Modulation of electron distribution and intermediate adsorption by C–O–Si sites for efficient oxygen reduction and lithium storage

Abstract

Biowaste-derived heteroatom-doped porous carbons have garnered substantial attention as nonmetal catalysts for electrocatalysis. However, various heteroatom species and/or diverse coordination states within the carbon framework complicate the understanding of their enhanced catalytic activity. This study employs hydroxyl-rich biomass, specifically banana peels, and photovoltaic waste SiCl₄ as precursors to synthesize porous carbon featuring highly uniform C–O–Si bonds. This material facilitates the elucidation of mechanisms underlying the improved oxygen reduction activity of C–O–Si active sites. Due to Si's substantially lower electronegativity (1.90) compared to that of O (3.44), Si atoms exhibit lower electron density and possess Lewis basicity. This characteristic allows Si to function as a novel active site, enhancing the oxygen reduction activity. The resulting Si–BP–Carbon catalyst demonstrates a low half-wave potential of 0.813 V (vs. RHE) and achieves a diffusion-limited current density of 4.69 mA cm⁻², coupled with good tolerance to methanol crossover and high selectivity toward the 4-electron reaction pathway. Density functional theory calculations corroborate that Si atoms can lose electrons in C–O–Si bonds, improving the adsorption of O₂ and *OOH intermediates. The abundance of C–O–Si bonds makes Si–BP–Carbon highly suitable as an anode material for lithium-ion batteries due to its enhanced capacity for lithium-ion (Li⁺) adsorption facilitated by Si heteroatoms.

---

Green foundation

1. In this study, using biomass waste (banana peels) and photovoltaic waste (SiCl₄) as raw materials, porous Si-doped carbons with abundant C–O–Si bonds were successfully synthesized. This innovative and sustainable strategy can regenerate waste resources to fabricate porous carbons utilized in green energy storage and conversion applications.

2. Using agricultural and industrial waste as raw materials can significantly reduce the environmental impact and economic cost during the fabrication stage. The as-prepared carbons demonstrate high ORR activity and excellent performance as lithium-ion battery anodes, thereby contributing to reduced energy consumption and carbon emissions during the application phase. Carbon-based materials are easier to recycle, further reducing the risk of pollution during the disposal stage.

3. Future research could explore other agricultural or industrial wastes to diversify raw sources and dopant elements to further promote the electrochemical performance and minimize the environmental impact.

Introduction

The oxygen reduction reaction (ORR) is critical in renewable energy conversion technologies. It is essential for electrochemical devices such as metal–air batteries and fuel cells, thereby driving sustainable energy development.¹ The intrinsic sluggish kinetics of the ORR can be addressed using high-loading Pt-based catalysts; however, their cost, limited availability, and poor durability pose challenges for their large-scale applications.² Consequently, extensive research has focused on developing alternative, cost-effective ORR catalysts to replace Pt-based materials in energy conversion and storage applications.³ Carbon materials have emerged as promising candidates due to their good electrical conductivity, tunable pore structure, surface chemistries, and stability, effectively addressing practical challenges.⁴ Biomass, a renewable precursor for advanced carbon materials, offers environmental benefits and is widely available.⁵ However, conventional carbon materials often exhibit inadequate electrocatalytic activity because their electronic structure is not conducive to efficient oxygen molecule adsorption and electron transfer. This limitation hinders their ability to meet the growing demands of markets. Consequently, current research predominantly focuses on strategically enhancing biomass-derived carbon materials through rational modification strategies.

In recent years, heteroatom-doped carbon materials have emerged as promising candidates for various applications, incorporating elements such as N,⁶ S,⁷ P,⁸ B,⁹ and halogens.¹⁰ These heteroatoms, each with distinct electronegativities, play a crucial role in altering the charge distribution of neighboring carbon atoms, generating more reactive sites and enhancing oxygen reduction activity. For instance, Nakamura et al.¹¹ designed highly oriented pyrolytic graphite catalysts with well-defined π-conjugation and controllable N doping to explore potential competitive sites and assess reaction mechanisms. Their carbon dioxide adsorption experiments revealed that the active sites responsible for oxygen reduction in nitrogen-doped carbons are not located at the pyridine nitrogen atoms but neighboring carbon atoms exhibiting Lewis base characteristics. Furthermore, Yang et al.¹² employed density functional theory (DFT) to analyze the most stable configuration of P-doped graphene concerning various adsorbates and reaction intermediates. Their findings highlighted that phosphorus atoms induce surface charge redistribution due to their lower electronegativity than that of carbon atoms, acting as pivotal active sites in the ORR. The charge of phosphorus atoms can be transferred to neighboring carbon atoms, enabling the electropositive phosphorus atoms to function as catalytically active centers. Table S1† summarizes the critical role of electronegativity differences between heteroatoms and carbon atoms in generating new active sites. When atoms with high electronegativity, such as N and F, are doped into carbon structures, the resulting electron loss from carbon atoms generates positively charged active sites. These sites facilitate enhanced oxygen adsorption and contribute to improved ORR activity. Conversely, heteroatoms such as B and P with low electronegativity lose electrons to become new active sites for the ORR. Therefore, modulating charge distribution by utilizing electronegativity differences between bonded heteroatoms and host atoms effectively enhances ORR activity. Oxygen naturally present in carbon materials has received limited attention for its potential to enhance the ORR activity. Carbonaceous materials such as biomass are rich in oxygen, which has an electronegativity of 3.44. In contrast, Si, a congener of carbon, has a substantially lower electronegativity of 1.90. Suppose that oxygen atoms are skillfully made to react with less electronegative Si atoms in the carbon skeleton. In this case, Si will lose more electrons and become strongly Lewis basic due to the substantial electronegativity difference. This will generate new active sites that enhance the adsorption of electronegative oxygen, increasing the ORR activity.

An Si-doped carbon material (Si–BP–Carbon) with abundant C–O–Si bonds was synthesized herein through thermal transformation and in situ doping using biomass waste (banana peel) and photovoltaic waste (SiCl₄). The material exhibited excellent ORR activity and lithium storage capability. DFT calculations confirmed that the formation of a C–O–Si bond resulted in Si transferring 3.27 e to O, making Si strongly Lewis basic and enhancing the adsorption of electronegative oxygen. Additionally, the C–O–Si bond introduced by Si doping imparted a superhydrophilic nature to the material,¹³ enabling better access of gas-phase O₂ molecules to the surface active sites, facilitating the ORR.¹⁴ As a result, the Si–BP–Carbon catalyst exhibited a half-wave potential of 0.813 V (vs. RHE) under alkaline conditions, marking a 0.155 V increase compared to BP–Carbon without Si doping. The diffusion-limited current density also reached 4.69 mA cm⁻², which was 1.5 times higher than that of the BP–Carbon catalyst. Experimental and theoretical calculations jointly demonstrated that Si doping improved the electron transfer pathway during the ORR, enhancing the selectivity and energy utilization efficiency. Moreover, the catalyst exhibited good electrocatalytic stability and tolerance to methanol crossover effects, positioning it as a promising metal-free catalyst. Furthermore, Si–BP–Carbon demonstrated exceptional performance in Li-ion batteries due to its C–O–Si bonds, which provided additional adsorption sites for Li⁺ storage.¹⁵ Consequently, Si–BP–Carbon achieved a superior specific capacity of 409.4 mA h g⁻¹ even after 100 cycles at 50 mA g⁻¹, alongside excellent rate capability, underscoring the advantages of C–O–Si bond formation for energy storage and conversion applications.

Experimental

Synthetic procedures of Si-doped carbon catalysts

For synthesis, 1.0 g of banana peel powder was dissolved in 3 mL of N-methyl-2-pyrrolidinone (NMP) and stirred for 30 min at room temperature. Subsequently, 1.2 mL of SiCl₄ was added dropwise to the stirred solution. After the reaction, the mixture was transferred into a household microwave oven and microwaved at medium-high heat (700 W) for 5 min to dry the solution. The resulting solid composite was then transferred into a tube furnace, heated under an argon atmosphere at a rate of 5 °C min⁻¹ to 900 °C, and maintained at this temperature for 3 h to obtain Si–BP–Carbon. BP–Carbon was also prepared for comparison using the same procedure but without adding SiCl₄.

Detailed methods of sample characterization, ORR performance testing, theoretical calculations, and lithium-ion battery performance testing are provided in the ESI.†

Results and discussion

Catalyst structure and morphology

As depicted in Fig. 1a and d, both the carbonized composites exhibit irregular bulk structures with sizes of around hundreds of nanometers. Compared to the dense and smooth surface of undoped BP–Carbon, the porous and interconnected morphology of Si–BP–Carbon reveals substantial structural changes following Si doping. The porous nature of Si–BP–Carbon likely arises from changes in the coefficient of thermal expansion and continuous release of gases during high-temperature treatment, leading to inhomogeneous expansion. Transmission electron microscopy (TEM) images further confirm the irregular bulk structures of the BP–Carbon and Si–BP–Carbon catalysts (Fig. 1b and e), illustrating the transformation into a porous structure after Si doping. This pore structure optimizes gas and electrolyte transport channels, shortening the charge diffusion path within the carbon material and enhancing the ion diffusion efficiency.¹⁶Fig. 1c and f show the successful transformation of banana peels into typical amorphous carbon during pyrolysis.¹⁷ Amorphous carbon exhibits a wide range of physical properties, including high electrical conductivity and flexibility,¹⁸ which enhance the electronic conductivity of the catalyst and buffer stress and volume changes during the ORR, improving the cycle life and stability of the catalyst. Additionally, the interlayer spacing of Si–BP–Carbon is 0.426 nm, which is much larger than that of BP–Carbon (0.400 nm; Fig. 1c, f and S1†), which facilitates rapid charge transfer within electrodes. Furthermore, EDS mappings reveal the uniform dispersion of C, O, and Si in Si–BP–Carbon (Fig. 1g), confirming the successful doping and uniform dispersion of Si species. The SiCl₄ additive reacts with abundant hydroxyl groups in the biomass (banana peels) to form stable C–O–Si chemical bonds, effectively preventing Si-species agglomeration during carbonization and ensuring their uniform dispersion.

Fig. 1 (a) Scanning electron microscopy (SEM) and (b and c) TEM images of BP–Carbon; (d) SEM and (e and f) TEM images of Si–BP–Carbon; and (g) SEM image and the corresponding energy-dispersive spectroscopy (EDS) mappings of Si–BP–Carbon.

The crystal structure and crystallographic information were analyzed using X-ray diffraction (XRD). BP–Carbon showed two distinct and broad peaks at 23.77° and 43.30° (Fig. 2a), corresponding to the (002) and (100) planes, indicative of low crystallinity and predominantly amorphous carbon with small crystals.¹⁹ Upon doping Si in the carbon matrix, the (002) peak of Si–BP–Carbon shifted to 21.91°—a noticeable negative shift. Simultaneously, the d-spacing of Si–BP–Carbon calculated using Bragg's equation was 0.413 nm, which is much larger than that of BP–Carbon (0.382 nm). XRD and TEM results confirmed that Si effectively expanded the carbon layer spacing due to its larger atomic radius, reducing charge transfer resistance within the composite.^20,21 As shown in Fig. 2b, the Raman spectrum shows a disorder-related D band at ∼1277 cm⁻¹ and a graphitization-related G band centered at ∼1551 cm⁻¹. The I_D/I_G peak intensity ratios for BP–Carbon and Si–BP–Carbon were calculated to be 1.04 and 1.03, respectively. Although Si doping as a heteroatom induces lattice cracking and edge distortion in banana peel-derived carbon materials, the microwave treatment of the precursor induces strong lattice rearrangement. This effect slightly decreases the overall I_D/I_G ratio after doping, thereby enhancing the graphitization of the carbon material.²² Increased graphitization enhances the formation of highly conductive internal networks, improving electron transport kinetics during the ORR.^23,24 Furthermore, Fourier-transform infrared (FT-IR) spectroscopy, as shown in Fig. 2c, revealed characteristic peaks at ∼3418 and 1624 cm⁻¹ for both BP–Carbon and Si–BP–Carbon, corresponding to –OH and C=C/C=O bonds, respectively.²⁵ The 800 cm⁻¹ signal corresponded to the stretching vibration of the C–Si bond.²⁶ Notably, the intense signal centered at 1104 cm⁻¹ for Si–BP–Carbon confirmed the presence of C–O–Si bonds,²⁷ indicating successful doping of silicon species into the carbon framework through stable chemical bonds.

Fig. 2 (a) XRD patterns, (b) Raman spectra, (c) FT-IR spectra, (d) contact angle results, and (e) N₂ adsorption–desorption isotherms of BP–Carbon and Si–BP–Carbon; (f) ²⁹Si solid-state NMR spectrum of Si–BP–Carbon; (g) XPS full spectra of BP–Carbon and Si–BP–Carbon; (h) C 1s and (i) O 1s spectra of BP–Carbon; and (j) Si 2p, (k) C 1s, and (l) O 1s spectra of Si–BP–Carbon.

As shown in Fig. 2d, water contact angle measurements indicate that water droplets quickly disperse in bulk Si–BP–Carbon, reducing the contact angle to 0° within 0.3 s. In contrast, BP–Carbon maintains a contact angle of 58.62° under the same conditions, highlighting that hydrophilicity is enhanced after Si doping. This improved hydrophilicity can be attributed to the highly hydrophilic nature of C–O–Si bonds formed after Si doping¹³ as well as the increased porosity of the carbon skeleton.²⁸ The superhydrophilic porous structure of Si–BP–Carbon provides a large accessible surface area, facilitating rapid infiltration of O₂-rich aqueous electrolytes and O₂ microbubbles, generating a rich three-phase interface that enhances contact between O₂ molecules and catalyst active sites.²⁹ N₂ adsorption–desorption tests (Fig. 2e) revealed that BP–Carbon and Si–BP–Carbon exhibit classical type-IV isotherms, indicating their mesoporous nature. The specific surface areas are calculated to be 34.3 and 114.9 m² g⁻¹ for BP–Carbon and Si–BP–Carbon, respectively. These substantially increases in specific surface area, average pore diameter, and pore volume upon Si doping are consistent with the SEM and TEM results, which show that Si doping alters the thermal expansion coefficient of carbon-based species, promoting structural evolution toward a porous morphology that enhances the availability of ORR catalytic active sites. Solid-state nuclear magnetic resonance (NMR) characterization was employed to analyze the valence bond states of Si species within the bulk carbon. The ²⁹Si NMR spectrum (Fig. 2f) showed two main peaks, Si(OC)₄ and CSi(OC)₃, corresponding to chemical shifts of −109.8 and −65.0 ppm, respectively.^30,31 The predominance of C–O–Si bonds in the carbon skeleton confirmed the formation of stable chemical bonds between the carbon framework and Si additive. X-ray photoelectron spectroscopy (XPS) further validated the presence and valence states of elemental species. The full spectrum (Fig. 2g) displayed peaks of Si 2p and Si 2s at 103 and 154 eV, respectively, indicating a Si doping percentage as high as 11.78 wt%. Fig. 4h and k show that the C 1s spectrum has three peaks located at 284.71, 285.80, and 289.35 eV, corresponding to the C=C, C–O, and C=O bonds,³² respectively. The Si 2p peaks of Si–BP–Carbon exhibited distinct signals at 103.29 and 104.03 eV (Fig. 2j), attributed to CSi(OC)₃ and Si(OC)₄ bonds,³³ respectively, confirming successful Si doping predominantly as C–O–Si bonds, which is consistent with the FT-IR and ²⁹Si NMR findings. Since the electronegativity difference between Si and O is 1.54, which exceeds the range of typical non-polar bonds, the Si–O bond exhibits characteristics of a partial ionic bond.^34,35 This indicates that the bonding electron cloud is significantly shifted toward the O atom, causing the O atom to carry a partial negative charge and the Si atom a partial positive charge, making the Si site strongly Lewis basic in nature. This characteristic optimizes charge distribution within the carbon network, enhancing oxygen adsorption by Si atoms and facilitating charge transfer to active sites, which in turn enhances the ORR activity. The O 1s spectrum of BP–Carbon (Fig. 2i) showed peaks at 531.37 and 532.78 eV corresponding to C=O and C–O bonds, respectively.³⁶ Meanwhile, the O 1s spectrum of Si–BP–Carbon (Fig. 2l) showed peaks at 530.93 and 532.84 eV assigned to C=O and C–O/Si–O bonds, respectively.^36,37 A comparison of O 1s peaks between BP–Carbon and Si–BP–Carbon revealed a substantial reduction in C=O bonds after Si doping. The C=O bonds are typically generated from C–O bonds during biomass carbon precursor carbonization, and their content is notably diminished in Si–BP–Carbon due to the formation of stable C–O–Si bonds. This finding further confirms the abundance of C–O–Si bonds and their exceptional stability, even under high-temperature conditions (900 °C) during carbonization.

ORR performance characterization

A preliminary assessment of the ORR activities of BP–Carbon and Si–BP–Carbon was conducted by examining cyclic voltammetry (CV) curves in O₂-saturated (solid line) and N₂-saturated (dashed line) 0.1 M KOH solutions. As depicted in Fig. 3a, the current response curve of BP–Carbon showed no notable difference for O₂-saturated and N₂-saturated electrolytes, while that of Si–BP–Carbon exhibited a more pronounced current peak in the O₂-saturated solution compared to the N₂-saturated solution. Therefore, it is evident from the CV curves that Si–BP–Carbon demonstrates substantially higher ORR activity than that of BP–Carbon, indicating that Si doping effectively enhances the ORR activity of carbon-based composites. To evaluate the change in active sites and area after Si doping, the double-layer capacitance (C_dl) (Fig. S2†) was used to estimate the electrochemical active surface area (ECSA) of the carbons. As shown in Fig. 3b, the C_dl value of Si–BP–Carbon is 46.7 mF cm⁻², which is significantly higher than that of BP–Carbon (35.7 mF cm⁻²). This indicates that Si doping provides a larger active surface area and more active sites. To further validate the ORR activity of the prepared catalysts, linear sweep voltammetry (LSV) tests were performed using a three-electrode electrochemical device. The LSV curves of BP–Carbon and Si–BP–Carbon showed an increasing trend of current density with increasing rotational speed (Fig. S3†), primarily due to improved dissolved-oxygen mass transfer efficiency at higher rotational speeds.³⁸Fig. 3c shows the LSV curves of BP–Carbon and Si–BP–Carbon at 1600 rpm, indicating their ORR activities. Si–BP–Carbon exhibits excellent catalytic activity with a half-wave potential of 0.813 V, which is substantially higher than that of BP–Carbon (0.658 V). Thus, the half-wave potential of the composite is enhanced by 155 mV after Si doping. Additionally, the diffusion-limited current density of Si–BP–Carbon is 4.69 mA cm⁻², which is notably higher than that of BP–Carbon (3.10 mA cm⁻²). The ORR activity of Si-doped carbon materials compares favorably or even surpasses those of other reported heteroatom-doped catalysts (Table S2†). The introduction of Si species substantially increases the half-wave potentials and diffusion-limited current densities of biomass-derived carbon materials. This enhancement is primarily attributed to Si doping, which generates new active sites and increases the interlayer spacing and pores, underscoring the advantages of Si doping in enhancing the intrinsic catalytic ORR activity of carbon materials.

Fig. 3 (a) CV curves of BP–Carbon and Si–BP–Carbon in O₂-saturated (solid line) and N₂-saturated (dashed line) 0.1 M KOH solutions; (b) C_dl plots of BP–Carbon and Si–BP–Carbon; (c) LSV curves at 1600 rpm; (d) RRDE voltammograms and (e) percentage of H₂O₂ produced and electron transfer number of BP–Carbon and Si–BP–Carbon; (f) Tafel slopes; (g) EIS curves of BP–Carbon and Si–BP–Carbon in O₂-saturated 0.1 M KOH at 1600 rpm; (h) the tolerance to methanol crossover effect of BP–Carbon, Si–BP–Carbon and commercial 20% Pt/C; and (i) chronoamperometric responses of BP–Carbon, Si–BP–Carbon and commercial 20% Pt/C in O₂-saturated 0.1 M KOH solution at 625 rpm.

To further validate catalyst reaction pathways, the hydrogen-peroxide current responses of the composites were evaluated via rotating ring-disk electrode (RRDE) tests. As shown in Fig. 3d, the H₂O₂ response current of Si–BP–Carbon is substantially lower than that of BP–Carbon, demonstrating the effective reduction of side reactions generating H₂O₂ due to Si doping. Additionally, Si–BP–Carbon shows a much lower H₂O₂ yield percentage (Fig. 3e), indicating that Si doping improves energy efficiency for complete oxygen reduction. Furthermore, the electron transfer number of BP–Carbon is calculated to be ∼3.0 (Fig. 3e), consistent with the existing results, indicating that oxygen-containing functional groups—especially the C=O bond—act as the main active sites for H₂O₂ generation.^39,40 Given the high C=O bond content in BP–Carbon as per the XPS analysis, its ORR pathways are thought to involve a complex reaction process with 4-electron and 2-electron transfer reactions. In contrast, the Si–BP–Carbon catalyst shows an integrated electron transfer number of ∼3.8, aligning with the 4-electron transfer mechanism. The improved reaction selectivity is likely due to the introduction of Si, which reduces the C=O bond content and introduces new Si active sites within the carbon skeleton, enhancing the catalytic ORR efficiency of the Si-doped biomass-based carbon catalyst. To further confirm the reaction kinetics, the electron transfer numbers of BP–Carbon and Si–BP–Carbon were calculated from the LSV curves at speeds of 400–2500 rpm using the Koutecky–Levich (K–L) equation (Fig. S4†). The electron transfer number obtained from the K–L equation generally agrees with the RRDE results, which is ∼3.9 for Si–BP–Carbon, and the catalytic ORR pathway is consistent with a 4-electron transfer mechanism. The electron transfer number of BP–Carbon is ∼2.9, further verifying that its reaction path is not limited to the traditional 4-electron transfer mechanism but represents a more complex electron transfer process with poor selectivity. Si doping improves the selectivity and energy efficiency of the catalyst in the ORR. Additionally, Tafel curves derived from the polarization curves indicate that Si–BP–Carbon exhibits a Tafel slope as low as 66.9 mV dec⁻¹ (Fig. 3f), demonstrating faster ORR reaction kinetics compared to those of BP–Carbon (152.4 mV dec⁻¹). Meanwhile, as shown in the electrochemical impedance spectroscopy results in Fig. 3g, BP–Carbon and Si–BP–Carbon catalysts show similar semicircular diameters, indicating that Si doping does not change the charge transfer efficiency of the catalysts.

Alcohol fuel cells demand high electrocatalytic activity and robust methanol crossover tolerance. In the methanol resistance test of BP–Carbon, Si–BP–Carbon, and commercial Pt/C in an O₂-saturated 0.1 M KOH electrolyte, a distinct difference was observed (Fig. 3h). Upon adding 1 mL of methanol to the solutions at 1500 s, BP–Carbon and Si–BP–Carbon exhibited stable current densities. In contrast, commercial Pt/C showed substantial current density fluctuations. This demonstrates that BP–Carbon and Si–BP–Carbon exhibit notable resistance to methanol poisoning, which is crucial for the long-term stability and practical applications of alcohol fuel cells. In addition, reaction stability is crucial in assessing fuel cell durability. Therefore, stability tests using chronoamperometry were conducted to compare the endurance of the prepared and commercial catalysts for the ORR. As depicted in Fig. 3i, the Si–BP–Carbon catalyst maintained 95.10% of its initial response current density after 30 000 s of continuous testing. In contrast, the commercial 20% Pt/C catalyst retained only 78.40% of its initial value over the same period. In addition, after 5000 cycles of accelerated degradation testing (ADT), the Si–BP–Carbon showed an active loss of 16 mV (Fig. S5†). XPS analysis of Si–BP–Carbon after the durability test (Fig. S6†) revealed that the peak positions and relative contents of CSi(OC)₃ and Si(OC)₄ in Si 2p remained almost unchanged, confirming the stability of the active site Si atoms in the catalyst. In addition, the durability-tested Si–BP–Carbon maintains an intact skeleton and pore structure, with no surface cracks (Fig. S7†). This underscores the ability of Si–BP–Carbon to maintain efficient catalytic activity over an extended period of time, potentially making it suitable for practical applications.

Reaction mechanism

Herein, the electronic structure modulation of active sites using Si species during the ORR was investigated via DFT calculations. Effective adsorption of oxygen molecules and oxygen-containing intermediates on surface active sites is crucial for achieving high oxygen reduction activity. Doping heteroatoms allows for the regulation of electronic distribution within carbon networks. Due to differences in electronegativity, atoms that typically lose electrons to Lewis bases become active sites for the ORR.^11,41 The trends of electron gains and losses for catalyst atoms were determined, as shown in Fig. 4a. In BP–Carbon, C atoms exhibited more loss of electrons to linked O atoms; therefore, a carbon atom losing 0.58 e in BP–Carbon was selected as the active site for calculations. In contrast, in Si–BP–Carbon, due to the substantial electronegativity difference between O (3.44) and Si (1.90), Si atoms lose 3.27 e in the Si(OC)₄ bond. In comparison, C atoms show lower loss of electrons of only 0.63 e. Thus, the greater Lewis basicity of the Si atom compared to that of the C atom makes it suitable as an active site in Si–BP–Carbon. More detailed charge transfer data can be found in Fig. S8.† To assess adsorption energies, those for oxygen molecules and oxygen-containing intermediates were calculated before and after Si doping. As shown in Fig. 4b and S9,† the adsorption energy of Si–BP–Carbon for oxygen molecules was −1.59 eV, substantially higher than that of BP–Carbon (−0.47 eV), highlighting the critical role of Si doping in achieving a high ORR activity. The selectivity of 4-electron and 2-electron ORR pathways depends on competing reactions of *OOH on the catalyst surface (*OOH + e⁻ → *O + OH⁻ and *OOH + e⁻ → OOH⁻).⁴² The 4-electron pathway requires high *OOH-adsorption energy for cleavage of the O–O bond in the *OOH intermediate.⁴³ Calculations of the *OOH-adsorption energy before and after Si doping revealed that the adsorption energy of Si–BP–Carbon was −2.59 eV, while that of BP–Carbon was −1.26 eV, indicating that the catalytic ORR pathway catalyzed by Si–BP–Carbon is more in accordance with the 4-electron transfer mechanism. According to the Bader charge analysis (Fig. 4c and S10†), after the adsorption of ORR intermediates onto the active sites of Si–BP–Carbon and BP–Carbon, electrons could transfer from the catalyst surface to reaction intermediates. Thus, *OOH, *O, and *OH intermediates gained 0.80, 1.27, and 0.78 e, respectively, from Si–BP–Carbon and 0.50, 0.80, and 0.49 e, respectively, from BP–Carbon, further confirming enhanced adsorption of oxygen-containing intermediates after Si doping. In summary, the intermediates gained more electrons from Si–BP–Carbon than from BP–Carbon, indicating better charge modulation and more rapid electron transfer during reactions catalyzed by Si–BP–Carbon, further improving the ORR activity.

Fig. 4 (a) Charge density difference diagrams for BP–Carbon and Si–BP–Carbon; (b) optimized configurations and adsorption energies of BP–Carbon and Si–BP–Carbon for O₂ molecules and intermediates *OOH, *O, and *OH; (c) charge density difference diagrams (purple: electron accumulation; blue: electron depletion) for *OOH, *O, and *OH adsorption on BP–Carbon and Si–BP–Carbon and the corresponding Bader charge transfer numbers; (d) Gibbs free energy profiles for the 2-electron ORR in the presence of BP–Carbon and Si–BP–Carbon for the standard electrode potential (U = 0 V) and equilibrium potential (U = 0.7 V); Gibbs free energy profiles for the 4-electron ORR in the presence of BP–Carbon and Si–BP–Carbon for the (e) standard electrode potential (U = 0 V); (f) equilibrium potential (U = 1.23 V); and (g) limiting potential.

The standard hydrogen electrode model was used to investigate the effect of electrode potential (U) on the free energy of each step. As shown in Fig. 4d and e, at U = 0 V (vs. RHE), the 4-electron transfer process is gradually thermodynamically decreasing for both the catalysts, implying an exothermic process that favors reactions.⁴⁴ BP–Carbon remains thermodynamically downward in the 2-electron transfer process, indicating the coexistence of 4-electron and 2-electron transfer processes during the ORR. However, after Si doping, the reaction step (*OOH + e⁻ → OOH⁻ (H₂O₂)) in the presence of Si–BP–Carbon becomes thermodynamically less favorable, demonstrating that the catalyst is not conducive to H₂O₂ generation and is more suitable for the 4-electron transfer pathway, consistent with the experimental results. As shown in Fig. 4d and f, at U = 1.23 V (vs. RHE), in the 4-electron transfer pathway, some steps in the free energy change are uphill, indicating that energy is required to overcome the positive change in free energy. The rate-determining steps (RDS) for BP-Carbon is the first step of the ORR of O₂ molecules into *OOH, which is also consistent with the adsorption energy calculations. Due to the low adsorption energy of oxygen molecules and *OOH on the catalyst, the ORR of O₂ into *OOH is the primary limiting step for BP–Carbon. After doping with Si species, the change in the RDS step is primarily due to the excessive adsorption energy of the *OH intermediate on Si–BP–Carbon, which limits the resolution of the intermediate and results in the RDS limiting the catalyzed ORR. However, the overall energy barrier of the RDS in the presence of Si–BP–Carbon is much lower than that in the presence of BP–Carbon, further confirming that Si doping favors catalytic oxygen evolution. In the 2-electron transfer pathway, when U = 0.70 V (vs. RHE), the formation of *OOH from BP–Carbon is an endothermic reaction while the formation of H₂O₂ is exothermic, further illustrating its poor selectivity. The applied potential that makes all ORR steps exothermic is the limiting potential; typically, higher limiting potentials lead to lower overpotentials. Fig. 4g shows the Gibbs free energy plots at the applied limiting potential. The limiting potential of 0.6575 V for Si–BP–Carbon is substantially higher than 0.2575 V for BP–Carbon. The overpotential values are 2.29 and 3.89 V for Si–BP–Carbon and BP–Carbon, respectively. Therefore, the low adsorption of oxygen molecules and *OOH (RDS) on BP–Carbon results in decreased ORR catalytic performance; meanwhile, the adsorption of oxygen intermediates is substantially improved after Si doping, enhancing the ORR catalytic activity.

Lithium-ion battery performance characterization

In addition to being used as catalysts for improving the ORR activity, the carbon-based composites can act as anodes for lithium-ion batteries, broadening their application range. As shown in Fig. S11,† in the first cycle, pronounced reduction peaks between 1.0 and 0.5 V are attributed to the formation of an SEI layer, while peaks at <0.5 V are mainly attributed to the Li+ storage process. In the differential capacitance curves during the second cycle (Fig. 5a), Si–BP–Carbon shows a different lithium insertion peak at 0.258 V, likely related to the reaction of lithium with the C–O–Si bond. Moreover, the shape and plateau of charge–discharge curves are consistent with the CV curves in the voltage range of 0.01–3.0 V (Fig. S12†). After the first cycle, the discharge capacity of BP–Carbon was ∼969.6 mA h g⁻¹, while the discharge capacity rapidly decreased to 281.9 mA h g⁻¹ in the subsequent cycle. The coulombic efficiency was calculated to be only 29.1%, mainly due to the formation of the SEI film. In comparison, the specific discharge capacity values of Si–BP–Carbon in the first and subsequent cycles were 1248.2 and 539.7 mA h g⁻¹, respectively, with a coulombic efficiency of 43.2%, which is substantially higher than that of BP–Carbon. Fig. S13† shows the Nyquist plots and related fitting results before and after Si doping; the charge transfer impedance between the electrolyte and the Si–BP–Carbon electrode was larger than that between the electrolyte and the BP–Carbon electrode. The contribution of various charge storage mechanisms of Si–BP–Carbon at different scan rates is shown in Fig. 5b, where the capacitive contribution tends to increase with increasing scan rate, reaching 76.5% at 1.0 mV s⁻¹. The capacitive mechanism is mainly attributed to the unique porous carbon framework, which provides abundant active sites and Si doping defects, leading to additional storage of Li⁺ ions.

Fig. 5 (a) Differential capacity curves of BP–Carbon and Si–BP–Carbon; (b) capacitance contribution at various scan rates for Si–BP–Carbon; (c) cycling performance of BP–Carbon and Si–BP–Carbon at 50 mA g⁻¹; (d) molecular dynamics (MD) simulation of Li⁺ adsorption on BP–Carbon and Si–BP–Carbon; and (e) rate capability of BP–Carbon and Si–BP–Carbon.

As seen from the long-term cyclic curve (Fig. 5c), Si–BP–Carbon can maintain an ultrahigh specific capacity of up to 409.4 mA h g⁻¹ after 100 cycles. The robust cycling stability may result from the high storage percentage of Li⁺ ions due to the large surface area and high Si content, which is much higher than that of BP–Carbon (233.9 mA h g⁻¹). Table S3† summarizes the anode capacity of heteroatom-doped carbon materials for lithium-ion batteries over the last three years, and the comparison reveals that the Si–BP–Carbon anode shows a moderate advantage in terms of reversible specific capacity. Fig. S14† shows the evaluation of the long-term cycling stability of the Si–BP–Carbon anode, and the specific capacity remained stable after 1000 cycles at a high current density of 1 A g⁻¹. MD simulations of Li⁺ adsorption by the catalysts were performed to investigate the mechanism of enhanced adsorption capacity due to Si doping. As shown in Fig. 5d, the Si–BP–Carbon electrode adsorbed up to 110 Li⁺ ions, while BP–Carbon could only adsorb 91 Li⁺ ions. The excess Li⁺ ions adsorbed by the former are mainly around the Si atoms, demonstrating the enhanced Li⁺-ion affinity of doped Si. In the rate capability evaluation (Fig. 5e), BP–Carbon and Si–BP–Carbon showed a decrease in the specific capacitance during cycling with increasing current densities, which was particularly evident during the first cycle. However, the capacitance gradually stabilized as the number of cycles increased, possibly due to electrolyte decomposition and the formation of a stable SEI film.⁴⁵ Additionally, the specific capacitance was recovered when the current density was restored to 0.1 C, demonstrating structural and cyclic stability. Overall, Si–BP–Carbon showed better lithium storage capacity, primarily due to its moderate Si doping, large surface area, and large d-layer spacing.

Conclusion

Herein, biomass waste (banana peels) and photovoltaic waste (SiCl₄) were used to prepare Si-doped porous carbon (Si–BP–Carbon) through an in situ doping strategy. The introduction of Si heteroatoms modulated the electron distribution, enhancing electrocatalytic activity. Si atoms were doped in the carbon skeleton, generating C–O–Si bonds, which widened the carbon layer spacing, alleviated carbon layer stacking, and endowed the catalyst with superhydrophilic properties. Under alkaline conditions, Si–BP–Carbon exhibited a half-wave potential of 0.813 V (vs. RHE) and a diffusion-limited current density of 4.69 mA cm⁻², with good tolerance to methanol crossover and long-term stability. Si doping substantially enhanced selectivity, making Si–BP–Carbon an efficient metal-free 4-electron ORR catalyst. DFT calculations revealed that the electron cloud on the Si atom was transferred to oxygen atoms, resulting in the formation of Lewis base sites. These new active sites enhanced the adsorption ability of oxygen molecules and intermediates, which was a key factor in improving the ORR activity and selectivity. Additionally, the catalyst showed excellent performance in lithium-ion batteries, achieving a high specific capacity of 409.4 mA h g⁻¹ after 100 cycles at 50 mA g⁻¹, attributed to the strong Li⁺-ion-storage ability attributed to C–O–Si bonds. Thus, the in situ doping of Si heteroatoms and formation of C–O–Si bonds provide a novel theoretical approach for designing new Si-doped carbon materials with modulated electron distribution for use as high-end catalysts and energy storage materials.

Author contributions

Shouhua Yang: conceptualization, data curation, methodology, and writing – original draft; Ying Tang: formal analysis, software, validation, and writing – original draft; Zhen Yang: data curation, formal analysis, investigation, methodology, and writing – original draft; Shengchao Yang: conceptualization, data curation, and writing – review & editing; Boqin Li: resources and visualization; Wencai Peng: methodology and resources; Banghua Peng: investigation and visualization; Gang Wang: methodology and visualization; Jie Liang: conceptualization and methodology; Junyi Ji: conceptualization, funding acquisition, and writing – review & editing; Feng Yu: conceptualization, funding acquisition, supervision, and writing – review & editing.

Data availability

The data that support the findings of this study are available in the main text and the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We appreciate the financial support from the Xinjiang Science and Technology Program (2023TSYCCX0118), the Bingtuan Science and Technology Program (2023CB008-21), and the National Natural Science Foundation of China (22278282).

References

1. W. Zhou, H. Su, W. Cheng, Y. Li, J. Jiang, M. Liu, F. Yu, W. Wang, S. Wei and Q. Liu, Nat. Commun., 2022, 13, 6414.
2. X. Wang, M. Yu and X. Feng, eScience, 2023, 3, 100141.
3. Q. Jing, Z. Mei, X. Sheng, X. Zou, Q. Xu, L. Wang and H. Guo, Adv. Funct. Mater., 2023, 34, 2307002.
4. Y. Yin, Q. Liu, Y. Zhao, T. Chen, J. Wang, L. Gui and C. Lu, Energy Fuels, 2023, 37, 3523–3554.
5. L. Ge, C. Zhao, M. Zuo, J. Tang, W. Ye, X. Wang, Y. Zhang and C. Xu, J. Anal. Appl. Pyrolysis, 2022, 168, 105747.
6. B. Wu, H. Meng, D. M. Morales, F. Zeng, J. Zhu, B. Wang, M. Risch, Z. J. Xu and T. Petit, Adv. Funct. Mater., 2022, 32, 2204137.
7. H. Yang, H. Wu, Y. Tian, G. Wei, L. Yao, X. Liu, F. Xu, P. Cai, H. Zhang, X. Yu, H. Peng and L. Sun, Mater. Today Sustainability, 2023, 24, 100506.
8. L. Song, H. Fan, X. Fan, H. Gong, T. Wang and J. He, Chem. Eng. J., 2022, 435, 134612.
9. M. Fan, Q. Yuan, Y. Zhao, Z. Wang, A. Wang, Y. Liu, K. Sun, J. Wu, L. Wang and J. Jiang, Adv. Mater., 2022, 34, e2107040.
10. S. G. Peera, R. S. Menon, S. K. Das, A. Alfantazi, K. Karuppasamy, C. Liu and A. K. Sahu, Coord. Chem. Rev., 2024, 500, 215491.
11. D. H. Guo, R. Shibuya, C. Akiba, S. Saji, T. Kondo and J. Nakamura, Science, 2016, 351, 361–365.
12. X. Zhang, Z. Lu, Z. Fu, Y. Tang, D. Ma and Z. Yang, J. Power Sources, 2015, 276, 222–229.
13. S. Y. Park, H.-W. Ahn, J. W. Chung and S.-Y. Kwak, Desalination, 2016, 397, 22–29.
14. W. Yao, A. Hu, J. Ding, N. Wang, Z. Qin, X. Yang, K. Shen, L. Chen and Y. Li, Adv. Mater., 2023, 35, e2301894.
15. K. Li, G. Yuan, X. Liu, Y. Guo, R. Huang, H. Li, H. Zhang, Q. Jia, Z. Xie, S. Zhang and W. Lei, Adv. Funct. Mater., 2023, 33, 2302348.
16. S. F. Huang, Y. Lv, W. Wen, T. Xue, P. Jia, J. Wang, J. J. Zhang and Y. F. Zhao, Mater. Today Energy, 2021, 20, 100673.
17. C. Chen, K. Sun, C. Huang, M. M. Yang, M. M. Fan, A. Wang, G. Y. Zhang, B. Li, J. C. Jiang, W. Xu and J. L. Liu, Biochar, 2023, 5, 51.
18. H. Tian, Z. Yao, Z. Li, J. Guo and L. Liu, ACS Nano, 2023, 17, 24468–24478.
19. H. Chen, Y.-c. Guo, F. Wang, G. Wang, P.-r. Qi, X.-h. Guo, B. Dai and F. Yu, New Carbon Mater., 2017, 32, 592–599.
20. J. Liu, Z. Wei, Z. Gong, M. Yan, Y. Hu, S. Zhao, G. Ye and H. Fei, Appl. Catal., B, 2023, 324, 122267.
21. Y. J. Sa, J. H. Kim and S. H. Joo, Angew. Chem., Int. Ed., 2019, 58, 1100–1105.
22. H. Liu, W. Yang, S. Che, Y. Li, C. Xu, X. Wang, G. Ma, G. Huang and Y. Li, Carbon, 2022, 196, 633–638.
23. D. Geng, Y. Huang, S. Yuan, Y. Jiang, H. Ren, S. Zhang, Z. Liu, J. Feng, T. Wei and Z. Fan, Small, 2023, 19, 2207227.
24. C. C. Yang, S. F. Zai, Y. T. Zhou, L. Du and Q. Jiang, Adv. Funct. Mater., 2019, 29, 1901949.
25. Y. Zhang, X. Guo, Q. Yang, Y. Shao, Y. Du, J. Qi, M. Zhao, Z. Shang, Y. Hao, Y. Tang, Y. Li, R. Zhang, B. Wang and J. Qiu, Proc. Natl. Acad. Sci. U. S. A., 2023, 120, e2314408120.
26. S. Dobos, G. Fogarasi and E. Castellucci, Spectrochim. Acta, Part A, 1972, 28, 877–887.
27. J. Habsuda, G. P. Simon, Y. B. Cheng, D. G. Hewitt, D. A. Lewis and H. Toh, Polymer, 2002, 43, 4123–4136.
28. C. Liang, H. Xia, L. Yin, C. Du, X. Wu, J. Wang, S. Li, J. Xu, X. Zhang, Y. Wang and W. Qu, Chem. Eng. J., 2024, 481, 148375.
29. K. Dong, J. Liang, Y. Wang, Z. Xu, Q. Liu, Y. Luo, T. Li, L. Li, X. Shi, A. M. Asiri, Q. Li, D. Ma and X. Sun, Angew. Chem., Int. Ed., 2021, 60, 10583–10587.
30. Y. Sato, R. Hayami and T. Gunji, J. Sol-Gel Sci. Technol., 2022, 104, 36–52.
31. C. M. Zheng, X. Liu, K. Xie and Y. F. Chen, Ionics, 2020, 26, 3853–3862.
32. Y. Luo, J. Xu, P. Mou, L. Li, Y. Chen, L. Yan, H. Yu, J. Shu and L. Zhang, Small, 2023, 20, 2304981.
33. S. Z. Fan, J. J. Zhang, S. Q. Cui, L. L. Chen, X. M. Chen, J. S. Rong, W. X. Wang, Y. R. Liu and J. T. Zhao, Chem. Eng. J., 2024, 479, 147785.
34. S.-G. Kang, W. Jeong, J. Paeng, H. Kim, E. Lee, G.-S. Park, S. Han, H. N. Han and I.-S. Choi, Mater. Today, 2023, 66, 62–71.
35. S. U. Ahmad, T. Szilvási, E. Irran and S. Inoue, J. Am. Chem. Soc., 2015, 137, 5828–5836.
36. W. Shen, C. Zhang, M. Alomar, Z. Du, Z. Yang, J. Wang, G. Xu, J. Zhang, J. Lv and X. Lu, Nano Res., 2023, 17, 1217–1224.
37. L. Wu, Y. Li, B. Zhou, J. Liu, D. Cheng, S. Guo, K. Xu, C. Yuan, M. Wang, G. J. H. Melvin, J. Ortiz-Medina, S. Ali, T. Yang, Y. A. Kim and Z. Wang, Carbon, 2023, 207, 36–48.
38. L. Zheng, M. Niu, T. Zeng, X. Ge, Y. Wang, C. X. Guo, W. Yuan, D. Cao, L. Y. Zhang and C. M. Li, eScience, 2024, 4, 100187.
39. Z. H. Sun, X. Zhang, X. D. Yang, W. N. Shi, Y. Q. Huang, Y. L. Men, J. Yang and Z. Y. Zhou, Chem. Commun., 2022, 58, 8998–9001.
40. S. Y. Chen, T. Luo, K. J. Chen, Y. Y. Lin, J. W. Fu, K. Liu, C. Cai, Q. Y. Wang, H. J. W. Li, X. Q. Li, J. H. Hu, H. M. Li, M. S. Zhu and M. Liu, Angew. Chem., Int. Ed., 2021, 60, 16607–16614.
41. J. Quílez-Bermejo, E. Morallón and D. Cazorla-Amorós, Carbon, 2020, 165, 434–454.
42. Y. Zheng, X. Xu, J. Chen and Q. Wang, Appl. Catal., B, 2021, 285, 119788.
43. J. Woo, J. S. Lim, J. H. Kim and S. H. Joo, Chem. Commun., 2021, 57, 7350–7361.
44. X. Guo, J. Shi, M. Li, J. Zhang, X. Zheng, Y. Liu, B. Xi, X. An, Z. Duan and Q. Fan, Angew. Chem., 2023, 135, e202314124.
45. L. Chen, Z. Y. Wang, C. N. He, N. Q. Zhao, C. S. Shi, E. Z. Liu and J. J. Li, ACS Appl. Mater. Interfaces, 2013, 5, 9537–9545.

---

Footnotes

† Electronic supplementary information (ESI) available: Supplementary figures, tables, and notes. See DOI: https://doi.org/10.1039/d4gc05650g

‡ These authors contributed equally to this work.

---