Reduced Graphene

Graphene Oxide Composites Enabling Stable and High-Capacity Sodium Storage. Xuming Yang†‡ , Jinkai Wang§ ... Publication Date (Web): July 11,...
0 downloads 0 Views 5MB Size
Download PDF
www.acsnano.org

Downloaded via TUFTS UNIV on July 11, 2018 at 22:46:14 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Vapor-Infiltration Approach toward Selenium/ Reduced Graphene Oxide Composites Enabling Stable and High-Capacity Sodium Storage Xuming Yang,†,‡ Jinkai Wang,§ Shuo Wang,∥ Hongkang Wang,§ Ondrej Tomanec,⊥ Chunyi Zhi,† Radek Zboril,⊥ Denis Y. W. Yu,*,∥ and Andrey Rogach*,†,‡ †

Department of Materials Science and Engineering, ‡Center for Functional Photonics (CFP), and ∥School of Energy and Environment, and Center of Super-Diamond and Advanced Films (COSDAF), City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong S.A.R. § Center of Nanomaterials for Renewable Energy (CNRE), State Key Lab of Electrical Insulation and Power Equipment, School of Electrical Engineering, Xi’an Jiaotong University, Xi’an 710049, China ⊥ Regional Centre of Advanced Technologies and Materials, Faculty of Science, Department of Physical Chemistry, Palacky University in Olomouc, 77146 Olomouc, Czech Republic S Supporting Information *

ABSTRACT: Emerging sodium−selenium batteries suffer from volume expansion of the selenium cathode and shuttling effects of soluble intermediates. Confining selenium within the carbon matrix is the most adopted strategy to address these two issues, which is generally realized via a melt-infusion method. Herein, we developed a vapor-infiltration method to fabricate selenium/carbon composites that are advantageous over the meltinfusion route in terms of several aspects: it relieves the requirement of intensive mechanical mixing and simplifies the ratio optimization between selenium and carbon; it avoids selenium aggregation and makes it possible to utilize all of the surface and pores of the carbon host. Utilizing this method, we fabricated a selenium/graphene composite from thermally reduced graphene oxide with a selenium loading equal to 71 wt %, thus approaching the record value. The obtained composite achieved the highest reported to date initial Coulombic efficiency of 88% among various selenium cathodes, with superior rate and cycle performance (410 and 367 mA h g−1 at 0.1 and 1 A g−1; capacity decay 70 wt %).27,28 In addition, the high surface area of the host carbon material for Se accommodation is often under-utilized because the amount of Se is limited due to competition between the complete Se infusion and Se aggregation. Alternatively, Se can be infiltrated into the carbon host at higher temperature so that it becomes vaporized and fills all the interstitial space of the carbon host. Mechanical mixing and optimization of feeding ratios are not the issues in this method anymore as Se and carbon are not mixed together but are placed separately in the same sealed vessel so that the excess Se is kept away from the carbon throughout. Herein, we demonstrate the feasibility of such a vapor-infiltration method of selenium on thermally reduced graphene oxide. The realized selenium/graphene composites with uniform and high Se loading (71 wt %) achieve high utilization of Se, high initial Coulombic efficiency, high-capacity retention in rate tests, and high stability in cycling in an ether electrolyte. They deliver a reversible capacity of 410 mA h g−1 with a Coulombic efficiency of 88% in the initial cycle at 0.1 A g−1 when coupled with Na metal as the anode, and the capacity increase to 420 mA h g−1, reaching nearly 90% of the theoretical value. The capacity retention from 0.1 to 1 A g−1 (410 and 367 mA h g−1) is as high as 89%, and the decay after 800 cycles at 2 A g−1 is less than 10%. In stark contrast, the control sample with a similar Se loading made by a traditional melt-infusion method suffered a severe capacity loss during the initial few cycles owing to the Se aggregation in the composite, and its reversible capacity was only 263 mA h g−1. Thus, the vapor-infiltration method not only simplifies the fabrication procedures of Se/ carbon composites but also brings vast improvements in the sodium battery performance; it thus has the potential to take over the place of the traditional melt-infusion method and become a leading approach toward Se/carbon composites, enabling stable and high-capacity sodium storage.

cathode side. Significant advances in sodium metal anodes have been achieved recently, which prompt extra momentum for the development of sulfur and selenium cathodes.9−11 Both sulfur and selenium suffer not only from drastic volume changes similar to that in the conversion-based anode materials but also from dissolution of intermediate polysulfides or polyselenides in electrolytes.12−14 Encapsulation is the most common strategy to address these two issues and stabilize battery cycling.7,13,15−21 However, the capacity is often reduced due to the incorporation of inactive host materials. The selection of electrolyte is also crucial for extending cycle lifespan. An ether-based electrolyte (NaSO3CF3 in diethylene glycol dimethyl ether) has been demonstrated to be useful in improving initial Coulombic efficiency and cycle stability of carbon and chalcogenide anode materials for sodium ion batteries.22,23 The improvements were ascribed to the thin and compact solid electrolyte interface (SEI) films which can reduce irreversible capacity caused by SEI formation on one hand and can prevent direct contact between electrolyte and active materials on the other hand, thus restraining further decomposition of the electrolyte as well as the dissolution of discharge/charge intermediates.22,24,25 Apart from good cyclability, efficient utilization of sulfur or selenium through high material loading is also required to maintain a competitive edge over layered and polyanionic cathode materials. This is particularly critical for Se, which possesses a comparable volumetric capacity to sulfur (3254 vs 3467 mA h cm−3) but falls short in terms of gravimetric capacity (678 vs 1675 mA h g−1). For Na−S batteries, the utilization of sulfur is usually low because of the insulative nature of sulfur and sulfides.13,26 In addition, the dissolution in the Na−S system still remains an issue. In contrast, Na−Se batteries are attractive, with several inspiring cases showing excellent stability and high Se loading (>50 wt % of the composite) in recent literature.18,27 This is attributed to two compelling properties of Se: its superior conductivity compared to that of S (1 × 10−3 vs 0.5 × 10−27 S m−1) and its compatibility with several kinds of electrolytes.28,29 Incorporating Se into the carbon matrix is an effective approach to prevent the loss of soluble intermediates, which is generally realized by a melt-infusion method, where the melted Se is infused into pores of carbon hosts via capillary forces.15−19 This method, however, suffers from two intrinsic drawbacks: the molten Se cannot wet the whole carbon host due to the limited volume, and the excess Se can easily form aggregated particles difficult to be removed. For these two reasons, Se and carbon should be fed with an appropriate ratio

RESULTS AND DISCUSSION Two fabrication approaches toward selenium/graphene composites utilized and compared in this work are schematically illustrated in Figure 1. An annealing treatment is applied to the commercial graphene oxide (GO) to create large B

DOI: 10.1021/acsnano.8b04114 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 2. SEM images of (a) GO and (b,c) TRGO at different magnifications. (d,e) TEM image of VSeG at different magnifications. EDS elemental mapping for Se (f) and C (g) performed on the selected area of the sample shown in (e).

Figure 3. (a) EDS spectra, (b) XPS survey spectra, and (c) TGA curves of TRGO, MSeG, and VSeG. (d) XRD patterns, (e) Raman spectra, and (f) core-level Se 3d XPS spectra of Se powder, MSeG, and VSeG.

amounts of defects as well as to reduce the number of layers through elimination of the oxygen-containing groups, resulting in a thermally reduced graphene oxide (TRGO) suitable for Se loading. Then, either the traditional melt-infusion method or the new vapor-infiltration method is applied to fabricate Se/ graphene composites which are denoted as MSeG and VSeG, respectively. In the melt-infusion route, the mixture of Se powders and TRGO is heated at 260 °C, allowing Se to be melted and wet TRGO. Molten Se cannot cover all surfaces and enter every pore due to its limited volume in a liquid state (4.83 g cm−3 for trigonal Se and 3.99 g cm−3 for Se liquid). On top of that, some Se droplets may turn into particles during cooling. For the vapor-infiltration method, the heating temperature (600 °C) is so high that Se can be vaporized to fill the whole inner space of the sealed vessel, and TRGO is totally immersed in Se vapor. In this case, selenium coating can

be presumably generated on all exposed carbon layers and inner surface of all open cavities. Following the above outlined synthetic methodology, we fabricated Se/graphene samples and characterized their morphologies using field emission scanning electron microscopy (FESEM) and transmission electron microscopy (TEM). Figure 2a shows the scanning electron microscopy (SEM) image of pristine GO, which are micron-sized, graphite-like bulk sheets. Characteristics of graphene are only revealed after annealing, with loosely stacked ultrathin graphene layers, as seen in the SEM images of TRGO (Figure 2b,c). It was previously suggested that the rapid emission of carbon monoxide/dioxide and H2O upon annealing creates high pressure within GO and thus gives rise to its explosive expansion.30 To contain the fast reaction, we first put GO in a large beaker and applied a preliminary annealing in a muffle C

DOI: 10.1021/acsnano.8b04114 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 4. Digital photos and SEM images of (a) pristine glass fiber separator and that in Na-MSeG and Na-VSeG cells using (b,c) DEGDMEbased and (d,e) PC-based electrolytes after one cycle.

furnace at 300 °C. Then, the expanded GO was collected and transferred into a tube furnace and annealed at 900 °C. The obtained TRGO was used to prepare MSeG and VSeG through melt-infusion and vapor-infiltration methods. Despite the large difference in volume (Figure S1), MSeG and VSeG appear alike TRGO under SEM (Figure S2a,b); loose layerstacking architecture is well-preserved after Se loading. TEM images (Figure 2d and Figure S2c,d) show that all three samples comprise wrinkled planes with void space between. Through the energy-dispersive X-ray spectroscopy (EDS) elemental mapping on a selected area (Figure 2e), we confirmed the uniform distribution of Se and C in the VSeG composite (Figure 2f,g, respectively). Such a morphology should be advantageous for an extended cyclability with sodium incorporation able to withstand about 290% volume increase from Se to Na2Se, as estimated from the difference in lattice constants of the two materials. Other than morphological features, Se contents and chemical forms for the TRGO, MSeG, and VSeG samples are further studied. EDS and X-ray photoelectron spectroscopy (XPS) survey data are presented in Figure 3a,b, respectively, and show the presence of Se in MSeG and VSeG. The Se loading of VSeG is determined to be 71 wt % by thermogravimetric analysis (TGA) (Figure 3c), which is nearly as high as the documented record of 72 wt % for melt-infusion methods (Table S1).27 Specific surface areas of TRGO, MSeG, and VSeG are estimated to be 360, 61, and 48 m2 g−1, respectively, with pore volumes of 3.2, 0.58, and 0.48 cm3 g−1 based on their nitrogen adsorption−desorption isotherms (Figure S3a,b). There is severe decrease in both the surface areas and pore volumes even when the values of MSeG and VSeG are normalized to the weight of constituent TRGO, but the change of the pore size (Figure S3b) is minor. This can be attributed to the thin Se deposited layers that smoothen the TRGO surface but do not notably reduce the pore size. X-ray diffraction (XRD) pattern shows that MSeG still features diffraction peaks of trigonal Se, whereas VSeG appears to be amorphous (Figure 3d). According to the Raman spectra (Figure 3e), only signals originating from helical Se chains (256 cm−1) are recorded for VSeG, whereas MSeG combines the signals from trigonal and helical Se (235 and 256 cm−1).31

For pure trigonal Se particles, the crystallinity is restored after melting and condensation (Figure S4). Previously, both crystalline and noncrystalline Se/C produced by melt-infusion methods were reported in the literature.15,17 If melted Se is well-dispersed in a carbon matrix, it may turn into Se rings or chains upon solidification, and if it aggregates, long-range periodic crystalline structure can be restored. The core-level C 1s XPS spectra of TRGO, MSeG, and VSeG (Figure S5) are almost the same, whereas Se 3d XPS spectra of Se powders, MSeG, and VSeG appear significantly different from each other (Figure 3f). The XPS spectrum of Se powders can be deconvoluted into Se 3d5/2 and 3d3/2 components with an area ratio of 3:2, but for MSeG and VSeG, one more pair of peaks is needed to achieve acceptable fitting goodness, which account for 28 and 33% of the total integral area for MSeG and VSeG, respectively. These additional contributions in the two Se/graphene composites cannot originate from chemical bonding between carbon and Se because the C 1s XPS spectra are unchanged. Instead, we ascribed them to selenium atoms with unsaturated coordination at the outermost surfaces and the ends of chains. Furthermore, the absence of Se 3d peaks from SeOx implies no air penetration into the sealed vessels. Sodium storage properties of MSeG- and VSeG-based electrodes are explored by pairing them with sodium foils in coin cells. We note that in most publications on Na−Se batteries, specific capacity is calculated based on the weight of Se only, but in this work, the capacity is normalized by the weight of the whole composites (Se/TRGO) in order to highlight the superiority in Se loading. The electrochemical characteristics of the samples are found to be strongly dependent on the type of electrolyte used in the cells. As shown in Figure S6, rapid capacity decay is observed for both MSeG and VSeG samples when a carbonate electrolyte (1 M NaClO4 in propylene carbonate, PC) is employed. The respective initial charging and discharging voltage profiles are shown in Figure S7. Both samples can deliver a capacity approaching the theoretical value of Se components (678 mA h g−1 × Se ratio), but initial Coulombic efficiency is calculated to be only 31 and 45% for MSeG and VSeG with PC as the electrolyte. When the ether electrolyte (1 M NaSO3CF3 in diethylene glycol dimethyl ether, DEGDME) was used, the D

DOI: 10.1021/acsnano.8b04114 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 5. Initial five cyclic voltammograms of (a) MSeG and (b) VSeG at 0.2 mV s−1; (c) cycle performance of VSeG and MSeG at 0.1 A g−1; (d) voltage profiles and (e) rate performance of VSeG from 0.1 to 5 A g−1; (f) long-term cycle stability of VSeG at 2 A g−1.

strongly decreases at the second cycle, and the peak around 2.0 V becomes far weaker. Such a decrease, which is more apparent in the CVs of Se particles (Figure S8a) but hardly visible for VSeG, is the result of the loss of aggregated Se particles. It can be quantified as one minus the ratio of integral areas in two successive cycles, delivering numbers of 75, 48, and 17% for Se powder, MSeG, and VSeG, respectively. The capacity is reversible in the subsequent cycles, as CV curves basically overlap since the second cycle. The trend is consistent with cycle performance at 0.2 A g−1, as shown in Figure 5c. The initial specific discharge capacity of the MSeG electrode is measured to be 473 mA h g−1, decreasing to 263 mA h g−1 in the next cycle. In contrast, VSeG delivered a slightly lower initial capacity of 450 mA h g−1 but remained at 390 mA h g−1 in the next cycle. In the following cycles, there is no substantial decay but rather a slight increase in capacity for both samples. The charge−discharge profiles of the VSeG electrode at different current rates are presented in Figure 5d, and the plot of capacity versus cycle number is shown in Figure 5e. The reversible capacity at 0.1 A g−1 reaches 410 mA h g−1, corresponding to 577 mA h g−1 based on the weight of Se only when accounting for the minor contributions from graphene (Figure S9). It is close to the best result reported in the literature, 613 mA h g−1 obtained at 36 mA g−1 from a composite with a Se content of 53 wt %.28 At the current rates of 0.2, 0.5, 1, 2, and 5 A g−1, the recorded capacity is 392, 377, 367, 356, and 320 mA h g−1, respectively. Capacity retention with respect to the value at 0.1 A g−1 is calculated to be 95, 91, 89, 86, and 77%. When the performance at 0.1 and 1 A g−1 is compared, only about 11% capacity is sacrificed when the current is increased 10-fold, and the voltage difference at the same charge or discharge capacity is less than 0.2 V. The VSeG electrode not only possesses exceptional rate capability but exhibits an extraordinary long-term durability, as shown in Figure 5f. The capacity retention is as high as 91% after 800 cycles at 2 A g−1.

Coulombic efficiencies of the MSeG and VSeG samples are significantly improved to 80 and 88%, respectively. The glass fiber separators used in these two electrolytes have been extracted from the cells after one discharge/charge cycle, and their photos and SEM images are shown in Figure 4 together with those of a pristine separator. Visually, the separators from the cells with carbonate-based electrolytes show more coloration than those with the ether electrolyte. SEM images further clarify the difference in the appearance. The white pristine separator is based on interwoven glass fibers with a clean surface (Figure 4a). After one cycle in the ether-based electrolyte, the surface is still clean (Figure 4b,c). In contrast, the separator used in the carbonate electrolyte has numerous particles covering the surface (Figure 4d,e), which are supposed to be intermediates detached from the electrode film. Active materials may break off from the current collector due to solubility (chemically) or insufficient confinement of binders and carbon matrices (mechanically) during sodium uptake and release processes, which usually leads to rapid deterioration during the cycling. In the present case, Se is not coated but deposited on graphene sheets, thus soluble intermediates cannot be effectively confined. We attribute the poor reversibility and stability in the carbonate electrolyte to the lack of encapsulation of polyselenides. Similar to the reported improvements for carbon and selenides as sodium ion anodes, the observed significant improvement in cycle reversibility and stability in the ether electrolyte is supposed to be related to a compact and thin organic layer formed at the exterior of the electrode, which restrains the loss of active materials.22,23 A control experiment with trigonal Se particles in ether electrolyte shows fast capacity decay due to the large volume expansion (Figure S7). This suggests that the graphene framework in our samples played an essential role in mechanically connecting the Se species during cycling. In the ether electrolyte, cyclic voltammograms (CVs) of MSeG and VSeG electrodes were recorded, as presented in Figure 5a,b. For MSeG, the area enclosed by the CV curves E

DOI: 10.1021/acsnano.8b04114 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

Figure 6. Electrochemical characteristics of VSeG: (a) differential specific capacity curve at 0.1 A g−1; cyclic voltammograms (b) from 0.1 to 0.4 and (c) from 0.4 to 3.2 mV s−1; (d) P1 to P5 peak intensity dependence on scan rate from 0.4 to 3.2 mV s−1; (e) EIS collected at various CV peaks; (f) pseudocapacitive contributions (shaded area) at 0.4 mV s−1.

1.8 and 2.1 V at the scan rate of 0.1 and 0.2 mV s−1, which merge into one when the scan rate is greater than 0.4 mV s−1. It happens because of kinetics limitation; in other words, there is not enough relaxation time for stepwise phase evolution during the rapid charging or quick scanning. For the five wellrecognizable peaks in the CV curves (denoted P1−P5) from 0.4 to 3.2 mV s−1, the dependence of the peak current on scan rates was analyzed according to the formula, i = avb or log(i) = log(a) + b log(v), where i is the peak current, v is scan rate, and a and b are parameters to be determined.32,33 The b value is commonly used to qualitatively describe the reaction mode, pseudocapacitive or diffusion-controlled. If CV curves at different rates hold the same shape and b value equals 1 (purely capacitive), the integral capacity will be independent of rates, which means the capacity will not decay as the discharge/charge current increases. The fitting results shown in Figure 6d reveal a linear relationship between log(i) and log(v), and the fitted b values vary between 0.7 and 0.97. The charge transfer resistance (the radius of the semicircle in the high-frequency region) from EIS collected at the five peaks (Figure 6e) is inversely correlated with the b value and suggests the same trends in kinetics. The CV data are further analyzed using the equation, i(V) = k1v + k2v1/2, to quantify pseudocapacitive (k1v) and diffusion-controlled (k2v1/2) contributions in the hybrid electrode processes.34 The pseudocapacitive fraction is shaded inside the CV curves (Figures 6f and S12), and the percentage is calculated as the ratio of the integral areas. The dominant contribution (around 70−80%) from pseudocapacitive effects account for the superior rate capability and is reasonably associated with the graphene@Se rather than Se@graphene structure, where sodium ions can unite with Se directly without diffusing through the carbon layer. The sodiation mechanisms of trigonal Se crystals or polymeric Se chains (Figure S13a,b) were both suggested to be multistep processes based on combination of experimental and theoretical studies.17,35 Trigonal Se experiences two

It is noteworthy that the CV curves of the initial and second cycle of the VSeG electrode are quite different in terms of peak positions and intensity. The reason may be the formation of a solid electrolyte interface layer and the activation of the electrodes in the initial cycle. The Raman peak (257 cm−1) of the VSeG electrode after one discharge/charge cycle (Figure S10) suggests that the structure of Se remains chain-like. Electrochemical impedance spectra (EIS) show that the charge transfer resistance is significantly reduced after cycling (Figure S11). From the second cycle onward, two pairs of sharp peaks appear on the CV curves, which are not observed in carbonatebased electrolytes. They correspond to two distinct voltage plateaus which are also visible in the differential capacity curves (Figure 6a): discharge plateaus (1.67 and 0.73 V) correspond to anodic peaks located at 1.63 and 0.64 V, and charge plateaus (1.45 and 1.99 V) correspond to cathodic peaks at 1.52 and 2.02 V. The discharge plateau at 0.73 V constitutes over onethird of the total capacity (∼150 out of 410 mAh g−1). We note that plateaus at such low potentials (below 0.8 V) were not observed when Na−Se batteries were tested in carbonatebased electrolyte.18,20,28 The estimated conversion potentials of Se into sodium polyselenides (NaSex, x ≥ 4) and sodium polyselenides into sodium selenide (Na2Se) via theoretical modeling are greater than 1.0 V, which means the corresponding reaction cannot be simply ascribed to conversion between different sodium (poly)selenides, and the electrolyte might be involved in the sodiation processes around 0.7 V.17 More research will be needed to shed light on the origin. The last charging plateau at 0.1 A g−1 (∼2.0 V) is higher than that at 2 A g−1 (∼1.9 V) (Figure 5d). Generally, the larger the current, the higher the charging plateau will be as polarization will increase as the current increases. The observed charging voltage profiles do not follow this trend because the corresponding electrochemical reactions are different. This is clearly reflected in the CV curves at different scan rates (Figure 6b,c). There are two cathodic peaks between F

DOI: 10.1021/acsnano.8b04114 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano

isothermally at 900 °C for 3 h under flowing N2 in a tube furnace to reduce the graphene oxide. Selenium/Graphene Composite via the Melt-Infusion Method. Thermal reduced graphene oxide (25 mg) and selenium powder (75 mg, 325 mesh) were ground together in an agate mortar, and then the mixture was transferred into a 25 mL glass bottle, sealed under vacuum, and heated in a muffle furnace at 260 °C for 12 h. The obtained sample is denoted as MSeG. Selenium/Graphene Composite via the Vapor-Infiltration Method. Thermally reduced graphene oxide (25 mg) was placed in a short one-side-closed glass tube together with 0.4 g of selenium powder, vacuum sealed in a 25 mL glass bottle, and heated in a muffle furnace at 600 °C for 9 h and then at 260 °C for 3 h. The obtained sample is denoted as VSeG. Note that the glass tube requires tight sealing, as the reaction temperature is much higher than that in the melt-infusion method (600 vs 260 °C). Characterization. XRD patterns were collected on a D2 PHASER X-ray diffractometer with Cu Kα (λ = 1.5406 Å) over a 2θ range of 10 to 80°. Sample morphologies were studied under a FESEM (FEI Quanta FEG 450) and a high-resolution transmission electron microscope (TEM TITAN operating at 60−300 kV) equipped with EDS used for chemical mapping of samples. Raman spectra were recorded on a Renishaw InVia Raman spectroscope using a 633 nm laser. TGA was conducted in N2 with a heating rate of 5 °C min−1 from ambient temperature to 600 °C following an isotherm at 600 °C for 1 h on a Mettler Toledo thermogravimetric analyzer. Nitrogen adsorption−desorption isotherms were collected using a Quantachrome surface area analyzer (Autosorb iQ-MP). XPS measurements were performed on a VG ESCALAB 220i-XL spectrometer equipped with a monochromatic Al Kα source. Battery Fabrication and Electrochemical Measurements. Selenium/graphene composites were mixed with acetylene black by grinding, followed by mixing with a certain amount of sodium carboxymethyl cellulose aqueous solution (5 wt %) to make a viscous slurry, so that the weight ratio of active materials, conductive reagent, and binder was 70:15:15. The slurry was cast onto copper foil and vacuum-dried at 110 °C for 4 h. 2032-Type coin cells were assembled by coupling the electrode film with sodium metal foil and glass fiber separator in an Ar-filled glovebox. We employed two types of electrolytes, namely, 1 M NaClO4 in propylene carbonate with 5 wt % fluoroethylene carbonate and 1 M NaSO3CF3 in DEGDME. Galvanostatic tests were performed on a Neware battery testing system. Cyclic voltammograms and EIS were collected on a multichannel VMP3 (Bio-Logic) electrochemical workstation.

intermediate crystalline phases (Na2Se4 and Na2Se2) and turns into Na2Se at the end of discharge; for Se chains, the Se atoms at both ends react first with sodium, thus leading to the direct formation of Na2Se and shortened chains. When the length of the chain is less than 4 atoms, Se can be directly transformed into Na2Se, with the claimed possibility of intermediate molecular phases including Na2Se4 and Na2Se2.17 It is speculated that the sharp redox peaks originate from the reaction between sodium and short Se chains or even small Se molecules, which features fast kinetics and prevents the formation of polyselenides.36 It is worth mentioning that the CV pattern is nearly the same as that observed for some transition metal selenides (FeSe2, CoSe2, Fe7Se8) tested in the ether electrolyte.23,37,38 Plausibly, identical CV patterns of the selenium- and selenide-based electrodes result from the phase segregation of the transition metals and selenium in selenides occurring upon cycling. We believe the interesting findings will trigger further in-depth studies to clarify the detailed electrochemical reactions through theoretical simulations and advanced in situ characterization techniques.39

CONCLUSIONS We developed a vapor-infiltration approach to achieve a high loading of Se onto thermally reduced graphene oxide, which provides several advantages over the traditional melt-infusion method. For the latter, both intensive mixing and the component ratio optimization are necessary, and the melted Se could only partially wet the carbon matrix; it is difficult to remove excess Se and realize uniform distribution of Se in the final product. In contrast, the vapor-infiltration method does not require any mixing, and the surplus Se does not adversely affect the sample; in addition, the carbon matrix can easily be immersed completely in Se vapor, thus ensuring a uniform Se deposition on all surface areas accessible to the vapor. The selenium/graphene composite (VSeG) obtained by the vaporinfiltration method has a Se content of 71 wt %, which is almost the largest recorded value compared to the values of most composites fabricated by the melt-infusion method (Table S1). Moreover, the vapor-infiltration method enables the deposition of Se on the surface rather than its encapsulation inside the carbon matrix. The lack of the surface protection layer excludes the use of traditional carbonate-based electrolytes where the capacity decay is expected due to the poor immobilization of intermediates. Thus, the cycling has been realized in the ether electrolyte (1 M NaSO3CF3 in diethylene glycol dimethyl ether), providing an important advantage, because sodium ions do not need to cross carbon layers before reacting with Se in the Se@graphene structure. As a result, the delivered rate capability and cycle stability of VSeG outperformed those in most previous works (Table S1). We anticipate that the simple and more efficient vaporinfiltration method may be preferentially considered for fabrication of Se-based composites with different host materials. Moreover, we introduced the ether electrolyte into Na−Se battery studies and demonstrated its prominent enhancement on both initial Coulombic efficiency and cycle stability.

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.8b04114. Digital photos of the samples, SEM and TEM images, nitrogen adsorption−desorption isotherms, additional battery data, Raman spectrum of cycled electrode, capacitive contributions at various scan rates, schematics of sodiation mechanism (PDF)

AUTHOR INFORMATION Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Hongkang Wang: 0000-0003-4893-5190 Chunyi Zhi: 0000-0001-6766-5953 Radek Zboril: 0000-0002-3147-2196 Denis Y. W. Yu: 0000-0002-5883-7087 Andrey Rogach: 0000-0002-8263-8141

METHODS Materials Preparation. Thermally Reduced Graphene Oxide. Commercial few-layer graphene oxide (Tanfeng Tech. Inc.) was first annealed in air from room temperature to 300 °C at a heating rate of 10 °C min−1 in a muffle furnace and then further annealed G

DOI: 10.1021/acsnano.8b04114 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano Notes

High Volumetric Energy Lithium and Sodium Metal Batteries. Adv. Energy Mater. 2018, 8, 1701918. (19) Yuan, B.; Sun, X.; Zeng, L.; Yu, Y.; Wang, Q. A Freestanding and Long-Life Sodium-Selenium Cathode by Encapsulation of Selenium into Microporous Multichannel Carbon Nanofibers. Small 2018, 14, 1703252. (20) Yang, X.; Wang, H.; Yu, D. Y. W.; Rogach, A. L. Vacuum Calcination Induced Conversion of Selenium/Carbon Wires to Tubes for High-Performance Sodium-Selenium Batteries. Adv. Funct. Mater. 2018, 28, 1706609. (21) Xin, S.; Yin, Y. X.; Guo, Y. G.; Wan, L. J. A High-Energy RoomTemperature Sodium-Sulfur Battery. Adv. Mater. 2014, 26, 1261−5. (22) Zhang, J.; Wang, D.-W.; Lv, W.; Zhang, S.; Liang, Q.; Zheng, D.; Kang, F.; Yang, Q.-H. Achieving Superb Sodium Storage Performance on Carbon Anodes through an Ether-Derived Solid Electrolyte Interphase. Energy Environ. Sci. 2017, 10, 370−376. (23) Zhang, K.; Hu, Z.; Liu, X.; Tao, Z.; Chen, J. Fese2Microspheres as a High-Performance Anode Material for Na-Ion Batteries. Adv. Mater. 2015, 27, 3305−9. (24) Hu, Z.; Zhu, Z.; Cheng, F.; Zhang, K.; Wang, J.; Chen, C.; Chen, J. Pyrite FeS2 for High-Rate and Long-Life Rechargeable Sodium Batteries. Energy Environ. Sci. 2015, 8, 1309−1316. (25) Peng, S.; Han, X.; Li, L.; Zhu, Z.; Cheng, F.; Srinivansan, M.; Adams, S.; Ramakrishna, S. Unique Cobalt Sulfide/Reduced Graphene Oxide Composite as an Anode for Sodium-Ion Batteries with Superior Rate Capability and Long Cycling Stability. Small 2016, 12, 1359−1368. (26) Carter, R.; Oakes, L.; Douglas, A.; Muralidharan, N.; Cohn, A. P.; Pint, C. L. A Sugar-Derived Room-Temperature Sodium Sulfur Battery with Long Term Cycling Stability. Nano Lett. 2017, 17, 1863−1869. (27) Wang, H.; Jiang, Y.; Manthiram, A. Long Cycle Life, Low SelfDischarge Sodium-Selenium Batteries with High Selenium Loading and Suppressed Polyselenide Shuttling. Adv. Energy Mater. 2018, 8, 1701953. (28) Ding, J.; Zhou, H.; Zhang, H.; Stephenson, T. J.; Li, Z.; Karpuzov, D.; Mitlin, D. Exceptional Energy and New Insight with Sodium-Selenium Battery Based on Carbon Nanosheet Cathode and Pseudographite Anode. Energy Environ. Sci. 2017, 10, 153−165. (29) Zeng, L.; Wei, X.; Wang, J.; Jiang, Y.; Li, W.; Yu, Y. Flexible One-Dimensional Carbon−Selenium Composite Nanofibers with Superior Electrochemical Performance for Li−Se/Na−Se Batteries. J. Power Sources 2015, 281, 461−469. (30) Pei, S.; Cheng, H.-M. The Reduction of Graphene Oxide. Carbon 2012, 50, 3210−3228. (31) Yannopoulos, S. N.; Andrikopoulos, K. S. Raman Scattering Study on Structural and Dynamical Features of Noncrystalline Selenium. J. Chem. Phys. 2004, 121, 4747−58. (32) Yang, X.; Zhang, J.; Wang, Z.; Wang, H.; Zhi, C.; Yu, D. Y. W.; Rogach, A. L. Carbon-Supported Nickel Selenide Hollow Nanowires as Advanced Anode Materials for Sodium-Ion Batteries. Small 2018, 14, 1702669. (33) Augustyn, V.; Come, J.; Lowe, M. A.; Kim, J. W.; Taberna, P.L.; Tolbert, S. H.; Abruña, H. D.; Simon, P.; Dunn, B. High-Rate Electrochemical Energy Storage through Li+ Intercalation Pseudocapacitance. Nat. Mater. 2013, 12, 518. (34) Brezesinski, T.; Wang, J.; Tolbert, S. H.; Dunn, B. Ordered Mesoporous [Alpha]-MoO3 with Iso-Oriented Nanocrystalline Walls for Thin-Film Pseudocapacitors. Nat. Mater. 2010, 9, 146−151. (35) Li, Q.; Liu, H.; Yao, Z.; Cheng, J.; Li, T.; Li, Y.; Wolverton, C.; Wu, J.; Dravid, V. P. Electrochemistry of Selenium with Sodium and Lithium: Kinetics and Reaction Mechanism. ACS Nano 2016, 10, 8788−8795. (36) Xin, S.; Gu, L.; Zhao, N.-H.; Yin, Y.-X.; Zhou, L.-J.; Guo, Y.-G.; Wan, L.-J. Smaller Sulfur Molecules Promise Better Lithium-Sulfur Batteries. J. Am. Chem. Soc. 2012, 134, 18510−18513. (37) Ge, P.; Hou, H.; Li, S.; Huang, L.; Ji, X. Three-Dimensional Hierarchical Framework Assembled by Cobblestone-Like CoSe2@C

The authors declare no competing financial interest.

ACKNOWLEDGMENTS This work was supported by the Research Grant Council of Hong Kong S.A.R. (project CityU 21202014) and by City University of Hong Kong. The authors gratefully acknowledge support from the Ministry of Education, Youth and Sports of the Czech Republic (Projects LO1305, CZ.02.1.01/0.0/0.0/ 16_019/0000754) and the assistance provided by the Research Infrastructure NanoEnviCz under Project LM2015073. REFERENCES (1) Goodenough, J. B. Electrochemical Energy Storage in a Sustainable Modern Society. Energy Environ. Sci. 2014, 7, 14−18. (2) Cao, X.; Tan, C.; Sindoro, M.; Zhang, H. Hybrid Micro-/NanoStructures Derived from Metal-Organic Frameworks: Preparation and Applications in Energy Storage and Conversion. Chem. Soc. Rev. 2017, 46, 2660−2677. (3) Hwang, J. Y.; Myung, S. T.; Sun, Y. K. Sodium-Ion Batteries: Present and Future. Chem. Soc. Rev. 2017, 46, 3529−3614. (4) Delmas, C. Sodium and Sodium-Ion Batteries: 50 Years of Research. Adv. Energy Mater. 2018, 8, 1703137. (5) Hou, H.; Qiu, X.; Wei, W.; Zhang, Y.; Ji, X. Carbon Anode Materials for Advanced Sodium-Ion Batteries. Adv. Energy Mater. 2017, 7, 1602898. (6) Huang, Z.; Hou, H.; Zhang, Y.; Wang, C.; Qiu, X.; Ji, X. LayerTunable Phosphorene Modulated by the Cation Insertion Rate as a Sodium-Storage Anode. Adv. Mater. 2017, 29, 1702372. (7) Wei, S.; Xu, S.; Agrawral, A.; Choudhury, S.; Lu, Y.; Tu, Z.; Ma, L.; Archer, L. A. A Stable Room-Temperature Sodium-Sulfur Battery. Nat. Commun. 2016, 7, 11722. (8) Yang, C. P.; Yin, Y. X.; Guo, Y. G. Elemental Selenium for Electrochemical Energy Storage. J. Phys. Chem. Lett. 2015, 6, 256−66. (9) Chi, S.-S.; Qi, X.-G.; Hu, Y.-S.; Fan, L.-Z. 3D Flexible Carbon Felt Host for Highly Stable Sodium Metal Anodes. Adv. Energy Mater. 2018, 8, 1702764. (10) Wang, H.; Wang, C.; Matios, E.; Li, W. Critical Role of Ultrathin Graphene Films with Tunable Thickness in Enabling Highly Stable Sodium Metal Anodes. Nano Lett. 2017, 17, 6808−6815. (11) Luo, W.; Lin, C. F.; Zhao, O.; Noked, M.; Zhang, Y.; Rubloff, G. W.; Hu, L. Ultrathin Surface Coating Enables the Stable Sodium Metal Anode. Adv. Energy Mater. 2017, 7, 1601526. (12) Peng, H. J.; Zhang, Q. Designing Host Materials for Sulfur Cathodes: From Physical Confinement to Surface Chemistry. Angew. Chem., Int. Ed. 2015, 54, 11018−11020. (13) Manthiram, A.; Yu, X. Ambient Temperature Sodium-Sulfur Batteries. Small 2015, 11, 2108−14. (14) Abouimrane, A.; Dambournet, D.; Chapman, K. W.; Chupas, P. J.; Weng, W.; Amine, K. A New Class of Lithium and Sodium Rechargeable Batteries Based on Selenium and Selenium-Sulfur as a Positive Electrode. J. Am. Chem. Soc. 2012, 134, 4505−4508. (15) Luo, C.; Xu, Y.; Zhu, Y.; Liu, Y.; Zheng, S.; Liu, Y.; Langrock, A.; Wang, C. Selenium@Mesoporous Carbon Composite with Superior Lithium and Sodium Storage Capacity. ACS Nano 2013, 7, 8003−8010. (16) Zeng, L.; Zeng, W.; Jiang, Y.; Wei, X.; Li, W.; Yang, C.; Zhu, Y.; Yu, Y. A Flexible Porous Carbon Nanofibers-Selenium Cathode with Superior Electrochemical Performance for Both Li-Se and Na-Se Batteries. Adv. Energy Mater. 2015, 5, 1401377. (17) Xin, S.; Yu, L.; You, Y.; Cong, H. P.; Yin, Y. X.; Du, X. L.; Guo, Y. G.; Yu, S. H.; Cui, Y.; Goodenough, J. B. The Electrochemistry with Lithium Versus Sodium of Selenium Confined to Slit Micropores in Carbon. Nano Lett. 2016, 16, 4560−4568. (18) Ding, J.; Zhou, H.; Zhang, H.; Tong, L.; Mitlin, D. Selenium Impregnated Monolithic Carbons as Free-Standing Cathodes for H

DOI: 10.1021/acsnano.8b04114 ACS Nano XXXX, XXX, XXX−XXX

Article

ACS Nano Nanospheres for Ultrastable Sodium-Ion Storage. ACS Appl. Mater. Interfaces 2018, 10, 14716−14726. (38) Wan, M.; Zeng, R.; Chen, K.; Liu, G.; Chen, W.; Wang, L.; Zhang, N.; Xue, L.; Zhang, W.; Huang, Y. Fe7Se8 Nanoparticles Encapsulated by Nitrogen-Doped Carbon with High Sodium Storage Performance and Evolving Redox Reactions. Energy Storage Mater. 2018, 10, 114−121. (39) Shadike, Z.; Zhao, E.; Zhou, Y.-N.; Yu, X.; Yang, Y.; Hu, E.; Bak, S.; Gu, L.; Yang, X.-Q. Advanced Characterization Techniques for Sodium-Ion Battery Studies. Adv. Energy Mater. 2018, 8, 1702588.

I

DOI: 10.1021/acsnano.8b04114 ACS Nano XXXX, XXX, XXX−XXX