aldehyde compound. The mixture was neutralized with NaHCO₃ solution, extracted with ethyl acetate, and purified by recrystallization using a hexane/ethyl acetate (3 : 1) mixture to yield pure Si-4CHO as a white crystalline solid (Scheme S1). Nitrogen-rich linkers, 1,4-benzenedicarboximidamide (TP-2NHNH₂) and [1,1′-biphenyl]-4,4′-bis(carboximidamide) (BP-2NHNH₂), were synthesized through a conversion of aromatic dinitriles into the corresponding dicarboximidamides (Schemes S2 and S3). In a representative procedure, benzene-1,4-dicarbonitrile was treated with lithium bis(trimethylsilyl)amide in dry THF under a nitrogen atmosphere, followed by acidification with 6 M HCl in ethanol. The resulting solid was collected, washed, and dried to yield TP-2NHNH₂ as an off-white powder. The same procedure was employed to prepare BP-2NHNH₂, replacing the starting material with [1,1′-biphenyl]-4,4′-dicarbonitrile. Both compounds were obtained in high yield and purity and were used directly for the subsequent synthesis of silicon-functionalized CTFs.

The successful syntheses of Si-4CHO, TP-2NHNH₂, and BP-2NHNH₂ were confirmed by FTIR, ¹H NMR, and ¹³C NMR analyses. As shown in Fig. S1, the FTIR spectrum of Si-4CHO displays a strong absorption band at 1705 cm⁻¹, corresponding to the C=O stretching vibration of the aldehyde groups, and a distinct band at 1210 cm⁻¹, attributable to the Si–C bond, confirming the incorporation of silicon into the aromatic framework. The ¹H NMR spectrum of Si-4CHO (Fig. S2) displays a distinctive signal at approximately 10 ppm assigned to the aldehydic –CHO proton, along with aromatic proton resonances at 8.02 and 7.72 ppm. The ¹³C NMR spectrum of Si-4CHO (Fig. S3) further corroborates the structure, exhibiting carbonyl carbon resonance at 194 ppm and aromatic carbon signals at 139.86, 138.15, 137.36, and 129.61 ppm. These consistent spectroscopic features collectively verify the successful synthesis and structural integrity of Si-4CHO. As shown in Fig. S4, the FTIR spectrum of TP-2NHNH₂ exhibits characteristic N–H stretching vibrations at 3240 and 3036 cm⁻¹, while strong absorption bands at 1688 and 1658 cm⁻¹ correspond to the C=N stretching of the carboximidamide groups. Additionally, the peak observed at 1480 cm⁻¹ is assigned to the C=C stretching of the aromatic ring. The ¹H NMR spectrum (Fig. S5) displays a prominent downfield resonance at 8.0 ppm, attributed to the aromatic C–H proton. The ¹³C NMR spectrum (Fig. S6) further validates the structure, showing a signal at 166.75 ppm for the C=N carbon, and aromatic carbon resonances at 133.85 and 129.00 ppm. These combined spectroscopic results confirm the successful formation and structural integrity of TP-2NHNH₂. As shown in Fig. S7, the FTIR spectrum of BP-2NHNH₂ displays characteristic N–H stretching bands at 3244 and 3079 cm⁻¹, while the strong absorption peaks at 1675 and 1665 cm⁻¹ correspond to the C=N stretching vibrations of the carboximidamide groups. Additionally, the band at 1475 cm⁻¹ is attributed to the C=C stretching of the aromatic ring. The ¹H NMR spectrum (Fig. S8) exhibits a distinct aromatic C–H signal at 8.0 ppm. The ¹³C NMR spectrum (Fig. S9) further supports the structure, showing a C=N carbon signal at 167.01 ppm, along with aromatic carbon peaks at 145.03 ppm and 130.03 ppm. These spectroscopic features collectively confirm the successful synthesis and structural purity of BP-2NHNH₂.

To the best of our knowledge, silicon-containing covalent triazine frameworks (Si-CTFs) have rarely been explored and have not been previously reported in this specific structural form. Introducing silicon directly as a framework node provides a fundamentally different design handle compared with conventional CTF chemistry, enabling the molecular-level incorporation of Si into the backbone together with the nitrogen-rich triazine network. This Si-CTF concept therefore offers a new precursor platform for generating N/Si co-doped porous carbons with programmable heteroatom chemistry and tunable pore architectures via linker selection. The Si-CTFs were synthesized through a condensation reaction between Si-4CHO and carboximidamide-based linkers under solvent-thermal conditions. For Si-CTF-1, Si-4CHO was condensed with TP-2NHNH₂ in DMSO using t-BuOK as the base catalyst. Likewise, Si-CTF-2 was obtained by reacting Si-4CHO with BP-2NHNH₂ under identical conditions (Fig. 1, S4, and S5). The formation of Si-CTF-1 and Si-CTF-2 was confirmed by FTIR spectroscopy (Fig. 2a and b). In comparison with the spectra of their respective monomers (Si-4CHO and TP-2NHNH₂ or BP-2NHNH₂), the characteristic aldehyde C=O stretching band at ∼1705 cm⁻¹ and the N–H stretching bands at 3240–3030 cm⁻¹ disappear in both frameworks, indicating the complete consumption of the formyl and amino groups during the polycondensation reaction. New absorption bands emerge at 1665 cm⁻¹ and 1353 cm⁻¹ for Si-CTF-1, and at 1660 cm⁻¹ and 1367 cm⁻¹ for Si-CTF-2, which are characteristic of C=N stretching vibrations of the newly formed triazine rings. Additionally, the peaks near 1210 cm⁻¹ correspond to Si–C stretching, confirming the retention of silicon linkages within the polymer backbone. These distinct spectral changes collectively demonstrate the successful formation of silicon-functionalized covalent triazine frameworks with extended conjugation and well-defined triazine linkages. The formation of the triazine frameworks was further confirmed by solid-state ¹³C CP/MAS NMR spectroscopy (Fig. 2c). Both Si-CTF-1 and Si-CTF-2 exhibit a distinct resonance at approximately 177 ppm, which is attributed to the C=N carbon of the triazine rings, providing clear evidence of successful cyclization during framework formation. Broad peaks centered around 140–125 ppm correspond to the aromatic carbons of the phenyl linkers. The similarity of these spectral features for both Si-CTF-1 and Si-CTF-2 indicates that the condensation reactions proceeded efficiently, leading to well-formed, conjugated silicon-containing covalent triazine networks. The thermal stability of the Si-CTFs was evaluated by thermogravimetric analysis (TGA) under a nitrogen atmosphere at a heating rate of 20 °C min⁻¹ from room temperature to 800 °C (Fig. 2d). Both frameworks exhibited excellent thermal resistance, with initial weight loss below 150 °C attributed to the removal of physically adsorbed moisture and residual solvents. The 10% weight-loss temperatures (T_d10%) were observed at 396.86 °C for Si-CTF-1 and 592.17 °C for Si-CTF-2, indicating their high structural robustness. The corresponding char yields at 800 °C were 53.77% and 63.86%, respectively (Table S1). The higher decomposition temperature and char yield of Si-CTF-2 suggest enhanced thermal stability, which can be ascribed to its greater aromatic content and higher degree of conjugation arising from the biphenyl-based linker.^39,40 These results confirm that both Si-CTFs possess outstanding thermal endurance suitable for high-temperature applications.

Fig. 2 FTIR spectra of (a) Si-CTF-1 and (b) Si-CTF-2 and their monomers. (c) Solid-state ¹³C NMR spectra of Si-CTF-1 and Si-CTF-2. (d) TGA curves of Si-CTF-1 and Si-CTF-2.

It is worth noting that most reported nitrogen–silicon co-doped porous carbons are prepared via organosilane/sol–gel additive routes using external additives (e.g., organosilanes/siloxanes/TEOS), followed by pyrolysis or biomass/self-doping approaches where the final Si content and dispersion can depend strongly on mixing/processing and post-treatment steps. In contrast, the present study adopts a precursor-defined framework-to-carbon conversion strategy in which both N and Si are intrinsically incorporated into a silicon-containing covalent triazine framework (Si-CTF) prior to carbonization. Because Si and N are built into the framework backbone at the molecular level, subsequent carbonization yields N/Si co-doped microporous carbons without requiring additional dopants, post-silylation, or template removal, thereby providing a more controllable and reproducible route to dual-heteroatom-doped carbons. Importantly, the linker-dependent evolution from NSi@C-1 to NSi@C-2 demonstrates that the framework design can tune pore accessibility/transport characteristics and interfacial kinetics, establishing a clear structure–property correlation rather than relying solely on a single performance metric. The obtained frameworks were then subjected to controlled carbonization to yield nitrogen- and silicon-doped microporous carbons (NSi@Cs) (Fig. 1). The polymer samples were placed in a ceramic crucible and heated under nitrogen at a ramp rate of 5 °C min⁻¹ to 700 °C, maintained for 7 h, and then allowed to cool naturally to room temperature. The choice of 700 °C as the carbonization temperature was guided by the TGA results of the Si-CTFs, which showed that significant decomposition and structural transformation occur above this threshold, marking it as the optimal onset for stable carbon network formation. Moreover, 700 °C is the most commonly employed carbonization temperature in previously reported studies on covalent triazine frameworks,⁴¹ offering an optimal balance between framework integrity and electrical conductivity. Compared to higher temperatures such as 800 °C or 900 °C, carbonization at 700 °C helps preserve nitrogen and silicon functionalities, minimizes excessive graphitization, and maintains a high degree of microporosity, resulting in heteroatom-enriched carbons with well-defined porous structures and excellent electrochemical activity.^42–44 This controlled treatment effectively converts the triazine frameworks into conductive, N/Si-doped microporous carbon materials with robust structural stability and uniform heteroatom distribution.⁴⁵ In contrast, higher carbonization temperatures improve conductivity but result in a pronounced decrease in heteroatom content and surface functional diversity.⁴⁶ Therefore, 700 °C offers an optimal compromise, ensuring the retention of heteroatom-rich active sites while generating conductive, stable, and porous N/Si-doped carbon frameworks.

The porous textural properties of the carbonized frameworks were characterized by N₂ adsorption–desorption isotherms and pore size distribution (PSD) analyses (Fig. 3a and b).^47–49 Both NSi@C-1 and NSi@C-2 display typical type I isotherms, confirming their predominantly microporous nature,^50–53 with a steep gas uptake at low relative pressures (P/P₀ < 0.1). However, the curve of NSi@C-2 exhibits a more gradual increase and a noticeable hysteresis loop at higher relative pressures compared to NSi@C-1, indicating the coexistence of micro- and meso-porous structures (Fig. 3a). This broader adsorption behavior is attributed to the biphenyl-based linker (BP-2NHNH₂) in NSi@C-2, which forms a more open and flexible framework during polymerization and carbonization, thereby allowing the development of larger and interconnected pores. In contrast, the terphenyl-based NSi@C-1 forms a more compact structure with fewer accessible pores, resulting in a flatter adsorption curve. The PSD curves derived from NLDFT analysis revealed narrow pore widths centered at 0.73 nm and 0.99 nm for NSi@C-1 and NSi@C-2, respectively, consistent with a microporous structure (Fig. 3b and Table S2). Correspondingly, the BET surface areas were 544 m² g⁻¹ for NSi@C-1 and 732 m² g⁻¹ for NSi@C-2, while the total pore volumes were 0.21 cm³ g⁻¹ and 0.59 cm³ g⁻¹, respectively (Table S2). The enhanced surface area and larger pore volume of NSi@C-2 arise from its extended π-conjugated biphenyl structure, which introduces higher free volume and reduces framework packing density. Notably, although microporosity is beneficial for achieving high capacitance by providing abundant adsorption sites, excessively narrow micropores can impose ion-transport limitations at high current densities; therefore, introducing interconnected pores and/or hierarchical porosity is often advantageous for improving rate capability.^54,55 In this context, the larger pore volume and micro/mesopore coexistence in NSi@C-2 are expected to facilitate electrolyte penetration and reduce diffusion resistance, consistent with prior reports on heteroatom-doped hierarchical porous carbons and 3D nanoporous graphene architectures for high-performance supercapacitors.^56,57

Fig. 3 (a) N₂ adsorption–desorption isotherms and (b) pore size distribution curves of NSi@C-1 and NSi@C-2. (c–e) TEM images of NSi@C-1 under different magnifications. (f–h) TEM images of NSi@C-2 under different magnifications.

The morphological characteristics of the carbonized materials were investigated using transmission electron microscopy (TEM). As shown in Fig. 3c–h, both NSi@C-1 (Fig. 3c–e) and NSi@C-2 (Fig. 3f–h) exhibit interconnected, irregularly stacked nanosheets with rough surfaces and disordered textures, confirming the formation of amorphous porous carbon frameworks. The low-magnification TEM images reveal aggregated carbon particles that assemble into sponge-like networks, providing continuous conductive pathways favorable for ion and electron transport. At higher magnification (Fig. 3d and g), the materials show closely packed nanostructures with uniform contrast, indicating a homogeneous distribution of nitrogen and silicon dopants within the carbon matrix. The high-resolution TEM (HR-TEM) images further confirm the highly uniform and microporous structures of both carbon materials, with pore diameters smaller than 2.0 nm (Fig. 3e and h). Such a hierarchical and porous architecture facilitates efficient diffusion and charge transport, thereby enhancing the rate capability of the supercapacitors. Notably, NSi@C-2 exhibits slightly more open and expanded porous regions compared to NSi@C-1, consistent with its larger BET surface area and pore volume. This morphological difference originates from the biphenyl linker (BP-2NHNH₂) used in the precursor Si-CTF-2, which produces a more extended and less densely packed framework during carbonization. Overall, the TEM observations corroborate the nitrogen adsorption–desorption results, demonstrating that NSi@C-2 possesses a more open, interconnected porous structure than NSi@C-1, which facilitates enhanced ion diffusion and provides more accessible active sites for electrochemical processes.

The surface morphology and elemental composition of the carbonized materials were examined using scanning electron microscopy (SEM) coupled with energy-dispersive X-ray spectroscopy (EDS) mapping, as shown in Fig. S10a and b. Both NSi@C-1 and NSi@C-2 exhibit irregularly shaped, aggregated particles composed of closely connected nanosheets, forming porous and rough surface textures. These morphological features are consistent with the porous carbon frameworks observed by TEM, confirming the successful carbonization of the Si-CTF precursors. In NSi@C-1 (Fig. S10a), the SEM image shows smaller, more compact agglomerates with a relatively dense structure, which can be attributed to the shorter conjugated terphenyl linker used in the precursor (TP-2NHNH₂). In contrast, NSi@C-2 (Fig. S10b) displays larger, more open clusters with a rougher and more corrugated surface, suggesting a more expanded and interconnected pore structure. This morphological difference aligns well with the BET results, in which NSi@C-2 exhibited a higher surface area and larger pore volume due to the more extended biphenyl linkage in its Si-CTF-2 precursor. The EDS elemental mapping clearly verifies that carbon (C, red), nitrogen (N, green), and silicon (Si, yellow) are uniformly dispersed throughout both materials. The uniform dispersion of nitrogen and silicon demonstrates their successful incorporation into the carbon matrix, with no detectable phase separation or aggregation. Overall, the SEM and EDS results confirm that both NSi@C-1 and NSi@C-2 possess porous, interconnected morphologies with uniformly distributed nitrogen and silicon dopants. The more open architecture and richer heteroatom content of NSi@C-2 suggest its greater potential for applications requiring efficient ion transport and surface-driven electrochemical reactions.

The crystalline structure and graphitic ordering of the carbonized materials were examined using powder X-ray diffraction (XRD), as shown in Fig. 4a. Both NSi@C-1 and NSi@C-2 display broad diffraction peaks centered around 2θ = 25°, which corresponds to the (002) reflection plane of disordered graphitic carbon.⁵⁸ This broad feature indicates the amorphous nature of the materials and the presence of short-range ordering due to π–π stacking between adjacent aromatic layers within the carbon framework.^59,60 Notably, the (002) peak of NSi@C-1 appears slightly upshifted and more intense than that of NSi@C-2, suggesting closer π–π stacking distances and a somewhat higher degree of structural ordering.⁶¹ This difference can be attributed to the more compact packing of the terphenyl-based precursor (TP-2NHNH₂) in NSi@C-1, which promotes tighter layer alignment during carbonization. In contrast, NSi@C-2, derived from the biphenyl-based linker (BP-2NHNH₂), exhibits a slightly broader and downshifted peak, reflecting larger interlayer spacing and weaker π–π interactions due to the more open and less dense molecular architecture. From an electrochemical perspective, this structural variation plays a crucial role in determining performance. The closer π–π stacking in NSi@C-1 facilitates efficient electron transport, improving electrical conductivity and charge transfer kinetics. Meanwhile, the wider interlayer spacing and more porous nature of NSi@C-2 enhance ion accessibility and electrolyte diffusion, which are beneficial for achieving higher capacitance and rate capability. Thus, although NSi@C-1 favors rapid electronic conduction, NSi@C-2 offers improved ion diffusion pathways, together revealing a balance between electronic conductivity and ion transport governed by the π–π stacking characteristics of the carbon frameworks. The Raman spectra of NSi@C-1 and NSi@C-2 (Fig. 4b) exhibit two characteristic bands of carbonaceous materials: the D band (∼1340 cm⁻¹) and the G band (∼1580 cm⁻¹).⁶² The D band corresponds to the disordered carbon (sp³-hybridized carbon atoms and structural defects), while the G band arises from the in-plane vibration of sp²-hybridized graphitic carbon domains.^63,64 The relative intensity ratio of these two peaks (I_D/I_G) is commonly used to evaluate the degree of graphitization and defect density in carbon materials. The calculated I_D/I_G ratios were 1.08 for NSi@C-1 and 1.04 for NSi@C-2 (Table S3), indicating that both materials possess partially graphitized carbon frameworks with moderate structural disorder. The slightly lower I_D/I_G value of NSi@C-2 suggests a higher degree of graphitic ordering and fewer structural defects compared to NSi@C-1. This improvement in graphitic order for NSi@C-2 can be attributed to the biphenyl-based linker (BP-2NHNH₂), which facilitates enhanced π–π conjugation and promotes a more ordered carbon structure upon carbonization. From an electrochemical perspective, this structural difference directly influences performance. The more disordered NSi@C-1 provides abundant defect sites that can serve as active centers for ion adsorption and pseudocapacitive behavior, while the higher graphitic order in NSi@C-2 enhances electrical conductivity, charge transport, and long-term cycling stability. The balance between preserved heteroatom sites and improved graphitization enables NSi@C-2 to achieve superior overall electrochemical performance, combining efficient ion diffusion with fast electron transfer.

Fig. 4 (a) XRD patterns, (b) Raman spectra, and (c) TGA curves of NSi@C-1 and NSi@C-2. High-resolution XPS spectra of NSi@C-1 for the (d) C 1s, (e) N 1s, and (f) Si 2p regions. High-resolution XPS spectra of NSi@C-2 for the (g) C 1s, (h) N 1s, and (i) Si 2p regions.

The thermal stability of the carbonized materials was evaluated by TGA under a nitrogen atmosphere, as shown in Fig. 4c. Both NSi@C-1 and NSi@C-2 exhibit excellent thermal stability, with negligible weight loss below 150 °C, which is attributed to the removal of residual moisture and adsorbed gases. A very slow and gradual mass decrease is observed throughout the heating process, demonstrating the strong carbon framework and high thermal endurance of the heteroatom-doped structures. The total weight loss of both materials remains below 10% even at 900 °C, confirming the high degree of carbonization and robust framework stability after pyrolysis (Table S4). The chemical composition and elemental states of the synthesized materials were investigated using X-ray photoelectron spectroscopy (XPS), as shown in Fig. S11. The XPS survey spectra of NSi@C-1 and NSi@C-2 confirm the coexistence of carbon (C 1s ≈ 285 eV), nitrogen (N 1s ≈ 400 eV), and silicon (Si 2p ≈ 103 eV and Si 2 s ≈ 155 eV) elements, verifying the successful incorporation of Si and N atoms into the carbon frameworks after carbonization. Both materials exhibit strong C 1s signals, which are consistent with the dominant carbon backbone, while the presence of Si and N peaks indicates that heteroatom doping was well retained during high-temperature treatment at 700 °C. Notably, NSi@C-1 exhibits slightly higher intensities for the Si 2 s, Si 2p, and N 1s peaks compared to NSi@C-2, suggesting a higher concentration of Si and N dopants. This can be attributed to its precursor structure containing more accessible reactive sites that favor Si–N linkage formation during polymerization and carbonization. The O 1s signal observed in both samples is likely due to surface oxidation or adsorbed oxygen species. Overall, these results confirm that both carbons are successful N- and Si-doped carbons. The high-resolution XPS spectra of NSi@C-1 and NSi@C-2 provide detailed insights into the chemical states of C, N, and Si elements within the doped carbon frameworks (Fig. 4d–i and Table S5). For the C 1s spectra, both materials show major peaks at 285.4–285.2 eV (C=C), corresponding to graphitic carbon, while additional peaks at 286.1–285.6 eV (C–Si) and 287.3–287.0 eV (C=N) indicate successful bonding of carbon with silicon and nitrogen atoms, respectively. A weaker component at 289.1–289.0 eV corresponds to C=O species, likely arising from surface oxidation (Fig. 4d and g). The slightly stronger C–Si and C=N peaks in NSi@C-2 suggest enhanced heteroatom interaction and hybridization between the dopants and carbon framework. The N 1s spectra resolve into pyridinic N (∼399.3–398.8 eV), pyrrolic N/amine (∼400.4–400.2 eV), graphitic (quaternary) N (∼401.7–401.4 eV), and a minor oxidized N (∼403.2–403.0 eV) component. NSi@C-1 shows features at 399.3, 400.4, 401.4, and 403.2 eV (Fig. 4e), while NSi@C-2 appears at ∼398.8, 400.2, 401.7, and 403.0 eV (Fig. 4h). Among these, pyridinic and pyrrolic nitrogen species dominate, particularly in NSi@C-2, accounting for ∼78% of the total N content (Table S6). These nitrogen functionalities are known to enhance pseudocapacitive behavior and improve electrochemical wettability by introducing active sites for ion adsorption and redox reactions. The Si 2p spectra further corroborate Si incorporation: NSi@C-1 shows 102.6 eV (Si–C) and 104.2 eV (Si–O) (Fig. 4f), while NSi@C-2 shows 103.7 eV (Si–C) and 105.2 eV (Si–O) (Fig. 4i). The dominant Si–C signal in both materials implies that silicon is covalently bonded within the carbon framework rather than existing as oxidized impurities. Overall, the XPS fitting results reveal that NSi@C-2 possesses a higher proportion of electrochemically active nitrogen species and a more balanced distribution of Si–C and C–N bonds. This composition, combined with its larger surface area, more open porosity (as determined by BET/TEM), and hierarchical porosity, provides abundant active sites and enhances charge transport, resulting in superior electrochemical performance compared to NSi@C-1. The synergistic effect of N- and Si-doping facilitates improved electron conductivity, ion diffusion, and redox activity, which are key factors responsible for the enhanced capacitance and rate capability observed in NSi@C-2.

The electrochemical performance of NSi@C-1 and NSi@C-2 was evaluated by cyclic voltammetry (CV) using a standard three-electrode setup, where a glassy carbon electrode was used as the working electrode, Hg/HgO served as the reference electrode, and platinum wire acted as the counter electrode, all immersed in a 1 M KOH electrolyte. As illustrated in the CV curves (Fig. 5a and b), both electrodes exhibit nearly rectangular profiles over scan rates ranging from 5 to 200 mV s⁻¹, indicating typical EDLC behavior with slight pseudocapacitive effects arising from the incorporated nitrogen and silicon dopants. The CV profiles maintain excellent symmetry and negligible distortion even at high scan rates, confirming the outstanding reversibility and rate capability of the electrodes. Notably, NSi@C-2 exhibits significantly larger enclosed areas than NSi@C-1 under identical conditions, indicating higher current response and superior charge storage capacity. This enhancement can be attributed to the synergistic effect of its higher surface area, hierarchical pore structure, and optimized distribution of nitrogen functionalities, which promote efficient ion diffusion and charge transfer. Furthermore, the presence of abundant pyridinic and pyrrolic nitrogen sites in NSi@C-2 enhances redox activity, while Si doping contributes to improved electronic conductivity. Consequently, NSi@C-2 demonstrates superior electrochemical performance, making it a promising electrode material for high-performance supercapacitors. The galvanostatic charge–discharge (GCD) behaviors of NSi@C-1 and NSi@C-2 were recorded in a three-electrode cell (glassy-carbon working, Hg/HgO reference, and Pt counter electrodes) in 1 M KOH, using current densities ranging from 0.5 to 20 A g⁻¹ (Fig. 5c and d). In both cases, the GCD curves show nearly linear isosceles triangles across the whole current range, indicating highly reversible capacitive behavior and low internal resistance. The discharge time of NSi@C-2 is markedly more extended than that of NSi@C-1 at identical currents, reflecting its larger charge-storage capacity. The specific capacitance (C_s) was obtained from the GCD profiles using C_s = IΔt/(mΔV), where I is the applied current, Δt the discharge time, m is the mass of the active material, and ΔV is the potential window. As summarized in the capacitance–current plot (Fig. 5e), NSi@C-2 exhibits about 403 F g⁻¹ at 0.5 A g⁻¹, retaining 160 F g⁻¹ at 20 A g⁻¹ (∼40% retention), while NSi@C-1 exhibits 133 F g⁻¹ at 0.5 A g⁻¹ and 40 F g⁻¹ at 20 A g⁻¹ (∼30% retention). The superior capacitance and slightly better rate capability of NSi@C-2 are consistent with its higher surface area, larger pore volume, and richer distribution of electroactive N sites (pyridinic/pyrrolic), which together facilitate rapid ion diffusion and efficient charge transfer within the N/Si-doped carbon framework. Although N₂ adsorption–desorption indicates that both carbons are mainly microporous, microporosity plays a dual role in supercapacitors: it provides an abundant ion-accessible surface area for charge storage, yet excessively narrow pores may impose ion-transport resistance at high current densities. In our case, NSi@C-2 shows a higher surface area (732 m² g⁻¹), a much larger total pore volume (0.59 cm³ g⁻¹), and a slightly wider characteristic pore size (∼0.99 nm) than NSi@C-1 (544 m² g⁻¹, 0.21 cm³ g⁻¹, and ∼0.73 nm, respectively), which collectively facilitate electrolyte penetration and ion accessibility. Accordingly, NSi@C-2 retains 160 F g⁻¹ at 20 A g⁻¹ from 403 F g⁻¹ at 0.5 A g⁻¹ (∼40% retention), compared with 40 F g⁻¹ from 133 F g⁻¹ for NSi@C-1 (∼30% retention). These results suggest that, despite the predominantly microporous nature, the more accessible pore network of NSi@C-2 alleviates diffusion limitations and supports improved high-rate performance.

Fig. 5 CV curves of (a) NSi@C-1 and (b) NSi@C-2 at various scan rates (5–200 mV s⁻¹). GCD curves of (c) NSi@C-1 and (d) NSi@C-2 at different current densities (0.5–20 A g⁻¹). (e) Specific capacitance versus current density plots for both electrodes. (f) Cycling stability of NSi@C-1 and NSi@C-2 over 10 000 cycles. Comparison of the specific capacitance of NSi@C-2 with previously reported (g) redox-active porous polymers and (h) heteroatom-doped carbon materials. (i) Nyquist plots with the corresponding fitting curves for NSi@C-1 and NSi@C-2. Inset in (i): enlarged high-frequency region and equivalent circuit model used for fitting.

The long-term cycling stability of NSi@C-1 and NSi@C-2 was assessed over 10 000 consecutive charge–discharge cycles at a constant current density (10 A g⁻¹) to evaluate their electrochemical durability. As shown in Fig. 5f, both NSi@C-1 and NSi@C-2 electrodes exhibit excellent capacitance retention, which maintains over 90% of their initial capacitance after 10 000 cycles, indicating outstanding structural robustness and chemical stability of the N/Si-doped carbon frameworks. The superior stability is attributed to well-integrated porous carbon skeletons, which effectively buffer volumetric changes during repeated ion adsorption/desorption, thereby preventing structural collapse or degradation. Interestingly, NSi@C-2 retains slightly higher capacitance than NSi@C-1 (95.25% after 10 000 cycles), demonstrating enhanced electrochemical durability. Overall, both materials exhibit remarkable stability and reversibility, confirming their strong potential as durable electrode materials for high-performance supercapacitors. Post-cycling characterizations were carried out to evaluate the structural and chemical stability of the heteroatom-doped carbons after long-term operation. After 10 000 charge–discharge cycles, the recovered NSi@C-1 and NSi@C-2 electrodes were examined by SEM and TEM. Compared with the pristine samples, both cycled electrodes retained their aggregated porous carbon morphology without noticeable particle pulverization, structural collapse, or severe cracking, indicating good mechanical robustness during repeated cycling (Fig. S12). In addition, XPS survey spectra collected before and after cycling showed that the characteristic signals of C 1s, N 1s, O 1s, and Si 2p/Si 2 s remained clearly detectable for both materials, confirming that the N- and Si-heteroatom functionalities are preserved during electrochemical operation (Fig. S13). Only minor variations in peak intensity were observed, which can be attributed to surface reorganization and/or adsorption of electrolyte-derived species after prolonged cycling. These post-cycling results provide direct evidence of the excellent electrochemical durability of NSi@C-1 and NSi@C-2, consistent with their high capacitance retention over 10 000 cycles. The specific capacitance of NSi@C-2 is remarkably higher than that of many reported redox-active and heteroatom-doped porous materials. For instance, it surpasses those of COF/rGO 269 F g⁻¹ at 0.5 A g⁻¹),⁶⁵ BTPP-DBTh (143.27 F g⁻¹ at 0.5 A g⁻¹),⁶⁶ BF-DTDO (95.62 F g⁻¹ at 0.5 A g⁻¹),⁶⁷ BF-Ph-DTDO (288.8 F g⁻¹ at 0.5 A g⁻¹),⁶⁷ TAT-CMP-1 (141 F g⁻¹ at 1 A g⁻¹),⁶⁸ TAT-CMP-2 (183 F g⁻¹ at 1 A g⁻¹),⁶⁸ TPA-By (78 F g⁻¹ at 1 A g⁻¹),⁶⁹ TPA-Bz (55.1 F g⁻¹ at 1 A g⁻¹),⁶⁹ BT-PDI (196 F g⁻¹ at 1 A g⁻¹),⁷⁰ c-DDSQ-MDA-BMI (73.6 F g⁻¹ at 0.5 A g⁻¹),⁷¹ cNPIM EA-TB₈₀ (46 F g⁻¹ at 1 A g⁻¹),⁷² CoPc-CMP/CNTs-2 (107.2 F g⁻¹ at 1 A g⁻¹),⁷³ GH-CMP (182.7 F g⁻¹ at 0.5 A g⁻¹),⁷⁴ CAP-2 (233 F g⁻¹ at 1 A g⁻¹),⁷⁵ An-TPP POP (38.12 F g⁻¹ at 1 A g⁻¹),⁷⁶ and An-TPA POP (117.7 F g⁻¹ at 1 A g⁻¹)⁷⁶ (Fig. 5g and Table S7). Furthermore, compared to typical heteroatom-doped carbons, such as N-, S-, or B-doped porous carbons, which generally deliver 150–350 F g⁻¹ under similar conditions, NSi@C-2 exhibits a distinctly higher charge-storage capability. For instance, it surpasses those of NPCC-550 (191 F g⁻¹ at 1 A g⁻¹),⁷⁷ NPCC-650 (200 F g⁻¹ at 1 A g⁻¹),⁷⁷ PC-R6A7 (366 F g⁻¹, at 20 A g⁻¹),⁷⁸ OPBNP (289 F g⁻¹ at 5 A g⁻¹),⁷⁹ MCNMs (282 F g⁻¹ at 0.2 A g⁻¹),⁸⁰ NPC-1.5 (371 F g⁻¹, at 1 A g⁻¹),⁸¹ OMC-2 (326 F g⁻¹ at 0.5 A g⁻¹),⁸² OSC (354 F g⁻¹, at 1 A g⁻¹),⁸³ S, N-PIC-1 (235.3 F g⁻¹ at 0.5 A g⁻¹),⁸⁴ PCNs/GCNs-5 (388 F g⁻¹ at 1 A g⁻¹),⁸⁵ N-doped C/rGO (234 F g⁻¹ at 0.8 A g⁻¹),⁸⁶ and BC-SA-0.5CP (324.85 F g⁻¹ at 0.5 A g⁻¹)⁸⁷ (Fig. 5h and Table S8). This outstanding performance can be attributed to the synergistic effects of its hierarchical porous structure, abundant nitrogen and silicon-dopant sites, and well-developed π-conjugated framework. The coexistence of Si–C and Si–N bonds enhances electrical conductivity. At the same time, pyridinic and pyrrolic nitrogen functionalities contribute additional faradaic activity, collectively leading to efficient ion diffusion, rapid charge transfer, and excellent capacitive behavior.

To further understand the charge transport and ion diffusion behavior of the heteroatom-doped carbons, electrochemical impedance spectroscopy (EIS) was performed at the open-circuit potential with a sinusoidal amplitude of 10 mV over a frequency range from 100 kHz to 0.01 Hz. As depicted in Fig. 5i, the Nyquist plots of NSi@C-1 and NSi@C-2 show the characteristic response of porous carbon supercapacitor electrodes, featuring a small semicircular arc in the high-frequency range and an oblique line in the low-frequency region. It should be emphasized that the semicircle does not correspond to a specific faradaic reaction; instead, it reflects the interfacial processes and distributed time constants at the electrode/electrolyte interface, which are commonly described by the combined contribution of contact/film resistance and charge-transfer resistance (R_ct) in porous electrodes. The high-frequency interception on the Z′ axis corresponds to the solution/series resistance (R_s), while the semicircle's diameter primarily reflects R_ct. The low-frequency region evolves toward a nearly vertical line, indicating a predominantly capacitive response with limited diffusion impedance, consistent with efficient ion transport within the porous carbon network. Experimentally, the R_s values were determined to be 12.91 Ω for NSi@C-1 and 10.01 Ω for NSi@C-2, suggesting that NSi@C-2 possesses lower interfacial resistance and faster charge-transfer kinetics, consistent with its enhanced electrical conductivity and superior electrochemical performance. The equivalent circuit model (inset of Fig. 5i) was used to fit the experimental data, where R_s represents the electrolyte and internal resistance, CPE1 represents the double-layer capacitance at the electrode/electrolyte interface, R_ct denotes the charge-transfer resistance, and W1 denotes the Warburg element associated with ion diffusion. The fitted R_s values are 13.40 Ω for NSi@C-1 and 10.47 Ω for NSi@C-2, indicating comparable series resistance in the two systems. Importantly, NSi@C-2 exhibits a significantly lower R_ct value compared to NSi@C-1, indicating superior electrical conductivity and faster charge-transfer kinetics. Specifically, the fitted R_ct values were 67.25 Ω for NSi@C-1 and 44.50 Ω for NSi@C-2, confirming that NSi@C-2 facilitates more efficient interfacial charge transport. Notably, the presence of a semicircle in EIS does not imply pseudocapacitance-dominated storage; pseudocapacitive contributions should be evaluated from CV features and kinetic analyses (e.g., b-value/Dunn/Trasatti), while EIS primarily reflects resistance and transport characteristics. This enhancement can be attributed to the more extended π-conjugation and well-developed porous structure in NSi@C-2, which collectively facilitate rapid electron transport and efficient ion accessibility. The improved interfacial charge dynamics and lower impedance confirm that NSi@C-2 exhibits better electrochemical kinetics and enhanced conductivity, which account for its superior specific capacitance and rate capability compared to NSi@C-1. Additionally, the electrical conductivity of the synthesized heteroatom-doped carbons was assessed to verify their efficient charge-transport capabilities. Conductivity measurements for NSi@C-1 and NSi@C-2 were carried out at room temperature using a four-probe setup, ensuring reliable data by minimizing the influence of contact resistance. The measured conductivities were 3.64 S cm⁻¹ for NSi@C-1 and 4.23 S cm⁻¹ for NSi@C-2, confirming their excellent electronic transport characteristics. These values are notably higher than those typically reported for porous carbon-based materials, demonstrating that the incorporation of nitrogen and silicon heteroatoms effectively enhances charge-transport pathways and the overall electrical conductivity of the carbon frameworks.

The power-law relationship in eqn (1) was employed to elucidate the dependence of the current response (i) on scan rate (v), thereby gaining insight into the charge-storage kinetics of the nitrogen- and silicon-co-doped microporous carbons as follows:⁸⁸

i = av^b, (1)

where a and b are adjustable parameters, and b is obtained from the slope of the linear plot of log(i) versus log(v). A b value approaching 1.0 indicates a predominantly surface-controlled (capacitive-controlled, typical of EDLC behavior) process, while a value close to 0.5 suggests a diffusion-controlled contribution. As shown in Fig. 6a and b, NSi@C-1 exhibits b values of 0.773 (anodic) and 0.692 (cathodic), while NSi@C-2 shows higher b values of 0.913 (anodic) and 0.845 (cathodic). These findings confirm that both co-doped carbon materials store charges via a synergistic combination of surface (capacitive)-controlled and diffusion-controlled processes. Notably, NSi@C-2 exhibits a greater surface-controlled effect, with enhanced ion accessibility and faster surface redox kinetics, improving its overall electrochemical performance. To further quantify the surface-controlled contribution, Dunn's method was applied using eqn (2) as follows:⁸⁹

i(V) = k₁v + k₂v^1/2, (2)

where k₁v represents the capacitive-controlled contribution and k₂v^1/2 represents the diffusion-controlled contribution at a given potential. At a scan rate of 5 mV s⁻¹, NSi@C-2 exhibits a significantly higher capacitive contribution of 63%, while NSi@C-1 exhibits a lower value of 30% (Fig. 6c and d), confirming the improved surface-controlled charge accumulation and ion accessibility of NSi@C-2. As the scan rate gradually increases to 200 mV s⁻¹, the capacitive contribution rises correspondingly, reaching approximately 73% for NSi@C-1 and 92% for NSi@C-2 (Fig. 6e and f). This indicates that charge storage is increasingly dominated by surface-controlled processes at higher scan rates. At elevated sweep rates, ion diffusion into the inner pore structure is kinetically restricted, and charge accumulation primarily occurs via rapid surface adsorption and desorption.⁹⁰ Conversely, at lower scan rates, ions have sufficient time to diffuse into the inner pores of the electrode framework, thereby enhancing the diffusion-controlled contribution and reducing the relative proportion of capacitive charge storage.⁹¹

Fig. 6 Log(i) vs. log(v) plots used to determine the b values for (a) NSi@C-1 and (b) NSi@C-2 carbons. Contributions from (c) capacitive and (d) diffusion-controlled processes assessed at 5 mV s⁻¹. Characteristics of charge-storage behavior governed by the capacitive–diffusion mechanisms at different scan rates for (e) NSi@C-1 and (f) NSi@C-2 carbons.

Furthermore, the Trasatti method was employed to provide a deep understanding of the contributions to charge storage. This analysis separates the overall capacitance into an outer (surface-accessible) contribution (C_Out, mainly associated with rapidly accessible surface sites/EDLC) and an inner (diffusion-limited) contribution (C_In, associated with charge storage within less-accessible pores and/or diffusion-limited/pseudocapacitance processes). These parameters were determined according to eqn (3) and (4) as follows:^92,93

C_s = C_Out + kν^−1/2, (3)

C_s⁻¹ = C_T⁻¹ + k′ν^1/2, (4)

where C_s represents the specific capacitance derived from the CV curves at different scan rates (ν), C_T is the total capacitance extrapolated at v → 0, and k and k′ are constants. Linear plots of C_s vs. ν^−1/2 and C_s⁻¹ vs. ν^1/2 were constructed to determine the C_Out and C_T, respectively (Fig. 7a and b for NSi@C-1 and Fig. 7d and e for NSi@C-2). The C_Out values were 68.70 F g⁻¹ and 316.10 F g⁻¹ for NSi@C-1 and NSi@C-2, respectively (Fig. 7a and d, respectively), while the C_T values were calculated to be 239.23 F g⁻¹ for NSi@C-1 and 480.76 F g⁻¹ for NSi@C-2 (Fig. 7b and e), respectively. Based on these results, the inner diffusion-controlled capacitance (C_In) can be obtained from eqn (5) as follows:⁹²

C_In = C_T − C_Out. (5)

Fig. 7 Trasatti analysis of NSi@C-1 and NSi@C-2 electrodes: (a and d) determination of the external (surface-accessible) capacitance (C_Out), (b and e) estimation of the total capacitance (C_T), and (c and f) quantitative contribution of C_Out and C_In to the overall charge storage.

The estimated C_In values for NSi@C-1 and NSi@C-2 were 170.53 and 164.66 F g⁻¹, respectively. As summarized in the pie charts (Fig. 7c and f), the maximum capacitance contributions of NSi@C-2 from electrical double-layer capacitance (surface-controlled) and pseudocapacitance (diffusion-controlled) processes were 65.74% and 34.25%, respectively, while those of NSi@C-1 were 28.71% and 71.28%, respectively. NSi@C-2 exhibits a much higher fraction of outer, surface-accessible capacitance (65.74%) than NSi@C-1 (28.71%), indicating that a larger portion of its charge storage originates from readily accessible sites at or near the electrode/electrolyte interface. This behavior is advantageous for rapid charge–discharge because ions can be adsorbed/desorbed (and the electric double layer can form) with minimal transport distance and reduced diffusion resistance, thereby improving the utilization of the active surface area at high scan rates/current densities. The markedly higher C_Out contribution of NSi@C-2 is consistent with its more favorable textural features (higher surface area/pore volume and slightly larger characteristic pore size) and lower interfacial resistance, which collectively facilitate electrolyte penetration and fast ion/electron transport. In contrast, NSi@C-1 shows a dominant inner, diffusion-limited contribution (71.28%), implying that a substantial fraction of charge storage occurs within less-accessible pore domains where ion migration is kinetically hindered. Consequently, under fast-charging conditions, these inner sites cannot be fully accessed within the limited timescale, leading to a larger diffusion-controlled fraction and a comparatively weaker rate capability.⁹⁴ Overall, the Trasatti analysis shows that the superior high-rate behavior of NSi@C-2 is primarily due to its higher proportion of surface-accessible charge storage and reduced diffusion constraints relative to NSi@C-1.

To explain the higher capacitance and lower charge-transfer resistance of NSi@C-2 in KOH, we analyzed electrolyte–surface interactions through adsorption-site screening and DFT calculations. The optimized molecular structures of the nitrogen- and silicon-co-doped microporous carbons are shown in Fig. 8a. Adsorption screening in aqueous KOH indicates that both NSi@C-1 and NSi@C-2 exhibit greater interfacial affinity than the Si-free model, consistent with their higher density of polar heteroatom environments (Fig. 8b). Under interfacial contact conditions, NSi@C-2, which has a more negative adsorption energy, exhibits the most favorable interaction with K-containing species (−0.777 kcal mol⁻¹) compared to NSi@C-1 (−0.599 kcal mol⁻¹) and N@C-2 (−0.548 kcal mol⁻¹) (Table S9). This suggests that Si incorporation further enhances cation affinity at the electrolyte–carbon interface. Importantly, NSi@C-2 shows a significantly lower deformation energy (0.035 kcal mol⁻¹) than NSi@C-1 (0.257 kcal mol⁻¹) and the Si-free model (0.211 kcal mol⁻¹), indicating that the adsorption environment in NSi@C-2 stabilizes electrolyte localization without requiring major structural distortion, which is beneficial for rapid ion rearrangement during charge/discharge. Additionally, H₂O adsorption is stronger for NSi@C-2 than for NSi@C-1, supporting improved wettability and the formation of a stable hydration layer that enhances ion accessibility within the porous framework. Notably, NSi@C-2 has more abundant polar adsorption motifs (e.g., Si–O/Si=O and O/N-rich environments) than NSi@C-1, as shown by XPS results. These motifs collectively promote stronger interfacial polarization and more effective hydrated-electrolyte adsorption, aligning with improved electrochemical performance kinetics. Furthermore, the ESP mapping (Fig. 8c) shows that NSi@C-2 has a more varied surface potential distribution, indicating stronger local polarization around Si/N-rich regions. This polarization likely helps stabilize hydrated electrolyte components through electrostatic interactions and H-bond networks, providing a mechanistic explanation for improved adsorption and lower interfacial resistance. Consistently, the DFT electronic-structure analysis (Fig. S14) further shows that NSi@C-2 exhibits a higher, more continuous density of states (DOS) in the frontier energy region than NSi@C-1, implying improved electronic accessibility and charge accommodation, consistent with the experimentally observed lower charge-transfer resistance. The dopant-associated (Si/O/N) contributions modulate the electronic structure and enhance interfacial charge localization. Finally, the enhanced adsorption and polarization at Si–O/N-rich sites, along with the improved electronic structure of the conjugated carbon framework, explain the lower interfacial resistance and higher capacitance of NSi@C-2 compared to NSi@C-1.

Fig. 8 (a) Optimized structures, (b) adsorption configurations for the aqueous KOH species over the optimized carbon cluster, and (c) ESP maps of the Si-free model, NSi@C-1, and NSi@C-2.

The results demonstrate that NSi@C-2 delivers a higher specific capacitance than NSi@C-1 because it combines a more accessible electrochemical surface, a more favorable pore architecture for hydrated-ion transport, and a stronger electrolyte–electrode affinity with faster interfacial kinetics, all of which are consistently supported by structural, spectroscopic, impedance, and computational results. Texturally, NSi@C-2 possesses a higher BET surface area (732 vs. 544 m² g⁻¹) and a much larger pore volume (0.59 vs. 0.21 cm³ g⁻¹), together with a slightly wider micropore size (∼0.99 vs. ∼0.73 nm), which increases electrolyte wettability/access and enables more effective formation of the electric double layer on internal surfaces. Consistently, XRD indicates a more disordered/expanded carbon stacking (broader/downshifted (002) feature), facilitating ion penetration through the carbon domains, while Raman shows a marginally lower I_D/I_G for NSi@C-2 (1.04 vs. 1.08), suggesting improved graphitic ordering that benefits electronic transport and stability. Surface chemistry further supports superior electrochemical utilization. XPS confirms stable incorporation of Si and N, and NSi@C-2 shows a high fraction of electrochemically active N species (pyridinic/pyrrolic N), while Si–C/Si–O environments increase surface polarity and electrolyte affinity, promoting higher effective capacitance. These structural advantages translate directly into improved electrochemical kinetics and reduced resistance. NSi@C-2 exhibits lower R_s and R_ct from EIS fitting (10.01 Ω and 44.50 Ω vs. 12.91 Ω and 67.25 Ω) and higher conductivity (4.23 vs. 3.64 S cm⁻¹), enabling rapid charge propagation and minimizing polarization. Importantly, the kinetic analyses confirm that NSi@C-2 stores charge in a more surface-accessible manner, with b-values closer to 1 (0.913/0.845 vs. 0.773/0.692), higher capacitive contribution by Dunn analysis (63% at 5 mV s⁻¹ and 92% at 200 mV s⁻¹), and a much larger outer-surface capacitance from Trasatti analysis (C_Out = 316.10 F g⁻¹; 65.74% surface-controlled contribution), demonstrating that its larger porous surface is effectively utilized even at high rates. Importantly, adsorption-site screening/DFT supports stronger electrolyte affinity for NSi@C-2. The adsorption of K-containing species is more favorable for NSi@C-2 (−0.777 kcal mol⁻¹) than for NSi@C-1 (−0.599 kcal mol⁻¹), and NSi@C-2 also shows a much lower deformation energy (E_def = 0.035 vs. 0.257 kcal mol⁻¹ for NSi@C-1), indicating stabilized ion localization with minimal framework distortion, consistent with improved wettability and faster interfacial charging. The electrochemical behavior of the symmetric supercapacitor assembled with NSi@C-2 electrodes was thoroughly examined in a two-electrode system employing 3 M KOH as the aqueous electrolyte, as depicted in Fig. 9a. Both the positive and negative electrodes were composed of the same heteroatom-doped carbon (NSi@C-2), with a porous filter paper separator placed between them to ensure ionic conductivity while preventing electrical short-circuiting. The device architecture provides a realistic assessment of the material's practical energy storage capability under operating conditions. The CV curves (Fig. 9b) recorded at various scan rates ranging from 5 to 200 mV s⁻¹ within the potential window of 0–1.2 V exhibit well-retained quasi-rectangular shapes, indicating ideal capacitive behavior and excellent reversibility during rapid charging and discharging processes. Even at high scan rates, the CV profiles remain nearly undistorted, suggesting superior rate capability and low internal resistance, which are essential for fast ion diffusion and charge propagation throughout the electrode matrix.

Fig. 9 (a) Schematic of the symmetric NSi@C-2 device in a 3 M KOH electrolyte. (b) CV curves of the device at different scan rates (5–200 mV s⁻¹) within a voltage window of 0–1.2 V. (c) GCD curves recorded at various current densities (0.2–8 A g⁻¹); inset: enlarged GCD profiles at higher current densities. (d) Specific capacitance values of the device at different current densities. Inset: three NSi@C-2-based SSC devices connected in series successfully powering a red LED, highlighting their practical real-world application.

The outstanding capacitive response can be attributed to the hierarchical porosity and rich heteroatom doping (N and Si), which provide both electrochemically active sites and well-connected conductive channels. The charge–discharge curves (Fig. 9c) exhibit nearly linear, symmetric profiles over the tested current densities (0.2–8.0 A g⁻¹), confirming the highly reversible and efficient capacitive behavior of the NSi@C-2 electrode. The absence of significant voltage drops or distortions during the charge–discharge process indicates low internal resistance and minimal polarization losses. This linear trend reflects the dominance of electric double-layer capacitance, supported by fast ion diffusion within the interconnected porous framework and excellent electrical conductivity from N, Si heteroatom doping. Together, these features enable efficient charge storage and release, ensuring excellent coulombic efficiency and long-term cycling stability. The specific capacitance determined from the GCD measurements reaches 211 F g⁻¹ at a current density of 0.2 A g⁻¹ and remains at 53% of this value even at 8 A g⁻¹ (Fig. 9d), indicating outstanding rate performance and rapid ion diffusion within the NSi@C-2 electrode. This remarkable performance primarily stems from the synergistic influence of nitrogen and silicon co-doping, which collectively create abundant redox-active sites and facilitate more efficient electron transport channels within the carbon framework. In contrast to previous studies on symmetric devices employing porous redox-active polymers and heteroatom-doped carbon materials, the NSi@C-2-based device exhibits significantly higher specific capacitance (Table S10).^95–100 This superior performance arises from its unique combination of hierarchical porosity, high surface area, and dual N, Si heteroatom doping, which synergistically enhances ion diffusion, charge transport, and active site accessibility, thereby outperforming conventional porous polymer electrodes reported in the literature. To further demonstrate practical applicability, three identical NSi@C-2-based symmetric supercapacitor devices were connected in series to extend the operating voltage. Each device was charged to 1.2 V, giving a total charging voltage of 3.6 V (1.2 V × 3) for the series-connected module. After charging, the module was directly connected to a red LED using alligator clips, and the LED was successfully illuminated (Fig. 9d, inset and S15), confirming that the assembled supercapacitor pack could deliver sufficient voltage and power to drive an external electronic load. This simple demonstration supports the feasibility of the NSi@C-2 devices for practical energy-storage applications and is consistent with their high capacitance, favorable rate capability, and excellent cycling stability.

The device demonstrates exceptional overall electrochemical characteristics, reflecting the synergistic influence of its hierarchical porous structure and N, Si heteroatom co-doping. The coulombic efficiency of 97.20% sustained throughout 10 000 continuous charge–discharge cycles at 0.3 A g⁻¹ indicates nearly perfect charge reversibility and minimal energy loss during repeated cycling (Fig. 10a). Such high efficiency arises from the robust electrode–electrolyte interface and the chemical stability of the doped carbon framework, which mitigates parasitic reactions and structural degradation. In addition, the capacitance retention of 91.50% confirms the remarkable mechanical integrity and durability of the electrodes, emphasizing the resilience of the porous architecture under long-term operational stress (Fig. 10b). To further evaluate the practical device stability, the symmetric two-electrode supercapacitor assembled with NSi@C-2 was subjected to long-term galvanostatic charge–discharge cycling for 10 000 cycles. After cycling, the device electrode was recovered and examined using TEM. As shown in Fig. S16, the cycled NSi@C-2 electrode retains its aggregated porous carbon framework without obvious particle pulverization, collapse of the porous texture, or severe structural fragmentation, indicating excellent mechanical integrity under repeated ion insertion/desorption and current flow. This preserved microstructure after prolonged cycling provides direct evidence that the NSi@C-2-based device exhibits robust structural stability during practical two-electrode operation, which is consistent with its high capacitance retention over 10 000 cycles. The Ragone plot further illustrates that the NSi@C-2 symmetric device achieves an impressive energy density of 152 W h kg⁻¹ at a power density of 1037 W kg⁻¹ (Fig. 10c), surpassing most conventional carbon-based supercapacitors. Even at higher power densities, the device maintains a stable energy output, demonstrating its excellent rate performance and suitability for high-power energy storage systems. Additionally, EIS reveals a nearly vertical line in the low-frequency region and a small semicircle at high frequencies, which are characteristics of low internal resistance and efficient charge transport. A charge-transfer resistance (R_ct) value of approximately 4.89 Ω indicates fast ion diffusion and superior electrical conductivity, arising from the synergistic effect of nitrogen and silicon dual-doping combined with the well-developed porous carbon structure (Fig. 10d and S17). Overall, the superior electrochemical behavior of the NSi@C-2-based symmetric device, including high specific capacitance, excellent rate performance, and outstanding cycling stability, can be ascribed to its hierarchically porous architecture, enhanced π–π conjugation, and heteroatom doping (N and Si), which collectively facilitate fast ion diffusion, efficient electron transport, and stable electrochemical interfaces. These findings confirm that NSi@C-2 is a highly promising electrode material for next-generation high-performance and long-life energy storage devices.

Fig. 10 (a) coulombic efficiency and (b) capacitance retention plots of the NSi@C-2-based symmetric supercapacitor device over 10 000 charge–discharge cycles. (c) Ragone plot and (d) Nyquist plot of the NSi@C-2-based symmetric supercapacitor device.

3. Conclusion

In summary, two nitrogen–silicon co-doped microporous carbons, NSi@C-1 and NSi@C-2, were successfully developed through a one-step carbonization of silicon-containing covalent triazine frameworks constructed from 4,4′,4″,4‴-silanetetrayltetrabenzaldehyde (Si-4CHO) and carboximidamide-based linkers (TP-2NHNH₂ and BP-2NHNH₂, respectively). Structural and electrochemical analyses revealed that NSi@C-2, derived from the biphenyl linker, exhibits a larger surface area (732 m² g⁻¹), higher pore volume (0.59 cm³ g⁻¹), and enhanced heteroatom distribution, leading to superior ion transport and charge-transfer kinetics. The dual N/Si doping and hierarchical porosity synergistically promote high conductivity, wettability, and electrochemical reactivity. As a result, NSi@C-2 achieved an outstanding specific capacitance of 403 F g⁻¹ at 0.5 A g⁻¹ and maintained 95.25% retention after 10 000 cycles. In a symmetric NSi@C-2//NSi@C-2 device, it delivered 211 F g⁻¹, an exceptional energy density of 152 W h kg⁻¹, and 91.50% retention after 10 000 cycles, outperforming most previously reported porous carbon electrodes. This work demonstrates a facile framework-to-carbon transformation strategy for constructing hierarchically porous, heteroatom-doped carbons with tunable structures and remarkable long-term stability, providing new insight into the rational design of advanced electrode materials for next-generation high-performance supercapacitors.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article are included in the main article and its Supplementary Information (SI). Supplementary information: contains synthetic routes, additional FTIR, NMR, TGA, SEM/EDS, XPS, BET/pore-size analyses, electrochemical measurements, comparison tables, computational details, and additional figures. See DOI: https://doi.org/10.1039/d5ta09202g.

Acknowledgements

This study was supported financially by the National Science and Technology Council, Taiwan, under the contract number NSTC 112-2221-E-110-005-MY3.

References

1. H. Jafarizadeh, E. Yamini, S. M. Zolfaghari, F. Esmaeilion, M. E. H. Assad and M. Soltani, Energy Rep, 2024, 12, 2179–2192.
2. S. Mandal, A. B. Mendhe, H. M. Rakhade, N. S. Barse, M. Roy, P. Rosaiah, T. Park, H. S. Lee, A. C. Mendhe and D. Kim, Chem. Eng. J. Adv., 2025, 21, 100690.
3. S. Liu, H. Zhang, X. Peng, J. Chen, L. Kang, X. Yin, Y. Yusuke and B. Ding, ACS Nano, 2025, 19, 13591–13636.
4. Y. M. Nabil, S. Abdelnaser, A. A. Mohammed, S. W. Kuo and A. F. M. EL-Mahdy, J. Mater. Chem. A, 2025, 13, 26337–26349.
5. F. Mashkoor, M. Shoeb, J. A. Khan, M. A. Gondal and C. Jeong, J. Energy Storage, 2024, 104, 114578.
6. S. V. Bhosale and S. V. Bhosale, Chem. Sci., 2025, 16, 10159–10227.
7. Z. Ayaganov, V. Pavlenko, S. F. B. Haque, A. Tanybayeva, J. Ferraris, A. Zakhidov, Z. Mansurov, Z. Bakenov and A. Ng, J. Energy Storage, 2024, 78, 110035.
8. L. Qi, J. Li, Y. Liu, K. Liu, M. Wang and H. Cui, Small, 2025, 21, 2408221.
9. M. Shoeb, F. Mashkoor, J. A. Khan, M. A. Gondal, B. J. Kim and C. Jeong, J. Energy Storage, 2025, 134, 118208.
10. L. R. Ahmed, J. Luder, C. H. Chuang and A. F. M. EL-Mahdy, ACS Appl. Mater. Interfaces, 2024, 16, 30408–30420.
11. S. Zhong, L. Dai, H. Xu, X. C. Yang, X. Zheng and S. Wang, J. Energy Storage, 2024, 97, 112867.
12. I. Hussain, K. Singh, A. C. Mendhe, M. R. Thalji, S. Mandal, I. Ali, A. F. M. EL-Mahdy, P. Rosaiah, M. K. Aslam and T. L. Tamang, Prog. Solid State Chem., 2025, 78, 100523.
13. J. Zhan and A. F. M. EL-Mahdy, Chem. Eng. J., 2023, 473, 145124.
14. S. K. Pati, S. Muduli, D. Patra, S. Oh, B. J. Kim, S. Mishra and S. Park, J. Mater. Chem. A, 2025, 13, 20504–20518.
15. C. Xue, Y. Zhang, L. Wang, C. Du, Z. Chang, Q. Geng, Q. Li, C. Du and J. Zhang, J. Colloid Interface Sci., 2025, 700, 138331.
16. F. Pileri, W. da Silva Freitas, A. D'Epifanio and B. Mecheri, J. Energy Storage, 2025, 113, 115677.
17. Y. Zhang, C. Kang, W. Zhao, B. Sun, X. Xiao, H. Huo, Y. Ma, P. Zuo, S. Lou and G. Yin, Energy Storage Mater., 2022, 47, 178–186.
18. M. G. Kotp, N. L. Torad, J. Lüder, A. A. El-Amir, W. Chaikittisilp, Y. Yamauchi and A. F. M. EL-Mahdy, J. Mater. Chem. A, 2023, 11, 764–774.
19. M. G. Kotp, C. L. Chang and A. F. M. EL-Mahdy, J. Water Process Eng., 2023, 53, 103675.
20. Y. Zhang, B. Zhang, L. Chen, T. Wang, M. Di, F. Jiang, X. Xu and S. Qiao, J. Colloid Interface Sci., 2022, 606, 1534–1542.
21. Y. Zheng, N. A. Khan, X. Ni, K. A. I. Zhang, Y. Shen, N. Huang, X. Y. Kong and L. Ye, Chem. Commun., 2023, 59, 6314–6334.
22. S. Raja, G. T. S. T. da Silva, E. A. Reis, J. C. da Cruz, A. B. Silva, M. B. Andrade, G. Periyasami, P. Karthikeyan, I. F. Perepichka, L. H. Mascaro and C. Ribeiro, Energy Fuels, 2024, 37, 19113–19123.
23. P. Song, J. Chen, Z. Yin, Z. Yang and L. Wang, ChemCatChem, 2025, 17, e202401441.
24. C. Jiang, H. Liu, J. Ye, N. Wang, Y. Tang, C. He, H. Zhang and B. Chen, Batter. Supercaps, 2025, 8, e202400563.
25. K. Yang, Q. Fan, C. Song, Y. Zhang, Y. Sun, W. Jiang and P. Fu, Green Energy Resour., 2023, 1, 100030.
26. D. R. L. Peralta, P. U. Okoye and C. Alegre, J. Power Sources, 2024, 617, 235140.
27. Y. Chen, G. Xie, R. Sun, X. Gao, Z. Pan, B. Cui, X. Gong and C. Yang, Mater. Today Commun., 2024, 39, 108814.
28. M. Ö. A. Çolak, A. L. Passos, A. Güngör, E. Erdem and R. Genç, J. Power Sources, 2025, 645, 237230.
29. P. Kour, S. Kour, A. L. Sharma and K. Yadav, Fuel, 2024, 358, 130173.
30. M. Saleem, U. Lassi, V. Srivastava and S. Tuomikoski, J. Power Sources, 2025, 641, 236879.
31. F. Niu, J.-M. Liu, L.-M. Tao, W. Wang and W.-G. Song, J. Mater. Chem. A, 2013, 1, 6130–6133.
32. B. Pérez-Román, M. A. Mazo, A. Merchán del Real, J. Rubio and F. Rubio-Marcos, Mater. Adv., 2025, 6, 8114–8130.
33. S. K. Ramasahayam, Z. Hicks and T. Viswanathan, ACS Sustain. Chem. Eng., 2015, 3, 2194–2202.
34. P. Ghimire, C. Gunathilake, N. P. Wickramaratne and M. Jaroniec, Carbon, 2017, 121, 408–417.
35. I. E. Monje, N. Sanchez-Ramirez, S. H. Santagneli, P. H. Camargo, D. Belanger, S. B. Schougaard and R. M. Torresi, J. Electroanal. Chem., 2021, 901, 115732.
36. Y. Xiao, D. Dong, J. Wang, J. Liu, T. Wang and Y. Zhang, Chem. Eng. J., 2024, 500, 157084.
37. J. Song, T. Mu, X. Xu, J. Jia, X. Ban, F. Liu, J. Gao and X. Zhu, Adv. Funct. Mater., 2026, 36, 25015805.
38. Z. Chen, J. Wang, J. Lu, C. Ma, X. Liu, J. Wang, W. Qiao and L. Ling, ACS Appl. Nano Mater., 2024, 7, 6418–6428.
39. Y. Zheng, C. Lin, P. Feng, P. Geng, S. Zhang, Y. Chen and Y. Shen, Chem. Commun., 2025, 61, 16354–16371.
40. M. G. Kotp, A. F. M. EL-Mahdy, T. L. Yang and S. W. Kuo, Microporous Mesoporous Mater., 2022, 331, 111669.
41. A. Maity, A. Jana, S. K. Dey, C. Janiak and A. Bhunia, Mater. Adv., 2025, 6, 4337–4344.
42. P. Yuan, J. Mo, X. Sha, X. Liu, L. Xiao and G. Chen, Diam. Relat. Mater., 2025, 152, 111921.
43. W. Li, Y. Xu, G. Wang, T. Xu and C. Si, Adv. Energy Mater., 2024, 14, 2403593.
44. X. Niu, R. Uchino, N. Moriyama, H. Nagasawa, T. Tsuru and M. Kanezashi, ACS Sustain. Chem. Eng., 2025, 13, 14140–14151.
45. Y. Gao, P. Cui, J. Liu, W. Sun, S. Chen, S. Chou, L. P. Lv and Y. Wang, ACS Appl. Energy Mater., 2021, 4, 4519–4529.
46. J. N. Pagaduan, A. Bhardwaj, T. Y. McCarty, S. Kraemer, C. J. Kearney, J. J. Watkins, T. Emrick and R. Katsumata, Adv, Funct. Mater, 2024, 34, 2310296.
47. A. F. Saber, C. C. Chueh, M. Rashad, S. W. Kuo and A. F. M. EL-Mahdy, Mater. Today Sustain., 2023, 23, 100429.
48. A. F. M. EL-Mahdy, H. A. Omr, Z. A. ALOthman and H. Lee, J. Colloid Interface Sci., 2023, 633, 775–785.
49. A. F. Saber, K. Y. Chen, A. F. M. EL-Mahdy and S. W. Kuo, J. Polym. Res., 2021, 28, 430.
50. A. F. Saber, S. U. Sharma, J. T. Lee, A. F. M. EL-Mahdy and S. W. Kuo, Polymer, 2022, 254, 125070.
51. T. L. Yang, J. Y. Chen, S. W. Kuo, C. T. Lo and A. F. M. EL-Mahdy, Polymers, 2022, 14, 3428.
52. J. H. Wang, A. E. Hassan, A. M. Elewa and A. F. M. EL-Mahdy, J. Mater. Chem. A, 2024, 12, 14005–14021.
53. G. Zhang, A. M. Elewa, M. Rashad, S. Helali, H. H. Chou and A. F. M. EL-Mahdy, Microporous Mesoporous Mater., 2024, 363, 112824.
54. F. Jiang, Y. Wang, T. Qiu, G. Yang, C. Yang, J. Huang, Z. Fang and J. Li, J. Power Sources, 2022, 523, 231041.
55. N. M. T. Tran, V. Lazeran, J. Santos-Pena and N. Berton, J. Mater. Chem. A, 2025, 13, 24933–24947.
56. R. Li, Z. Shen, H. Zheng, L. Jin, Y. Zhang, W. Yuan and X. Wang, J. Energy Storage, 2023, 73, 109521.
57. X. Wu, C. Zhong, L. Y. Zhang, J. Lu, Q. He, Q. Zhang, W. Yuan and C. M. Li, J. Mater. Chem. A, 2025, 13, 25082–25092.
58. M. Devi, P. Pandey and S. Sharma, J. Appl. Phys., 2025, 137, 025104.
59. N. T. Luu, R. Anbazhagan, A. S. Ivanov, S. E. Lin, C. H. Chang, C. Y. Hung, I. Popovs, J. C. Lee, T. H. Chen and W. Kaveevivitchai, Chem. Eng. J., 2025, 522, 167782.
60. T. A. Gaber, L. R. Ahmed and A. F. M. EL-Mahdy, J. Mater. Chem. A, 2023, 11, 19408–19417.
61. Y. Z. Xiao, A. A. Mohammed, S. W. Kuo and A. F. M. EL-Mahdy, Sep. Purif. Technol., 2025, 356, 129950.
62. S. Gokul Eswaran, M. Rashad, A. Santhana Krishna Kumar and A. F. M. EL-Mahdy, Chem.–Asian. J., 2025, 20, e202401181.
63. B. Nagappan, K. Mehar, S. Prashanth, D. R. Vaghela, S. Samantaray, C. P. Sahu, A. Smerat and K. K. Priya, Carbon Trends, 2025, 20, 100584.
64. N. Kostoglou, P. Andreou, G.-M. Christodoulou, V. Terziyska, A. Hofer-Roblyek, S. Hinder, M. Baker, G. Constantinides, K. Kostoglou and C. Rebholz, Mater, Chem. Phys., 2025, 348, 131607.
65. C. Li, J. Yang, P. Pachfule, S. Li, M. Y. Ye, J. Schmidt and A. Thomas, Nat. Commun., 2020, 11, 4712.
66. S. Abdelnaser, S. W. Kuo and A. F. M. EL-Mahdy, J. Power Sources, 2025, 635, 236535.
67. S. X. Liao and A. F. M. EL-Mahdy, ACS Appl. Energy Mater., 2025, 8, 3074–3086.
68. X. C. Li, Y. Zhang, C. Y. Wang, Y. Wan, W. Y. Lai, H. Pang and W. Huang, Chem. Sci., 2017, 8, 2959–2965.
69. M. G. Kotp, S. U. Sharma, J. T. Lee, A. F. M. EL-Mahdy and S. W. Kuo, J. Taiwan Inst. Chem. Eng., 2022, 134, 104310.
70. S. K. Pati, D. Patra, S. Muduli, S. Mishra and S. Park, J. Mater. Chem. A, 2024, 12, 21165–21175.
71. Z. Y. Chen, W. C. Chen and S. W. Kuo, Polym. Chem., 2024, 15, 553–564.
72. J. W. Jeon, J. Shin, J. Lee, J. H. Baik, R. Malpass-Evans, N. B. McKeown, T. H. Kim, J. C. Lee, S. K. Kim and B. G. Kim, Appl. Surf. Sci., 2020, 530, 147146.
73. L. Mei, X. Cui, Q. Duan, Y. Li, X. Lv and H. g. Wang, Int. J. Hydrogen Energy, 2020, 45, 22950–22958.
74. M. Zhang, T. Zhao, J. Dou, Z. Xu, W. Zhang, X. Chen, X. Wang and B. Zhou, ChemElectroChem, 2019, 6, 5946–5950.
75. W. Liu, M. Ulaganathan, I. Abdelwahab, X. Luo, Z. Chen, S. J. R. Tan, X. Wang, Y. Liu, D. Geng, Y. Bao, J. Chen and k. P. Loh, ACS Nano, 2018, 12, 852–860.
76. A. F. Saber, S. Abdelnaser, A. F. M. EL-Mahdy and S. W. Kuo, Electrochim. Acta, 2025, 511, 145397.
77. L. Zhang, X. Li, L. Li, X. Cheng, H. Wu and J. Zheng, J. Energy Storage, 2024, 96, 112680.
78. Y. Li, Y. Zhang, Q. Zhang, X. Shen, Y. Li, J. Wang, R. Gao and R. Li, J. Energy Storage, 2024, 102, 114065.
79. Y. Li, X. Zou, S. Li, Y. Chen, G. Wang, H. Yang and H. Tian, J. Mater. Chem. A, 2024, 12, 18324–18337.
80. C. Ma, G. Song, Z. Li, H. Wu, C. Wang, Y. Wang, X. Zhang, Y. Song and J. Shi, J. Energy Storage, 2024, 88, 111465.
81. A. Liu, L. Yan, Y. Zhang, R. Ma, N. Guo, L. Wang, B. Zhang, D. Jia and R. Sheng, Surf. Interfaces, 2024, 44, 103754.
82. R. Zhang, H. Liu, Z. Cui, Y. Zhang and Y. Wang, Fuel, 2024, 364, 131062.
83. H. Li, T. Du, Q. Wang, J. Guo, S. Zhang and Y. Lu, J. Energy Storage, 2023, 66, 107397.
84. S. Wu, X. Yan, X. Sun, S. Tian, J. Wang, C. Liu, S. Sun, L. Wu, X. Zhao and Q. Yang, J. Energy Storage, 2023, 71, 108152.
85. G. Zhang, Y. Zhang, J. Wang, J. Yu, K. Wang, G. Li and T. Guan, J. Colloid Interface Sci., 2024, 660, 478–489.
86. S. A. Al Kiey and H. N. Abdelhamid, J. Energy Storage, 2022, 55, 105449.
87. W. Yue, Z. Yu, X. Zhang, H. Liu, Y. Zhang and X. Ma, J. Anal. Appl. Pyrolysis, 2024, 179, 106525.
88. G. Nabi, A. Siddiqa, M. Tanveer, A. Nadeem and K. S. Ahmad, Inorg. Chem. Commun., 2024, 168, 112850.
89. Y. Cai, G. Zhao, Q. Yuan and J. Zhao, Chem. Eng. J., 2024, 484, 149491.
90. S. Ghosh, Z. Qi, H. Wang, S. K. Martha and V. G. Pol, Electrochim. Acta, 2021, 383, 138339.
91. N. A. M. Barakat, A. Khalid, T. Melegy and M. M. Abdel-Aty, Mater. Adv., 2025, 6, 9024–9042.
92. M. N. Khan, M. Shoeb, F. Mashkoor, K. J. Kim, M. S. Kang, J. Yu, H. S. Bae, I. s. Jang and C. Jeong, J. Energy Storage, 2025, 127, 117108.
93. A. M. Abdelrahim, M. G. A. El-Moghny, H. H. Abdelhady, H. S. Wali, M. M. Gamil, S. R. Fahmy, T. M. Abdel-Hamid, G. K. Mohammed, Y. A. Ahmed and M. S. El-Deab, Sci. Rep., 2024, 14, 16807.
94. R. Raavi, S. Archana, P. A. Reddy and P. Elumalai, Energy Adv., 2023, 2, 385–397.
95. S. C. Abbas, Z. Hua, Q. Deng, M. S. Ahommed, J. Guo, H. Huang, X. Ma, S. Cao, H. Huang, X. Ma, S. Cao and Y. Ni, EcoMat, 2024, 6, e12434.
96. J. Zheng, T. Cao, B. Ding, X. Zhang, H. Wu and X. Li, J. Energy Storage, 2025, 110, 115297.
97. A. A. Wavhal, S. Dilwale, P. P. Puthiyaveetil, N. R. Kakde, M. Prakash, S. Kurungot and S. K. Asha, J. Phys. Chem. C, 2024, 128, 12808–12821.
98. D. Patra, S. K. Pati, S. Muduli, S. Mishra and S. Park, J. Chem. Eng., 2024, 482, 149162.
99. D. T. Bakhoum, S. Sarr, V. M. Maphiri, N. F. Sylla, N. M. Ndiaye, M. Diop, B. D. Ngom, M. Chaker and N. Manyala, J. Energy Storage, 2024, 80, 110353.
100. A. F. Saber, Y. F. Chen, L. Mabuti, S. V. Chaganti, S. U. Sharma, J. Lüder, J. T. Lee, S. W. Kuo and A. F. M. EL-Mahdy, Mater. Adv., 2025, 6, 607–616.

---