Constructing Carbon Nanobubbles with Boron Doping as Advanced Anode for Realizing Unprecedently Ultrafast Potassium Ion Storage

Carbonaceous material with favorable K+ intercalation feature is considered as a compelling anode for potassium‐ion batteries (PIBs). However, the inferior rate performance and cycling stability impede their large‐scale application. Here, a facile template method is utilized to synthesize boron doping carbon nanobubbles (BCNBs). The incorporation of boron into the carbon structure introduces abundant defective sites and improves conductivity, facilitating both the intercalation‐controlled and capacitive‐controlled capacities. Moreover, theoretical calculation proves that boron doping can effectively improve the conductivity and facilitate electrochemical reversibility in PIBs. Correspondingly, the designed BCNBs anode delivers a high specific capacity (464 mAh g−1 at 0.05 A g−1) with an extraordinary rate performance (85.7 mAh g−1 at 50 A g−1), and retains a considerable capacity retention (95.2% relative to the 100th charge after 2000 cycles). Besides, the strategy of pre‐forming stable artificial inorganic solid electrolyte interface effectively realizes high initial coulombic efficiency of 79.0% for BCNBs. Impressively, a dual‐carbon potassium‐ion capacitor coupling BCNBs anode displays a high energy density (177.8 Wh kg−1). This work not only shows great potential for utilizing heteroatom‐doping strategy to boost the potassium ion storage but also paves the way for designing high‐energy/power storage devices.

Recently, the beforementioned problem has been remarkably alleviated by doping heteroatoms (N, S, and B) into carbon-based anode, with the aim of extending the interlayer spacing and facilitating reversible K + diffusion. [30][31][32][33][34] For example, Chen et al. developed N-doped hollow carbon nanospheres for PIBs, in which the hollow structure alleviated the volume expansion and achieved superior K + storage performance (154.0 mAh g −1 at 1.0 A g −1 upon 2500 cycles). [35] Mitlin et al. demonstrated that S-grafted carbon can minimize capacity fading and Coulombic efficiency (CE) loss, maintaining 93.0% capacity retention from the 5th to 1000th cycle at 3.0 A g −1 , with a stable CE of 100%. [36] Lu et al. reported boron doping porous carbons, in which boron doping ensures fast K + ion diffusion kinetics and excellent rate capability (428 mAh g −1 at 100 mA g −1 ) for PIBs. [34] From recent studies, we know that the three valence electrons of boron enable the Fermi level shift toward the valance band, leading to a higher carrier concentration and density of states at the Fermi level, and jointly enhancing the supercapacitive properties, which is not reflected in N and/or S-doped composites. [37][38][39] Moreover, the boron heteroatoms feature a similar atomic radius to carbon, allowing them to easily enter the carbon lattice. Despite some progresses realized in the realm of boron doping sofar, the reported synthetic routes, such as chemical vapor deposition, [40] and hydrothermal reaction, [41] are too complicated to meet practical demands. Therefore, advanced design of boron doping carbon anode toward the practical PIBs/PICs remains a formidable challenge.
Here, inspired by the doping strategy, we develop a template strategy to synthesize a boron doping porous carbon anode with interconnected nanobubbles-like structure using commercial MgO powder as the sacrificial template, which endows the PIBs with relatively high specific capacity (464 mAh g −1 at 0.05 A g −1 ), superior rate properties (85.7 mAh g −1 at 50 A g −1 ), and outstanding cycle stability (156.3 mAh g −1 at 10 A g −1 after 2000 cycles). In addition, theoretical calculation reveals that the introduction of B atoms can achieve high electrical conductivity and ensure superior electrochemical reversibility in PIBs, harvesting high intercalation-and capacitive-controlled capacities. Moreover, the as-built PIC by employing boron doping carbon nanobubbles (BCNBs) as anode achieves a high energy density of 177.8 Wh kg −1 with low capacity decay of merely 0.0147% per cycle during 2000 cycles at 20 A g −1 .

Results and Discussions
A typical synthetic route for BCNBs is exemplified in Figure 1a. The proposed strategy was achieved by simple ball milling of phenylboronic acid (C 6 H 7 BO 2 ) precursor and magnesium oxide (MgO) template with a fixed ratio of 2:1, followed by annealing at 1000°C for 2 h in Ar atmosphere. Ball-milling was employed to guarantee the uniform mixing of the C 6 H 7 BO 2 precursor and MgO template. C 6 H 7 BO 2 in its oblique crystal form was applied as carbon/boron sources, which can be decomposed into B-and O-doped carbon by growing on the surface of MgO at the target carbonization temperature. The MgO template can be beneficial for catalyzing the growth of carbon layer, and the removal of the template leads to the formation of BCNBs with hollow interiors. Analogously, carbon nanobubbles (CNBs) without boron doping were also produced for comparison.
The scanning electron microscopy (SEM) images of BCNBs illustrate a uniform nanobubble structure (Figure S1 Supporting Information, Figure 1b), and the overlapped nanobubbles can be beneficial for accommodating the volume expansion caused by K + insertion/extraction, so as to further improve the structural stability. For comparison, the CNBs also have similar nanobubble-like structure to BCNBs ( Figure S2, Supporting Information). The typical transmission electron microscope (TEM) image in Figure 1c further demonstrates the porous nanobubble nature of BCNBs. A representative high-resolution TEM image of BCNBs is described in Figure 1d, revealing lattice fringes with a spacing of 0.35 nm as the result of the expansion of the (002) crystal plane of graphite, which is conducive to the insertion of K ions. [42] Figure 1e-h present the energy-dispersive spectroscopy (EDS) elemental mapping of the BCNBs, illustrating that the B atoms are successfully incorporated into the BCNBs.
X-ray diffraction (XRD) is conducted to investigate the structure of as-prepared carbons. As displayed in Figure 2a, the (002) peak of BCNB is centered at 25.4 o , and the related interlamellar distance is 0.35 nm as evaluated by Bragg's law (2dsinθ ¼ nλ), in good agreement with the TEM result. For CNBs, the (002) peak shows a more pronounced left shift in comparison with that of BCNBs, with a larger interlayer spacing of 0.36 nm. In addition, the (002) peak of BCNBs is sharper than that of CNBs, suggesting that the improved graphitization degree can be ascribed to the incorporation of B atoms. Further investigation on the graphitization degree of the BCNBs and CNBs was carried out by using Raman spectra. As shown in Figure 2b and Figure S3, Supporting Information, the two obvious peaks located at~1340 and 1580 cm −1 are ascribed to the D and G band signals of carbon species. [31,[43][44][45] Also, the intensity ratio of I G /I D can reflect the order degree of carbon materials. [46,47] Notably, a significant increase of I G /I D value from 0.56 for CNBs to 0.77 for BCNBs suggests the enhanced structural order of carbon network by B doping, [48] and the increased graphitization degree can improve the conductivity.
To further recognize the porosity information of as-designed products, nitrogen adsorption-desorption analysis was conducted (Figure 2c, Figure S4, Supporting Information). Both samples display a typical IVtype sorption curve, indicating a mesoporous structure. The specific surface area of CNBs is 578 m 2 g −1 , while it decreases to 271 m 2 g −1 for BCNBs (Table S1, Supporting Information). C 6 H 7 BO 2 acts as a crosslinker to solidify the epoxy resin, which can enhance the order degree of the precursor, accommodate the volume expansion during annealing and elevate the degree of graphitization after pyrolysis, thus contributing to a lower specific surface area. Intriguingly, the total pore volume of BCNBs is as high as 1.8 cm 3 g −1 , which is much higher than that of CNBs (1.0 cm 3 g −1 ). The pore size distribution curves of the two samples ( Figure 2c, Figure S4, Supporting Information, inset) demonstrate that plenty of mesoporous and macroporous exist in BCNBs while CNBs mainly composed of micropores. The porous structure results from the generation of a vast amount of gas during the pyrolysis process. Moreover, the appearance of relatively large pores in BCNBs can ensure the rapid diffusion of large K + in the electrolyte, thus improving the performance at higher charge/discharge rates. The lowsurface area of BCNBs enables a favorable higher packing density of 0.92 g cm −3 than that of CNBs (0.68 g cm −3 ).
The surface composition of BCNBs and CNBs samples were further analyzed by X-ray photoelectron spectrometer (XPS) analysis.  [42,[48][49][50] proving the successful introduction of B. The sp 2 / (sp 2 + sp 3 ) ratio increases from 66.4% of CNBs to 74.1% of BCNBs (Table S2, Supporting Information; Figure S5c, Supporting Information), further demonstrating the increase of graphitic region by boron doping, which is in accordance with XRD and Raman analysis. Additionally, the O 1 s spectra of BCNBs can be fitted by three peaks, which are centered at around 530.9, 532.5, and 535.1 eV, corresponding to the C=O, C-OH/C-O-C, and COOH bonds, [5,49] respectively, as shown in Figure 2e and Figure S5d, Supporting Information. Compared with CNBs (5.1 at.%), the mass ratio of O in BCNBs is significantly decreased (2.7 at.%), and the B-doping with the low content of O atoms can be beneficial for enhancing the electrical conductivity. [51] The B 1 s XPS spectrum of BCNBs can be deconvoluted into four peaks, including BC 3 (189.1 eV), BC 2 O (190.8 eV), BCO 2 (192.9 eV), and B 4 C (187.2 eV), [42,48] accounting for around 54.6, 18.6, 18.6, and 8.2 at.%, respectively, as shown in Figure 2f. It is clear that BC 3 is the main configuration, in which electron-deficient B substitutes C in graphene. The BC 3 structure can perform as electron acceptors which modify the electronic structure of the original graphene, leading to enhanced charge storage properties of the BCNBs. [52] Coin cells were assembled to evaluate the potassium ion storage capability of BCNBs and CNBs anodes using K foil as the counter electrode. Figure 3a reveals the first five cyclic voltammetry (CV) profiles of the BCNBs electrode between 0.001 and 3.0 V at a scan rate of 0.1 mV s −1 . The observed cathodic peak (0.85 V) in the first cycle disappears in the following cycle, which is associated with the occurrence of irreversible reactions and the formation of a solid electrolyte interface (SEI) film. [43] Also, there is a pair of reduction/oxidation peaks at 0.2/ 0.5 V, which is ascribed to the insertion/extraction of K + . The anodic scan features an invertible peak at~1.2 V related to the interaction of K + with B-containing functional group. In the subsequent cycles, the CV profiles almost overlap, manifesting the excellent reversibility of BCNBs. As a contrast, CNBs anode shows a similar behavior to BCNBs, but there is no oxidation peak observed at~1.2 V ( Figure S6a, Supporting Information). Figure 3b shows the discharge-charge profiles of BCNBs with traditional SEI film at a current density of 0.05 A g −1 . The first charge capacity was recorded to be 452.7 mAh g −1 , which surpassed the CNBs of 231.8 mAh g −1 ( Figure S6b). Most obviously, boron doping achieves a considerable improvement in electrochemical performance. It is noteworthy that the BCNBs and CNBs exhibit fairly low initial CE of 23.1% and 32.0%, respectively. The low initial CE is mainly ascribed to increased contact between electrode material and electrolyte induced by high-specific surface area and porosity, causing some side reactions. These side reactions and the irreversible formation of SEI films consume K + from the electrolyte, causing capacity decay. Therefore, it is of great interest to improve the initial CE in the K-based storage systems. Following this idea, an artificial SEI (A-SEI) was constructed on the BCNBs electrode to tackle this issue based on previous research. [53] A-SEI film was obtained by keeping BCNBs electrode in direct contact with K metal foil and soaking them in 2 M KFSI-DME electrolytes for 12 h. In the above process, A-SEI film forms spontaneously on the BCNBs electrode surface through the reaction of K metal and concentrated KFSI in DME. The above experimental process was carried out in an argon-stuffed glove box ((O 2 ) < 0.01 ppm, (H 2 O) < 0.01 ppm). A-SEI processing is very effective to improve initial CE. For BCNBs with A-SEI film, the discharge/charge capacity of 656/516 mAh g −1 was obtained (Figure 3c), corresponding to a 79.0% CE in the first cycle. Notably, the initial CE significantly improves from 23.1% to 79.0%, which is attributed to formation of a relatively intact SEI film in the pretreatment stage.
The rate performance of BCNBs as well as the counterparts of A-SEI BCNBs and CNBs are demonstrated in Figure 3d. BCNBs present a high discharge capacity of 464 mAh g −1 at 0.05 A g −1 , which declines slowly to 374.1, 318.4, 271.2, 248.7, 234.2, 213.7, 199.5, 139.7, 85.7, and 39.2 mAh g −1 as the current density increases to 0.1, 0.2, 0.5, 1, 2, 5, 10, 20, 50, and 100 A g −1 , respectively. Once the current density returns to 50, 20, 10, 5, 2, 1, 0.5, 0.2, 0.1, and 0.05 A g −1 , the discharge capacity can be resumed to its initial value without any sign of capacity loss. Such extraordinary rate capability has been rarely achieved for carbon anodes in metal-ion batteries. Moreover, A-SEI BCNBs and BCNBs exhibit similar rate performance, further revealing the advantage of the formation of an artificial SEI film for improving initial CE. The excellent rate performance for the BCNBs electrode is superior to the previously reported works, as stated in Figure 3e and Table S3, Supporting Information. In contrast, the CNBs electrode only achieves a relative low discharge capacity of 232.5 mAh g −1 at a current density of 0.05 A g −1 . Moreover, the volumetric specific capacity of BCNBs can reach 426.9 mAh cm −3 at a current density of 0.05 A g −1 and even 36.1 mAh cm −3 at 100 A g −1 ( Figure S7, Supporting Information), far surpassing the corresponding values of CNBs.
Then, taking 0.1 V versus K/K + as the boundary, the potassiation mechanisms of BCNBs and CNBs can be divided into two stages. The partition of 0.1 V is a little less rigorous, but it is quite effective for discussing different storage mechanisms. The sloping voltage region above 0.1 V is ascribed to the defect adsorption process (termed as stage a, capacitive contribution) and stage b is attributed to the K + intercalation into a given graphite interlayer (intercalated contribution) ( Figure S8, Supporting Information). [54,55] The capacities related to the two different storage mechanisms are then quantified and analyzed in detail (Figure S9, Supporting Information). As for stage b of BCNBs, intercalation is favored due to the increased graphitization and conductivity compared to that of CNBs, resulting in a higher capacity. Furthermore, the BCNBs electrode achieves a higher potassium storage capability in stage a, manifesting that the doping B atoms could adjust the electronegativity of carbon, thus presenting a higher capacity ration of K + adsorption. Interestingly, from the perspective of proportion, a higher percentage of K-ion intercalation at current densities ranging from 0.05 to 0.5 A g −1 is also observed owing to more orderly carbon structure and higher conductivity of BCNBs than CNBs. The pivotal disparity between BCNBs and CNBs is the B content, which is 1.1 at.% versus nil. Hence, it is reasonable to speculate that boron doping plays an essential role in enhancing K-storage, although CNBs have larger interlayer spacing and specific surface area. The electrochemical reaction mechanism would be explored via a series of ex-situ characterizations and density functional theory (DFT) in the next section.
Based on the 3D nanobubbles-like structure, both BCNBs and CNBs exhibit stable cycling performance. However, the BCNBs electrode presents a much higher reversible capacity of 156.3 mAh g −1 after 2000 cycles at 10 A g −1 (Figure 3f), showing a considerable capacity retention of 95.2% relative to the 100th charge. Electrochemical impedance spectroscopy (EIS) analysis was carried out to illustrate the kinetic characteristics ( Figure S10a,b, Supporting Information), and the equivalent circuit models are displayed in Figure S10c, Supporting Information, where R e represents the internal resistance, and R ct + R f links to the resistance correlated with the charge transfer resistance from the electrolyte/SEI interface and the transfer resistance of potassium through the SEI layer. [49] As listed in Table S4, Supporting Information, the R ct + R f values of BCNBs and CNBs present a significant downward trend with cycling, from 5167/4907 Ω in the initial state to 503/ 909 Ω after 100 cycles and 231/660 Ω after 2000 cycles, which are ascribed to the entire infiltration of the electrode material by the electrolyte and the completion of the activation process.
Besides, the D K þ is further studied by EIS with the following formula: where R is gas constant, T is absolute temperature, S is surface area, n is the number of charge transfers, F is Faraday's constant, and C is the potassium-ion concentration. The Weber factor σ is determined by the slope of Z 0 against ω À1=2 . As demonstrated in Figure S11, Supporting Information, σ of BCNBs electrode is  In views of the above analysis, high K + diffusion ability could be established for BCNBs, enabling superior electrochemical reaction kinetics. Moreover, galvanostatic intermittent titration technique (GITT) analysis was also conducted to elucidate the kinetics of BCNBs and CNBs anodes for potassium storage ( Figure S12, Supporting Information, Figure 3g). The BCNBs electrode presents higher K + diffusion coefficients (D K þ ) (2.8 × 10 −12 -3.6 × 10 −11 cm 2 s −1 for the discharged state; 6.8 × 10 −12 -5.3 × 10 −11 cm 2 s −1 for the charged state) than those of the CNBs electrode (1.4 × 10 −12 -1.5 × 10 −11 cm 2 s −1 for the discharged state; 2.6 × 10 −12 -1.9 × 10 −11 cm 2 s −1 for the charged state). These results demonstrate that BCNBs harvests faster K + diffusion kinetics in conformity with the EIS analysis. Moreover, the diffusion coefficient outperforms other reported electrode materials for potassium storage, such as H-FeS 2 @3DCS (1.4 × 10 −13 -5.3 × 10 −12 cm 2 s −1 for the discharged state; 8.1 × 10 −13 -4 × 10 −12 cm 2 s −1 for the charged state), [13] N/O co-doped porous carbon spheres (2 × 10 −13 -8.3 × 10 −12 cm 2 s −1 for the discharged state; 1.1 × 10 −12 -9.2 × 10 −12 cm 2 s −1 for the charged state), [56] and S-doped N-rich carbon (6.0 × 10 −14 -2.7 × 10 −11 cm 2 s −1 for the discharged state; 6.1 × 10 −13 -2.0 × 10 −11 cm 2 s −1 for the charged state), [43] leading to a conclusion that BCNBs achieve superior kinetics. The kinetic analysis of BCNBs/CNBs anode was also performed using CV curves at various scan rates ( Figure S13, Supporting Information). Apparently, the current intensity increases with the increase of the scan rates, further accounting for the surface-dominated characteristics. [47,57] Commonly, the peak current i ð Þ and scan rate v ð Þ can be related to the power-law relationship of i ¼ av b , [58][59][60][61] where a, b are defined as variable constants and the b-value can be collected by the slope of the log v ð Þ and log i ð Þ. To be precise, the b-value of 0.5 implies an ideal diffusion-determined process, whereas the b-value of 1 means a reaction-controlled capacitive behavior. [62,63] The calculated anodic bvalues are 0.80 (peak 1) and 0.86 (peak 2) for BCNBs, signifying a cocontrolled process of capacitive-and diffusion-controlled reactions. The higher b-values of CNBs (0.90 for peak 1 and 0.89 for peak 2) suggest that the capacitive-controlled process is dominant to a larger extent. Furthermore, the reaction-controlled contribution ratio under varying scan rates can be quantified by separating the response current (i) at a fixed voltage (V) into capacitive (k 1 v) and diffusion-controlled reactions (k 2 v 1=2 ), as the following equation: [64][65][66] For BCNBs, as the scan rate increases, the ratio of capacitive contribution rises, that is, 57.0%, 65.0%, 71.2%, 73.1%, and 75.4% at the scan rates of 0.2, 0.4, 0.6, 0.8, and 1 mV s −1 , as depicted in Figure 3h. As for CNBs, the ratio of capacitive contribution is higher, that is, 70.3%, 73%, 75.6%, 81.6%, and 84.1% with increasing scan rates. The relatively low fraction of capacitive-dominated processes for BCNBs is related to the increased graphitization and conductivity, that is in consonance with the higher proportion of intercalation behavior analyzed previously.
To delve into the K-ion storage behavior of BCNBs, ex situ XPS analysis was performed. Figure 4b shows the K 2p signal for different (de) potassiation states. For the pristine electrode, no K signal is observed. The two featured K 2p peaks located at 293.2 and 296.0 eV during potassiation processes are ascribed to the formation of KC x compounds, proving a successful K-ion insertion. [67] During the depotassiation process, the intensity of K 2p peak decreases, reflecting the reversible release of K + from carbon framework. Figure 4c exhibits the B 1 s signal for these different states to illuminate the potassiation process of the diverse B groups. At 0.75 V, the XPS spectra indicated that four initial peaks remain, along with four extra peaks (BC 3 /K at 188.6 eV, BC 2 O/ K at 189.9 eV, BCO 2 /K at 192.3 eV, and B 4 C/K at 186.6 eV), confirming the potassiation of the functional group of boron. At 0.001 V, potassiation of all of the B groups was accomplished, and the binding energies of BC 3 , BC 2 O, BCO 2 , and B 4 C transfer from the primary values of 189. 1, 190.8, 192.9, and 187.2 eV to 188.6, 189.9, 192.3, and 186.6 eV, respectively, denoting the formation of K-protonated B structures in the carbon matrix. As the BCNBs electrode is gradually depotassiated to 0.5 V, BCO 2 /K peaks still exists, while the other three peaks revert to the pristine states owing to K + -ion deintercalation. Upon charging to 3 V, all of the B groups are in accordance with the initial state, which provides a forceful evidence of good reversibility during the cycling process. To further disclose the potassium ion storage mechanisms, ex situ high-resolution TEM was performed on BCNBs electeode at fully potassiation state (0.001 V) and fully depotassiation state (3 V). An obvious expanded lattice fringe of 0.41 nm was observed when the BCNBs electrode was discharged from open-circuit voltage to 0.001 V (Figure 4d), certificating that the electrode undergoes K + intercalation, agreeing with the ex situ XPS results. In the fully depotassiation state, the lattice fring is reduced to 0.36 nm, suggesting the good reversibility of the process (Figure 4e).
Based on the above characterization and analysis, it can be concluded that the superior electrochemical performance of BCNBs is ascribed to the improved intercalation and capacitive behavior due to boron atom doping. To further assess the effect of boron doping on the K-ion storage and conductivity of carbon, DFT calculation was carried out. According to the graphene structure, three configurations (carbon, Bdoped carbon, and K-intercalated B-doped carbon, see Figure S14a-c, Supporting Information) were established with BC 3 functional group as the model for investigating the effect the boron atoms doping. The DOS of CNBs tends to zero near the Fermi level (Figure 4f), while the intensity improves greatly after boron incorporation (Figure 4g), indicating that B-doping effectively improves the conductivity. It should be noticed that, when K + isintercalated, the DOS is nearly unchanged compared with the initial state of B-doped carbon, as shown in Figure 4h, indicating the ultra-high stability of our designed carbon-based electrode material. Based on the theoretical analysis, it can be confirmed that B-doped carbon electrode that we designed increases both electrical conductivity and reversibility.
In order to assess the role of B concentration in electrochemical performance, BCNBs were prepared using different calcination temperatures. As indicated by the XPS analysis ( Figures S5, S15, S16, Supporting Information), it is worth noting that the amount of B doping decreases with the increase of temperature, that is, 1.3 at.% in BCNBs-900, 1.1 at.% in BCNBs, and 0.2 at.% in BCNBs-1100. It should be pointed out that the lower doping content of B induces a reduced contribution to the electrochemical properties. Moreover, we Energy Environ. Mater. 2023, 6, e12559 6 of 10 can also observe that the degree of graphitization is increased with the increase of calcination temperature based on the XPS (Table S2, Supporting Information) and Raman results ( Figures S3 and S17, Supporting Information). The improved degree of graphitization can enhance the electrical conductivity of BCNBs electrodes and increase K-ion intercalation capacity. However, the higher degree of graphitization usually indicates the narrower lattice spacing of graphitic layer, which can also hinder K-ion intercalation kinetics. As a result, the balance between high B content and adequate degree of graphitization provides an opportunity to boost the electrochemical performance. In view of these advantageous features, the resulting BCNBs prepared at 1000°C achieves optimal K + storage performance as an anode for PIBs (Figure S18, Supporting Information).
To further evaluate the potential application of BCNBs, a PIC is fabricated using BCNBs as anode and N,S co-doped porous carbon (NSPC) as cathode with 1 M KFSI/EC-DEC electrolyte. As described in Figure  5a, during the charging-discharging process, the BCNBs anode achieves reversible K + storage, while the NSPC cathode undergoes capacitive behavior with FSI − anions. Based on the previous result, [13] the NSPC cathode with high surface area (>2000 m 2 g −1 ) and hierarchical porous structure can realize a high and fast adsorption capability of FSI − ions. In addition, the presence of a large number of Ncontaining and S-containing functional groups can also promote the electronic conductivity and provide extra adsorption sites to enhance the capacity. Therefore, the NSPC is selected as a prospective cathode material for PICs in the present work. The NSPC cathode operated in the voltage window of 2.0-4.0 V achieves excellent rate capacity of 69.6, 52.7, 46.0, 40.9, 38.4, and 37.2 mAh g −1 at current densities of 0.05, 0.1, 0.2, 0.5, 1.0, and 2.0 A g −1 , respectively, as presented in Figure S19, Supporting Information. The symmetrical and highly linear charge/discharge curves without obvious platform denote that NSPC electrode possesses a typical capacitive behavior as ideal capacitor-type cathode material. Figure 5b shows typical CV profiles for the BCNBs// NSPC-based PICs at various scan rates from 5 to 50 mV s −1 between 0.0 and 4.0 V, which displays a nonideal rectangular shape owing to the combined storage mechanism. Impressively, the BCNBs//NSPC PICs can be effectively cycled under different current densities ranging from 0.5 to 10 A g −1 , realizing high energy densities of 177.8 Wh kg −1 at 1000 W kg −1 and 55.2 Wh kg −1 at 20000 W kg −1 (Figure 5c). The performance of BCNBs//NSPC PICs is superior to previously reported energy storage devices, as listed in Figure 5d and Table S5, Supporting Information. [24,27,[68][69][70][71][72][73][74] Noticeably, the cycling performance of the BCNBs//NSPC PICs is highly stable (Figure 5e), showing a high capacity retention ratio of 70.6% after 2000 cycles at 20 A g −1 with a low capacity decay of 0.0147% per cycle.

Conclusion
To summarize, we have realized high-performance potassium ion storage with both excellent rate capability and cycle stability on the basis of boron doping carbon nanobubbles, which can be synthesized by a facile template-assisted strategy with boron-containing precursor. The obtained BCNBs material affords hollow interior carbon structure with coexistence of sp 3 -defective carbon and sp 2 graphitic carbon, and integrates the merits of porous nanobubbles-like morphology and high conductivity, which provide fast transport pathways for both potassium ions and electrons, leading to excellent electrochemical performance in terms of high-rate properties (85.7 mAh g −1 at 50.0 A g −1 ) and cycling stability (156.3 mAh g −1 after 2000 cycles at 10.0 A g −1 ). It is worth noting that carbon anode with such a high rate capability up to 100 A g −1 is rarely reported. Most importantly, the formation of artificial SEI film on the surface of BCNBs electrode leads to the realization Energy Environ. Mater. 2023, 6, e12559 8 of 10 of a high initial CE of 79.0%. Certificated by the ex situ XPS/TEM analysis and DFT calculation, boron doping plays a vital role in realizing superior electrochemical reversibility and kinetics dynamics. We believe our findings provide a promising strategy for designing high-rate carbon materials for energy storage devices.

Experimental Section
Materials: Analytical grade chemicals were utilized without any further purification. C 6 H 7 BO 2 and benzoic acid (C 7 H 6 O 2 ) were acquired from Sigma-Aldrich. MgO and HCl were obtained from Sinopharm Chemical Reagent Co., Ltd.
Synthesis of BCNBs: A mixture of C 6 H 7 BO 2 and MgO (50 nm) with the mass ratio of 2:1 was first ball-milled for 3 h and then carbonized in a tube furnace under Argon atmosphere at 1000°C for 2 h. The resulting product was washed in 2 M HCl for 24 h, followed by filtering and freeze-drying. As a control experiment, CNBs were synthesized under the same condition by replacing C 6 H 7 BO 2 with C 7 H 6 O 2 . In addition, BCNBs-900 and BCNBs-1100 were prepared using a similar procedure to that of BCNBs, with calcination temperature being 900 and 1100°C, respectively.
Materials characterizations: The crystal phases of the products were characterized by XRD using a Bruker D8 Advance powder diffractometer equipped with a Cu Kα radiation source. SEM (Hitachi S4800) and TEM (JEOL 2010F) analyses were performed to observe the morphologies and microstructures. Raman spectra was conducted by a micro-Raman spectroscope (LabRam HR800). XPS spectra were recorded with Thermo ESCALAB 250XI. Nitrogen adsorption-desorption analysis was carried out using a Micromeritics TristarII 3020 instrument. The specific surface area and pore size distribution are calculated by Brunauer-Emmett-Teller (BET) method and the nonlocal DFT model. The packing density is the apparent powder density obtained by pressing the material in a container with the pressure of 10 MPa.
Electrochemical measurements: To carry out the electrochemical measurements, CR2032 coin cells were assembled in a glovebox filled with Ar. A slurry was fabricated by mixing 75 wt% BCNBs, 15 wt% acetylene black, and 10 wt% carboxyl methylated cellulose (CMC) in~20% aqueous alcohol, and then the slurry was coated on copper foil and dried at 80°C for 12 h. For battery assembly, pure potassium foil was used as the counter and the reference electrode, a Whatman glass fiber disk was used as a separator, and 1 M potassium bis(fluorosulphony)imide (KFSI) in a mixture of ethylene carbonate (EC) and diethyl carbonate (DEC) was used as the electrolyte. CV and EIS were carried out using a Gamry Interface 1000 workstation. In addition, a LAND CT2001A battery system was applied to execute GITT and galvanostatic charge-discharge tests.
For the fabrication of PICs, BCNBs and NSPC were employed as anode and cathode materials, in which NSPC was prepared according to the previously described procedure. [13] For pretreatment, half cells using the active material as the anode, were performed in 0.001-3 V for 5 cycles at 0.05 A g −1 . The cutoff voltage of PICs was 4.0 V for charging and 0 V for discharging. The energy (E, Wh kg −1 ) and power (P, W kg −1 ) densities of PICs were calculated as per the following formulas: [26,75] where t represent the discharge time (s), I stands for the charge/discharge current (A), V max and V min are the potential at the beginning and the end of the charge (V), and m denotes the total mass both cathode and anode (kg). DFT calculations: In density functional theory calculations, structural optimization was performed by Vienna Ab-initio Simulation Package (VASP) with the projector augmented wave (PAW) method. [76,77] The exchange-functional was treated using the Perdew-Burke-Ernzerhof (PBE) [78] functional, in combination with the DFT-D3 correction. [79] Cutoff energy of the plane-wave basis was set as 450 eV. For optimization of both geometry and lattice size, the Brillouin zone integration is performed with 2 × 1 × 1 Monhorst-Pack5 k-point sampling. The self-consistent calculations applied a convergence energy threshold of 10 −5 eV. The equilibrium geometries and lattice constants were optimized with maximum stress on each atom within 0.02 eV/Å. Both band structure and density of state were both obtained by vaspkit interface. [80]