PxSy nanoparticles encapsulated in graphene as highly reversible cathode for sodium ion batteries

Ting Wng,Zhixiong Hung,Donghung Wng,Jiqi Wu,Junjie Lu,Zihn Jin,Shojun Shi,*,Yongqi Zhng,d,**

a Jiangsu Laboratory of Advanced Functional Material,School of Chemistry and Materials Engineering,Changshu Institute of Technology,Changshu 215500,China

b Yangtze Delta Region Institute (Huzhou),University of Electronic Science and Technology of China,Huzhou 313001,China

c Jiangsu Institute of Metrology,Nanjing 210023,China

d Institute of Fundamental and Frontier Sciences,University of Electronic Science and Technology of China,Chengdu 611731,China

Keywords:Sulfur Graphene Phosphosulfide Cathode Sodium ion battery

ABSTRACT Herein,phosphorus-mediated sulfur nanoparticles encapsulated in reduced graphene oxide nanosheets(P-SrGO-T) were successfully synthesized as the cathode for sodium ion battery by a ball milling and the following thermal treatment.A series of covalent bonds,such as P–S,C–S–C,C–O–P and C–S–P,are formed in this process,which are in favor of fixing the sulfur and suppressing the parasitic shuttle effect of polysulfide.Benefiting from the graphene sheets and these covalent bonds,a high reversible capacity of 637.4 mAh/g was achieved in P-SrGO-T after 100 cycles at the current density of 0.2 A/g.In addition,P-SrGO-T also delivers a high-rate capacity (330.7 mAh/g at 5 A/g) attributing to low charge transfer resistance and faster ion diffusion kinetic.This work pushes the progress forward in developing phosphosulfide cathode for sodium ion batteries.

Sodium ion batteries (SIBs) are emerged as one of the viable alternatives to lithium-ion batteries (LIBs) due to the suitable electrochemical potential (-2.71 Vvs.standard hydrogen electrode for Na+/Na) and the abundant reserves of Na (thousands of times over Li) [1,2].As one of most proposing candidates for the nextgeneration rechargeable batteries,SIBs,has received much attention [3].To match the Na metal anode,sulfur is regarded as a promising cathode for SIBs owing to its high theoretical capacity [4].However,the application is extremely hindered by its poor electronic conductivity,serious dissolution potential,parasitic shuttle effect of polysulfide and large volume expansion,all of which result in the limited reversible capacity and inferior rate capability[5].

Incorporating sulfur-based materials with carbon-based materials is an effective strategy to address these fatal weaknesses [7–9].In the last decade,many carbon-based materials have been studied to probe their feasibility,such as carbon nanotube,mesoporous carbon,graphene oxide (GO),and reduced GO (rGO) [10–12].For instance,Ji’s group prepared a sulfur–carbon (S–C) complex for SIBs that delivered a high reversible capacity of ～750 mAh/g at 0.25 A/g,with an outstanding rate capability of 526 mAh/g at 8 A/g [9].However,it is hard to suppress the dissolution/shuttle effects onlyviacompositing sulfur with carbon matrixes simply.Hence,S host matrixes with inherent polarization,such as metal sulfides (MSn) [12],phosphosulfides (PxSy) [13,14],and metal phosphosulfides (MPSn) [15,16],have been introduced into the composite cathode to fix the polar polysulfideviastrong chemical interaction.Meanwhile,the study of these PxSy[13,17] and MPSncomposites have been proposed as new class electroactive materials with higher theoretical capacities and better rate performance than the corresponding MSn[12].The strong interaction between P and S fixes Sviaforming the P-S and/or P=S bond [14,18] in MPSnand PxSyto inhibit the shuttle effects.The Na storage ability of NiPS3reported by Yan’s group exhibits an excellent capacity retention of 88% after 200 cycles at 2.0 A/g [18].The phosphorus sulfide (P4S40)molecules provide a stabilized lithiation capacity retention at approximately 720 mAh/g at 0.5 A/g after 95 cycles [17].In addition,it has been demonstrated that the graphene sheets provide highly reactive functional groups that can serve to hold the P and S atoms through hydrogen and covalent bonds,which could further improve the electrochemical performance of PxSy/MPSnbased anodes [19,20].For example,Wu’s group reported that the rGOFePS3maintains a much higher capacity (250 mAh/g) than bare FePS3(～20 mAh/g) under 0.1 A/g [16].

Fig.1.Schematic diagram of the experimental procedure to synthesis P-SrGO-T.

Inspired by the merits of involving graphene sheet and P atoms into the sulfur matrix,herein,the design and fabrication of the Pmediated S,hierarchically anchoring to layered rGO (P-SrGO)viaa simply mechanically ball milling and following annealing process(Fig.1).The P-mediated S can be obtained during ball millingviathein situcovalent binding between P and S,as reported by Wanget al.[13].Further thermal treatment was applied to trigger the formation of a C-S-P coupling bond (named P-SrGO-T for short).At the same time,the mechanical ball milling and thermal reduction within an inert atmosphere could improve the electronic conductivity of electrode materials obviously by several orders of magnitude [21–23].When serving as cathode for SIBs,the P-SrGO-T delivered a good reversible capacity of 637.4 mAh/g at 0.2 A/g with a capacity retention of 63.4% after 100 cycles,and an enhanced rate capacity of 340.0 mAh/g at 5 A/g.Therefore,from a practical point of view,the upgraded strategy followed in this work is beneficial for the application of naturally abundant sulfur-based materials in large-scale EESs.

The synthetic routes of nanosized P-mediated S particles incorporated with reduced graphene oxide (P-SrGO) and its derivate(P-SrGO-T) are illustrated in Fig.1.Firstly,0.5 g red phosphorus(RP,Alfa Aesar,98.9%),1 g sulfur (S,Sinopharm,99.5%) and 1.5 g graphene oxide were mixed with 10 mL 5% PVP (Sigma Aladdin,MW: 130,000) solution.This mixture slurry was transferred to an agate environment filled with Ar gas quickly,and continuous ball milling at 400 rpm for 12 h,followed by suction filtration and then drying directly at 60°C overnight in a vacuum oven,named as P-SrGO shortly.Secondly,the as-synthesized P-SrGO powder was sealed in a vacuum quartz tube,then annealed at 600°C for 3 h,subsequently cooling down to 280°C at a rate of 1°C/min for 24 h,the new obtained sample was named as P-SrGO-T for short.This thermal treatment utilized can trigger the formation of C-S-P on the carbon skeleton,further reduction of GO and the pyrolysis of PVP,concurrently.Other two counterpart samples named as SrGOT and PrGO-T were also fabricatedviathe similar process,but without RP and S during their individual ball milling process,respectively.Two more samples with the maximum loading ratio of active materials by controlling the ration of red P to sulfur is 1:1 and without the addition of GO during ball milling process were obtained,which named as PS and PS-T.

Fig.2.(a) Lower magnified and (b) high magnified SEM images of P-SrGO.(c) Mapping images of P-SrGO.(d) Lower magnified and (e) high magnified SEM images of P-SrGO-T.(f) Mapping images of P-SrGO-T.(g) TEM and (h) HRTEM images of PSrGO.(i) TEM and (j) HRTEM images of P-SrGO-T.

X-ray diffraction (XRD) patterns were recorded by a Rigaku D/max 2500 X-ray diffractometer with Cu Kαradiation.Thermal stability analysis was carried out on a thermogravimetric analyser (TGA; SDTA851) under N2atmosphere at a range of 25–700°C with 5°C/min Fourier transform infrared spectroscopy (FTIR) were performed by a Nicolet 6700 spectrometer.Raman spectra were carried out by Raman microscopy (DXR,Thermo Fisher Scientific,532 nm).X-ray photoelectron spectroscopy (XPS) was tested on an Escalab250Xi spectrometer using Al Kαradiation (Thermo Scientific).An Al target is used as the excitation source with an electron emission angle of 60°.The size of analyzed area with a dimeter of 400 m is performed during the test under the base pressure below 5×10-8mBar.The charge neutralizer is on during the whole process.Argon ion clusters applied during etching process with a speed of ～1 nm/s.The content of carbon,sulfur and oxygen were detectedviaelemental analysis (EA,Flash EA 1112).The morphology and structure were studiedviascanning electron microscopy (SEM,SIGMA,ZEISS microscope with EDS) and highresolution transmission electron microscopy (HRTEM,TECNAI G20).The Brunauer–Emmett–Teller (BET) specific surface area was calculated based on the N2adsorption-absorption (ASAP 2020).The electronic conductive (σ) was tested at room temperatureviaa four-probe resistance machine (Keithley 4200 SCS).

All electrochemical measurements were carried out in coincell configurationviaassembling 2032-type coin cells with Na foil as counter-electrode.The slurry for electrode was prepared by mixing as-prepared samples,super P and polyvinylidene fluoride (PVDF) binder at a weight ratio of 80%,10% and 10% inNmethylpyrrolidone (NMP).Then this slurry was spread onto copper foil followed by drying at 80°C in vacuum oven overnight.The mass loading of the electrode was controlled around 3.0 mg/cm2.The solution of 1 mol/L NaPF6in diethyl carbonate and ethylene carbonate (1:1,v/v) with 5% fluoroethylene carbonate was used as the electrolyte.The glass fiber membrane was served as the separator.The assembling of batteries was performed in an Ar-filled glovebox.The galvanostatic charge/discharge tests and the galvanostatic intermittent titration technique (GITT) were performed between 0.01 V and 3.0 Vvs.Na+/Na on a Land battery tester (CT2001A,Wuhan Land,China) to evaluate the sodiation/desodiation performance of electrode.The detailed GITT test parameters for each electrode are as following,the current pulse is 10 mA/g,pulse time is 10 min and relaxation time is 1 h.The cyclic voltammetry (CV) was conducted by an electrochemical workstation (CHI 600E) between 0.01 V and 3 Vvs.Na+/Na at various scan speed range from 0.1 mV/s to 5 mV/s.Electrochemical impedance spectroscopy (EIS) was performedviaCHI 600E electrochemical workstation within a frequency range of 0.01 Hz to 105Hz at an amplitude of 5 mVversusthe different potentials.

Field-emission SEM and EDS measurements were carried out to characterize the morphology and elemental distribution of P-SrGO and P-SrGO-T.As shown in Fig.2a,most of the particles in P-SrGO are irregularly shaped on a nanoscale,and encapsulated in the layered nanosheets.From the magnified SEM images (Fig.2b),the layered nanosheets and nanosized particles in a compact multilayered structure can clearly be seen,indicating the dense composition between the P-mediated S and layered nanosheets.Figs.2d and e display the morphology of P-SrGO-T,rounded and smoothed particles embedded under the nanosheets,indicating that most of the particles underwent a melting but not vaporizing process at high temperature due to the protection of layered nanosheets.No obvious separated nanosheets with large areas in these two samples further indicates that the most flexible nanosheets intimately contact the particles [24].A maximum weight loss of 4.64% is observed at 600 °C,as is shown in the TGA result (Fig.S1a in Supporting information).It suggests that only a small quantity of free or separated particles that were superficially exposed underwent vaporizing during the calcination process,which corresponds well to the morphology results observed in SEM.Furthermore,an elemental analysis was performed before and after thermal treatment,with the corresponding results summarized in Table S1 (Supporting information).P-SrGO contains 39.4 wt%,31.1 wt% and 12.0 wt% of C,S and O,respectively,and the P content can be calculated to around 17.5 wt%.The accurate contents of C,S and O in P-SrGO-T were determined to be 48.0 wt%,27.4 wt% and 9.4 wt%,respectively,and the P content was calculated to be 15.2 wt%.Compared with P-SrGO,the decreasing content of S,O and P in P-SrGO-T underpins the further thermal reduction of GO and the pyrolysis of PVP combined with a small part of the S and P vaporization at high temperature [25,26].The lower O content in the two asprepared samples than that in raw GO suggests the conversion of GO to rGO during the ball milling and thermal reduction processes[21,25].The EDS mappings in Figs.2c and f confirm the uniform distribution of S,P,O,and C in the as-prepared P-SrGO and PSrGO-T,respectively.Furthermore,the TEM images (Figs.2g and i) indicate the nanosized particles and layered nanosheets structure in P-SrGO and P-SrGO-T.The layered nanosheets and particles in the two samples were initially detected in the HRTEM images(Figs.2h and j).Apart from a lattice fringe related to the (002)plane for the rGO nanosheet,the absence of other lattice planes proves the low-crystallinity nature of particles in both samples.The similar low-crystallinity nature and totally amorphous structure of S/P-based materials illustrated by HRTEM are also reported previously [26,27].Compared with the larger lattice fringe around 10 ˚A in typical GO [28],a reduction of the (101) lattice fringe in P-SrGO (3.7 ˚A) and a further decrease in P-SRGO-T (3.4 ˚A) was observed,which illustrates the reduction of GO during the synthesis process and matches well with the above elemental analysis results.The lower surface areas of as-prepared samples compared with that of raw GO (219.2 m2/g) indicates that S and P are filled in the space among layered nanosheets during the ball milling and further high-temperature treatment processes (Fig.S1b and Table S1 in Supporting information) [21].

Fig.3a displays the XRD patterns of the as-synthesized materials,as well as the raw materials of GO,S,and P.The diffraction pattern of P-SrGO shows a broad peak around 24° that corresponds to the typical reflection from the (002) plane of disordered carbon in rGO [28,29].This broad diffraction underwent a right shift to a high angle after thermal treatment,indicating the decrease in the lattice spacing of the P-SrGO-T [29,30].Apart from the broad peaks in both samples,no other sharp diffractions illustrate their amorphous state [27],which is in agree with the HRTEM results.The absence of features for GO,RP,and S8in the XRD patterns of both as-prepared samples also indicates the chemical interaction among these raw materials during the synthesis process.For further insight,Raman spectroscopy was performed to illuminate the status of defects and levels of disorder in carbon materials.As is shown in Fig.3b,spectra show a typical D band (1342 cm-1) and G band (1585 cm-1) [31,32].Compared with theID/IGvalues of 0.82 in raw GO,theID/IGvalues for P-SrGO are increased to 1.07,suggesting that the disordered carbon domains increased with the incorporation of P and S.A further increase inID/IGin P-SrGO-T(1.43) was observed,indicating more defect formation and the rebuilding of sp2carbon in the following annealing process.It has been well documented that the formation of a large quantity of defects and disordered carbon domains can be beneficial for energy storage due to the extra electrochemically active sites and the fast diffusion kinetics of sodium ions [33].

Further structural information on GO,P-SrGO and P-SrGO-T was investigated by FTIR measurement (Fig.3c).In the pattern of GO,it was found that there are many functional groups located around 2925/2850,1710,1630,1460/1380,1050 and 895 cm-1,which are associated with the asymmetrical/symmetrical stretching vibration of–CH2,carboxyl (–COOH) stretching,aromatic (C=C) stretching,asymmetrical/symmetrical–CH2bending,C–O stretching,and C–H out-of-plane bending,respectively [34].Some of these functional groups favor its functionalization with heteroatoms doping (such as N,P,S and F) [33–35],which is also reflected in this work.Compared with raw GO,several new vibrations were observed in P-SrGO and P-SrGO-T.Clearly,these two samples exhibit a sharp band around 550 cm-1,which is indexed to P–S [16,36],suggesting an apparent P-mediated S in P-SrGO and P-SrGO-T.Notably,C–SOx–C,C–O–P interfacial covalent bonds,and P–O dangling bonds located around 750,1055 and 950 cm-1were observed in the two samples owing to the strong covalent binding among P,S and O in some oxygenous functional groups outside the carbon plane[37,38].These bonds in the P-SrGO and P-SrGO-T heterointerface are vital to constructing the architecture by stabilizing the interaction among P,S,and the layered skeleton.In the P-SrGO-T sample,two new bonds ascribed to the coupling C–S–P bond located between C–SOx–C and P–S are observed,indicating a further interaction among S,P and C [39].The P–P bond relating to unreacted P was only detected in the FTIR curve of P-SrGO,meaning that the ball milling cannot facilitate a complete interaction.In general,more surface defects induced by P,S heteroatoms and less oxygenous group could accelerate the charge transfer [13,33].As is summarized in Table S1,the four-probe point measurement reflects that the electrical conductivity in P-SrGO-T is up to 6.52×10-3S/cm,which is much higher than that of GO (3.75×10-6S/cm)and P-SrGO (4.09×10-5S/cm).

The low-resolution XPS survey (Fig.S1c in Supporting information) verifies the presence of P,S and C elements in the P-SrGO and P-SrGO-T.The S 2p deconvoluted results illustrate three chemical bonding regions (Fig.3d),corresponding to P–S bonding (163.6 and 164.8 eV),–C–SOx–C–bonding (168.5 and 170.0 eV),and terminal sulfur atoms ST(161.7 eV) [6,19,20].This STshould correspond to linear sulfur molecules trapped inside the GO matrix (e.g.,inside closed pores or an enclosure space) rather than in open nano-porosity or on the exposed surface [5].Significantly,the existence of P–S bonding in an S 2p spectrum confirms the formation of P-mediated S during the ball milling process and survival after thermal treatment [16,19].In particular,the P–O–C and P–C bonds centered around 134.9 and 132.5 eV in the P 2p spectra (Fig.3e) that can be attributed to part of the P atoms were doped into GO nanosheets during the synthesis process [40].The observed–C–SOx–C–,P–O–C and P–C chemical bonds in the S 2p and P 2p spectra of P-SrGO and P-SrGO-T were also in good agreement with the observation of their C 1s core level.The high-resolution C1s(Fig.3f) spectra of P-SrGO and P-SrGO-T exhibits similar chemical states,with five peaks around the binding energies of 283.8,284.6,285.7,286.8 and 288.7 eV,which correspond to C–P,C sp2,C–S,C–O and C=O bonds,respectively,demonstrating that both S and P covalently bond with C atoms [6,41,42].P-SrGO and PSrGO-T possess similar primary surficial chemical bonds,confirming that these bonds formed during the ball milling process and preserved during the thermal treatment process.Drawing on all of the aforementioned characterization analyses,the carbon skeleton of two samples is compared graphically in Fig.S2 (Supporting information).More defects are displayed for P-SrGO-T because of additional C–S–P coupling bonds.

Fig.3.(a) XRD patterns of the as-prepared P-SrGO and P-SrGO-T synthesized from raw materials of S8,red phosphorus and graphene oxide.(b) Raman spectra and (c) FTIR spectra of as-prepared two samples and graphene oxide.(d) S 2p,(e) P 2p and (f) C 1s spectra of as-prepared P-SrGO and P-SrGO-T without etching; XPS bonding content comparison of P-SrGO and P-SrGO-T in (g) S 2p,(h) P 2p and (i) C 1s region before etching and after etching with 120 s.

XPS testing after Ar etching with 120 nm depths (Fig.S1 in Supporting information) was performed to further compare the chemical state of C,S and P between the surficial and internal parts.The content of each of the bonds in S 2p,P 2p and C 1s of two samples before and after etching are summarized in Figs.3g–i,respectively.The S 2p,P 2p and C 1s spectra (Figs.S1d-f) from the internal part of two samples have similar peaks to those on the surficial part,but with changing content.As shown in the content comparison of S 2p,the sharp decreasing of–C–SOx–C–after etching demonstrates that the formation of–C–SOx–C–mainly due to exposure and oxidization in the air.A little content of–C–SOx–C–still maintained after etching could indicate a part of S covalent with the oxygenous function group in the carbon skeleton [41].The P residue content internal to P-SrGO increased significantly,indicating that part of the P cluster was embedded deeply in the rGO nanosheets without involving interaction with rGO and S during the ball milling process.The absence of P residue in the surficial part combined with only a small trace of P residue and increased P–S content in the internal part of P-SrGO-T may relate to the further formation of the C–S–P coupling bond during the thermal treatment [39].As is shown in the C 1s spectrum,C–P,C sp2,C–S,C–O and C=O could still be observed in the internal part of the two samples and the corresponding content was almost the same as that in the surficial part,indicating that these bonds have high homogeneity from the external to the internal parts of the two samples.Hence,it can be concluded that the covalent Pmediated S uniformly distributes among the rGO nanosheets in the P-SrGO and P-SrGO-T at the molecular level with C–S,P–S and C–O–P bonds.These bonds would be beneficial for intimate electrical contact between Sn(n＜4),P,PxSy,and the carbon skeleton,which can not only improve the utilization ratio but also remit the dissolution of polysulfide during cycling by directly forming solid polysulfide,improving the reversible capacity and cycle life [43].Thermal treatment could guarantee further enhancement of the interaction strengthviathe formation of C–S–P coupling bonds and increasing the electronic conductivity in the P-SrGO-T sample,and these improvements will be of benefit for higher capacities and better rate performance.

The electrochemical storage mechanism was investigated by applying cyclic voltammetry (CV) and galvanostatic discharge/charge in the potential range of 0.01–3.0 V.As is shown in the second CV scan curve of SrGO-T (Fig.4a),one redox couple of around 1.0/2.0 V was observed,which contributes to the reversible Na+sodiation/desodiation reaction to form Na2Sx(x＞4)/Na2Sy(y＜4)[5,44].As for the PrGO-T electrode,several anodic peaks in the range of 1.2–0.01 V were ascribed to the stepwise alloying reaction of P to form Na3P7,NaP,Na2P and Na3P [45,46].A main oxidation peak of around 1.5 V was observed in the subsequent cathodic scan of the PrGO-T electrode,which relates to the decomposition of Na3P to form Na3P7[47].As is shown in Figs.4b and c,several reduction peaks were observed in the initial cathodic scan of the P-SrGO and P-SrGO-T electrodes,and these peaks were assigned to stepwise reactions that the formation of Na2S and Na3P,along with the SEI film.Two oxidation peaks of around 1.44/1.88 and 1.42/1.93 V were noted in the initial anodic scan of P-SrGO-T and P-SrGO,which could possibly relate to the stepwise Na+desodiation from Na3P and solid Na2S phase to form solid NaP7and liquid Na2Sx(x=4,8) with a long chain,respectively.The following cathodic scans showed nearly overlapping CV profiles with a broad peak at 1.1 V in P-SrGO and two separated peaks located at 1.2 and 0.9 V in P-SRGO-T,corresponding to reversible Na+sodiation to form Na3P7and short-chain Na2Sy(y＜4) [44].The peaks in the CV curves of the four electrodes match well with the corresponding platforms in their galvanostatic charge/discharge curves.(Figs.4d–f and Fig.S3 in Supporting information).

Fig.4.(a) CV curves at the second cycle of the SrGO-T and PrGO-T electrodes at a scan rate of 0.1 mV/s from 0.0 V to 3.0 V.CV curves at the initial three cycles of the (b)P-SrGO and (c) P-SrGO-T electrodes at a scan rate of 0.1 mV/s from 0.0 V to 3.0 V.(d) Second charge-discharge curves of SrGO-T and PrGO-T electrodes at 50 mA/g.The initial three galvanostatic charge/discharge curves of (e) P-SrGO and (f) P-SrGO-T electrodes at 50 mA/g.

Fig.5.HRTEM images of (a,a1,a2) P-SrGO and (b,b1,b2) P-SrGO-T electrodes after the initial full discharge.HRTEM images of (c) P-SrGO and (d) P-SrGO-T electrodes after the initial full charge.

The HRTEM results of P-SrGO and P-SrGO-T electrodes after initial discharge and charge states further evince the reaction mechanism outlined above.As is shown in Figs.5a and b,the crystal lattice fringe of 0.25 nm ascribed to the (110) plane of Na3P is presented in both the P-SrGO-T and P-SrGO electrodes following initial discharge [48].Apart from the observation of the lattice for Na3P(Figs.5a1 and b1),some broad and blurry lattice fringes were observed that can be related to the formation of solid Na2S (Figs.5a2 and b2),and the low crystallinity of Na2S has been shown in many previous studies [49,50].As displayed in the HRTEM images of the two electrodes after initial desodiation (Figs.5c and d),most unrecognized blurry lattice fringes relate to the amorphous state of recovered materials.No breaks or cracks were observed in the TEM images (Fig.S4 in Supporting information) of two electrodes after discharging/charging processes implies that the self-buffer effect benefited the structural integrity of the overall electrode during the desodiation/sodiation process.The HRTEM results,combined with the CV results,demonstrate the same sodiation/desodiation mechanism in the two electrodes; however,P-SrGO-T owns a high level of reversibility and high specific capacity,as implied by its almost immobile and high-intensity peaks.

The cycling performance in Fig.6a at 0.2 A/g shows that P-SrGO-T exhibits an initial discharge/charge capacity of 1004.2/612.2 mAh/g with a coulombic efficiency (CE) of 60.9%,and P-SrGO exhibiting an initial discharge/charge capacity of 651.8/407.4 mAh/g with a CE of 62.5%.The low efficiency in both two electrodes is mainly accounted for the formation of SEI [4,51].In the subsequent cycles,an average CE above 93.0% could be realized in both electrodes.Beyond 100 cycles,the reversible discharge capacity of the P-SrGO and P-SrGO-T was around 532.1 and 637.4 mAh/g with 81.6% and 63.4% capacity retention,respectively.The continued increase in capacity after the initial cycle due to the continued activation of sodiation to decreasing its inner resistance [44].Promisingly,the absence of large capacity decay in both electrodes indicates their highly reversible properties,which contributed to the remission of the volumetric change and polysulfide dissolution during repeated sodium-ion insertion/extraction.

As shown in Fig.6b and Fig.S5b (Supporting information),the P-SrGO-T electrode delivered a superior rate performance of 564.8,561.0,546.6,518.8,484.5,441.3 and 330.7 mAh/g at 0.05,0.1,0.2,0.5,1,2 and 5 A/g than that of the P-SrGO electrode 486.3,478.8,467.0,443.8,410.6,366.1 and 248.6 mAh/g,respectively.The better rate performance of P-SrGO-T contributed to more content of the electrode involved in the reaction.A high capacity of 540 and 605 mAh/g could be restored in P-SrGO and P-SrGO-T when the current density recovered to 0.05 A/g,demonstrating that both electrodes could endure high current densities.Furthermore,the PSrGO and P-SrGO-T electrodes could retain 90.4 and 90.9% within 50 cycles (from the 51stto the 100thcycle),exhibiting the good stability of active materials under a high current density of 0.5 A/g(Fig.6b).The good cycling stability of P-SrGO and P-SrGO-T could mainly attribute to the integrative advantages from the synergistic effect among rGO,sulfur and red phosphorus,as illustrated by the poor cycling stability of PrGO-T (Fig.S3c) as well as PS (Figs.S5c and d in Supporting information) and low capacity of SrGO-T(Fig.S3c) as well as PS-T (Figs.S5c and d).The summarized performance comparison between target materials in this work and other carbon-based materials for sodium battery from previous works are shown in Table S2 (Supporting information) [18,22,38,42,52–56].It can find the P-SrGO-T shown the comparable electrochemical performance when compare the electrochemical performance with other carbon-based materials for SIBs.This comparison result could reflect the design strategy and practical synthesis route in this work have competitive advantage to some extent.The shuttle/dissolution effect of the two electrodes was compared by a voltage recording with a long resting time after the 100thfull charge.As is shown in Fig.6c,there was less voltage decay in PSrGO-T compared with P-SrGO,demonstrating that the construction in P-SrGO-T was more effective at stabilizing soluble products.

Fig.6.Electrochemical performance of P-SrGO and P-SrGO-T electrodes.(a) Cycling performance at current density of 0.2 A/g.(b) Rate capability at different current densities.(c) Voltage decay recording after full charge.(d) Rf, Rct and W-R value variation after various cycles from 3rd to 100th cycle.Diffusion coefficients of Na+ during (e) charge process and (f) discharge process.(g) CV curves of P-SrGO-T at different scan rates ranging from 0.1 mV/s to 5 mV/s.The pseudocapacitive contribution of (h) P-SrGO and (i)P-SrGO-T at different scan rates.

The improvement in capacity and rate capability of P-SrGO-T was analyzed using electrochemical impedance spectroscopy (EIS)and galvanostatic intermittent titration technique (GITT) testing.The EIS testing of two electrodes were conducted in a pristine state,after the 3rd,10th,25th,50thand 100thcycle,to study their resistance changes during continuous cycling.Fig.S6 (Supporting information) displays Nyquist plots with the typical characteristics of semicircles in the medium–high frequency region and sloping straight lines in the low-frequency range,which can be fittedviathe equivalent circuit (Fig.S6d).The semicircle in the middle frequency range corresponds to the sum resistance,which includes the charge transfer resistance on the electrode (Rct) and through the electrode/electrolyte interface (Rf) [57].The sloping straight lines in the low-frequency range indicate the solid-state diffusion of sodium in the electrode material,corresponding to the Warburg resistance (W-R).Fig.6d graphically depicts theRct,Rfand W-R values based on the fitted equivalent circuit,with all fitted results being listed in Table S2.Prior to cycling (fresh state),the sum value ofRctandRfin P-SrGO-T (1557.1Ω) was much smaller than that of the P-SrGO mixture (4037.0Ω),indicting the higher electronic conductivity in P-SrGO-T than P-SrGO.After the 3rdcycle,the sum value was decreased to 389.9 and 256.0Ω,whereas the W-R value was increased to 375.6 and 42.0Ωin the P-SrGO and P-SrGO-T electrodes,respectively.The decreased sum value in the two electrodes was attributed to the inserted active Na+in the host materials,which decreased the electrical resistance and obtained a better charge transfer capability [58].The formation of SEI limited the solid-state pathway for subsequent Na+diffusion during the initial dynamic activation cycles [59],which is the main reason for the increased W-R in both electrodes.Fortunately,this situation did not undergo further deterioration,as evidenced by the simultaneous decreasing ofRf,Rct,and W-R in both electrodes after the 10th,25th,and 50thcycles,which guarantees their high reversibility following the activation process (Fig.6d and Table S3).After 100 times repeating discharging/charging,Rf,Rctand W-R in both electrodes increased inversely,which related to the inevitable degradation of the active material and the accumulation of devitalized products during the repeating cycles [60,61].The comparison of the Na+diffusion coefficients was further evaluated by means of the GITT.Fig.S7 (Supporting information) shows the GITT curves of two samples during the third discharge/charge processes under a small current density of 10 mA/g.As shown in Fig.S7,the P-SrGO-T shows a lower mid-voltage value of 0.71 V in the GITT curve compared with the 1.01 V in P-SrGO,proving its decreased redox reaction polarity.Compared with P-SrGO,the P-SrGO-T exhibited higher diffusion coefficients by almost one order of magnitude,demonstrating a faster sodium diffusion ability in P-SrGOT (Figs.6e and f) [60].The accelerated migration of sodium ions is mostly ascribed to the synergistic effect of the more compact interaction and increased electronic conductivity,which provides shorter pathways and eases the transfer of sodium diffusion.

To reveal the kinetic reason behind high reversibility and the good rate capability of P-SrGO-T,the sodium storage behavior of two electrodes was compared by analyzing their CV profiles at various scan speeds,ranging from 0.1 mV/s to 5 mV/s (Fig.6g,Figs.S8 and S9 in Supporting information).It is reported that the overall charge storage mechanism (k1ʋ) and capacitive contribution (k2ʋ1/2) can be determined by the following equation:(V=baʋ=k1ʋ+k2ʋ1/2),whereaandbare tunable parameters,ʋmeans various scan rates,andk1andk2are adjustable values that can be determined from the slope and y-axis intercept of a linear equation through plotting [39,51].Figs.6h and i exhibit the summarized contribution of the non-diffusion controlled and diffusion-controlled capacities of P-SrGO and P-SrGO-T at various scan speeds.Clearly,the non-diffusion controlled contribution of both electrodes grows gradually with the increasing scan rates,and P-SrGO-T represents a comparable non-diffusion controlled capacities contribution (75.6% of total specific capacitance) with that of P-SrGO (66.3% of the total specific capacitance).Two conclusions can be drawn from this: (i) the fast surface-induced non-diffusion controlled behavior is the dominant reaction at high current densities [15,51]; (ii) rather than the higher specific area,the higher electronic conductivity may play a key role in accelerating the separation of Na+and e-for avoiding the Na+/e-aggregation on the surficial of the electrode,as indicated by the higher percentage of non-diffusion controlled capacities contribution but lower specific BET area in the P-SrGO-T electrode compared with those in the PSrGO electrode.

Fig.7.XPS spectra of S 2p,P 2p and C 1s region in (a,c,e) P-SrGO and (b,d,f)P-SrGO-T electrodes after the 100th cycle,respectively.

After 100 cycles,the chemical states and morphology of two asprepared electrodes were detectedviaXPS and SEM to illustrate the stability of composites and whole electrodes.To avoid disturbance from the SEI layer,Ar etching (～1 nm/s) with various etching times from 0 to 240 s was utilized during the XPS testing process.As shown in the S 2p spectrum of two electrodes (Figs.7a and b),with the etching time increasing,the P-S covalent bond centered at 164.5 eV disappearing,which indicates that the covalent P–S bond underwent breakage during repeated cycling [15].Moreover,at the initial state,Na2S was located at around 162.1 eV in the P-SrGO electrode,but a Na2S signal was absent in the P-SrGO-T electrode.This shows that the conversation reaction of Sn(n＜4) in P-SrGOT is confined beneath the rGO nanosheets,whereas some exposed Sn(n＜4) in the P-SrGO is entailed by sodiation without any protection.As is shown in the P 2p spectrum of the two electrodes(Figs.7c and d),the appearance of an extra P–F around 138 eV was observed prior to etching and could be assigned to the decomposition of NaPF6in the electrolyte [59–62].This P–F disappeared in the P-SrGO-T electrode after etching,whereas it continued to exist in the P-SrGO electrode,even after etching with 240 s,indicating that the salt decomposition problem was more serious in the P-SrGO electrode during electrochemical cycling.The same result was also concluded through the C 1s results of the two electrodes(Figs.7e and f),with the peak of around 290 eV being related to the organic–COx–group from the decomposed electrolyte solution[60].The discovery of a higher content of the organic–COx–group in the P-SrGO electrode than that in P-SrGO-T electrode suggests that more electrolyte solution decomposed in the P-SrGO electrode during charge/discharege process,which is unfavorable for maintaining the electrode stability [63].The morphology images of the two electrodes were further compiled to illustrate their condition after 100 cycles.As shown in Fig.S10 (Supporting information),there are clearly big cracks in the P-SrGO electrode in the crosssection images,whereas the P-SrGO-T electrodes maintain good electrode integrity,almost without cracks.Furthermore,the SEM images of active materials that were scratched from the current collector clearly show more evidence of good integrity in the PSrGO-T electrode.As shown in Fig.S10c,intense contact between the particle and nanosheets could still be observed in P-SrGO-T;however,dilapidated nanosheets and exposed particles were found in the P-SrGO (Fig.S10d).

Fig.8.The schematic illustration of P-SrGO and P-SrGO-T electrodes before cycling and after 100 cycles.The white lines indicate the occurrence of cracks in the P-SrGO and P-SrGO-T electrodes after 100 repeating cycles.

Based on the characteristic insights into the composition and morphology evolution combined with the electrochemical performance and kinetics of the two samples,the as-prepared P-SrGO-T offers an obvious structural advantage for improving Na storage capability with a high degree of reversibility.The schematic diagram comparison for sodium storage stability in P-SrGO and P-SrGO-T is presented in Fig.8.Fragmentation and material degradation are observed in the P-SrGO electrode after the 100thcycle,whereas these problems were not obvious in the P-SrGO-T electrode.The P-SrGO-T electrode was able to retain a high capacity upon cycling without significant deterioration in the structure and morphology of the electrodes,due to the soluble intermediate Na2Sx(x＞4)being confined between the graphene nanosheet layers,as is displayed in Fig.8.

In summary,we successfully designed and synthesized a Pmediated S and rGO nanosheet (P-SrGO-T) composite as cathode for SIBs.P-SrGO-T presents a highly reversible capacity(637.4 mAh/g under 0.2 A/g) after 100 cycles and a superior rate capability of 340 mAh/g at 5 A/g.Such excellent performances were attributed to the plentiful covalent bonds and physical leakproof structural.First,the interaction among P,S and rGO in P-SrGO-T are mostly based on chemical covalently bonds,rather than “free” physical mixing.Second,increased electronic conductivity was achieved in P-SrGO-T.Third,most of the particles in PSrGO-T were sealed into the rGO nanosheets and well protected.The strategy employed in this work will broaden horizons with respect to the optimization of sulfur-based materials for large-scale energy storage systems.

Declaration of competing interest

The authors declare that they have no known competing financial interests that could influence the work reported in this paper.

Acknowledgment

This work was sponsored by National Natural Science Foundation of China (Nos.21701017,52002052).

Supplementary materials

Supplementary material associated with this article can be found,in the online version,at doi:10.1016/j.cclet.2022.02.021.