lable at ScienceDirect

Journal of Power Sources 369 (2017) 138e145

Contents lists avai

Journal of Power Sources

journal homepage: www.elsevier .com/locate/ jpowsour

Buffer layer enhanced stability of sodium-ion storage

Xusheng Wang a, b, Zhanhai Yang b, Chao Wang a, Dong Chen b, Rui Li c, Xinxiang Zhang a,Jitao Chen a, *, Mianqi Xue b, **

a Beijing National Laboratory for Molecular Sciences, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, Chinab Institute of Physics and Beijing National Laboratory for Condensed Matter Physics, Chinese Academy of Sciences, Beijing 100190, Chinac School of Advanced Materials, Peking University Shenzhen Graduate School, Shenzhen 518055, China

h i g h l i g h t s

� Layered SnSe2 single crystal with integrated SeeSe buffer layers is prepared.� SeeSe buffer layers could accommodate the intercalation process.� Autogenous Na2Se layers could confine the structural damage of tin sequences.� SnSe2 single crystal realizes efficient, fast and long-term sodium-ion storage.� Sodiation/desodiation processes of the SnSe2 single crystal are investigated.

a r t i c l e i n f o

Article history:Received 1 August 2017Received in revised form12 September 2017Accepted 30 September 2017Available online 9 October 2017

Keywords:Tin diselenideBuffer layerSodium-ion storageHigh rateSuperior cycling stability

* Corresponding author.** Corresponding author.

E-mail addresses: chenjitao@pku.edu.cn (J. Chen),

https://doi.org/10.1016/j.jpowsour.2017.09.0880378-7753/© 2017 Elsevier B.V. All rights reserved.

a b s t r a c t

Se�Se buffer layers are introduced into tin sequences as SnSe2 single crystal to enhance the cyclingstability for long-term sodium-ion storage by blazing a trail of self-defence strategy to structural pul-verization especially at high current density. Specifically, under half-cell test, the SnSe2 electrodes couldyield a high discharge capacity of 345 mAh g�1 after 300 cycles at 1 A g�1 and a high discharge capacityof 300 mAh g�1 after 2100 cycles at 5 A g�1 with stable coulombic efficiency and no capacity fading. Evenwith the ultrafast sodium-ion storage at 10 A g�1, the cycling stability still makes a positive response anda high discharge capacity of 221 mAh g�1 is demonstrated after 2700 cycles without capacity fading. Thefull-cell test for the SnSe2 electrodes also demonstrates the superior cycling stability. The flexible andtough SeeSe buffer layers are favourable to accommodate the sodium-ion intercalation process, and theautogenous Na2Se layers could confine the structural pulverization of further sodiated tin sequences bythe slip along the Na2SeeNaxSn interfaces.

© 2017 Elsevier B.V. All rights reserved.

1. Introduction

Lithium-ion batteries (LIBs) technology has driven the storage ofelectrical energy on the path to the indispensable regulation foreveryday life in the past decades [1e3], now that sodium-ion bat-teries (SIBs) technology is bouncing back to resurge the vitality ofnext-generation batteries by addressing to the sustainable sodiumreserves [4e7]. Similar to the operational mechanism of LIBs, theSIBs could balance the energy storage and release through theextraction and insertion of sodium ions along with the circuit ofelectrons [8]. However, direct extrapolation from LIBs technology is

xuemq@iphy.ac.cn (M. Xue).

insufficient to build the SIBs system. For instance, the commer-cialized graphite anode in the LIBs field cannot provide enoughsodium-ion storage in the specific electrolytes for its relativelynarrow interlayer spacing [9e11], thus causing very poor stability.And the prosperous studies of silicon material for LIBs (typically viaCui's group [12,13]) are relatively deserted for the inactive Na-Sialloys [14,15]. Hence, developing feasible SIB anodes with long-term cycling stability is looming ahead.

At present, strategies to improve the cycling stability of theelectrode materials have focused on the structure and space de-signs which mainly based on the nanocomposite technology withthe strengths of reducing the path length of inward ion diffusionand enhancing the resistance to the structural pulverization[16e25]. Take the tin-enriched materials as example, Jahel et al.[26] confined crystalline SnO2 nanoparticles in mesoporous carbon

X. Wang et al. / Journal of Power Sources 369 (2017) 138e145 139

as SIB anode with a capacity retention of 417 mAh g�1 at 1.8 A g�1

even after 4000 cycles. A SnS2 NC/EDA-RGO (NC, nanocrystal; EDA,ethylenediamine; RGO, reduced graphene oxide) nanocompositeprepared by Jiang et al. [27] exhibited a high capacity of 480 mAhg�1 at 1 A g�1 after 1000 cycles. Zhou et al. [28] reported aSnS@graphene hybrid SIB anode with an excellent sodium-ionstorage of 308 mAh g�1 at 7.29 A g�1 after 250 cycles. Althoughthese achievements have shown remarkable progresses in thecycling stability, the simple production processes, satisfactorycompacted density, and long-term operation at high current den-sity are still the eternal goals of battery materials [4,29e33]. In thisview, blazing a trail of self-defence strategy to structural damagecould be feasible to implement the long-term cycling stability.

Inspired by the Se-doping induced improvement in the cyclingstability of tin microparticles [34], the exfoliation characteristic ofSnS2-xSex (0 � x � 2) single crystal materials [35], and the incom-plete ionization in the Li-FeSe system [36], here we introduce theSeeSe layers as the buffer device in fabricating the SnSe2 singlecrystal to solve the stubborn nature of poor cycling stability for SIBs.In the study of the SnSe2/RGO nanocomposites [37], due to theenhancement in electrical conductivity and structural stabilityderived from graphene, the SnSe2/RGO nanocomposites demon-strated excellent capacity delivery and cycling stability at 0.1 A g�1,which is much better than that of the pure SnSe2 nanosheets,despite the high-rate and long-term performances are unavailableto us. Further study on the two-dimentional SnSe2 with confinednanostructures and patterns [38] may indeed help us recognize thecharacteristics of SnSe2. For this matter, owing to the high integrityand regularity of single crystal, the encoded SeeSe buffer layersshould possess admirable flexibility and toughness, which isfavourable to accommodate the sodium-ion intercalation process,and the Na2Se layers generated from the conversion reaction couldconfine the structural pulverization of further sodiated tin se-quences by the slip along the Na2SeeNaxSn interfaces during thealloying process [21], thus leading to the superior long-termcyclability at high current density. Under half-cell test, the SnSe2electrodes, prepared by simple ball mill mixing technology, are ableto yield a high discharge capacity of 345mAh g�1 after 300 cycles at1 A g�1 and a high discharge capacity of 300 mAh g�1 after 2100cycles at 5 A g�1 without capacity fading. Even with the ultrafastsodium-ion storage at 10 A g�1, a high discharge capacity of221 mAh g�1 is still delivered after 2700 cycles with no capacityfading. The full-cell test for the SnSe2 electrodes also demonstratesthe superior cycling stability. The encoded SeeSe buffer layerscould work along with the published nanocomposite technologyfor the tin-enrichedmaterials or even go beyond them in long-termcycling stability at high current density. The concept of encodingintegrated and regular SeeSe buffer layers into tin sequencesshould be applied to other SIB anodes [21,22], or even other batterysystems. Furthermore, the repertoire of the highly ordered andextended layers with superior resistance to the structural pulveri-zation could be amplified to build the graphene- and even polymer/protein-mimetic materials [39,40].

2. Experimental

2.1. Synthesis of SnSe2, SnS2, and Na3V2(PO4)3 material

The SnSe2material was synthesized by solid-state reaction of tin(Alfa Aesar, 99.9%) and selenium (Alfa Aesar, 99.999%) at 400 �C for20 h and then 900 �C for another 20 h in an evacuated quartz tube.The SnS2 material was synthesized by the same procedure ascomparison. The Na3V2(PO4)3 material was synthesized by using asol-gel method with NaOH, NH4VO3, NH4H2PO4, and citric acid asraw materials.

2.2. Material characterizations

The crystal structures of SnSe2 and other intermediate productswere measured by X-ray diffraction (XRD, PANalytical diffractom-eter) with Cu Ka radiation (l ¼ 1.5416 Å). The morphologies andmicrostructures of SnSe2 and other intermediate products werevisualized by scanning electron microscopy (SEM, S4800, Hitachi),transmission electron microscopy (TEM, JEM-2100F, JEOL), andenergy dispersive spectroscopy (EDS, JEM-2100F, JEOL). The surfaceelemental analyses were performed by X-ray photoelectron spec-troscopy (XPS, Axis Ultra, Kratos Analytical Ltd.).

2.3. Preparation of the SnSe2, SnS2, and Na3V2(PO4)3 electrodes

The SnSe2 electrodes were prepared by mixing the SnSe2 pow-ders, carbon black, and sodium carboxymethylcellulose (CMC) at aweight ratio of 75: 15: 10. The slurry was spread on a copper foiland then dried in an electric thermostatic drying oven at 90 �C. Thedried copper foil was cut into disks with the diameter of 11mm (theaverage loading of SnSe2 is 1.8 mg cm�2) as the SnSe2 electrodesand then dried in an oven at 80 �C for 12 h under vacuum. The SnS2electrodes were prepared by the same procedure as comparison.The Na3V2(PO4)3 electrodes were prepared by mixing theNa3V2(PO4)3 powders (90 wt% of pure Na3V2(PO4)3 and 10 wt% ofcarbon), carbon black, and polyvinylidene fluoride (PVDF) at aweight ratio of 75: 15: 10. The slurry was spread on the aluminumfoils and then dried in an electric thermostatic drying oven at 90 �C.The dried aluminum foils were cut into disks (11 mm) as theNa3V2(PO4)3 electrodes and then dried in an oven at 120 �C for 12 hunder vacuum.

2.4. Electrochemical performances of the SnSe2 and SnS2 electrodes

The half-cell performances of the SnSe2 and SnS2 electrodeswere characterized with sodium metal foils as the counter elec-trodes, glass fibers as the separators, and 1 mol L�1 sodium tri-fluomethanesulfonate (NaCF3SO3) in diethylene glycol dimethylether (DEGDME) as the electrolyte, and then assembled intoLIR2032-type coin cells in an argon-filled glove box in which themoisture and oxygen contents were below 0.1 ppm. The half-cellperformances of Na3V2(PO4)3 material were characterized by1 mol L�1 sodium perchlorate (NaClO4) in propylene carbonate (PC)with 5 vol% fluoroethylene carbonate (FEC) as the electrolyte. Thefull-cell performances of the SnSe2 electrodes were characterizedby the similar procedure only with the differences of theNa3V2(PO4)3 as the cathode and 1 mol L�1 sodium perchlorate(NaClO4) in DEGDME as the electrolyte. The cycling tests of theassembled cells were performed on a Land CT2001A battery testingsystem within the voltage range of 0.1e3.0 V versus Naþ/Na(Moreover, the voltage range of 2.2e4.0 V was used for theNa3V2(PO4)3 half-cell test and the voltage range of 0.3e2.9 V wasused for the full-cell test). Cyclic voltammogram (CV) and electro-chemical impedance spectroscopy (EIS) were measured on anAutolab electrochemical workstation. The cycle number of thecycling performances for the SnSe2 half cells in the text starts afterseveral discharge-charge processes, in which a small current den-sity (0.1e0.5 A g�1) was used to activate the electrodes to achievethe stable cycling state.

3. Results and discussion

Fig. 1a presents the X-ray diffraction (XRD) pattern of the SnSe2crystal. All the peaks located at 14.5�, 29.1�, 44.3�, 60.4�, and 77.8�

could be severally indexed to the (001), (002), (003), (004), and(005) crystal planes of the hexagonal SnSe2 P-3m1 crystal system

Fig. 1. Structure and morphology of the SnSe2 crystal. (a) XRD pattern of the SnSe2 crystal. Only the oriented peaks distribution along the c-axis appears. (b) Appearancemorphology of the SnSe2 crystal in bulk form. (cef) SEM and HRSEM images of the SnSe2 crystal. The adjacent SnSe2 nanosheets (c,e) stacked and tiled together in a terraced wayexhibit a typical layered structure. The scroll structure (d,f) should be attributed to the fold of ordered and ultrathin SnSe2 nanosheets with flexible and tough SeeSe buffer layers.The scale bars are 3 mm, 2.0 mm, 1.0 mm, 0.2 mm, and 0.2 mm in panels (bef), respectively.

X. Wang et al. / Journal of Power Sources 369 (2017) 138e145140

(ICSD no. 98-004-3857). There only appears the oriented peaksdistribution along the c-axis, indicating a preferred crystal growthof regular arrangement of SeeSe buffer layers, where the positivestructural response against sodium-ion intercalation happens[41,42]. The introduction of selenium instead of sulfur is favourableto the increase of interlayer spacing. This is verified by the com-parison between the calculated 0.611 nm for SnSe2 and 0.585 nm(ICSD no. 98-002-9012) for SnS2, thus allowing for better cushionfor the electrostatically driving sodium-ion intercalation [43]. InFig. 1b, the SnSe2 crystal in the bulk form shows a glistening surfacemorphology, indicating a high crystalline quality and hence a highintegrity of SeeSe buffer layers, which is consistent with the nar-row and intensive XRD peaks. Fig. 1cef displays the scanningelectron microscopy (SEM) images of SnSe2 crystal from differentperspectives. As shown in Fig. 1c, the adjacent SnSe2 nanosheetsstacked and tiled together in a terraced way exhibit a typicallayered structurewith planar configuration, which could enable theSeeSe buffer layers to extend laterally. The high-resolution SEM(HRSEM) image shown in Fig. 1e illustrates the translucent SnSe2nanosheets with rectangular edges and smooth planes, in whichthe closely packed SeeSneSe layers are bound by the van derWaalsforce [35,39,41,44,45]. A scroll structure within the plane on thesurface is also displayed in Fig. 1d. The scroll dimension lies within500 nm in diameter and 5 mm in length. The scroll structure shouldbe attributed to the fold of the ordered and ultrathin SnSe2 nano-sheets, which relies on the flexibility and toughness of the SeeSe

buffer layers. Thus it could allow better resistance to the sodium-ion intercalation. The HRSEM image of the stacked SnSe2 lamellasin curled configuration is exhibited in Fig. 1f. Although the lamellasare carved into ravines, they accumulate together compactly intoclusters by the strain tolerance from the SeeSe buffer layers. Thelow-resolution SEM image of the SnSe2 crystal is exhibited inFig. S1. Fig. S2a and 2b illustrate the transmission electron micro-scopy (TEM) images of the SnSe2 crystal. Fig. S2a demonstrates thesignificant interface distribution of the SnSe2 lamellas, indicatingthat the SnSe2 nanosheets assemble together by a highly orderedway. In accordance with our previous research [35], the SnSe2nanosheet shown in Fig. S2b exhibits a silky planar backbonewhichis atom-thin but extends energetically in two dimensions, and itsresilience to deformation should be generated from the flexibilityand toughness of SeeSe buffer layers. Fig. S2c and 2d display thehigh-resolution TEM (HRTEM) images of the SnSe2 crystal, whichreveal the high crystallinity with lattice spacing of 0.329 and0.190 nm for the (010) and (110) planes, respectively. The inset inFig. S2c reveals the selected area electron diffraction (SAED)pattern, and the regular diffraction spots indicate the single-crystalnature of SnSe2. Precise temperature control is the key to ensurethe composition, morphology, and crystallinity of SnSe2 crystal[40,46].

As shown in Fig. 2a, the cyclic voltammograms (CV) of the SnSe2and comparative SnS2 electrode were performed between 0.1 and3.0 V at different scan rates of 0.1, 0.2, 0.5, 0.75, and 1.0 mV s�1

Fig. 2. Electrochemical performances of the SnSe2 and comparative SnS2 electrodes. (a) CV curves of the SnSe2 and comparative SnS2 electrodes between 0.1 and 3.0 V. The CVcurves of SnSe2 electrode retain the initial shape at 0.1 mV s�1 along with the increase of scan rate, and the intercalation peaks in zone A are invisible for the SnS2 electrode.Galvanostatic voltage profiles of the SnSe2 electrode (b) at 1 A g�1 and the comparative SnS2 electrode (c) at 0.5 A g�1 from the 2nd to the end cycle between 0.1 and 3.0 V (0.1 A g�1

in the 1st cycle for both the electrodes). The SnSe2 electrode shows much higher 1st-cycle coulombic efficiency at 97.5% compared to the SnS2 electrode at 75.2% at the same currentdensity. The intercalation plateaus are intact and overlapped in zone B (b) and yet almost disappear in zone C (c). (d) Voltage drops of the SnSe2 and comparative SnS2 electrodealong with setting time. Both the instantaneous and gradual voltage drops of the SnSe2 electrode are smaller than that of the SnS2 electrode.

X. Wang et al. / Journal of Power Sources 369 (2017) 138e145 141

(1.0 mV s�1 for the SnS2 electrode). In the sodiation process of theSnSe2 electrode at 0.1 mV s�1, the cathodic scan curve shows fourpeaks located at 2.20, 1.63, 1.06, and 0.64 V separately. The peak at2.20 V is not common and could be assigned to the reaction be-tween sodium ion and the trace amount of selenium [47]. Based onthe study of FeSe2, SnS2, and SnSe2 anode materials [22,37,41,42],the peak at 1.63 V should correspond to the sodium-ion intercala-tion into the SeeSe buffer layers, which is distinct from the sub-sequent electrochemical reactions, i.e., only NaxSnSe2 compoundappears and no phase decomposition happens. The peaks at 1.06and 0.64 V could be severally indexed to the conversion andalloying reactions (NaxSnSe2 / Sn þ Na2Se / NaxSn). The Na2Selayers generated from the conversion reaction around the fresh tinsequences could function as the protective sheath to confine thealloying sodiation process. The schematic illustration of the sodia-tion process for the SnSe2 material is shown in Fig. S3. In the anodicprocess, the peak at 1.51 V should correspond to the dealloyingreaction, and the peak at 1.94 V could be identified as the restitu-tion of tin selenides. Along with the increase of scan rate, the CVcurves retain the initial shape at 0.1 mV s�1, indicating an excellentrate capability. Despite the increase of scan rate, the intercalationpeaks in zone A keep vigorous for the SnSe2 electrode, while theyare invisible for the SnS2 electrode. Therefore, the flexible andtough SeeSe buffer layers with larger interlayer spacing than thatof the SnS2 material is favourable to accommodate the sodium-ionintercalation process. The galvanostatic voltage profiles of theSnSe2 electrode at 1 A g�1 and the comparative SnS2 electrode at0.5 A g�1 are illustrated in Fig. 2b and c, respectively. The SnSe2electrode shows a much higher 1st-cycle coulombic efficiency of97.5% than that of 75.2% for the SnS2 electrode, indicating the goodstructural stability of SnSe2. This is favourable to reduce the amountof “dead sodium” stored in the anode to improve the capacity

delivery of full batteries. The overlapped curves of the SnSe2 elec-trode indicate an excellent cycling stability, and the coulombic ef-ficiency has always remained at ~100%. The plateaus in thesodiation process exhibit great tenacity, which should be attributedto the accommodation of the SeeSe buffer layers and theconfinement of the autogenous Na2Se layers. In contrast, despitethe higher 1st-cycle discharge capacity, the SnS2 electrode exhibitsserious capacity fading. The intercalation plateaus are intact andoverlapped in zone B (Fig. 2b) and yet almost disappear in zone C(Fig. 2c), which is consistent with the result of zone A in Fig. 2a.Fig. 2d demonstrates the voltage drops of the SnSe2 electrode(1 A g�1) and the comparative SnS2 electrode (0.5 A g�1) during thesetting time. For the SnSe2 electrode, the instantaneous voltagedrops are almost the same value of 0.1 V for the 50th, 100th, and200th cycle, and the gradual voltage drops show a small growth to0.3 V. While for the SnS2 electrode, the two values are severally 0.4and 0.5 V for the 100th cycle. The instantaneous voltage dropscould be ascribed to the ohmic polarization and activation polari-zation, and the gradual voltage drops during the setting time arecontrolled by the concentration polarization [8]. In this sense, thelimitations of ohmic resistance, charge-transfer hindrance, andsodium-ion diffusion for the SnSe2 electrode are smaller than thatof the SnS2 electrode. These results agree well with the superiorcyclability of the SnSe2 electrode. The improvement of kineticsdepends partly on the structural stability [23,48], which is sup-ported by the tough SeeSe buffer layers and autogenous Na2Selayers, and vice versa, rapid and steady charge-transfer and ion-diffusion processes could diminish the structural damage of SnSe2single crystal.

The specific half-cell performances of the SnSe2 electrodes be-tween 0.1 and 3.0 V are exhibited in Fig. 3. As shown in Fig. 3a, ahigh discharge capacity of 345mAh g�1 is delivered after 300 cycles

Fig. 3. Specific half-cell performances of the SnSe2 electrodes between 0.1 and 3.0 V. (a) Cycling performance of the SnSe2 electrode at 1 A g�1. The electrode delivers a highdischarge capacity of 345 mAh g�1 after 300 cycles without capacity fading. (b) Long-term cycling performance of the SnSe2 electrode at 5 A g�1. The electrode maintains a highdischarge capacity of 300 mAh g�1 after 2100 cycles without capacity fading. And the coulombic efficiency maintains ~100% in the whole cycling process. (c) Long-term cyclingperformance of the SnSe2 electrode at 10 A g�1. The electrode shows a high discharge capacity of 221 mAh g�1 after 2700 cycles. The 2nd, 21th, 41th, 61th, …, 2101th (2701th inpanel (c)) cycle are present in panels (b) and (c). All the cycling performances do not include the 1st cycle (0.1 A g�1).

X. Wang et al. / Journal of Power Sources 369 (2017) 138e145142

at 1 A g�1 without capacity fading. The coulombic efficiencymaintains ~100% in the whole cycling process. When the currentdensity is adjusted to 0.3 A g�1 (Fig. S4), a high discharge capacity of392 mAh g�1 is achieved after 130 cycles. The capacity increasemight be assigned to the process of activation and stabilization. Theblocked sodium-ion diffusion pathways in the initial cycles couldinduce the remaining inactive SnSe2 species and then the repeatedcycling process should yield the more smooth and homogeneousdiffusion of sodium ions to realize the full use of the former inactiveSnSe2 species. Fig. 3b demonstrates the long-term cycling perfor-mance at 5 A g�1. After 2100 cycles, the SnSe2 electrode could stillmaintain a high discharge capacity of 300 mAh g�1 with stablecoulombic efficiency and no capacity fading, indicating the superiorlong-term cyclability of the SnSe2 electrode at high current density.Even with the ultrafast sodium-ion storage at 10 A g�1 shown inFig. 3c, the cycling stability still makes a positive response and ahigh discharge capacity of 221 mAh g�1 (0.395 mAh cm�2) isdemonstrated after 2700 cycles without capacity fading. Comparedto our prior research about the SnSSe material [35], despite theweakness in capacity output at low current density, the SnSe2electrode realizes the transcendence at 10 A g�1, which could beattributed to the highly integrated and regular SeeSe buffer layersin the single-crystal SnSe2 instead of the defective SeSe bufferlayers in the SnSSe material. Therefore, these superior half-cellperformances especially at high current density suggest that theencoded SeeSe buffer layers and autogenous Na2Se layers couldstabilize the sodiation process without extra protective measures.Table S1 presents the comparison between the SnSe2 electrodesand some reported tin chalcogenides for SIB anodes[21,24,26e28,37,41,42,49e52]. In the column of morphology, theSnSe2 crystal is among the minority of noncomposites, which couldreduce the inert components and simplify the production tech-nology. Yet with a little imagination the introducing of grapheneand even polymer materials via self-assembly or in situ polymeri-zation [53] may give such materials another boost. As for the col-umns of cycle number, current density, and capacity retention, theSnSe2 electrodes demonstrate the maximums, indicating that theencoded SeeSe buffer layers are comparable to the adscititiousnanocomposite technology in improving the long-term cycling

stability, and even better at high current density. To evaluate thepracticability of the SnSe2 electrodes, full-cell test is performedwith Na3V2(PO4)3 material (the half-cell performances ofNa3V2(PO4)3 is shown in Fig. S5) as the cathode between 0.3 and2.9 V (Fig. S6). Satisfactory capacity delivery and cycling stability areachieved, indicating the feasibility of practical application for theSnSe2 electrodes.

To better understand the positive effect enabled by the trans-formation from the SeeSe buffer layers to Na2Se layers in thesodiation process, the SEM and HRSEM images of the SnSe2 elec-trodes before the battery test (initial electrode), discharged to 1.0 V(1.0 V electrode), and discharged to 0.1 V (0.1 V electrode) areillustrated in Fig. S7. In Fig. S7a and 7b, the initial SnSe2 particles arefeatured of micron-size distribution and irregular shape withdistinct edges and smooth surfaces. The morphology of the 1.0 Velectrode shown in Fig. S7c changes significantly compared to thatof the initial electrode, manifesting as the disappearance of the bulkstructure of SnSe2 and the appearance of rough surface with manyhollows. As for the 0.1 V electrode shown in Fig. S7d, the bumpsdisplace the hollows and the SnSe2 particles could hardly be foundon the surface. This phenomenon should be caused by the coverageof sodium carboxymethylcellulose (CMC), which could be provenby the TEM and its corresponding energy dispersive spectroscopy(EDS) images shown in Fig. S8. More specifically, the interceptedpart in Fig. S8a consists of relatively large particles in deep black, ifthe assumption of CMC coverage is invalid, the mappings of tin andselenium elements shown in Fig. S8b and 8c should be moreintensive upon the deep black particle, but the uniform distributionof tin and selenium elements should support the assumption. Theuniform distribution of sodium element shown in Fig. S8d againstthe uneven particle distribution illustrates that the main contri-bution of sodium element comes from the covered CMC, which is inaccordance with the morphology in Fig. S7d. The feature HRSEMimage of the 1.0 V electrode is exhibited in Fig. 4a. The bulkstructure of SnSe2 changed into the flower-like lamellar nano-structure, and the thickness of these ultrathin lamellas are less than50 nm. The sodium-ion intercalation process is over in the node of1.0 V, and it only separates but does not destroy the nanosheets.This unique transformation of the morphology could be attributed

Fig. 4. (a) HRSEM of the 1.0 V SnSe2 electrode. The bulk structure of SnSe2 changed into the flower-like lamellar nanostructure. (b) Feature SEM image of the 0.1 V SnSe2 electrode.The flower-like lamellar nanostructure continues into the sodiated 0.1 V SnSe2 only with the thickness change. (c) XPS curves of the 1.0 V SnSe2 electrode for tin and seleniumelements. (d) XPS curves of the 0.1 V SnSe2 electrode for tin and selenium elements. (e) HRTEM image of the discharged SnSe2 electrode. Both the NaxSn and Na2Se could be found inthe image. (f) Ex situ XRD patterns of the initial and discharged SnSe2 electrodes. Four peaks within five degrees appear for the discharged electrode. The scale bars are 2 mm,200 nm, and 5 nm in panels (a), (b), and (e), respectively.

X. Wang et al. / Journal of Power Sources 369 (2017) 138e145 143

to the accommodation of the SeeSe buffer layers with admirableflexibility and toughness. Despite the CMC coverage, the flower-likelamellar nanostructure continues into the sodiated 0.1 V SnSe2 onlywith the thickness change, which is displayed in Fig. 4b. This sug-gests that the Na2Se substrate could confine the alloying reaction ofthe tin sequences to support the integrated lamellar structure evenwith a deep sodiation process to 0.1 V. Superior structural stabilityis favourable to the long-term cycling stability especially at highcurrent density [51].

The X-ray photoelectron spectroscopy (XPS) tests of the initial,1.0 V, and 0.1 V SnSe2 electrodes were also launched to consolidatethe accommodation of the encoded SeeSe buffer layers and theconfinement of the autogenous Na2Se layers in the sodiation pro-cess. Fig. S9a and 9b illustrate the original XPS curves of the initial,1.0 V, and 0.1 V SnSe2 electrode for tin and selenium elements,respectively. In Fig. S9a, the Sn 3d3/2 peaks where the shaded arealies are interfered by the strong Auger peaks of sodium element,only the Sn 3d5/2 peaks could be analysed. The shaded area inFig. S9b represents the interferential peaks of Na 2s. The peak in-tensity of the 0.1 V electrode is much lower than that of the initialand 1.0 V electrodes, indicating the absence of tin and seleniumelements on the surface of the electrode, which is consistent withthe EDS images in Fig. S8. The simulated XPS curves of the initialelectrode for Sn 3d5/2 and Se 3d are exhibited in Fig. S9c and 9d,respectively. The Sn 3d5/2 curve could be deconvoluted into twopeaks: the main peak at 487.0 eV could be assigned to the Sn4þ in

SnSe2 [41], and the small shoulder peak at 484.6 eV could beindexed to the low valence state of Sn. The Se 3d curve could bedeconvoluted into two peaks of Se 3d5/2 in Se2� (54.5 eV) and Se3d3/2 in Se2� (55.4 eV). In Fig. 4c, the Sn 3d5/2, Se 3d5/2, and Se 3d3/2of the 1.0 V electrode are severally simulated to 487.1, 54.1, and55.0 eV, that is, a small blue shift for Sn 3d and a red shift for Se 3dcompared to the initial electrode, which should be caused by thesodium-ion intercalation into the SeeSe buffer layers. In Fig. 4d, theSn 3d5/2, Se 3d5/2, and Se 3d3/2 of the 0.1 V electrode are simulatedto 487.4, 54.3, and 55.2 eV, respectively. The Se 3d in Se2� could beassigned to the Na2Se product [54]. The NaxSn product in the 0.1 Velectrode belongs to the group of electrides [55], and it does notfollow the bond-valence theory. The introduction of sodiumelement (one valence electron) could decrease the concentration ofthe valence electrons of tin element (four valence electrons), i.e.,the shielding effect from the valence electrons would decrease,thus leading to the increase of binding energy for the inner elec-trons (Sn 3d). As shown in Fig. 4e, the NaxSn and Na2Se products arevisible in the HRTEM image of the discharged SnSe2 electrode. Fourdifferent kinds of lattice spacings for NaxSn product could be foundin Fig. 4e and Fig. S10, and the NaxSn particles are surrounded bythe Na2Se product, which could confine the volume expansion ofthe alloying reaction. The corresponding ex situ XRD pattern isexhibited in Fig. 4f. Four peaks within five degrees appear for thedischarged SnSe2 electrode, despite the weakness compared to the(001) and (002) peaks of the initial electrode. A previous study of

X. Wang et al. / Journal of Power Sources 369 (2017) 138e145144

SnS2 material as SIB anode [52] suggested that there is no obviousNa2S peak in the range of 20e30� during the discharge process,which is consistent with our result, despite the Na2Se product isdocumented in Fig. 4e and previous study [37]. Therefore, the fourpeaks should be indexed to different NaxSn products, which agreeswell with the HRTEM images in Fig. 4e and Fig. S10.

The XPS measurements for the SnSe2 electrode recharged to3.0 V (recharged electrode) are performed to evaluate the resilienceof the SeeSe buffer layers. In Fig. S11, the Sn 3d5/2, Se 3d5/2, and Se3d3/2 are severally simulated to 487.2, 54.1, and 55.0 eV. Thesevalues are close to that of the 1.0 V electrode, which could be causedby the concentration polarization, i.e., the sodium-ion dein-tercalation is incomplete. The electrochemical impedance spec-troscopy (EIS) measurement for the recharged SnSe2 electrode isimplemented to confirm the electrochemical stability. As shown inFig. S12a, the semicircle at high frequency represents the charge-transfer resistance (Rct) between the electrode and electrolyte[56,57] and the line at low frequency relates to the sodium-iondiffusion process. The Rct value, which is simulated to be 15.2 U,is far smaller than that of most ester-based electrolyte systems,indicating a rapid charge-transfer process. Fig. S12b demonstratesthe relationship between Zre and u�1/2 (u is angular frequency;u ¼ 2pf) among the low-frequency region. By analogy to thesuperlattices of graphene oxide and titanium oxide [58], thealloying reaction could be regarded as the sodiation of tin se-quences within an interminable sandwich structure of“Na2SeeSneNa2Se”, and unlike the heterogeneous integration ofgraphene into nanocomposite [37], the autogenous Na2Se layerscould not only guarantee the cycling stability but also improve thecontent of active materials.

4. Conclusions

In summary, the Se�Se buffer layers are encoded into tin se-quences as the SnSe2 single crystal to ensure the long-term sodium-ion storage by blazing a trail of self-defence strategy to structuraldamage especially at high current density. The highly integratedand regular Se�Se buffer layers with admirable flexibility andtoughness are favourable to accommodate the sodium-ion inter-calation process, and the autogenous Na2Se layers generated fromthe conversion reaction could confine the structural pulverizationof further sodiated tin sequences by the slip along the Na2SeeNaxSninterfaces during the alloying process. The encoded SeeSe bufferlayers and autogenous Na2Se layers could work along with thenanocomposite technology in long-term cycling stability or even dobetter at high current density. The encoding of SeeSe buffer layerswith the autogenous Na2Se layers could offer a promising route toimprove the cycling stability in the next-generation SIB anodes andother battery systems such as LIB anodes, magnesium-ion battery(MIB) cathodes, and even the intercalation cathodes for the LIB andSIB fields. We also propose that the repertoire of the highly orderedand extended buffer layers with admirable resistance to structuraldamage could be derived to encode the graphene- and evenpolymer/protein-mimetic materials.

Acknowledgements

The authors wish to thank T. Zhang, Prof. G. Chen, Prof. Z. Renand Prof. Y. W. Zhang for their strong supports. This work wassupported by the National Natural Science Foundation of China(21622407, 21673008), the National Key Research and Develop-ment Program of China (2016YFB0700604), Guangdong Innovativeand Entrepreneurial Research Team Progress (2013N080) and theShenzhen Science and Technology Research Grant(JCYJ20160531141109132, JCYJ20170412150450297).

Appendix A. Supplementary data

Supplementary data related to this article can be found athttps://doi.org/10.1016/j.jpowsour.2017.09.088.

References

[1] J. Billaud, F. Bouville, T. Magrini, C. Villevieille, A.R. Studart, Nat. Energy 1(2016) 16097.

[2] M. Armand, J.-M. Tarascon, Nature 451 (2008) 652e657.[3] A.R. Armstrong, C. Lyness, P.M. Panchmatia, M.S. Islam, P.G. Bruce, Nat. Mater.

10 (2011) 223e229.[4] D. Larcher, J.-M. Tarascon, Nat. Chem. 7 (2015) 19e29.[5] K.-T. Kim, G. Ali, K.Y. Chung, C.S. Yoon, H. Yashiro, Y.-K. Sun, J. Lu, K. Amine, S.-

T. Myung, Nano Lett. 14 (2014) 416e422.[6] X. Xiang, K. Zhang, J. Chen, Adv. Mater. 27 (2015) 5343e5364.[7] M.D. Slater, D. Kim, E. Lee, C.S. Johnson, Adv. Funct. Mater. 23 (2013) 947e958.[8] M. Winter, R.J. Brodd, Chem. Rev. 104 (2004) 4245e4270.[9] P. Ge, M. Fouletier, Solid State Ionics 28 (1988) 1172e1175.

[10] D. Wu, X. Li, B. Xu, N. Twu, L. Liu, G. Ceder, Energy Environ. Sci. 8 (2015)195e202.

[11] Y. Wen, K. He, Y. Zhu, F. Han, Y. Xu, I. Matsuda, Y. Ishii, J. Cumings, C. Wang,Nat. Commun. 5 (2014) 4033.

[12] C.K. Chan, H. Peng, G. Liu, K. McIlwrath, X.F. Zhang, R.A. Huggins, Y. Cui, Nat.Nanotechnol. 3 (2008) 31e35.

[13] Y. Li, K. Yan, H.-W. Lee, Z. Lu, N. Liu, Y. Cui, Nat. Energy 1 (2016) 15029.[14] J. Song, Z. Yu, M.L. Gordin, X. Li, H. Peng, D. Wang, ACS Nano 9 (2015)

11933e11941.[15] S.W. Kim, D.H. Seo, X. Ma, G. Ceder, K. Kang, Adv. Energy Mater. 2 (2012)

710e721.[16] B. Luo, L. Zhi, Energy Environ. Sci. 8 (2015) 456e477.[17] C. Wang, W. Wan, Y. Huang, J. Chen, H.H. Zhou, X.X. Zhang, Nanoscale 6 (2014)

5351e5358.[18] H. Li, L. Yang, J. Liu, S. Li, L. Fang, Y. Lu, H. Yang, S. Liu, M. Lei, J. Power Sources

324 (2016) 780e787.[19] (a) F. Li, J. Chen, X. Wang, M. Xue, G. Chen, Adv. Funct. Mater. 25 (2015)

4601e4606;(b) M. Xue, D. Chen, X. Wang, J. Chen, G. Chen, J. Mater. Chem. A 3 (2015)7715e7718.

[20] L. Wu, D. Buchholz, D. Bresser, L.G. Chagas, S. Passerini, J. Power Sources 251(2014) 379e385.

[21] Y. Kim, Y. Kim, Y. Park, Y.N. Jo, Y.-J. Kim, N.-S. Choi, K.T. Lee, Chem. Commun.51 (2015) 50e53.

[22] K. Zhang, Z. Hu, X. Liu, Z. Tao, J. Chen, Adv. Mater. 27 (2015) 3305e3309.[23] C. Zhu, X. Mu, P.A. van Aken, Y. Yu, J. Maier, Angew. Chem. Int. Ed. 53 (2014)

2152e2156.[24] Q. Li, Q. Wei, W. Zuo, L. Huang, W. Luo, Q. An, V.O. Pelenovich, L. Mai, Q. Zhang,

Chem. Sci. 8 (2017) 160e164.[25] X. Ma, M. Xue, F. Li, J. Chen, D. Chen, X. Wang, F. Pan, G. Chen, Nanoscale 7

(2015) 8715e8719.[26] A. Jahel, C.M. Ghimbeu, A. Darwiche, L. Vidal, S. Hajjar-Garreau, C. Vix-Guterl,

L. Monconduit, J. Mater. Chem. A 3 (2015) 11960e11969.[27] Y. Jiang, M. Wei, J. Feng, Y. Ma, S. Xiong, Energy Environ. Sci. 9 (2016)

1430e1438.[28] T. Zhou, W.K. Pang, C. Zhang, J. Yang, Z. Chen, H.K. Liu, Z. Guo, ACS Nano 8

(2014) 8323e8333.[29] P. Thomas, D. Billaud, Electrochim. Acta 47 (2002) 3303e3307.[30] S.W. Kim, D.H. Seo, X. Ma, G. Ceder, K. Kang, Adv. Energy Mater. 2 (2012)

710e721.[31] K. Su, C. Wang, H. Nie, Y. Guan, F. Liu, J. Chen, J. Mater. Chem. A 2 (2014)

10000e10006.[32] A. Vlad, A.L.M. Reddy, A. Ajayan, N. Singh, J.-F. Gohy, S. Melinte, P.M. Ajayan,

Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 15168e15173.[33] A. Magasinski, P. Dixon, B. Hertzberg, A. Kvit, J. Ayala, G. Yushin, Nat. Mater 9

(2010) 353e358.[34] H.X. Dang, K.C. Klavetter, M.L. Meyerson, A. Heller, C.B. Mullins, J. Mater.

Chem. A 3 (2015) 13500e13506.[35] (a) Z. Yang, H. Liang, X. Wang, X. Ma, T. Zhang, Y. Yang, L. Xie, D. Chen, Y. Long,

J. Chen, Y. Chang, C. Yan, X. Zhang, X. Zhang, B. Ge, Z. Ren, M. Xue, G. Chen, ACSNano 10 (2015) 755e762;(b) X. Wang, D. Chen, Z. Yang, X. Zhang, C. Wang, J. Chen, X. Zhang, M. Xue,Adv. Mater. 28 (2016) 8645e8650.

[36] D. Chen, X. Wang, J. Chen, Z. Ren, M. Xue, G. Chen, Adv. Mater. 27 (2015)4224e4228.

[37] F. Zhang, C. Xia, J. Zhu, B. Ahmed, H. Liang, D.B. Velusamy,U. Schwingenschl€ogl, H.N. Alshareef, Adv. Energy Mater. 6 (2016) 1601188.

[38] (a) M. Xue, Fengwang Li, Tingbing Cao, Nanoscale 4 (2012) 1939e1947;(b) J. Zhu, M. Xue, D. Zhao, M. Zhang, L. Duan, Y. Qiu, T. Cao, Angew. Chem. Int.Ed. 50 (2011) 12478e12482.

[39] (a) R.V. Mannige, T.K. Haxton, C. Proulx, E.J. Robertson, A. Battigelli,G.L. Butterfoss, R.N. Zuckermann, S. Whitelam, Nature 526 (2015) 415e420;(b) Y. Liu, H. Liang, Z. Xu, J. Xi, G. Chen, W. Gao, M. Xue, C. Gao, ACS Nano 11(2017) 4301e4306.

X. Wang et al. / Journal of Power Sources 369 (2017) 138e145 145

[40] (a) M. Xue, Y. Wang, X. Wang, X. Huang, J. Ji, Adv. Mater. 27 (2015)5923e5929;(b) M. Xue, F. Li, D. Chen, Z. Yang, X. Wang, J. Ji, Adv. Mater. 28 (2016)8265e8270.

[41] Y. Zhang, P. Zhu, L. Huang, J. Xie, S. Zhang, G. Cao, X. Zhao, Adv. Funct. Mater.25 (2015) 481e489.

[42] B. Qu, C. Ma, G. Ji, C. Xu, J. Xu, Y.S. Meng, T. Wang, J.Y. Lee, Adv. Mater. 26(2014) 3854e3859.

[43] (a) M. Li, S. Xiao, R. Yan, S. Vishwanath, S. Fullerton-Shirey, D. Jena, H.G. Xing,in: Device Research Conference (DRC), 2016 74th Annual, IEEE, 2016, pp. 1e2;(b) X. Ma, D. Zhao, M. Xue, H. Wang, T. Cao, Angew. Chem. Int. Ed. 49 (2010)5537e5540;(c) D. Zhao, L. Duan, M. Xue, W. Ni, T. Cao, Angew. Chem. Int. Ed. 48 (2009)6699e6703.

[44] (a) S.Z. Butler, S.M. Hollen, L.Y. Cao, Y. Cui, J.A. Gupta, H.R. Gutierrez, T.F. Heinz,S.S. Hong, J.X. Huang, A.F. Ismach, E. Johnston-Halperin, M. Kuno,V.V. Plashnitsa, R.D. Robinson, R.S. Ruoff, S. Salahuddin, J. Shan, L. Shi,M.G. Spencer, M. Terrones, W. Windl, J.E. Goldberger, ACS Nano 7 (2013)2898e2926;(b) H. Liang, X. Ma, Z. Yang, P. Wang, X. Zhang, Z. Ren, M. Xue, G. Chen, Carbon99 (2016) 585e590.

[45] B. Luo, Y. Fang, B. Wang, J. Zhou, H. Song, L. Zhi, Energy Environ. Sci. 5 (2012)5226e5230.

[46] C. Li, L. Huang, G.P. Snigdha, Y.F. Yu, L.Y. Cao, ACS Nano 6 (2012) 8868e8877.[47] A. Abouimrane, D. Dambournet, K.W. Chapman, P.J. Chupas, W. Weng,

K. Amine, J. Am. Chem. Soc. 134 (2012) 4505e4508.[48] C.S. Park, H. Kim, R.A. Shakoor, E. Yang, S.Y. Lim, R. Kahraman, Y. Jung,

J.W. Choi, J. Am. Chem. Soc. 135 (2013) 2787e2792.[49] D. Su, X. Xie, G. Wang, Chem. Eur. J. 20 (2014) 3192e3197.[50] Z. Zhang, X. Zhao, J. Li, Mater. Lett. 162 (2016) 169e172.[51] Y. Liu, H. Kang, L. Jiao, C. Chen, K. Cao, Y. Wang, H. Yuan, Nanoscale 7 (2015)

1325e1332.[52] W. Sun, X. Rui, D. Yang, Z. Sun, B. Li, W. Zhang, Y. Zong, S. Madhavi, S. Dou,

Q. Yan, ACS Nano 9 (2015) 11371e11381.[53] (a) M. Xue, X. Ma, Z. Xie, L. Duan, Y. Jiang, M. Zhang, T. Cao, Chem. Asian J. 5

(2010) 2266e2270;(b) M. Xue, F. Li, Y. Wang, X. Cai, F. Pan, J. Chen, Nanoscale 5 (2013)1803e1805;(c) S.H. Lee, H.W. Kim, J.O. Hwang, W.J. Lee, J. Kwon, C.W. Bielawski, R.S. Ruoff,S.O. Kim, Angew. Chem. Int. Ed. 122 (2010) 10282e10286.

[54] X. Zhu, Z. Zhou, Y. Wang, L. Zhang, A. Li, F. Huang, Sol. Energy Mater. Sol. Cells101 (2012) 57e61.

[55] A.S. Ichimura, J.L. Dye, M.A. Camblor, L.A. Villaescusa, J. Am. Chem. Soc. 124(2002) 1170e1171.

[56] E.M. Lotfabad, J. Ding, K. Cui, A. Kohandehghan, W.P. Kalisvaart, M. Hazelton,D. Mitlin, ACS Nano 8 (2014) 7115e7129.

[57] Y. Li, Z. Wang, L. Li, S. Peng, L. Zhang, M. Srinivasan, S. Ramakrishna, Carbon 99(2016) 556e563.

[58] X. Cai, T.C. Ozawa, A. Funatsu, R. Ma, Y. Ebina, T. Sasaki, J. Am. Chem. Soc. 137(2015) 2844e2847.