Atomic Sn–incorporated subnanopore-rich hard carbon host for highly reversible quasi-metallic Li storage

Abstract

The practical application of Li metal anodes has been hindered by severely irreversible side reactions for low Coulombic efficiency, uncontrollable growth of Li dendrites, and large volume change. Herein, we report subnanopore-rich carbon spheres encapsulated with Sn single atoms (Sn/CS@SC) as a Li host to address these challenges. Owing to the high Li affinity of Sn single atoms, Sn/CS@SC can promote storage of quasi-metallic Li within the inner void space other than direct plating of metallic Li on the outer surface. Moreover, the subnanopores with a strong spatial confinement effect can prevent the penetration of ester electrolyte for reduced side reactions. As expected, the Sn/CS@SC host demonstrates a high Coulombic efficiency of 99.8% over 600 cycles. Moreover, a full cell using a prelithiated Sn/CS@SC anode and LiNi_0.8Co_0.1Mn_0.1O₂ cathode shows high capacity retention (~80%) over 500 cycles at high current density.

Subnanopore-rich carbon spheres decorated with Sn single atoms are designed as hosts for reversible quasi-metallic Li storage.

INTRODUCTION

It is urgent to explore next-generation Li-ion batteries with high energy density to meet the ever-growing needs for electrified transportations and portable devices (1). However, a traditional graphite anode is approaching its theoretical capacity limit of 372 mAh g⁻¹ (2, 3). Metallic Li is a promising anode because of its high theoretical specific capacity (3860 mAh g⁻¹) (4–6). However, the practical utilization of Li metal anode (LMA) is hindered by several intrinsic issues including undesired Li dendrites and huge volume variations (7–10). The low plating/stripping Coulombic efficiency (CE) of LMA is far below the performance goalpost for the commercialization (5, 11). The main pathway for irreversible consumption of metallic Li is attributed to the thermodynamic instability of a Li/organic electrolyte interface through the repetitive formation/decomposition of a solid-electrolyte interphase (SEI) (12, 13). Delicate controls for the nucleation and growth of metallic Li can suppress the dendrite growth and address the volume expansion (14–16). Moreover, physical barrier or passivation treatment is highly necessary to prevent the direct contact between the active LMA surface and the organic electrolyte (8, 17, 18).

Numerous approaches have thus been developed to solve the above challenges, such as electrolyte optimization, construction of artificial SEI layers, and three-dimensional (3D) hosts (15, 19, 20). Among them, a 3D porous carbon–based host is highly attractive to deploy an electric field for Li plating and accommodate volume change for long cycling stability (21, 22). However, most 3D porous carbon–based hosts with abundant micropores/mesopores are accessible to electrolytes, leading to severe underlying side reactions, especially in the commercial ester electrolyte (23–25). Recently, hard carbon (HC) is considered to be a promising Li host to address this issue. Apart from metallic Li, HC can store quasi-metallic Li (QM-Li) in subnanopores because of the confinement effect (21, 24, 26). Although QM-Li has similar features to metallic Li, limited sizes of subnanosized pores within HC can block electrolyte penetration and subsequent side reactions for high CE (21, 24). Despite great potential, HC still needs effective structural design for subnanopores to achieve the high reversible storage of QM-Li. Moreover, carbon hosts without functional groups suffer poor Li affinity, which require surface engineering to reduce overpotential for Li deposits (27). Therefore, it is critical to construct porous HC–based hosts with a well-designed pore structure and ingredient for highly reversible QM-Li storage.

Modulation approaches have been developed for the design of HCs including pretreatment before the carbonization process, heteroelement doping, templating methods, and chemical etching/deposition for pore regulation (28–32). As a typical example, Yang’s group proposed sieving carbon with highly tunable nanopores with tightened pore entrances via chemical vapor deposition of methane on commercial porous carbon (23). Shrunken pores restrain the entrance of electrolyte to impede the formation of SEI inside, indicating the importance of pore structure in HCs. Apart from pore structures, the interfacial/bulk thermodynamics and kinetics for Li plating/stripping on carbon hosts are usually optimized by the introduction of metallic clusters and heteroatoms (27, 33, 34). Specially, incorporation of M-N_x-C species (M refers to metal) with high Li affinity inside the pores of carbon hosts could avoid the formation of metallic Li on the surface to maintain high CE (34, 35). Recently, a self-engaged strategy has attracted increased attention for porous materials, which can control the structure and composition within one design (36, 37).

On the basis of the above considerations, we herein report a combined strategy to achieve both regulation of subnanopores and interfacial chemistry of HC-based host for highly reversible QM-Li storage. Notably, SnO₂ is imported during the precursor preparation as the self-engaged template for pore formation and compositional optimization. To be specific, the consumption of carbonaceous species during consecutive thermal reduction of SnO₂ and soft-carbon (SC) coating processes is executed to create subnanopore-rich HC spheres (CS) encapsulated with Sn single atoms (Sn/CS@SC). Notably, Sn single atoms can store Li⁺ ions through an alloy-type reaction, facilitating the subsequent formation of QM-Li with high CE (38, 39). Moreover, the obtained Sn single atoms demonstrate high binding energy with Li⁺ ions, which can store QM-Li within the inner void space instead of direct Li plating on the surface. Benefitting from the confinement effect of subnanopores, Sn/CS@SC can prevent electrolyte penetration and the accompanying side reactions. In addition, plenty of void space within Sn/CS@SC can alleviate volume change. As expected, the Sn/CS@SC host demonstrates high Li storage capacity over 600 cycles with high CE in the ester electrolyte. In addition, the full cell based on the LiNi_0.8Co_0.1Mn_0.1O₂ (NCM811) cathode and prelithiated Sn/CS@SC anode presents high capacity retention over 500 cycles at 5 C (1 C = 190 mAh g⁻¹).

RESULTS

Formation of Sn/CS@SC

Figure 1A illustrates the multistep preparation of Sn/CS@SC through consecutive thermal reduction and SC coating treatments. Resorcinol formaldehyde spheres (RFSs) with SnO₂ (SnO₂/RFSs) prepared via a modified Stöber strategy are selected as the initial precursors (40). A follow-up carbonization is carried out to convert the SnO₂/RFS precursor into monodispersed Sn atoms within carbon spheres (denoted as Sn/CS; fig. S1). Afterward, a thin SC coating layer is introduced to regulate the micropore/mesopore structure. Owing to the confinement effect, the pore sizes within Sn/CS@SC are narrow enough to prevent the penetration of solvent molecules in the electrolyte. Field-emission scanning electron microscopy (FESEM) and transmission electron microscopy (TEM) images show that the Sn/CS@SC particles are smooth and uniform in size (Fig. 1, B and C). High-resolution TEM (HRTEM) images illustrate that there are no agglomerated particles within Sn/CS@SC and the thickness of the SC coating layer is quite thin with an interlayer distance (d₀₀₂) of 0.348 nm (Fig. 1D and fig. S2). Figure 1E shows the homogeneous distribution of C, N, and Sn elements on Sn/CS@SC. Furthermore, the aberration-corrected high-angle annular dark-field scanning TEM (HAADF-STEM) image (Fig. 1F) and corresponding elemental mappings (fig. S3) demonstrate the high dispersity of Sn dots on the Sn/CS@SC substrate. Inductively coupled plasma optical mass spectrometry indicates that the weight percentage of Sn is determined to be about 1 wt % (table S1). X-ray diffraction (XRD) patterns of Sn/CS@SC, Sn/CS, and CS (Fig. 1G) display typical (002) and (100) planes of carbon without impurities, suggesting similar long-range disordered graphitic structures. Raman spectra (fig. S4) present that the intensity ratio of the D-band to the G-band (I_D/I_G) of Sn/CS@SC (1.27) is lower than that of Sn/CS (1.62), indicating less defects after SC coating.

Fig. 1. Formation of Sn/CS@SC.

(A) Schematic illustration of the synthetic route for Sn/CS@SC. (I) Polymerization. (II) Reduction. (III) Coating. (B) FESEM image, (C) TEM image, and (D) HRTEM image of Sn/CS@SC. (E) HAADF-STEM image and corresponding elemental mappings of Sn/CS@SC. (F) Aberration-corrected HAADF-STEM image of Sn/CS@SC. (G) XRD patterns of different samples. a.u., arbitrary units.

Composition characterizations and theoretical calculation of Sn/CS@SC

To elucidate the chemical states and coordination modes of Sn atoms in Sn/CS@SC, extended x-ray absorption fine structure (EXAFS) and x-ray absorption near-edge structure (XANES) tests are conducted. As demonstrated in Fig. 2A, the Sn K-edge XANES profile of Sn/CS@SC is between Sn foil and SnO₂, which reveals that the oxidation states of Sn species within Sn/CS@SC are between 0 and +4. Fourier-transformed EXAFS (FT-EXAFS) analysis (Fig. 2B) demonstrates that there is no Sn-Sn coordination within Sn/CS@SC, confirming the atomic dispersity of Sn species. Wavelet-transformed EXAFS (WT-EXAFS) data (Fig. 2C) reflect only one intensity maximum at 4.5 Å⁻¹ arising from Sn-N coordination, further supporting the monodispersity of Sn species in Sn/CS@SC. As shown in the quantitative least-squares EXAFS fitting (Fig. 2, D and E), the coordination configuration confirms the Sn-N₄ coordination within Sn/CS@SC (table S2). In addition, x-ray photoelectron spectroscopy (XPS) spectra (fig. S5) and Fourier transform infrared (FTIR) spectrum (fig. S6) of Sn/CS@SC confirm the presence of Sn─N and C─N bonds and the interaction between Sn and N (27, 41). Density functional theory simulations are carried out to study the adsorption energy (E_ad) of Li⁺ ion on different hosts. As shown in Fig. 2F, Sn/CS@SC with Sn-N₄ coordination exhibits much larger E_ad with Li⁺ ions (−5.42 eV) than graphene (−1.37 eV), demonstrating higher Li affinity.

Fig. 2. Composition characterizations and theoretical calculation of Sn/CS@SC.

(A) Sn K-edge XANES spectra of various samples and corresponding (B) FT k²-weighted Sn K-edge EXAFS spectra. (C) WT-EXAFS signals of Sn/CS@SC, SnO₂, and Sn foil. FT-EXAFS fitting curves of Sn/CS@SC at the Sn K-edge in (D) R-space and (E) k-space. (F) Adsorption energy and charge-density models of the Li⁺ ion on graphene and Sn/CS@SC.

Structural characterizations of Sn/CS@SC

We further carry out a series of N₂/CO₂ adsorption-desorption tests to investigate the pore structure of Sn/CS@SC (fig. S7 and table S3). When N₂ is utilized as an adsorbate, the incorporation of SnO₂ and the subsequent thermal reduction expand the pore size from a micropore in CS to a mesopore for Sn/CS (Fig. 3A), whereas SC coating notably reduces the pore entrance diameter of Sn/CS@SC to block N₂ penetration. Therefore, the specific surface area (SSA) of Sn/CS@SC (4.66 m² g⁻¹) is markedly reduced compared with that of Sn/CS. Meanwhile, CO₂ can be absorbed by subnanopores to present a large SSA (182.7 m² g⁻¹) for Sn/CS@SC with a pore distribution from 0.4 to 0.8 nm (Fig. 3B). These results reveal that SC coating successfully converts the mesopores of Sn/CS into subnanopores for Sn/CS@SC. Furthermore, the density values of Sn/CS@SC obtained from He and dimethyl carbonate (DMC) skeletal density analyses are 2.018 and 1.397 g cm⁻³, respectively (table S4). On the basis of the ideal density of graphite (2.26 g cm⁻³), the volume of pores that are inaccessible to DMC within Sn/CS@SC is calculated to be 0.273 cm³ g⁻¹. Meanwhile, the volume of pores that are inaccessible to DMC but accessible to He in Sn/CS@SC is estimated to be 0.22 cm³ g⁻¹. Small-angle x-ray scattering (SAXS) and wide-angle x-ray scattering (WAXS) are used to analyze the pore structure that cannot be penetrated by N₂ and CO₂ gases (Fig. 3C and table S5) (42). Compared to Sn/CS (SSA: 678.6 m² g⁻¹; closed pore body diameter: 0.51 nm), Sn/CS@SC has slightly reduced SSA (540.1 m² g⁻¹) and smaller closed pore body diameter (0.50 nm). On the basis of the above results, it can be deduced that Sn/CS@SC has plenty of subnanopores inaccessible to CO₂ (Fig. 3D). As shown in SAXS patterns (Fig. 3E), the wide peak in the intermediate Q range of Sn/CS disappears after cycles, indicating potential filling of pores by SEI or inactivated Li (23, 42). On the contrary, the broad peak in the intermediate Q range of Sn/CS@SC remains unchanged after cycles, confirming the inaccessibility of electrolyte for reduced side reactions (Fig. 3F).

Fig. 3. Structural characterizations of Sn/CS@SC.

Pore distributions of various samples in (A) N₂ and (B) CO₂ adsorption-desorption tests. (C) Full-range SAXS-WAXS analysis of Sn/CS@SC and Sn/CS. (D) Comparison of SSA deduced from SAXS and CO₂ adsorption-desorption tests. SAXS patterns of (E) Sn/CS and (F) Sn/CS@SC anodes before and after cycling at a current density of 1 mA cm⁻².

Lithiation processes in different hosts

To understand the process of Li storage, we study the voltage profiles of Sn/CS@SC at different Li storage capacities. As shown in Fig. 4A, the potential drops fast upon Li insertion in Sn/CS@SC, while a notable plateau can be found after discharge below 0 V. Different from the normal nucleation process for metallic Li, there is no sharp potential decrease or characteristic “V” shape in the discharge curve of Sn/CS@SC, even at the storage capacity of 700 mAh g⁻¹. Meanwhile, FESEM observations show a slightly morphology change for Sn/CS@SC, indicating potential Li storage within the subnanopores (fig. S8). As control, CS and Sn/CS present obvious Li plating behaviors on the surface below 0 V with pronounced nucleation overpotentials. On the basis of the above data, we deduce that subnanopore-rich Sn/CS@SC could store extra Li below 0 V in the form of QM-Li rather than metallic Li (24). To further verify the superiority of SnO₂ colloid as a self-engaged template, Sn/CS@SC using SnCl₂ as a Sn source (denoted as Sn/CS@SC-SnCl₂) is prepared as control. As shown in the voltage profile, Sn/CS@SC-SnCl₂ demonstrates a pronounced “V” shape at the storage capacity of about 470 mAh g⁻¹ (fig. S9). We deduce that SnCl₂ could form larger pores to exceed the critical value for the nucleation of metallic Li. Ex situ XRD tests are carried out for Sn/CS@SC and Sn/CS@SC-SnCl₂ at different Li storage capacities (fig. S10). The large angular shift of the typical (002) diffraction peak of Sn/CS@SC-SnCl₂ at a Li storage capacity of 700 mAh g⁻¹ suggests the reduced interlayer distance and extruded structure. On the contrary, there is no obvious shift during the lithiation process for Sn/CS@SC, indicating its robust structure. To further explore the specific role of Sn in Li storage, Cu/CS@SC using CuCl₂ as a Cu source is prepared as a control sample (fig. S11). As shown in the voltage profile, Cu/CS@SC exhibits a pronounced “V” shape at the Li storage capacity of about 350 mAh g⁻¹ (fig. S12). This result can be attributed to the poor Li⁺ ion affinity of Cu species for homogeneous nucleation of metallic Li on the surface of Cu/CS@SC (43). Furthermore, CS@SC is prepared as another control sample to investigate the role of Sn single atoms (fig. S13). The voltage profile of Li storage in CS@SC displays an obvious “V” shape at the Li storage capacity of about 300 mAh g⁻¹ (fig. S14). It can be seen that the insufficient inner void space and poor Li⁺ ion affinity of CS@SC lead to faster Li growth on the surface. To get the deep understanding of the components and element valence of Sn/CS@SC and Sn/CS in the lithiated state, XPS tests are carried out for these two samples at an etching depth of 90 nm (Fig. 4B and fig. S15). Notably, a specific characteristic peak of QM-Li located between the Li metal (53.0 eV) and Li-C (55.5 eV) at 54.3 eV can be observed for Sn/CS@SC (21). To further explore the feature of QM-Li, the reaction of the lithiated Sn/CS@SC electrode (storage voltage above and below 0 V) with ethanol is conducted and tested by phenolphthalein solution (fig. S16). The color of product solution turns red for the lithiated Sn/CS@SC electrode below 0 V, revealing the similar feature of QM-Li to that of metallic Li (44). On the basis of above results, we can deduce the Li storage mechanisms in different carbon hosts (Fig. 4C). To be specific, the in situ generation of atomic Sn species effectively creates plenty of subnanopores with high Li⁺ ion affinity in the Sn/CS@SC host. During the lithiation process, Li_xSn species are priorly generated before the formation of LiC_x species in subnanopores. Both Li_xSn and LiC_x belong to QM-Li species. Owing to the spatial confinement effect, the ester-based electrolyte is blocked outside to prevent side reactions. As a comparison, CS without Sn-N modification is prone to inducing the formation of metallic Li on the surface, whereas Sn/CS with large pores can facilitate the nucleation and growth of metallic Li inside.

Fig. 4. Lithiation processes in CS, Sn/CS, and Sn/CS@SC hosts.

(A) Voltage profiles of Li storage in different anodes. (B) XPS spectra of Sn/CS@SC and Sn/CS with a Li storage capacity of 700 mAh g⁻¹ at the etching depth of 90 nm. (C) Schematic diagram of lithiation processes in different hosts.

Li storage behaviors on different hosts

Galvanostatic intermittent titration technique (GITT) tests are carried out to evaluate the Li⁺ diffusion coefficients ($D_{{\text{Li}}^{+}}$) and the Li⁺ storage kinetics in various electrodes (fig. S17). Among them, the corresponding average $D_{{\text{Li}}^{+}}$ of the Sn/CS@SC anode is the highest during discharge and charge processes (figs. S18 and S19). Compared to other porous structures, the small subnanopores in Sn/CS@SC enable faster reaction kinetics (24). To validate the reversibility for the Li storage in Sn/CS@SC during cycling, postmortem XPS examinations are performed. As shown in Fig. 5A, there are no distinguishable changes within the compositions and valences for the surface SEI layer of Sn/CS@SC after 100 cycles, indicating good interface stability. In contrast, more C-F and CO₃²⁻ groups can be found on the surface of Sn/CS, indicating the severe side reactions between the ester-based electrolyte and Li deposits (Fig. 5B) (21). Besides, there are more C-F groups on the surface of Li-CS after 30 cycles, demonstrating an unstable interface without Sn single atoms and the SC layer (fig. S20). Furthermore, HRTEM observations reveal that the SEI layer formed on Li-Sn/CS@SC is thin and smooth after 100 cycles (Fig. 5C). In sharp contrast, SEI on Sn/CS is quite rough, indicating the continuous formation and decomposition (Fig. 5D). Moreover, Nyquist plots verify the stable electrolyte-electrode interface for Sn/CS@SC during cycling, superior to that for Sn/CS with a rapid impedance increase (fig. S21). FESEM images further confirm the structural consistency for Li-Sn/CS@SC after long-term test (Fig. 5E) and remarkable Li deposits on Li-Sn/CS and Li-CS (Fig. 5F and fig. S22). In addition, in situ optical microscopy observations reflect volume evolutions for different samples during cycling. Apparent mossy Li can be seen on Cu foil (Fig. 5G) and Sn/CS (Fig. 5H), whereas Sn/CS@SC well maintains the original shape without dendrites owing to the unique subnanopore-rich structure (Fig. 5I).

Fig. 5. Li storage behaviors on different hosts.

XPS spectra of (A) Sn/CS@SC and (B) Sn/CS after 10 and 100 cycles. HRTEM images of (C) Li-Sn/CS@SC and (D) Li-Sn/CS after 100 cycles. FESEM images of (E) Li-Sn/CS@SC and (F) Li-Sn/CS after 100 cycles. (G to I) Optical microscopy images of half cells after discharging for different times at a current density of 5 mA cm⁻².

Performance of half cells, full cells, and pouch cells

Half cells are used to evaluate the long-term electrochemical cycling stability of the Sn/CS@SC host. As a result, the Sn/CS@SC host demonstrates a reversible Li storage capacity of 700 mAh g⁻¹ over 600 cycles with an average CE of about 99.8% at a current density of 1 mA cm⁻² with a high areal capacity of about 1.4 mAh cm⁻² (Fig. 6A and figs. S23 and S24). This performance is comparable or superior to the recently reported modified strategies for LMAs (table S6) (1, 5, 7, 8, 14, 15, 17, 20, 45, 46). Moreover, the Sn/CS@SC anode also exhibits outstanding cycling performance with an average CE of about 99.2% over 230 cycles at a current density of 3 mA cm⁻² (fig. S25). Even at a higher specific capacity of 800 mAh g⁻¹ (equivalent to 1.6 mAh cm⁻²), Sn/CS@SC still demonstrates an average CE of about 99.0% over 200 cycles (figs. S26 and S27). With the increase in Li storage capacity to 900 and 1000 mAh g⁻¹, Sn/CS@SC hosts present low CE and poor cycling stability, indicating potential side reactions between extra deposited Li and electrolyte (figs. S28 and S29). We further verify the practicability of Sn/CS@SC in symmetric cells and full cells. As expected, the Li-Sn/CS@SC∥Li-Sn/CS@SC symmetric cell displays a flat voltage plateau over 1000 hours at a current density of 1 mA cm⁻² with a high depth-of-discharge value of about 60% (Fig. 6B). Where the current density is increased to 3 mA cm⁻², the Li-Sn/CS@SC∥Li-Sn/CS@SC symmetric cell still maintains stable cycling over 500 hours (fig. S30). Moreover, the Li-Sn/CS@SC electrode with a Li storage capacity of 700 mAh g⁻¹ exhibits better rate performance than those with higher Li storage capacities of 900 and 1000 mAh g⁻¹ (fig. S31). Moreover, a full cell based on the NCM811 cathode and Li-Sn/CS@SC hybrid anode is assembled. The Li-Sn/CS@SC∥NCM811 full cell with an N/P ratio of about 1.4 presents a stable cycling over 500 cycles with a high capacity retention of about 80% at 5 C (Fig. 6C and fig. S32), which outperforms the full cells using the Li-Sn/CS or Li-CS anode. Moreover, this result is close to or better than the recently reported modified strategies for Li∥NMC811 full cells (table S7) (1, 5, 7, 8, 14, 19, 20, 45, 47, 48). Even at lower N/P ratios of 1.1 and 1.2, Li-Sn/CS@SC∥NCM811 full cells still exhibit good cycling performance over 500 cycles with a capacity retention of about 75% at 5 C (fig. S33). Furthermore, the Li-Sn/CS@SC∥NCM811 full cell also delivers good rate performance with high capacity retention (fig. S34). Meanwhile, the Li-Sn/CS∥NCM811 and Li-CS∥NCM811 cells suffer fast capacity decay upon increased current densities (fig. S35). Impressively, the Sn/CS@SC∥NCM811 pouch cell can power a light-emitting diode bulb with a capacity retention of 90% after 300 cycles (Fig. 6D and fig. S36).

Fig. 6. Performance of half cells, full cells, and pouch cells.

(A) CE profiles of half cells at a current density of 1 mA cm⁻². (B) Voltage profiles for symmetrical cells at a current density of 1 mA cm⁻² with a fixed capacity of 1 mAh cm⁻². h, hours. (C) Cycling performance of full cells. (D) Cycling performance of the Sn/CS@SC∥NCM811 pouch cell.

DISCUSSION

In summary, we report a unique design for the Sn/CS@SC host with highly reversible QM-Li storage via consecutive thermal reduction and SC coating. The incorporation of in situ–formed Sn single atoms could realize the simultaneous regulation of subnanopores and interfacial chemistry. Benefitting from the confinement effect of subnanopores, the side reactions between the electrolyte and active QM-Li are greatly suppressed. Furthermore, the strong interaction between Sn-N₄ increases the Li-ion affinity to enable the storage of QM-Li within subnanopores. Consequently, the Sn/CS@SC host achieves high Li storage capacity over 600 cycles with a high average CE of about 99.8% in the ester electrolyte. Furthermore, the corresponding full cell using the prelithiated Sn/CS@SC anode and NCM811 cathode shows a good capacity retention of about 80% after 500 cycles at 5 C.

MATERIALS AND METHODS

Synthesis of SnO₂ colloid solution

SnCl₂·2H₂O (0.9 g, Alfa Aesar Co., Ltd) and thiourea (0.3 g, Alfa Aesar Co., Ltd) were added in 30 ml of deionized (DI) water, and the mixture was vigorously stirred until it became transparent yellow. Then, another 90 ml of DI water was added under strong stirring condition.

Synthesis of CS

Ammonia aqueous solution (3 ml, Alfa Aesar Co., Ltd) was mixed with a solution containing absolute ethanol (70 ml, Alfa Aesar Co., Ltd) and DI water (10 ml) and then stirred for more than 1 hour. Subsequently, resorcinol (0.4 g, Alfa Aesar Co., Ltd) was added and continually stirred for 30 min. Then, the formaldehyde solution (0.56 ml, Alfa Aesar Co., Ltd) was added to the reaction solution and stirred for 24 hours at 30°C. The product was collected by centrifugation, dried in an oven at 60°C, and then carbonized at 1000°C for 2 hours with a heating rate of 1°C min⁻¹ under a N₂ atmosphere as CS.

Synthesis of Sn/CS

Ammonia aqueous solution (3 ml) was mixed with a solution containing ethanol (70 ml) and DI water (10 ml) and then stirred for more than 1 hour. Subsequently, resorcinol (0.4 g) and SnO₂ colloid solution (150 μl) were added and continually stirred for 30 min. Then, the formaldehyde solution (0.56 ml) was added to the reaction solution and stirred for 24 hours at 30°C. The product was collected by centrifugation, dried in an oven at 60°C, and then carbonized at 1000°C for 2 hours with a heating rate of 1°C min⁻¹ under a N₂ atmosphere as Sn/CS.

Synthesis of Sn/CS@SC

Sn/CS and pitch powders were added in tetrahydrofuran solution and continually stirred for 24 hours at 30°C. Then, the solvent from the solution was evaporated using a rotary evaporator. After annealing at 1000°C for 2 hours in a N₂ atmosphere, Sn/CS@SC was obtained.

Synthesis of Sn/CS@SC-SnCl₂

Ammonia aqueous solution (3 ml) was mixed with a solution containing absolute ethanol (70 ml) and DI water (10 ml) and then stirred for more than 1 hour. Subsequently, resorcinol (0.4 g) and SnCl₂ solution (150 μl) were added and continually stirred for 30 min. Then, the formaldehyde solution (0.56 ml) was added to the reaction solution and stirred for 24 hours at 30°C. The product was collected by centrifugation, dried in an oven at 60°C, and then carbonized at 1000°C for 2 hours with a heating rate of 1°C min⁻¹ under N₂ as Sn/CS-SnCl₂. Afterward, the prepared Sn/CS and pitch powders were added in tetrahydrofuran solution and continually stirred for 24 hours at 30°C. Then, the solvent from the solution was evaporated using a rotary evaporator. After annealing at 1000°C for 2 hours in a N₂ atmosphere, Sn/CS@SC-SnCl₂ was obtained. The ramping rate was kept at 5°C min⁻¹.

Synthesis of Cu/CS@SC

Ammonia solution (3 ml) was mixed with a solution containing ethanol (70 ml) and DI water (10 ml) and then stirred for more than 1 hour. Subsequently, resorcinol (0.4 g) and CuCl₂ solution (150 μl) were added and continually stirred for 30 min. Then, formaldehyde solution (0.56 ml) was added to the reaction solution and stirred for 24 hours at 30°C. The product was collected by centrifugation, dried in an oven at 60°C, and then carbonized at 1000°C for 2 hours with a heating rate of 1°C min⁻¹ under a N₂ atmosphere as Cu/CS. Cu/CS and pitch powders were added in tetrahydrofuran and continually stirred for 24 hours at 30°C. Then, the solvent from the solution was evaporated using a rotary evaporator. After annealing at 1000°C for 2 hours in a N₂ atmosphere, Cu/CS@SC was obtained.

Synthesis of CS@SC

Ammonia solution (3 ml) was mixed with a solution containing ethanol (70 ml) and DI water (10 ml) and then stirred for more than 1 hour. Subsequently, resorcinol (0.4 g) was added and continually stirred for 30 min. Then, formaldehyde solution (0.56 ml) was added to the reaction solution and stirred for 24 hours at 30°C. The product was collected by centrifugation, dried in an oven at 60°C, and then carbonized at 1000°C for 2 hours with a heating rate of 1°C min⁻¹ under a N₂ atmosphere as CS. CS and pitch powders were added in tetrahydrofuran and continually stirred for 24 hours at 30°C. Then, the solvent from the solution was evaporated using a rotary evaporator. After annealing at 1000°C for 2 hours in a N₂ atmosphere, CS@SC was obtained.

Materials characterization

The morphology of the samples was investigated by FESEM (TESCAN Clara; JEOL JSM-7800F), TEM (JEOL, HT7700), HAADF-STEM (FEI-Talos F200S), and aberration-corrected HAADF-STEM (JEM-ARM 200F) equipped with energy dispersive x-ray spectroscopy. The crystal structure was analyzed by XRD on a Bruker D2 Phaser x-ray diffractometer with Ni-filtered Cu Kα radiation (λ = 1.5406 Å) at a voltage of 30 kV and a current of 10 mA. The surface chemical states of the samples were examined by an ESCALAB 250 XPS system with an Al Kα radiation source. Raman spectra were collected through a Renishaw inVia Raman microscope equipped with a Leica DMIRBE inverted optical microscope using the laser excitation source at 514 nm. Inductively coupled plasma optical emission spectrometry data were collected on iCAP 7000 (Thermo Fisher Scientific) to determine the content of elements in solution. The FTIR spectrum was collected in a transmission mode on a Nicolet Antaris II spectrophotometer (Thermo Fisher Scientific). The measurements of XAS at the Sn K-edge containing XANES and EXAFR were performed at the beamline BL14W1 of Japan Synchrotron Radiation Facility. SSA and pore diameter distribution were analyzed with the Brunauer-Emmett-Teller method and the nonlocal density functional theory method, respectively. SAXS and WAXS were performed with a Cu x-ray source of 30 W (wavelength of 0.1542 nm) and an x-ray spot size of 0.8 mm by 0.8 mm. Skeletal density analysis using He was operated by AccuPyc 1330, and the impregnation method was operated by an electronic densimeter (JHY-120 wt) and pycnometer using DMC. The pore volume can be calculated by a simplified equation (49)

$$V_{\text{pore}}=\frac{1}{{\mathrm{\rho }}_{\text{tested}}}–\frac{1}{{\mathrm{\rho }}_{\text{baseline}}}$$ (1)

Electrochemical characterizations

The active material, carbon black, carboxymethyl cellulose (MTI Kejing Group), and styrene butadiene rubber (MTI Kejing Group) were mixed in DI water with a mass ratio of 90:4:3:3 as the anode slurry. Then, the anode slurry was coated onto a Cu foil and vacuum dried at 100°C for 10 hours to reach a mass loading of about 3 mg cm⁻². NCM811 (MTI Kejing Group) powder was mixed with carbon black and poly(vinylidene fluoride) (MTI Kejing Group) in a mass ratio of 8:1:1 in 1-methyl-2-pyrrolidinone (Macklin) as the cathode slurry. Then, the cathode slurry mixture was coated onto Al foil and vacuum dried at 120°C for 6 hours to reach a mass loading of about 5 mg cm⁻². CR2032-type coin cells and pouch cells (4.5 cm by 3.5 cm) were accomplished in an Ar-filled glove box with O₂ and H₂O content below 0.5 ppm (parts per million). A Celgard 2325 membrane was used as the separator. Electrodes for coin cells were cut into round disks with a diameter of 10 mm. The electrolyte for all the cells was 1 M LiPF₆ in ethylene carbonate/DMC/diethyl carbonate in a volume ratio of 1:1:1 with a 5 vol % fluoroethylene carbonate additive. Metrohm Autolab was used for the electrochemical impedance spectroscopy test. The Neware CT-4008T battery test system was used for cycling and GITT test. Before the GITT test, prelithiation was carried out for samples with a Li storage capacity of 700 mAh g⁻¹ at a pulse current density of 70 mA g⁻¹. The test voltage window was 0.01 to 2 V. The $D_{{\text{Li}}^{+}}$ was calculated using a simplified equation of Fick’s second law (25)

$$D=\frac{4}{\mathrm{\pi \tau }}{\left(\frac{m_{\mathrm{B}}{V}_{\mathrm{M}}}{M_{\mathrm{B}}S}\right)}^2{\left(\frac{\mathit{∆}{E}_{\mathrm{s}}}{\mathit{∆}{E}_{\mathrm{t}}}\right)}^2$$ (2)

where $\mathrm{\tau }$ is the pulse duration (s), m_B is the active mass of the electrode (g), M_B is the molar mass of HC (g mol⁻¹), V_M is the molar volume (cm³ mol⁻¹), and S is the active surface area of the HC electrodes (m² g⁻¹). In addition, ∆E_s (20 min) and ∆E_t (120 min) can be obtained from the GITT curves. For the CE test, the half cells were discharged with a Li storage capacity of 1.4 mAh cm⁻² on different samples at the current density of 1 mA cm⁻², followed by charging to 2 V versus Li⁺/Li. Before the CE test, two activation cycles were performed at a current density of 0.2 mA cm⁻². For the symmetric cell test, all the anodes went through a prelithiation process with a capacity of 1.65 mAh cm⁻² before long-term cycling. For the coin-based full cells, prelithiation processes were taken on Sn/CS@SC, Sn/CS, and CS with a capacity of 0.35 mAh cm⁻². The full cells were assembled using the NCM811 cathode and prelithiated anodes with a capacity of 700 mAh g⁻¹. The full cells were cycled at a constant rate of 1 to 10 C with a cutoff voltage window of 4.3 to 3.0 V (1 C ≈ 190 mA g⁻¹). The pouch cell was assembled using the NCM811 cathode and Sn/CS@SC anode. The pouch cell was cycled at a constant rate of 2 C with a cutoff voltage window of 4.2 to 3.0 V.

Optical microscopy observation

Optical microscopy observations were carried out in an optical cell with a quartz window (E001, Tianjin IDA) using an optical microscope (BX53MRF-S, Japan). The assembly of the optical cell was performed at room temperature in an Ar-filled glove box with the water and oxygen contents less than 0.5 ppm.

Theoretical computation method

Vienna Ab initio Simulation Package software was adopted for optimizing all the structures. Projected augmented wave formalism by Perdew, Burke, and Ernzerhof was used for the electron-ion interactions (50, 51). Values of 450 eV, 10⁻⁴ eV, and 0.02 eV/Å were used for the cutoff energy and the convergence criteria of energy and force, respectively. The binding energy was defined as follows

$$E_{\mathrm{b}}=E_{\text{system}}-E_{\text{surface}}-E_{\text{Li}}$$ (3)

where E_system and E_surface stand for energies of the total system and the surface, respectively, and E_Li is the energy of the Li atom.

Supplementary Materials

This PDF file includes:

Note S1

Figs. S1 to S36

Tables S1 to S7