Molecular grafting towards high-fraction active nanodots implanted in N-doped carbon for sodium dual-ion batteries

Abstract Sodium-based dual-ion batteries (Na-DIBs) show a promising potential for large-scale energy storage applications due to the merits of environmental friendliness and low cost. However, Na-DIBs are generally subject to poor rate capability and cycling stability for the lack of suitable anodes to accommodate large Na+ ions. Herein, we propose a molecular grafting strategy to in situ synthesize tin pyrophosphate nanodots implanted in N-doped carbon matrix (SnP2O7@N-C), which exhibits a high fraction of active SnP2O7 up to 95.6 wt% and a low content of N-doped carbon (4.4 wt%) as the conductive framework. As a result, this anode delivers a high specific capacity ∼400 mAh g−1 at 0.1 A g−1, excellent rate capability up to 5.0 A g−1 and excellent cycling stability with a capacity retention of 92% after 1200 cycles under a current density of 1.5 A g−1. Further, pairing this anode with an environmentally friendly KS6 graphite cathode yields a SnP2O7@N-C||KS6 Na-DIB, exhibiting an excellent rate capability up to 30 C, good fast-charge/slow-discharge performance and long-term cycling life with a capacity retention of ∼96% after 1000 cycles at 20 C. This study provides a feasible strategy to develop high-performance anodes with high-fraction active materials for Na-based energy storage applications.

However, unlike Li + and K + ions, it is difficult for traditional graphite materials to act as the anode for the intercalation of Na + ions [48,49]. Further, the large ionic radius of Na + (1.02Å vs. 0.76Å for Li + ) results in sluggish reaction kinetics and large volume changes of the anode materials such as Sn [50,51], MoS 2 [52][53][54], TiO 2 [55] and FePO 4 [56], and thus leads to poor rate capability and unsatisfied cycling stability [57][58][59]. Several approaches have been applied to improve the electrochemical performance C The Author(s) 2020. Published by Oxford University Press on behalf of China Science Publishing & Media Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.  of these anodes, including nanoscale modification and carbon-based composite construction [60][61][62][63][64][65], for example, carbon-based tin pyrophosphate (SnP 2 O 7 ) composite with a high carbon content (16.8%) has been demonstrated to exhibit enhanced cycling stability for Na + -ion storage [66]. Although the carbon matrix can improve the electronic conductivity and provide a buffer framework for alleviating the volume expansion of these anodes, the excessive carbon content (commonly >15 wt%) would decrease the fraction of active material and thus reduce the energy density of batteries. Therefore, it is necessary to increase the fraction of active materials as high as possible and reduce the content of inactive carbon without compromising the conductivity of composite anodes, so that anodes can effectively deliver their specific capacities.
Herein, we propose a molecular grafting strategy to in situ implant SnP 2 O 7 nanostructure in N-doped carbon (SnP 2 O 7 @N-C) as the anode for Na-DIBs. Such a strategy enables high-fraction (95.6 wt%) active materials to uniformly embed in the carbon matrix and to effectively prevent the exfoliation of active materials, while the N doping leads to high conductivity even at a low C content. It exhibits a high specific capacity of 400 mAh g −1 at 0.1 A g −1 and excellent cycling stability with a capacity retention of 92% after 1200 cycles under 1.5 A g −1 . Consequently, pairing this anode with an environmentally friendly graphite cathode yields a SnP 2 O 7 @N-C||KS6 Na-DIB, which shows excellent rate performance up to 30 C, good fast-charge/slow-discharge ability and long-term cycling life with a capacity retention of 96.3% after 1000 cycles, showing a promising potential for Na-based energy storage devices. Figure 1 schematically illustrates the synthesis procedure of SnP 2 O 7 @N-C via the molecular grafting method. Owing to the complexing interaction between radical groups (e.g. phosphate groups) and metal cations (e.g. tin ions) and the hydrogen bond between organic precursors, many precursor agents can molecularly graft into precursor composite with a three-dimensional framework, accompanied by a full mixing procedure. In this case, we chose phytic acid as the phosphorous source to strengthen the adhesion between active nanodots and carbon matrix due to sufficient O-C bonds and strong complexing ability of phosphate groups to tin cations. Simultaneously, the low atomic ratio of C to P can avoid residual carbon content in the formed composite. Besides, we chose the melamine as the N doping source because its high atomic ratio of N to C can result in a high concentration of nitrogen in the carbon matrix. After the filtration and drying processes, composite precursor nanoparticles were achieved. Finally, the SnP 2 O 7 /N-C nanoparticles were synthesized via calcining the precursor composite under an Ar atmosphere.

RESULTS AND DISCUSSION
In order to investigate the molecular grafting process, Fourier transform infrared spectra (FTIR) measurements of precursors and their composites were carried out. The characteristic absorption peaks originating from phosphate radical (980 and 1140 cm −1 ), phosphate hydrogen radical (1631 cm −1 ) and stretching vibration of O-H (3329 cm −1 ) are observed in the FTIR spectrum of the phytic acid solution ( Supplementary  Fig. S1a). To confirm the complexing interaction between phosphate groups and tin cations, phytic acid and SnCl 2 were applied to synthesize a precursor composite without the addition of melamine. Compared with that of phytic acid, the FTIR spectrum of the composite ( Supplementary Fig. S1b) presents an obvious peak shift of phosphate to 1035 cm −1 , which should be ascribed to the complexing interaction between phosphate groups and tin cations. For the pure melamine, some characteristic absorption peaks involving the out-of-plane ring bending vibration of triazine ring (810 cm −1 ), stretching vibration of C-N (1431 cm −1 ), stretching vibrations of triazine ring (1526 cm −1 ), scissoring vibration of NH 2 (1626 cm −1 ) and stretching vibrations of NH 2 (3100-3500 cm −1 ) were observed in its FTIR spectrum ( Supplementary Fig. S1c). Once melamine had been added, its three typical absorption peaks at 773, 1440 and 1529 cm −1 were detected in the FTIR spectrum of the precursor composite (Supplementary Fig. S1d). The obvious peak shift of melamine at 810 to 773 cm −1 should be attributed to the formation of intermolecular hydrogen bonds between melamine and phosphate groups [67].
Compared with the precursor composite (Supplementary Fig. S2), the synthesized SnP 2 O 7 @N-C composite features a stable morphology without structural collapse after calcination treatment ( X-ray diffraction (XRD) pattern and Raman spectrum were carried out to provide more structural information. As observed in Fig. 2d, all sharp diffraction peaks can be indexed to cubic-phase SnP 2 O 7 (JCPDS Card No. , in accordance with the HRTEM observations. In contrast, those samples calcined at 500 • C and 700 • C (Supplemen-tary Fig. S4) do not present similar characteristic diffraction peaks of SnP 2 O 7 . Note that a bump peak located at ∼26 • should originate from the amorphous carbon matrix. Its amorphous feature is also confirmed by the Raman spectrum (Fig. 2e), where two characteristic peaks of carbon situated at ∼1360 and ∼1585 cm −1 can be observed. Both peaks are individually attributed to D band of disordered carbon and G band of graphitic carbon. The ratio of I D to I G approximates 1.0, implying the carbon matrix's dominant defective and disordered nature. Further thermogravimetric analysis (TGA) measurement ( Fig. 2f) indicates that the fractions of SnP 2 O 7 nanodots and N-doped carbon are 95.6 wt% and 4.4 wt%, respectively, which is the highest fraction of active material among previously reported Sn-based compound/carbon composites [66,[68][69][70]. Both the XRD pattern (Supplementary  Fig. S6) shows that the carbon content of the SnP 2 O 7 @C composite is ∼3.7%, close to that of SnP 2 O 7 @N-C (∼4.4%), which suggests that the addition of melamine slightly increases the carbon content, ascribable to its high atomic ratio of N to C. The nitrogen adsorption/desorption isotherm of SnP 2 O 7 @N-C ( Supplementary Fig. S7) reveals that its Brunauer-Emmert-Teller (BET) specific surface area is ∼9.0 m 2 g −1 .
The chemical components of the SnP 2 O 7 @N-C sample were analyzed by X-ray photoelectron spectroscopy (XPS). As shown in Supplementary  Fig. S8a, the survey XPS spectrum suggests the existence of O, P, Sn, C and N elements in the sample, consistent with the energy dispersive X-ray spectroscopy (EDX) elemental mappings where these elements uniformly distribute in the SnP 2 O 7 @N-C composite ( Supplementary Fig. S9). High-resolution Sn 3d XPS spectrum (Fig. 2g) presents a pair of characteristic peaks at 495.4 and 487.0 eV, corresponding to Sn 3d 3/2 and Sn 3d 5/2 of Sn 4+ in SnP 2 O 7 , respectively. Besides, only one peak at 134.0 eV referring to the P 2p can be observed (Fig. 2h), indicating a complete transformation of P source to SnP 2 O 7 without P doping. The deconvoluted O 1s spectrum ( Supplementary  Fig. S8b) includes two peaks. The dominant peak is assigned to the SnP 2 O 7 , and another involves O-C bonding. Moreover, the high-resolution C 1s spectrum ( Supplementary Fig. S8c) can be fitted into three peaks at 284.6, 285.5 and 286.5 eV, individually originating from C-C, C-N and C-O, exhibiting that the carbon matrix is doped with  dominant nitrogen and slight oxygen [54]. The high-resolution N 1s spectrum (Fig. 2i) shows the existence of pyridinic N (398.5 eV), pyrrolic N (400.0 eV) and quaternary N (401.1 eV) [71]. Such structural and chemical features confirm the homogeneous implantation of SnP 2 O 7 nanodots in the N-doped carbon matrix, which is expected to optimize its charge transfer kinetics and electrochemical stability for SIBs.
We firstly carried out the cyclic voltammogram (CV) measurement to investigate the Na + -storage behavior of the SnP 2 O 7 @N-C anode. Figure 3a exhibits the first three CV curves at 0.1 mV s −1 in the potential range of 0.01-3.0 V. In the first sodiation process, there are multiple peaks situated at 1.55, 1.10, 0.58, 0.39 and 0.07 V. According to previous reports, the Na-Sn alloying reactions occurred at po-tentials below 0.9 V [70,72]. Thus, the first two peaks should involve the conversion process of SnP 2 O 7 to metallic Sn, and the others stem from the Na-Sn alloying reactions [70,71]. In the following cycles, only a broad and strong peak at 1.18 V is observed for the conversion reaction. However, the desodiation processes in different cycles always exhibit five peaks at 0.23, 0.69, 0.86, 1.38 and 1.84 V. Such behavior suggests the conversion reaction probably refers to a two-step reduction/oxidation reaction of Sn 4+ /Sn 2+ and Sn 2+ /Sn 0 , and the Na-Sn alloying/dealloying reaction is also associated with a multi-step reaction process.
To further get insight into its Na + -ion storage mechanism, the sodiation/desodiation process of the SnP 2 O 7 @N-C anode was detected with synchrotron X-ray absorption near edge structure (XANES) spectra of the Sn L 3 -edge ( Fig. 3b and c) at different charging/discharging states (Supplementary Fig. S10), where two peaks are assigned to the 2p 3/2 -5s 1/2 transition [73]. As observed in Fig. 3b, the intensities of the characteristic peaks decrease as the discharging process proceeds, associated with the transformation of Sn 4+ to Sn 0 [73,74]. A reverse evolution of the peak intensities is obviously observed during the charging process (Fig. 3c), demonstrating the good reversibility of Sn state during cycling. Note that the slight difference of Sn L 3 -edge XANES spectra at 0.01 V and 1.2 V should be ascribed to the de-alloying reactions of the sample without obvious variation in the valence state of Sn element. Further, HRTEM characterizations at different charging/discharging states were performed to verify its sodiation/desodiation mechanism. For the pristine sample, an interplanar spacing of 0.398 nm is clearly distinguished (Supplementary Fig. S11a), which corresponds to (200) plane of the SnP 2 O 7 . When the sodiation process proceeds until 0.55 V, there are some lattice fringes with lattice spacing of 0.248 and 0.217 nm (Fig. 3d), which match well with (131) plane of Sn 2 P 2 O 7 (ICSD No. 170846) and (208) plane of Na 9 Sn 4 (PDF No. 31-1326), respectively. And the fully sodiated state clearly contains Na 4 P 2 O 7 and Na 15 Sn 4 two crystal phases (Fig. 3e), further confirming that the sodiation process of SnP 2 O 7 @N-C anode involves both conversion and alloying reactions. Conversely, as the desodiation process is conducted to 1.2 V, the presence of metallic Sn is verified by the HRTEM image in Fig. 3f. The completed desodiation process at 3.0 V is accompanied by the formation of SnP 2 O 7 (Fig. 3g), indicating a good sodiation/desodiation reversibility of SnP 2 O 7 . It is also noteworthy that, differently from the reported results [66], there are some lattice fringes with an interplanar spacing of 0.304 nm ( Supplementary Fig. S11b), corresponding to the (−131) plane of P-1 Sn 2 P 2 O 7 (ICSD No. 170846), which implies the presence of Sn 2 P 2 O 7 during the desodiation process. Therefore, the HRTEM result is greatly consistent with the analyses of CV result during sodiation/desodiation processes.
The electrochemical properties of the SnP 2 O 7 @N-C anode were evaluated in a coin-type half-cell. As observed in Fig. 4a, an abnormal shape of the galvanostatic charge-discharge profile at the first cycle is attributed to the incompletely reversible sodiation process of SnP 2 O 7 and the formation of solid-electrolyte interphase (SEI) layer [54,75]. Although a pulverization phenomenon of SnP 2 O 7 is observed after the first sodiation/desodiation process ( Supplementary  Fig. S12), there is a stable shape of galvanostatic charge-discharge profiles after the first cycle. It shows a specific discharge capacity of ∼400 mAh g −1 at 0.1 A g −1 with a Coulombic efficiency of ∼100% (Fig. 4b), indicating a good stability during the following sodiation/desodiation processes. Further, the EDX mappings of the anode at fully discharged state (Supplementary Fig. S13) verify uniform distributions of O, P, Sn, Na, C and N elements, implying a homogeneous sodiation reaction during discharging process. Such robust charging/discharging behavior was also confirmed by the electrochemical impedance spectroscopy (EIS, Fig. 4c). No obvious variation in its EIS spectra is observed before and after 100 cycles, ascribable to the strong adhesion between SnP 2 O 7 nanodots and N-doped carbon matrix. Figure 4d and e present the rate performance of SnP 2 O 7 @N-C anode at current densities from 0.1 to 5.0 A g −1 . It exhibits specific capacities of 400, 381, 354, 335, 305, 295, 261 and 210 mAh g −1 at current densities of 0.1, 0.2, 0.3, 0.5, 1.0, 1.5, 3.0 and 5.0 A g −1 , respectively. The specific capacities are recoverable as the current density is returned to 0.1 A g −1 . It should be noted that the rate capability of the SnP 2 O 7 @N-C anode is much better than that of pure SnP 2 O 7 (58 mAh g −1 at 1.5 A g −1 ), SnP 2 O 7 @C without N doping (176 mAh g −1 at 1.5 A g −1 ) (Supplementary Fig. S14) and previously reported SnP 2 O 7 composite with 16.8 wt% carbon nanosheets [66]. The excellent rate capability can be attributed to that N doping enhances the conductivity of carbon framework and facilitates diffusion kinetics of Na + ions [71,76], and the active nanodots shorten the diffusion path of Na + ions [77,78]. Figure 4f and Supplementary Fig. S15 show the composite anode's cycling performance under a current density of 1.5 A g −1 , exhibiting excellent cycling stability with a capacity retention ∼92% after 1200 cycles and the corresponding Coulombic efficiency close to 100%. In contrast, much lower  capacity retentions of ∼79% and ∼49% are obtained for SnP 2 O 7 @C and pure SnP 2 O 7 after 400 cycles ( Supplementary Fig. S16), respectively. Among the reported Sn-based compound/carbon composite anodes for SIBs (Supplementary Table S1), the SnP 2 O 7 @N-C with the lowest carbon content delivers a competitive specific capacity and superior cycling performance. Consequently, we paired this anode with an environmentally friendly KS6 graphite cathode to construct a proof-of-concept Na-DIB to further explore its practical sodium storage capability in the full cell. Figure 5a schematically illustrates its working mechanism, where Na + cations and PF 6 − anions separately move to the SnP 2 O 7 @N-C anode and KS6 graphite cathode during the charging process, while both cations and anions return back to the electrolyte from the anode and cathode during discharging process, respectively. Its typical galvanostatic charge-discharge profile (Fig. 5b) in the voltage range of 1.0 to 4.0 V at 3 C (1 C = 100 mA g −1 ) exhibits several voltage plateaus, corresponding to the different intercalation/de-intercalation stages of PF 6 -anions. According to the dQ/dV differential curve (Fig. 5c), the charging process (Fig. 5b) can be roughly separated into three voltage regions of 2.6-3.25 V (stage I), 3.25-3.55 V (stage II) and 3.55-4.0 V (stage III), corresponding to three different stages of anion intercalation into KS6 graphite cathode [50,52,53]. In order to get insight into the electrochemical process of the SnP 2 O 7 @N-C||KS6 Na-DIB during the charging process, the galvanostatic charge-discharge profile of Na||KS6 half-cell and corresponding dQ/dV differential curve ( Supplementary Fig. S17) were provided. The dQ/dV differential curve also presents three different stages, indicating the dominant role of anion intercalation into KS6 graphite cathode. Conversely, a reverse evolution accompanies the discharging process, where different de-intercalation stages occur in voltage ranges of 4.0-2.6 V (stage III ), 2.6-2.06 V (stage II ) and 2.06-1.30 V (stage I ) in the discharging process ( Fig. 5c and Supplementary  Fig. S17). Such intercalation/de-intercalation behavior of PF 6 anions was further confirmed by the in situ XRD measurements during the charging/discharging process (Fig. 5d). The original XRD pattern presents a characteristic (002) peak of KS6 graphite cathode at 26.7 • . In the charging process, the characteristic peak becomes weak and splits into two peaks individually shifting towards lower (main peak) and higher 2θ degrees, corresponding to the stage I of anion intercalation into KS6 graphite cathode. The stage II involves the formation of a stable intercalation phase at 23.6 • . Then, the stage III relates to a sharp transition of diffraction peaks and the formation of another stable phase at 22.1 • . Such peak evolution is ascribable to the successful intercalation of PF 6 − anions into graphite cathode [79,80]. A reverse evolution occurs in the discharging process, and the two peaks gradually merge into the initial peak at 26.7 • at the end of the discharging, indicating excellent reversibility of the intercalation/de-intercalation process of PF 6 anions into/from KS6 graphite cathode. Figure 6a presents the rate capability of the SnP 2 O 7 @N-C||KS6 Na-DIB, which delivers a reversible discharge capacity of 78 mAh g −1 at 3 C.
Even at 30 C, a specific capacity of 65 mAh g −1 can be obtained (83.3% capacity retention) with ∼100% Coulombic efficiency. The galvanostatic charge-discharge profiles at different current densities show similar shapes and a slight shift of voltage plateaus, indicating negligible electrochemical polarization (Fig. 6b). Besides, it can be rapidly charged at 30 C and slowly discharged down to 3 C (Fig. 6c and d). The discharge profiles exhibit a slight variation, and the corresponding specific capacity can be stably delivered even at different current densities, exhibiting a good fast-charge/slow-discharge ability. Moreover, the Na-DIB shows an excellent cycling performance with a capacity retention of ∼96% and a Coulombic efficiency of ∼100% after 1000 cycles under a high rate of 20 C (Fig. 6e). The galvanostatic charge-discharge profiles at 10th, 100th, 500th and 1000th cycles have the same shape and voltage plateaus (Fig. 6f), further verifying its stable cycling ability. As shown in Supplementary  Table S2, the SnP 2 O 7 @N-C||KS6 Na-DIB presents superior cycling performance, rate capability and Coulombic efficiency to previously reported Na-DIBs based on different anode materials [50,[52][53][54][55][56]75,[80][81][82][83][84][85][86].

CONCLUSION
In summary, high-fraction (up to 95.6 wt%) SnP 2 O 7 active anode material was successfully in situ implanted in the N-doped carbon matrix via a molecular grafting strategy. Such a synthesis strategy effectively enhanced the adhesion between active materials and carbon matrix, while the N doping led to high conductivity even at low C content. As a result, the anode showed a high specific capacity of ∼400 mAh g −1 at 0.1 A g −1 , good rate performance up to 5.0 A g −1 and excellent cycling stability with a capacity retention of 92% after 1200 cycles at 1.5 A g −1 . Furthermore, this anode was paired with an environmentally friendly KS6 graphite cathode to yield a proof-of-concept Na-DIB, showing a superior rate capability with a capacity retention of ∼83% even at a high current density of 30 C, good fast-charge/slow-discharge ability and long-term cycling life with a capacity retention of ∼96% after 1000 cycles at 20 C, exhibiting a great potential for high-performance Na-based energy storage devices.

Synthesis of SnP 2 O 7 @N-C
SnCl 2 ·2H 2 O powder was dissolved in deionized water under stirring, and then phytic acid solution and melamine powder were sequentially added into the above solution and subsequently stirred vigorously. Then the mixture was transferred to a two-necked flask and absolute ethanol was added and refluxed under stirring. Next, the obtained reaction product was collected by centrifugation, successively washed with deionized water and ethanol several times and dried under vacuum. Finally, the powder product was calcined in an Ar atmosphere to obtain a SnP 2 O 7 @N-C sample.

Materials characterization
The morphological and elemental features were characterized using field-emission scanning electron microscope (FE-SEM). The FEI Tecnai G2 F30 was applied to acquire the transmission electron microscope (TEM) images, elemental mappings and SAED pattern. XRD analyses were implemented on a Rigaku D MiniFlex 600 diffractometer. Raman spectra were collected on Horiba LabRAM HR800. N 2 physical adsorption-desorption analysis was carried out on ASAP 2020M. The chemical composition of SnP 2 O 7 /N-C sample was determined using XPS with monochromatic aluminum Kα radiation. TGA were conducted from 100 • C to 700 • C. FTIR of precursors and their composites were acquired using a PerkinElmer Frontier FTIR spectrophotometer. Tests about XANES were carried out at Synchrotron Light Research Institute (SLRI), Thailand.

Electrochemical measurement
The electrochemical performance of the half-cells and DIB was carried out using CR2032 coin-type cells. The SnP 2 O 7 @N-C electrode was prepared by coating mixture slurry of the SnP 2 O 7 @N-C, Ketjenblack and carboxy methyl cellulose with a weight ratio of 70:20:10. For the half cells, the electrodes were pressed and punched into circular sheets 10 mm in diameter. The KS6 graphite cathode was prepared by mixing 80 wt% KS6 graphite, 10 wt% conductive carbon black and 10 wt% polyvinylidene fluoride (PVDF) to form a homogeneous slurry. In order to boost the full utilization of cathode material, the cathode sheet was punched into circular sheets 10 mm in diameter. The mass loading ratio of active anode/cathode materials for Na-DIB was ∼1 : 1 and the corresponding size of the anode sheet was 12 mm in diameter. Glass fabric was used as the separator, and 1 M NaClO 4 in propylene carbonate (PC) with 5 wt% fluoroethylene carbonate (FEC) was used as the electrolyte for half cells. The electrolyte for the SnP 2 O 7 @N-C||KS6 DIB was 1 M NaPF 6 dissolved in a mixture of ethylene carbonate (EC)/dimethyl carbonate (DMC)/ethyl methyl carbonate (EMC) (4 : 3 : 2 in volume). Cells were assembled in a glove box with water and oxygen content below 0.1 ppm and tested at room temperature. Galvanostatic charge-discharge tests and rate tests were conducted with a battery test system. EIS and CV were performed on an Autolab electrochemical workstation. All chemical reagents were used as received without any further purification. The capacity is calculated based on the mass of SnP 2 O 7 @N-C for half cells. The mass of KS6 is used to calculate the specific capacity of the DIB. More detailed materials are available in the online supplementary data.

SUPPLEMENTARY DATA
Supplementary data are available at NSR online.

AUTHOR CONTRIBUTIONS
Y.T. conceived and designed the study. S.M. and Q.L. carried out the synthesis and most of the structural characterizations and electrochemical tests. P.K. and X.Z. carried out XANES measurement. S.M., Q.L. and Y.T. co-wrote the manuscript. S.M., Q.L., P.K., X.Z., W.W. and Y.T. discussed the results and participated in analyzing the experimental results.
Conflict of interest statement. None declared.