English

• Sodium-ion batteries (SIBs) have received widespread attention due to its similar working principles as LIBs, low expense and abundant resources. However, sodium ions have a larger radius compared to lithium ions (1.02 Å vs. 0.76 Å) and the sodium ion intercalation in graphite are not thermodynamically favourable. Therefore, finding suitable anode materials remains a challenge in SIBs research. Hard carbons are considered as one of the most promising candidates due to their large interlayer spacing (0.37–0.42 nm), numerous nanopores, and abundant active sites [1-3]. Generally, the discharge/charge curves of hard carbons can be divided into a sloping region (>0.1 V) and a plateau region (<0.1 V) [4]. To increase the energy density and operating voltage window of SIBs, it is vital to improve the reversible capacity based on the plateau region below 0.1 V, which is universally believed to be closely related to the Na⁺ filling in closed pores and Na⁺ intercalation between carbon layers [5]. As a result, the pore structures and microcrystalline structures of hard carbons play key roles in improving the plateau capacity. As for the microcrystalline structure, hard carbons with maximized pseudo-graphitic phase are highly desirable, which could not only provide more Na⁺ storage sites but also increase the microstructure stability. As for the pore structure, the number, size and volume of closed pores in a hard carbon are important factors affecting the filling behavior in closed pores. Mai et al. recently found that not all the closed pores in hard carbons can be filled with Na⁺ [4]. Only when the interlayer spacings are larger than 0.37 nm, sodium ions can pass through to reach closed pores for further filling. Furthermore, Song et al. indicated that the size of pseudo-graphite domains, especially the lateral width (L_a), plays a more essential role than the interlayer spacings in determining whether sodium ions can diffuse into the closed pores and a concept of accessible closed pores is proposed [6]. As a result, apart from constructing enough closed pores in hard carbons, it is also crucial to regulate the microcrystalline structures (the content of pseudo-graphitic phase and the size of pseudo-graphite domains) of hard carbons to improve the sodium ion storage in closed pores. Nevertheless, an effective method of increasing the content of pseudo-graphitic phase and simultaneously reducing the lateral width of pseudo-graphite domains is still lacking.

  Sodium carboxymethyl cellulose (CMC–Na), as a derivative of cellulose, is an abundant and low-cost polysaccharide, widely used as emulsifier, dispersing agent and binder in industries [7-9]. As a hard carbon resource, CMC–Na can form sodium carbonate/hard carbon composite after carbonization in inert atmosphere [10]. Thus, the sodium carbonate (Na₂CO₃) can act as a salt template to introduce large amount of uniformly dispersed closed pores in hard carbon as the sodium element in CMC–Na exists at the molecular level. Furthermore, the uniform Na₂CO₃ particles can also act as a catalyst to promote the local graphitization of hard carbons around them and regulate the microcrystalline structure (the content and the size of pseudo-graphite domains) of hard carbons [10,11]. In addition, the removing of Na₂CO₃ template can be achieved just by water washing because of the ultrahigh solubility of Na₂CO₃ in water, which is more environmentally benign and cost effective.

  Based on the above analysis, CMC–Na was used as a precursor to fabricate hard carbons with large number of closed pores. The microstructures of hard carbons are regulated by esterification modification and tuning the pre-carbonization temperature. The esterification modification effectively improves the structural stability of CMC–Na during the pre-carbonization process and hinders the overgrowth and agglomeration of Na₂CO₃ particles in the carbon matrix. High pre-carbonization temperature is also conducive to form smaller and more uniform-sized Na₂CO₃ particles, which further exert a different catalytic effect on the graphitization process of hard carbons during the pre-carbonization process. Ultimately, hard carbons with abundant uniformly dispersed closed pores, maximized pseudo-graphitic content (73%), and minimized lateral width (L_a) of pseudo-graphitic domains were fabricated to improve the plateau capacity. Benefiting from the high pseudo-graphitic content and improved accessibility of sodium ions to the intercalation sites and filling sites, the optimized hard carbon (EHC-500) delivers a high reversible sodium storage capacity of 340 mAh/g at 30 mA/g with a high plateau capacity of 236.7 mAh/g (below 0.08 V). Furthermore, the assembled EHC-500//NVP full cell delivers a high energy density of 203 Wh/kg and a high average working voltage of 3.3 V. The current study offers a novel and efficient route to the rational microstructure regulation of hard carbons towards enhanced sodium storage performance.

  Two preparation routes from CMC–Na precursor to prepare hard carbons with different pore structures and pseudo-graphite phases are presented in Fig. 1. The hard carbon with direct carbonization of CMC–Na is denoted as HC-x, x of which refers to pre-carbonization temperature. During the pre-carbonization process, Na₂CO₃ particles tend to be formed in the resulting carbon material. After removing the salt template (Na₂CO₃) and high-temperature carbonization, closed pores are introduced into the hard carbon. In an altered preparation pathway, CMC–Na was firstly esterified with maleic anhydride to improve the structural stability during the pre-carbonization process and further regulate the pore structure and pseudo-graphitic phase (domain size and content) of the final hard carbon (denoted as EHC-x, x refers to pre-carbonization temperature). Benefitting from the improved structural stability of the esterified CMC–Na, smaller and more uniform-sized Na₂CO₃ particles are uniformly distributed in the carbon matrix after the pre-carbonization process. After removing the Na₂CO₃ particles, abundant smaller and more uniform-sized closed pores are introduced into the hard carbons. During the high-temperature carbonization, the closed pore size was further reduced and the pre-formed nanopores would hinder the long-range graphitization of carbon microcrystals. As a result, compared with that of HC, the pseudo-graphitic content is increased and the pseudo-graphitic domain size is also different in EHC due to the catalytic effect of smaller Na₂CO₃ particles.

  Figure 1

  

  Figure 1.  Schematic illustration of the routes to fabricate CMC–Na-based hard carbons with different pore structures and pseudo-graphite phases.

  DownLoad: Full-Size Img PowerPoint

  The functional groups of CMC–Na and the esterified CMC–Na were examined by the Fourier transform infrared spectroscopy (FTIR). As shown in Fig. S1 (Supporting information), the FTIR spectra of CMC–Na illustrates that peaks centered around 3327, 2913 and 1034 cm⁻¹ correspond to the stretching vibrations of -OH, the stretching vibrations of C–H, and the stretching vibrations of C–O-C, respectively. Due to the asymmetric and symmetric stretching of the carboxylate group (-COONa) in CMC–Na, two adsorption peaks at 1587 and 1414 cm⁻¹ appear in the FTIR spectra [12,13]. After esterification, two new absorption peaks attributed to the ester group and the butyl ester arise at 1699 and 854 cm⁻¹, respectively [14,15]. In addition, the esterified CMC–Na exhibits a decreased intensity of the -OH adsorption peak, further demonstrating the successful esterification reaction between CMC–Na and maleic anhydride.

  In order to explore the impact of esterification modification on the pyrolysis behavior of CMC–Na, thermogravimetric analysis (TG) was conducted under nitrogen atmosphere from room temperature to 900 ℃ (Fig. 2a and Fig. S2 in Supporting information). As shown in Fig. S2a, CMC–Na demonstrates two primary stages of mass loss. Firstly, a mass loss at about 100 ℃ is due to the loss of H₂O adsorbed on the CMC–Na surface. The second mass loss at about 300 ℃ is corresponds to the loss of CO₂ originated from polysaccharide degradation and decarboxylation of COO⁻ group in CMC–Na [7]. Maleic anhydride exhibits a rapid weight loss from 100 ℃ to 200 ℃ and was totally converted to small gas molecules when the temperature reached 200 ℃ (Fig. S2b). Based on the TG curves of CMC–Na and maleic anhydride, a curve based on the mass ratio of CMC–Na and maleic anhydride (5:2) was fitted in Fig. 2a. In comparison with the fitting curve, the TG curve of the esterified CMC–Na shows a higher thermal decomposition temperature at the decomposition stage below 300 ℃. And the carbon yield of the esterified CMC–Na (29.07%) at the pyrolysis temperature of 900 ℃ is significantly higher than the fitting curve (23.90%). Combined with FTIR, these results show solid evidence for the successful cross-linking reaction between CMC–Na and maleic anhydride and the improved structural stability of the esterified CMC–Na in comparison with CMC–Na.

  Figure 2

  

  Figure 2.  (a) TG curves of esterified CMC–Na and the mixture of CMC–Na and maleic anhydride. (b) XRD patterns. (c) Corresponding phase contents from the XRD fitting results. SEM images of (d) EHC-500 and (g) HC-500. HRTEM images and SAED images of (e) EHC-500 and (f) HC-500. Schematic representation of the pseudo-graphite domain sizes in (h) EHC-500 and (i) HC-500.

  DownLoad: Full-Size Img PowerPoint

  To investigate the effect of preparation pathway on the microstructures of hard carbons, EHC-500 and HC-500 were chosen as the typical samples. And the correlation between the pre-carbonization temperatures and the microstructures of the final hard carbons (i.e., EHC-300, EHC-400 and EHC-500) was further studied. X-ray diffraction (XRD) patterns of EHC-300, EHC-400, EHC-500 and HC-500 are shown in Fig. 2b. All patterns exhibit two broad peaks at 23° and 43°, corresponding to the (002) and (100) lattice planes of hard carbons. According to the XRD result, the interlayer spacing (d₀₀₂), lateral widths (L_a) and thickness (L_c) of pseudo-graphite domains are calculated and listed in Table S1 (Supporting information). The interlayer spacings of EHC-300, EHC-400, EHC-500 and HC-500 are 0.378, 0376, 0.374 and 0.377 nm calculated by the Bragg equation, respectively, which are all larger than 0.37 nm and favorable for the insertion and extraction of Na⁺ [16]. It is worth noting that the L_a values of hard carbons with esterification (EHC-300, EHC-400 and EHC-500) are much lower than that of hard carbon without esterification (HC-500), and EHC-500 exhibits the lowest L_a value (4.943 nm) (Table S1 and Fig. S3 in Supporting information). The L_c values of hard carbons with esterification increase as the pre-carbonization temperature increases. Under the same pre-carbonization temperature, the L_c value of EHC-500 is also higher than that of HC-500. Therefore, both the high pre-carbonization temperature and the cross-linked structure after esterification could not only hinder the lateral growth of pseudo-graphite domains but also promote the growth in the stacking direction [17]. The (002) peak of hard carbon in XRD can be divided into two peaks attributed to the highly disordered carbon phase and pseudo-graphitic phase (Fig. S4 in Supporting information), respectively, and the phase composition was calculated [18]. As shown in Fig. 2c, EHC-500 exhibits the highest proportion of pseudo-graphitic phase (73%) and the lowest proportion of highly disordered carbon phases (27%) among the three EHC samples. By contrast, EHC-300 holds the lowest proportion of pseudo-graphitic phases (57%) and the highest proportion of highly disordered carbon phases (43%), which indicates that the highly disordered carbon phase gradually evolves towards a pseudo-graphitic phase as the pre-carbonization temperature increases. And HC-500 possesses a lower composition of the pseudo-graphite phase (66%) when compared to the esterified sample EHC-500, which can be attributed to the distinct catalytic effect of different Na₂CO₃ particle sizes formed in the pre-carbonization stage. The XRD pattern of commercial CMC–Na shows a characteristic peak at 20.1°, corresponding to the crystal structure of cellulose Ⅱ (Fig. S5 in Supporting information) [12]. After pre-carbonization, a large number of spherical particles were generated in the samples pre-EHC-300, pre-EHC-400, pre-EHC-500 and pre-HC-500, respectively, and are confirmed to be Na₂CO₃ (PDF#37–0451) particles (Figs. S6-S8 in Supporting information). As shown in Fig. S6, compared with pre-HC-500, pre-EHC-300 and pre-EHC-400, pre-EHC-500 demonstrate obviously smallest and most uniform-sized Na₂CO₃ particles, suggesting that more catalytic sites are exposed to promote the local graphitization of the hard carbon around them, ultimately leading to the formation of the highest proportion of pseudo-graphitic phase in EHC-500. In addition, the reduced size of Na₂CO₃ particles would reduce the catalytic area between Na₂CO₃ particles and carbon layers. Furthermore, the highest number of nanopores in EHC-500 would hinder the long-range graphitization of carbon microcrystals during the high-temperature carbonization and result in the lowest L_a value in EHC-500. During the pre-carbonization process, the Na₂CO₃ particles formed near the outer surface would be partly vaporized and condensed on the outer surface. As a result, we can observe the Na₂CO₃ particles adhering to the outer surface, of which the size is apparently larger than that embedded in the carbon matrix. As the increase of pre-carbonization temperature, the amount of vaporized Na₂CO₃ steam would increase and condense into Na₂CO₃ particles with larger size on the outer surface. As most of the XRD signals were detected from the outer surface, the intensity of the Na₂CO₃ peaks show a gradually increasing tendency with the increase of pre-carbonization temperature (Fig. S7).

  Moreover, the Na₂CO₃ particles are uniformly distributed in the carbon matrix of the pre-carbonized samples with esterification, while agglomerated in the carbon matrix of pre-HC-500. It reveals that the cross-linking can effectively improve the structural stability of CMC–Na during the pre-carbonization process and hinder the overgrowth and agglomeration of Na₂CO₃ particles. After washing with deionized water to remove Na₂CO₃ particles, abundant pores are introduced into the hard carbons according to the scanning electron microscopy (SEM) images (Figs. 2d and g, Figs. S9 and S10 in Supporting information). Compared with HC-500, EHC-500 obtained by esterification treatment shows obvious smaller and more uniform-sized pores in the carbon matrix. Transmission electron microscopy (TEM) was conducted and the pores are further demonstrated to be existed not only on the surface of but also inside the carbon particles (Fig. S11 in Supporting information). Figs. 2e and f exhibit the high-resolution transmission electron microscopy (HRTEM) images of EHC-500 and HC-500, which both consist of pseudo-graphitic and highly disordered domains. Meanwhile, the related SAED patterns exhibit dispersing rings in the insets of Figs. 2e and f, further revealing their amorphous structure. Closed nanopores can also be observed both in EHC-500 and HC-500 from the HRTEM images. In line with the XRD result, the lateral widths of pseudo-graphitic domains of EHC-500 are apparently smaller than that of HC-500. Schematic representation of the pseudo-graphite domain sizes in EHC-500 and HC-500 are shown in Figs. 2h and i.

  To further investigate the pore structures of hard carbons, N₂ sorption measurement, true density analysis and small-angle X-ray scattering (SAXS) were conducted. According to the N₂ adsorption desorption measurement (Fig. 3a), the surface areas and open pore volumes of these samples are listed in Table S1. Apparently, not only the surface areas but also the open pore volumes of the hard carbons with esterification (EHC-300, EHC-400 and EHC-500) are much lower than that of the hard carbon without esterification (HC-500). As shown in Fig. 3b, the open pore size shows a shift towards smaller sizes with the increase of pre-carbonization temperature. Consistent with the SEM results (Figs. 2d and g), EHC-500 shows smaller and more uniform-sized pores in contrast to HC-500. Apart from the open pores, closed pores were tested by true density analysis and SAXS. The true density was tested using helium as the analyzing gas according to Archimedes' principle, as helium can enter all pore structures except closed pores. The true density of ideal graphite is 2.26 g/cm³. Therefore, the closed pore volumes of hard carbons can be calculated by the following equation [17]:

  $ V_{\text {Closed pore }}=\frac{1}{\rho_{\text {ture }}}-\frac{1}{2.26} $

  (1)

  Figure 3

  

  Figure 3.  (a) N₂ adsorption and desorption isotherms. (b) Pore size distribution. (c) SAXS patterns of EHC-500 and HC-500. (d) Raman spectra and fitted curves of EHC-500 and HC-500. (e) C 1s spectra and fitted curves of EHC-500 and HC-500. (f) O 1s spectra and fitted curves of EHC-500 and HC-500.

  DownLoad: Full-Size Img PowerPoint

  As shown in Table S1, not only the open pore volumes but also the closed pore volumes of the hard carbons with esterification are lower than that of the hard carbon without esterification (HC-500). However, the ratio of closed pore volumes to total volumes (V_{closed pore}/V_{total pore}) for the hard carbons with esterification are much higher than that of the hard carbon without esterification and more than 95% are closed pores. In addition, both the SAXS curves of EHC-500 and HC-500 show a broad peak close to 0.1 Å⁻¹ (Fig. 3c), which is typical of the presence of closed pores [19]. EHC-500 exhibits higher scattering intensity and higher shoulder Q values than HC-500, suggesting more closed pores and smaller closed pore size in EHC-500 (3.65 nm) than those in HC-500 (5.46 nm) (Fig. S12 in Supporting information) [20]. Based on the above analysis, it can be indicated that it is easy to obtain hard carbons with a large number of closed pores by using CMC–Na as carbon precursor and the pore structures can be effectively regulated by esterification and tuning the pre-carbonization temperature. Esterification can effectively construct crosslinking structure and improve the structure stability of precursor CMC–Na, hinder the release of small gas molecules and overgrowth of Na₂CO₃ particles during the pre-carbonization process, and promote the formation of high ratio of smaller and more uniform-sized closed pores in the resulting hard carbons. Besides, increasing the pre-carbonization temperature can promote the shift of the size of pores to the smaller ones.

  As shown in Fig. 3d and Fig. S13 (Supporting information), according to Lorentz function, Raman spectra can be deconvoluted into four subpeaks at approximately 1200 cm⁻¹ (D₄ band), 1320 cm⁻¹ (D₁ band), 1500 cm⁻¹ (D₃ band) and 1575 cm⁻¹ (G band), respectively [21-23]. The G band corresponds to the E_2g vibration of ideal graphite lattice, the D₁ band corresponds to the A_1g symmetric vibration of defects in graphene edge [24,25]. Thus, the band area ratio of the D₁ band to the G band (I_D1/I_G) can be used to characterize the defect concentration of the carbon [26,27]. As listed in Table S1, the values of I_D1/I_G are 1.52, 1.32, 1.79 and 1.46 for EHC-300, EHC-400, EHC-500 and HC-500, respectively. EHC-500 exhibits the highest I_D1/I_G value among them, and it could be partly attributed to the numerous defects on the pore surface exposed by the smallest pores in EHC-500. In addition, it should simultaneously originate from its highest amount of oxygen element in EHC-500 (Fig. S14 in Supporting information). To analyze the surface elemental compositions, X-ray photoelectron spectroscopy (XPS) analysis was investigated. From the full XPS spectra shown in Fig. S14, only carbon elements and oxygen elements are detected in these hard carbons. The oxygen contents in EHC-300, EHC-400, EHC-500 and HC-500 are 6.15, 5.45, 7.38 and 5.10 wt%, respectively, which reveal that the oxygen contents in the hard carbons with esterification are generally higher than that of the hard carbon without esterification. It could be attributed to the improved structural stability of esterified CMC–Na. The C 1s spectra (Fig. 3e and Fig. S15a in Supporting information) can be classified into four peaks with binding energies (B.E.) of 284.7, 285.3, 286.7, and 288.9 eV, which correspond to the C–C, C=C, C–O and C=O, respectively [28,29]. As indicated in Table S2 (Supporting information), the proportion of C–O in HC-500 is obviously higher than that in the hard carbons with esterification, which would cause irreversible adsorption of Na⁺ and decrease the initial Coulombic efficiency (ICE). In contrast, the proportion of C = O in HC-500 is obviously lower than that in the hard carbons with esterification, which would be active sites for sodium storage [30]. The O 1s spectra (Fig. 3f and Fig. S15b in Supporting information) shows two peaks of C=O (532.1 eV) and C–O (533.4 eV). The content of these two oxygen configurations in HC-500 is both lower than that in the hard carbons with esterification. Again, these results indicate that esterification can improve the structural stability of CMC–Na during the pyrolytic process..

  The electrochemical properties of these hard carbons as anodes of SIBs were investigated in half cells. Fig. 4a shows the cyclic voltammetry (CV) curves of EHC-500 at 0.1 mV/s. In the first scan cycle, a broad peak centered at 0.6 V is observed, mainly due to the decomposition of the electrolyte on the electrode surface and the formation of solid electrolyte interphase (SEI) layer [31]. A sharp reduction/oxidation peak pair at 0.01 and 0.11 V corresponds to the sodiation/desodiation process between the adjacent carbon layers and/or in the closed nanopores [32]. The subsequent scanning curves of EHC-500 almost overlap, demonstrating its excellent electrochemical reversibility. The galvanostatic charge-discharge (GCD) curves of each sample are shown in Fig. 4b. The ICE of the EHC-300, EHC-400, EHC-500, and HC-500 electrodes are 70.4%, 70.6%, 72.1%, and 67.5%, respectively. As HC-500 possesses the highest open pore surface area and proportion of C–O, it shows the lowest ICE. The reversible capacities of EHC-300, EHC-400, EHC-500, and HC-500 are 313, 298, 340, and 303 mAh/g. All of them are composed of a slope capacity above 0.1 V and a plateau capacity below 0.1 V. The former usually are attributed to the adsorption of Na⁺ by oxygen-containing functional groups (e.g., C=O) and defect, whereas the latter are corresponding to pore filling and/or interlayer insertion mechanisms of Na⁺ [33]. As shown in Fig. 4c, EHC-500 shows the highest plateau capacity (238.3 mAh/g) while HC-500 shows the lowest plateau capacity (174.1 mAh/g). When compared to some other recently reported hard carbon materials (Table S3 in Supporting information), EHC-500 shows excellent electrochemical performance. The sodium storage mechanism in hard carbons will be discussed in the following section.

  Figure 4

  

  Figure 4.  (a) Cyclic voltammetry curves of EHC-500 at a scanning rate of 0.1 mV/s. Electrochemical Na-storage performances of the carbon samples: (b) GCD curves at 30 mA/g; (c) Plateau and slope capacity contribution corresponding to the second cycle discharge curve; Cycling performance at the current densities of (d) 30 and (e) 300 mA/g; (f) Rate performance.

  DownLoad: Full-Size Img PowerPoint

  The cycling stability of these samples was evaluated at the current densities of 30 and 300 mA/g, respectively (Figs. 4d and e). All the samples exhibit good cycling stability and EHC-500 shows the highest capacity at the current density of not only 30 mA/g but also 300 mA/g. At the current density of 30 mA/g, EHC-500 retains a high capacity of 309 mAh/g after 100 cycles with a capacity retention of 91.1%, demonstrating the excellent structural stability during the continuous insertion and extraction of Na⁺. EHC-500 can also maintain a high capacity of 275 mAh/g with a capacity retention of 90.9% after 250 cycles at the current density of 300 mA/g [34]. Meanwhile, the rate performance of these samples was further tested at different current densities (Fig. 4f). The reversible specific capacities of EHC-500 at current densities of 0.03, 0.06, 0.12, 0.3, 0.6 and 1.2 A/g are 342, 324, 309, 285, 257, and 207 mAh/g, respectively, superior to that of HC-500. When the current density come back to 0.03 A/g, a capacity of 321 mAh/g can be retained, indicating a high reversibility of the discharge-charge process. Compared with EHC-300, EHC-400 and HC-500, EHC-500 delivers the highest reversible capacity within the current densities below 0.12 A/g, whereas shows a more inferior capacity retention at the current densities above 0.12 A/g. It could be attributed to the fact that EHC-500 exhibits the highest plateau capacity and the kinetics of the insertion reaction and pore filling in the plateau region is sluggish in dynamics (Fig. S16 in Supporting information) [35]. In addition, the relative higher amounts of mesopores in EHC-300, EHC-400 and HC-500 provide short transport pathways for rapid Na⁺ migration resulting in the superior rate performance [36].

  In order to gain a deep understanding of sodium storage behaviors in the hard carbons, EHC-500 and HC-500 were selected as the typical samples to be further investigated. Firstly, CV analyses were performed at various scan rates (0.1–2.0 mV/s), the curves of which are shown in Figs. 5a and d. The relationship between scan rate (v) and peak current (i) can be expressed by the following equation [37], i.e., i = av, where a and b are constants and the value of b can be obtained from the slope of ln(i) versus ln(v). A value of b close to 0.5 indicates diffusion-controlled process and a value of b close to 1 indicates a surface-controlled process [38]. The linear relationship between the logarithm of the peak current and the logarithm of scan rate is shown in Figs. 5b, c, e and f, and the curves have a good linear relationship with R_A² = 0.98, R_B² = 0.99 for EHC-500 and R_A² = 0.99, R_B² = 0.99 for HC-500. The b value of peak A is 0.33 and 0.42 for EHC-500 and HC-500, respectively, indicating a sodium storage process of diffusion control at the low potential region. Whereas the b value of peak B is 1.16 and 1.10 for EHC-500 and HC-500, respectively, indicating a sodium storage process of surface control at the high potential region, which corresponds to adsorption on defects. Electrochemical impedance spectroscopy (EIS) was conducted to further analyze these hard carbons (Fig. 5g). All the samples show similar EIS spectra composed of a semicircle in the high-frequency region and a sloping line in the low-frequency region, which are attributed to the charge transfer resistance (R_ct) and Warburg impedance (Z_w), respectively. The Z_w is related to the Na⁺ diffusion in electrode. The corresponding equivalent circuit and fitting resistance values are shown in Fig. S17 and Table S4 (Supporting information). The R_ct values of EHC-300 (1.77 Ω), EHC-400 (1.96 Ω), EHC-500 (1.83 Ω), are much smaller than that of HC-500 (16.76 Ω), suggesting that the high proportion of pseudo-graphite phase and small L_a value improve charge transfer [34,39].

  Figure 5

  

  Figure 5.  EHC-500: (a) CV curves at different scan rates; (b, c) Linear relationship between ln(i) and ln(v). HC-500: (d) CV curves at different scan rates; (e, f) Linear relationship between ln(i) and ln(v). (g) EIS. GITT curves of (h) EHC-500 and (i) HC-500.

  DownLoad: Full-Size Img PowerPoint

  The diffusion coefficient of Na⁺ in materials was measured by the galvanostatic intermittent titration technique (GITT). The samples were subjected to a pulse time for 15 min and intervals of 1 h (Fig. S18 in Supporting information). The corresponding diffusion coefficient of Na⁺ (D_Na) was calculated according to Fick's second law equation and the relationship between the D_Na and potential is shown in Figs. 5h and i. During the process of sodiation, the diffusion coefficient of EHC-500 is considerably greater than that of the non-esterified sample, HC-500, in line with the EIS result. Furthermore, the calculated D_Na values of EHC-500 and HC-500 exhibit a similar changing trend. At first, the values of D_Na remain a high level within the high potential region but decline rapidly at 0.08 V. Another turning point is happened at 0.04 V and the value of D_Na suddenly rise. The sodium storage mechanism in hard carbons remains controversial, especially for the plateau region. According to different experimental results, three prevailing models are proposed on the sodium storage mechanism of the plateau region, i.e., “intercalation” model [40], “intercalation-filling” model [41], and “filling” model [42]. To study the sodium ion storage behavior at the plateau region, the reaction phenomena between naphthalene and the EHC-500 electrode after discharged and charged to different potentials were investigated. Quasi-metallic Na will react with naphthalene in dimethylether (DME) solution to form Na-naphthalene compound, which appears color in DME solution. As shown in Fig. 6a, the naphthalene solution remains transparent and colorless when the electrode discharged to 0.07 V. However, it appears light yellow when discharged to 0.03 V, and the solution color deepens when discharged to 0.001 V, suggesting that the potential scope of 0.08–0.04 V is corresponding to the insertion of Na⁺ into the hard carbon layers and the sodium storage mechanism below 0.04 V can be correlated with the filling of Na⁺ into the closed pores to form quasi-metallic Na (Fig. 6a) [43]. Due to the voltage polarization, quasi-metallic Na still exists in the closed pores of EHC-500 after charged to 0.07 V. And the naphthalene solution with the EHC-500 electrode remains yellow after charged to 0.07 V. Based on the above analysis, the capacity of the plateau region can be separated into two parts. One part is the capacity between 0.08 V and 0.04 V attributed to the sodium ion intercalation, and the other part is the capacity below 0.04 V due to the closed pore filling (Fig. 6b). The former is 134.4 and 126.0 mAh/g for EHC-500 and HC-500, respectively. And the latter is 102.3 and 66.3 mAh/g for EHC-500 and HC-500, respectively (Fig. 6c), indicating that EHC-500 is superior to HC-500 in terms of the capacity of sodium ion intercalation as well as the capacity of closed pore filling. As the proportion of pseudo-graphitic phase in EHC-500 is higher than that of HC-500, more sodium storage sites are existed in EHC-500. In addition, EHC-500 exhibits more closed nanopores than HC-500 according to the SAXS results, which implies higher capacity of closed pore filling for EHC-500 than HC-500. Moreover, the L_a value is proven to be an important factor influencing the accessibility of sodium ions to the intercalation sites in carbon layers and filling sites in closed pores [39]. Therefore, the smaller L_a value in EHC-500 would greatly promotes the full utilization of the sodium storage sites of plateau region in comparison with HC-500. Based on the above analysis, schematic illustration of the effect of microstructure variation on the sodium storage behaviors of EHC-500 and HC-500 is shown in Fig. 6d.

  Figure 6

  

  Figure 6.  (a) The discharge curves of EHC-500 and the corresponding color changes of DME containing naphthalene after reaction with the EHC-500 electrode at different potentials. (b) The second cycle discharge curves of EHC-500 and HC-500. (c) The plateau capacity calculated from the second cycle of the discharge curves. (d) Schematic illustration of the effect of microstructure variation on the sodium storage behaviors of EHC-500 and HC-500.

  DownLoad: Full-Size Img PowerPoint

  To explore the potential of EHC-500 in practical application, the full cell was assembled by using Na₃V₂(PO₄)₃ (NVP) as the cathode and EHC-500 as the anode (Fig. S19 in Supporting information). The full cells are evaluated in a voltage range of 1.5–3.9 V and the optimized capacity ratio of the positive electrode to negative electrode is 1.2. The full cell exhibits a high average working voltage of 3.3 V and achieves a high reversible capacity of 297 mAh/g (based on the mass load of anode material) after 15 cycles at 30 mA/g. Accordingly, the energy density based on anode and cathode mass loading is determined to be 203 Wh/kg. These outcomes suggest that EHC-500 displays significant potential for practical SIBs application.

  In summary, hard carbon materials with tunable microcrystalline structures (the content and the size of pseudo-graphite domains) and pore structures were synthesized by using CMC–Na as the precursor. The effect of esterification modification and pre-carbonization temperature on the microstructures of the final hard carbons was studied. The esterification modification can effectively improve the structural stability of CMC–Na during the pre-carbonization process and hinder the overgrowth and agglomeration of Na₂CO₃ particles in the carbon matrix, which can act as a salt template to introduce large amount of smaller and more uniform-sized closed pores in hard carbons and as a catalyst to promote the local graphitization of hard carbons around them and then regulate the microcrystalline structures of hard carbons. High pre-carbonization temperature is also conducive to the formation of smaller and more uniform-sized Na₂CO₃ particles in the carbon matrix. The optimized sample EHC-500 presents abundant uniformly dispersed closed pores, maximized pseudo-graphitic content (73%), and minimized lateral width (L_a) of pseudo-graphitic domains, which will not only greatly increase the sodium storage sites of plateau region but also improve the accessibility of sodium ions to the intercalation sites in carbon layers and filling sites in closed pores. As an anode material, EHC-500 exhibits a high reversible capacity of 340 mAh/g, an ICE value of 72.1%, and excellent cycling performance. Furthermore, the assembled EHC-500//NVP full cell delivers a high energy density of 203 Wh/kg and a high average working voltage of 3.3 V. This work provides a novel route to the rational regulating of microcrystalline structures and pore structures of hard carbons towards enhanced sodium storage performance, and promotes the understanding of sodium storage behaviors in hard carbons.

  CRediT authorship contribution statement

  Guang Zeng: Writing – original draft, Methodology, Investigation, Conceptualization. Yue Zeng: Writing – review & editing, Investigation. Huamin Hu: Writing – review & editing, Investigation. Yaqing Bai: Writing – review & editing, Investigation. Fangjie Nie: Writing – review & editing, Investigation. Junfei Duan: Writing – review & editing, Methodology. Zhaoyong Chen: Writing – review & editing, Methodology, Conceptualization. Qi-Long Zhu: Writing – review & editing, Investigation, Conceptualization.

  Acknowledgments

  The authors are grateful for the financial support of the National Natural Science Foundation of China (NSFC, No. 21905278), the Natural Science Foundation of Hunan Province (No. 2023JJ30015).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110122.

  
  
• 1. [1]

     Q. Meng, Y. Lu, F. Ding, et al., ACS Energy Lett. 4 (2019) 2608–2612. doi: 10.1021/acsenergylett.9b01900

  2. [2]

     Q. He, H. Chen, X. Chen, et al., Adv. Funct. Mater. 34 (2024) 2310226.

  3. [3]

     H. Wang, H. Chen, C. Chen, et al., Chin. Chem. Lett. 34 (2023) 107465.

  4. [4]

     C. Cai, Y. Chen, P. Hu, et al., Small 18 (2022) 2105303.

  5. [5]

     M. Song, L. Xie, F. Su, et al., Chin. Chem. Lett. 35 (2024) 109266.

  6. [6]

     H. Yamamoto, S. Muratsubaki, K. Kubota, et al., J. Mater. Chem. A 6 (2018) 16844–16848. doi: 10.1039/c8ta05203d

  7. [7]

     Y. Zhang, Y. Meng, Y. Wang, et al., ChemElectroChem 4 (2017) 500–507.

  8. [8]

     X. Qi, K. Huang, X. Wu, et al., Carbon 131 (2018) 79–85.

  9. [9]

     Y. Xu, Y. Zhang, Mater. Lett. 139 (2015) 145–148. doi: 10.14801/jkiit.2015.13.3.145

  10. [10]

      N. Cheng, W. Zhou, J. Liu, et al., Nano-Micro Lett. 14 (2022) 146.

  11. [11]

      X. Zhou, L. Guo, Q. Wang, et al., J. Electroanal. Chem. 889 (2021) 115179.

  12. [12]

      Y. Zhao, Z. Hu, C. Fan, et al., Small 19 (2023) 2303296.

  13. [13]

      Y. Zuo, X. He, P. Li, et al., Polymers 11 (2019) 72. doi: 10.3390/polym11010072

  14. [14]

      Y. Zuo, J. Gu, L. Yang, et al., Int. J. Biol. Macromol. 64 (2014) 174–180.

  15. [15]

      M.X. Song, L.J. Xie, J.Y. Cheng, et al., J. Energy Chem. 66 (2022) 448–458.

  16. [16]

      Y. Cao, L. Xiao, M.L. Sushko, et al., Nano Lett. 12 (2012) 3783–3787. doi: 10.1021/nl3016957

  17. [17]

      M. Yuan, B. Cao, H. Liu, et al., Chem. Mat. 34 (2022) 3489–3500. doi: 10.1021/acs.chemmater.2c00405

  18. [18]

      L. Wang, Q. Zhu, J. Zhao, et al., Microporous Mesoporous Mat. 279 (2019) 439–445.

  19. [19]

      Y. Li, Y. Lu, Q. Meng, et al., Adv. Energy Mater. 9 (2019) 1902852.

  20. [20]

      D.A. Stevens, J.R. Dahn, J. Electrochem. Soc. 147 (2000) 4428.

  21. [21]

      A. Sadezky, H. Muckenhuber, H. Grothe, et al., Carbon 43 (2005) 1731–1742.

  22. [22]

      M.W. Smith, I. Dallmeyer, T.J. Johnson, et al., Carbon 100 (2016) 678–692.

  23. [23]

      C. Hu, S. Sedghi, A. Silvestre-Albero, et al., Carbon 85 (2015) 147–158.

  24. [24]

      Z. Xing, Y. Qi, Z. Tian, et al., Chem. Mat. 29 (2017) 7288–7295. doi: 10.1021/acs.chemmater.7b01937

  25. [25]

      S. Zhang, Q. Liu, H. Zhang, et al., Carbon 157 (2020) 714–723.

  26. [26]

      P. Bai, Y. He, X. Zou, et al., Adv. Energy Mater. 8 (2018) 1703217.

  27. [27]

      L. Kitsu Iglesias, E.N. Antonio, T.D. Martinez, et al., Adv. Energy Mater. 13 (2023) 2302171.

  28. [28]

      J. Xia, N. Zhang, S. Chong, et al., Green Chem. 20 (2018) 694–700. doi: 10.1039/c7gc03426a

  29. [29]

      S. Wang, L. Xia, L. Yu, et al., Adv. Energy Mater. 6 (2016) 1502217.

  30. [30]

      P. Yang, Z. Wu, S. Wang, et al., Angew. Chem. Int. Ed. 62 (2023) e202311460.

  31. [31]

      M. Wang, Z. Yang, W. Li, et al., Small 12 (2016) 2559–2566. doi: 10.1002/smll.201600101

  32. [32]

      Y. Lu, C. Zhao, X. Qi, et al., Adv. Energy Mater. 8 (2018) 1800108.

  33. [33]

      M. Song, Z. Yi, R. Xu, et al., Energy Storage Mater. 51 (2022) 620–629.

  34. [34]

      Y.P. Gao, Z.B. Zhai, K.J. Huang, et al., New J. Chem. 41 (2017) 11456–11470.

  35. [35]

      P. Zhang, D. Wang, Q. Zhu, et al., Nano-Micro Lett. 11 (2019) 81.

  36. [36]

      G. Zeng, B. Zhou, L. Yi, et al., Sustain. Energ. Fuels 2 (2018) 855–861. doi: 10.1039/c7se00517b

  37. [37]

      X. Pu, D. Zhao, C. Fu, et al., Angew. Chem. Int. Ed. 60 (2021) 21310–21318. doi: 10.1002/anie.202104167

  38. [38]

      C. Nita, B. Zhang, J. Dentzer, et al., J. Energy Chem. 58 (2021) 207–218.

  39. [39]

      L. Yang, M. Hu, Q. Lv, et al., Carbon 163 (2020) 288–296.

  40. [40]

      S. Qiu, L. Xiao, M.L. Sushko, et al., Adv. Energy Mater. 7 (2017) 1700403.

  41. [41]

      Y. Jin, S. Sun, M. Ou, et al., ACS Appl. Energ. Mater. 1 (2018) 2295–2305. doi: 10.1021/acsaem.8b00354

  42. [42]

      D.S. Bin, Y. Li, Y.G. Sun, et al., Adv. Energy Mater. 8 (2018) 1800855.

  43. [43]

      X. Chen, C. Liu, Y. Fang, et al., Carbon Energy 4 (2022) 1133–1150. doi: 10.1002/cey2.196

• 

  Figure 1  Schematic illustration of the routes to fabricate CMC–Na-based hard carbons with different pore structures and pseudo-graphite phases.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) TG curves of esterified CMC–Na and the mixture of CMC–Na and maleic anhydride. (b) XRD patterns. (c) Corresponding phase contents from the XRD fitting results. SEM images of (d) EHC-500 and (g) HC-500. HRTEM images and SAED images of (e) EHC-500 and (f) HC-500. Schematic representation of the pseudo-graphite domain sizes in (h) EHC-500 and (i) HC-500.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) N₂ adsorption and desorption isotherms. (b) Pore size distribution. (c) SAXS patterns of EHC-500 and HC-500. (d) Raman spectra and fitted curves of EHC-500 and HC-500. (e) C 1s spectra and fitted curves of EHC-500 and HC-500. (f) O 1s spectra and fitted curves of EHC-500 and HC-500.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) Cyclic voltammetry curves of EHC-500 at a scanning rate of 0.1 mV/s. Electrochemical Na-storage performances of the carbon samples: (b) GCD curves at 30 mA/g; (c) Plateau and slope capacity contribution corresponding to the second cycle discharge curve; Cycling performance at the current densities of (d) 30 and (e) 300 mA/g; (f) Rate performance.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  EHC-500: (a) CV curves at different scan rates; (b, c) Linear relationship between ln(i) and ln(v). HC-500: (d) CV curves at different scan rates; (e, f) Linear relationship between ln(i) and ln(v). (g) EIS. GITT curves of (h) EHC-500 and (i) HC-500.

  下载: 全尺寸图片 幻灯片
  

  Figure 6  (a) The discharge curves of EHC-500 and the corresponding color changes of DME containing naphthalene after reaction with the EHC-500 electrode at different potentials. (b) The second cycle discharge curves of EHC-500 and HC-500. (c) The plateau capacity calculated from the second cycle of the discharge curves. (d) Schematic illustration of the effect of microstructure variation on the sodium storage behaviors of EHC-500 and HC-500.

  下载: 全尺寸图片 幻灯片
•