Na₃V₂(PO₄)₂F₃–SWCNT: a high voltage cathode for non-aqueous and aqueous sodium-ion batteries

Abstract

Due to the merits of low cost, safety, environmental friendliness, and abundant sodium reserves, non-aqueous and aqueous sodium-ion batteries are wonderful alternatives for large-scale energy storage. Na⁺ super ionic conductor structured Na₃V₂(PO₄)₂F₃ is considered as a potential high-capacity cathode material for Na-ion batteries. However, its insufficient cyclability remains a challenge for battery applications. In addition, the narrow electrochemical stability window (1.23 V) of aqueous electrolytes does not allow us to make the best use of it. All of these pose obstacles to the practical voltage and energy output. Therefore, we designed and synthesized a Na₃V₂(PO₄)₂F₃–SWCNT composite, which demonstrated superior Na⁺-storage performance with a high reversible capacity of 117 mA h g⁻¹ at a moderate current of 0.5C in non-aqueous electrolyte. In addition, we developed an aqueous rechargeable sodium ion battery using Na₃V₂(PO₄)₂F₃–SWCNT as the cathode and NaTi₂(PO₄)₃–MWCNT as the anode, and this battery can deliver a high energy density of 150 W h kg⁻¹ with an operating voltage of up to 2.0 V by using the high concentration electrolyte, 17 m NaClO₄. The outstanding electrochemical performance makes Na₃V₂(PO₄)₂F₃–SWCNT a promising cathode for Na⁺ storage and encourages more investigations into practical sodium-ion battery applications.

---

Introduction

Despite the great success of lithium-ion batteries (LIBs) in portable electronics and electric vehicles in recent years, the insufficient reserves, unbalanced distribution as well as increasing cost of lithium resources restrict their practical application in large-scale electrical energy storage systems.^1–3 Sodium-ion batteries (SIBs) are considered as an excellent alternative to LIBs due to the natural abundant reserves of sodium.^4–6 However, Na⁺ is about 42% larger in ionic radius than Li⁺, which makes it difficult to find suitable host materials that are capable of accommodating Na⁺ and allowing reversible and rapid Na⁺ insertion and extraction.^7,8 Three-dimensional (3D) host framework structured materials could provide enough interstitial channels for Na⁺ transit and permit the reversible intercalation/de-intercalation of large Na⁺, and have attracted great interest for SIBs in the past years.^9–13 Among them, Na⁺ super ionic conductor (NASICON) structured Na₃V₂(PO₄)₂F₃ has attracted extensive attention for its high potential plateaus (3.6 V/4.1 V),¹³ good specific capacity (128 mA h g⁻¹), and high thermal stability. However, its poor electronic conductivity (∼10⁻¹² S cm⁻¹) usually causes low utilization of active material and restricts high-rate performance. Constructing nanosized particles or conductive surface coatings, such as by nanocrystallization, particle formation from liquid dispersion systems, and particle coating, can effectively shorten the ionic transfer pathway and increase electronic conductivity, hence realizing a high electrochemical performance.^14–16 In addition, compared with expensive and flammable organic liquid electrolytes, aqueous electrolytes have advantages such as low cost, safety, and environmental friendliness. Therefore, aqueous rechargeable sodium-ion batteries (ARSBs) have received increasing attention in the field of electrochemical energy storage (EES) technology and represent a promising alternative in future power grids.^17,18

In recent years, various ARSBs have been reported. Unfortunately, they are limited by the narrow electrochemical stability window (1.23 V) of the aqueous electrolyte, which restricts the output voltage of ARSBs to under 1.50 V and leads to compromised energy densities,^19,20 for instance, Na_0.66[Mn_0.66Ti_0.34]O₂//NaTi₂(PO₄)₃ (1.0 V),²¹ Alizarin//PPy (1.06 V),²² Na_0.44MnO₂//NaTi₂(PO₄)₃ (1.1 V),²³ NaVPO₄F//Polyimide (1.3 V),²⁴ Na₃MnTi(PO₄)₃//Na₃MnTi(PO₄)₃ (1.4 V),²⁵ Na₂NiFe(CN)₆//NaTi₂(PO₄)₃ (1.4 V),²⁶ Na₃V₂O₂(PO₄)₂F–SWCNT//NaTi₂(PO₄)₃–MWCNT (1.5 V),²⁷etc. For ARSBs, increasing the voltage is more effective than increasing the capacity of the electrode material in terms of enhancing the energy density, since the electrode materials only account for about 30% of the total weight of the assembled complete cell.²⁸ Therefore, expanding the electrochemical stability window of the aqueous electrolyte, eliminating the O₂ in it and suppressing the water activity are of vital importance. Several approaches have been developed to resolve these problems; for example, increasing the concentration of the electrolyte enlarges the potential window, such as 9.26 m sodium trifluoromethane sulfonate aq. (NaCF₃SO₃, or NaOTf, 2.5 V),²¹ and 17 m NaClO₄ aq. (2.8 V),²⁹ where m represents molality (molality [m] = moles of solute/weight of solvent [mol kg⁻¹]). Modification with a surfactant such as by the addition of sodium dodecyl sulfate (SDS) to the dilute aqueous electrolyte expanded the window to about 2.5 V.³⁰ These strategies, especially the use of high-concentration electrolytes, can significantly improve the performance of ARSBs.

In this work, we synthesize Na₃V₂(PO₄)₂F₃–single walled carbon nanotube (Na₃V₂(PO₄)₂F₃–SWCNT) composites through a simple solvothermal method at low temperature, and the composites demonstrate superior Na⁺-storage performance with a high reversible capacity of 117 mA h g⁻¹ at a moderate current of 0.5C in non-aqueous electrolyte. In addition, a high voltage 2.0 V ARSB full-cell is assembled using Na₃V₂(PO₄)₂F₃–SWCNT composites as the cathode, highly stable NaTi₂(PO₄)₃ microblock-multi walled carbon nanotube (NaTi₂(PO₄)₃–MWCNT) composites as the anode, and 17 m NaClO₄ aq. as the electrolyte, which demonstrates the great potential of the ARSBs for grid-scale energy storage systems.

Experimental

Preparation of the Na₃V₂(PO₄)₂F₃–SWCNT composites

For the preparation of Na₃V₂(PO₄)₂F₃–SWCNT composites, 0.434 mmol vanadium(III) 2,4-pentanedionate (C₁₅H₂₁O₆V) (98%, energy) is added to 12 mL ethanol solution and stirred for 1 h. At the same time, 0.651 mmol sodium dihydrogen phosphate (NaH₂PO₄·2H₂O) (99%, energy) is dissolved in 4 mL water, and then 0.72 mmol sodium fluoride (NaF) (99%, Meryer) is added into it. After they are simultaneously stirred at room temperature for 0.5 h, a certain amount of SWCNT (0.5 mg mL⁻¹) aqueous solution is added to the latter aqueous solution and stirred for 0.5 h. The two solutions are transferred to a 100 mL polytetrafluoroethylene-lined hydrothermal autoclave and reacted at 120 °C for 10 h, and then the product is filtered and washed with ethanol and water repeatedly, dried in a vacuum oven at 80 °C, and collected. To synthesize products with different SWCNT content, different amounts of SWCNT aqueous solution (0.5 mg mL⁻¹), where the quantities of SWCNTs are 5 wt%, 12.5 wt% and 20 wt% based on the yield of bare Na₃V₂(PO₄)₂F₃, are added into the mixed solution of NaH₂PO₄·2H₂O and NaF; all other operations are the same as above. A slight excess of NaF is employed instead of the stoichiometric amount in order to guarantee that the vanadium precursor can react completely. The procedure for preparing bare Na₃V₂(PO₄)₂F₃ is the same except for the addition of SWCNTs. The formation of Na₃V₂(PO₄)₂F₃ can be explained according to the following chemical reaction (formula (1)):

2V(C₅H₇O₂)₃ + 3NaH₂PO₄ + 3NaF → Na₃V₂(PO₄)₂F₃ + 6C₅H₈O₂ + Na₃PO₄ (1)

Preparation of the NaTi₂(PO₄)₃–MWCNT composites

NaTi₂(PO₄)₃–MWCNT composites are prepared by the solid state method. In a typical synthesis, stoichiometric amounts of NaH₂PO₄·2H₂O (1.02 mmol, Aladdin), TiO₂ (2 mmol, 99.8% metals basis, 5–10 nm, Aladdin), and (NH₄)₂HPO₄ (1.96 mmol, 99.9%, Aladdin), mixed with 10 wt% multi-walled carbon nanotubes (MWCNTs, CNano Technology Ltd.) based on the theoretical yield of NaTi₂(PO₄)₃ are ball-milled in a planetary ball mill (PM 200) using zirconia milling media in ethanol at 400 rpm for 12 h, followed by drying at 80 °C in a vacuum. The precursor is pre-sintered at 400 °C for 2 h in a flowing argon atmosphere, and then reground in an agate mortar. A final heat treatment is done at 850 °C for 12 h in a flowing argon atmosphere as well.

Characterization

The structure of the as-prepared samples is characterized by powder X-ray diffraction (XRD, Rigaku Mini Flex 600 X-ray generator, Cu Kα radiation, λ = 1.5406 Å) at a scanning rate of 4° min⁻¹. The morphology of the as-prepared products is examined by field-emission scanning electron microscopy (SEM, JEOL JSM-6700F Field Emission, operating at 5 kV) and high-resolution transmission electron microscopy (HRTEM, JEOL JEM-2010 FEF TEM, operating at 200 kV) coupled with energy dispersive spectroscopy (EDS). The crystal structure is obtained using a confocal Raman microscope (DXR, Thermo-Fisher Scientific, 532 nm excitation from an argon-ion laser), and a Fourier transform infrared spectrometer (FTIR, Bruker Tensor II). The electronic states of the samples are investigated by X-ray photoelectron spectroscopy (XPS, Perkin Elmer PHI 1600 ESCA). The content of carbon is measured using an elemental analyzer (EA, Vario EL CUBE). The specific surface area is measured using Brunauer–Emmett–Teller (BET) nitrogen adsorption–desorption measurements (BELSORP-mini instrument) at 77 K. The porosity is estimated from the adsorption isotherms by the Barrett–Joyner–Halenda (BJH) method.

Electrochemical measurements

The cathode, anode and SWCNT electrodes are all fabricated by mixing 70 wt% of the active material with 20 wt% of Super P and 10 wt% polyvinylidene difluoride (PVDF) in a N-methylpyrrolidone (NMP) solvent to make a slurry.

For the non-aqueous electrolyte tests, the obtained Na₃V₂(PO₄)₂F₃–SWCNT slurry is coated on Al foil, dried at 80 °C for 12 h in a vacuum oven and cut into circular electrodes with diameters of 10 mm. The loading is ∼0.3 mg cm⁻². CR2032 coin cells are assembled in an argon-filled dry glove box (with less than 0.1 ppm of H₂O and O₂) using sodium metal as the counter electrode and reference electrode, borosilicate glass fibers (Whatman GF/D) as the separator, and 1 M NaPF₆ in propylene carbonate (PC) with 5 vol% of fluoroethylene carbonate (FEC) as the electrolyte.

For the aqueous electrolyte tests, the prepared Na₃V₂(PO₄)₂F₃–SWCNT and NaTi₂(PO₄)₃–MWCNT slurries are respectively coated on carbon paper and titanium foil that are 10 mm in diameter. The loading for the Na₃V₂(PO₄)₂F₃–SWCNT and NaTi₂(PO₄)₃–MWCNT electrodes is ∼0.52 and ∼0.38 mg cm⁻², respectively. Both the as-prepared electrodes are dried at 80 °C for 12 h in a vacuum oven. The electrochemical properties of the electrodes are obtained using a three electrode system, where platinum foil is used as the counter electrode, 17 m NaClO₄ aq. (purged with Ar before use) is used as the electrolyte, and an Ag/AgCl electrode (0.2412 V vs. standard hydrogen electrode) is used as the reference electrode. The performance of the Na₃V₂(PO₄)₂F₃–SWCNT//NaTi₂(PO₄)₃–MWCNT full-cell is tested using CR2032 coin cells, which are assembled in an argon-filled dry glove box (with less than 0.1 ppm of O₂).

Cyclic voltammograms (CVs) are recorded using a Parstat 263A electrochemical workstation (AMETEK). Galvanostatic charge/discharge tests are performed on a LAND battery-testing instrument (CT2001A). Electrochemical impedance spectroscopy (EIS) measurements are carried out using a Parstat 2273 electrochemical workstation over a frequency range from 2 Hz to 100 kHz and at an AC amplitude of 5 mV. All electrochemical tests are performed at room temperature.

Results and discussion

The Na₃V₂(PO₄)₂F₃–SWCNT composites are synthesized by a simple solvothermal route. By adjusting the mass ratio of SWCNTs, three products with different additive amounts of SWCNTs are obtained. The elemental analysis results (Fig. 1a) suggest that the carbon content is 9.56, 13.45, and 54.61 wt%, respectively. The three samples with different carbon content are denoted as Na₃V₂(PO₄)₂F₃–SWCNT-1, Na₃V₂(PO₄)₂F₃–SWCNT-2, and Na₃V₂(PO₄)₂F₃–SWCNT-3, respectively.

Fig. 1 (a) CHN content, (b) XRD spectra, (c) Raman spectra, and (d) FTIR spectra of Na₃V₂(PO₄)₂F₃, Na₃V₂(PO₄)₂F₃–SWCNT serial samples, and SWCNTs.

The structure of Na₃V₂(PO₄)₂F₃–SWCNT is confirmed by the XRD technique. As shown in Fig. 1b, all the peaks are readily assigned to the space group P4₂/mnm (no. 136), being consistent with the standard values of the tetragonal Na₃V₂(PO₄)₂F₃ (JCPDS no. 89-8485). Five main peaks appearing at 16.12°, 16.52°, 27.76°, 28.68° and 32.54° can be perfectly matched to the (111), (002), (220), (113) and (222) crystal planes, respectively.^31,32 Furthermore, it is worth noting that the (111), (002), (220) and (113) characteristic peaks are obviously sharp for Na₃V₂(PO₄)₂F₃–SWCNT-1 and Na₃V₂(PO₄)₂F₃–SWCNT-2, compared with those of the bare Na₃V₂(PO₄)₂F₃ (Fig. S1a†), indicating that the SWCNTs significantly increase the crystallinity of the materials. In addition, the intensities of the (111), (002), (220) and (113) characteristic peaks of Na₃V₂(PO₄)₂F₃–SWCNT-3 decrease evidently, indicating a decreased crystallinity and the negative impact of excessive SWCNTs on Na₃V₂(PO₄)₂F₃. All these results indicate that SWCNTs can not only contribute to the formation of a pure phase, but can also be used as seed crystals to improve the crystallinity of the material.

This phenomenon can also be verified from the SEM results. Bare Na₃V₂(PO₄)₂F₃ exhibits low crystallinity with the agglomeration of particles (Fig. S1b†). However, things are quite different when it is combined with SWCNTs. In Na₃V₂(PO₄)₂F₃–SWCNT-1 (Fig. S2a and b†), there are hexahedral blocks which are connected by a small amount of threadlike SWCNTs, while in Na₃V₂(PO₄)₂F₃–SWCNT-2 (Fig. S2c and d†), large sized cuboids appear. Moreover, the cuboids are wrapped and linked with each other by a SWCNT network, which would be quite helpful in increasing the electronic conductivity; however, for Na₃V₂(PO₄)₂F₃–SWCNT-3 (Fig. S2e and f†), individual SWCNT bundles can be observed due to the poor dispersibility of the excessive SWCNTs. The particles and SWCNTs are randomly mixed without strong interactions, and deposited separately from each other. As a result, the aggregation of Na₃V₂(PO₄)₂F₃ particles could not be avoided.³³ All of these results demonstrate the importance of the SWCNT content in crystal formation. Considering its regular structure and Na₃V₂(PO₄)₂F₃ content, the Na₃V₂(PO₄)₂F₃–SWCNT-2 composite is chosen for further study.

The Raman spectra are analyzed to further characterize the as-prepared samples. Fig. 1c displays the Raman spectra of the three samples. The band at 945 cm⁻¹ is assigned to the P–O symmetrical stretching vibration of PO₄, and the band at 1050 cm⁻¹ is attributed to the P–O asymmetrical stretching vibration of PO₄.^34–37 The bands observed below 400 cm⁻¹ (269 cm⁻¹ and 192 cm⁻¹) are the result of the external modes of PO₄.^34,35 Besides, the characteristic bands of SWCNTs (Radial Breathing Mode (RBM) band of 158.8 cm⁻¹, D band of 1345 cm⁻¹, G band of 1590 cm⁻¹ and 2D band of 2632 cm⁻¹) can be observed. Similar results can be verified from the FTIR spectra of Na₃V₂(PO₄)₂F₃–SWCNT (Fig. 1d). The band at 1072 cm⁻¹ belongs to the asymmetrical stretching vibration of the PO₄ tetrahedron. The bands at 674 cm⁻¹ and 560 cm⁻¹ are due to the stretching vibration mode and symmetrical bending vibration of P–O, respectively. Bands appearing at 912 cm⁻¹ and 955 cm⁻¹ correspond to the stretching vibration mode of V–O and V–F.^38–41 On the FTIR curve of SWCNTs, the band at 1128 cm⁻¹ represents the tensile vibration of the C–O bond.^42–44 The above characteristic FTIR peaks can be observed on all the FTIR curves of the Na₃V₂(PO₄)₂F₃–SWCNT series products. Combining the results of Raman and FTIR spectroscopy, it can be concluded that all the as-prepared samples possess the structure of Na₃V₂(PO₄)₂F₃, and Na₃V₂(PO₄)₂F₃ and SWCNTs have been well combined in the solvothermal process.

In order to further explore the structural properties of Na₃V₂(PO₄)₂F₃–SWCNT, refinement of Na₃V₂(PO₄)₂F₃–SWCNT-2 is performed as shown in Fig. 2a. All the diffraction peaks can be readily indexed to the NASICON structure in the single phase P4₂/mnm (no. 136) space group, which indicates the high purity of the prepared material. The lattice parameters are calculated as a = b = 9.047 Å and c = 10.705 Å (R_p = 12.96%, R_wp = 9.28%). The crystal structure of Na₃V₂(PO₄)₂F₃–SWCNT-2 is illustrated in Fig. 2b. The 3D framework structure of Na₃V₂(PO₄)₂F₃ can be described in terms of [V₂O₈F₃] bi-octahedral and [PO₄] tetrahedral units as illustrated in Fig. 2b.⁴⁵ The [V₂O₈F₃] bi-octahedron is bridged with two [VO₄F₂] octahedral units by one fluorine atom, whereas the oxygen atoms are all interconnected through the [PO₄] units. Two different interstitial sites (Na₁ and Na₂) are occupied by sodium ions. The sodium ions are statistically distributed in the network where the NASICON framework grants Na ions significant mobility. This property confers to Na₃V₂(PO₄)₂F₃ the ability to intercalate/deintercalate Na⁺ reversibly, making it a favorable positive electrode material choice for SIBs.^45–47

Fig. 2 (a) XRD patterns along with fitted results, (b) crystal structure, (c) SEM images, (d) HRTEM image (the inset is the selected area electron diffraction (SAED) pattern), (e) TEM image, and (e–k) EDS elemental mapping images of the Na, V, P, O, F and C elements of Na₃V₂(PO₄)₂F₃–SWCNT-2.

Fig. 2c shows the SEM images of the Na₃V₂(PO₄)₂F₃–SWCNT-2 composites. The prepared materials have a nanocube-like morphology with a height of 0.5–1 μm and width and length of 300–500 nm, and these cubes are well dispersed in the network of SWCNTs. The typical HRTEM image of Na₃V₂(PO₄)₂F₃–SWCNT-2 in Fig. 2d displays clear and continuous lattices with a d-spacing of 5.35 Å corresponding to the (002) facets, indicating the high crystallinity of the Na₃V₂(PO₄)₂F₃–SWCNT-2 sample. The SAED pattern in the inset of Fig. 2d can be indexed to the (002), (200), and (220) planes, which are attributed to the tetragonal phase of Na₃V₂(PO₄)₂F₃–SWCNT-2. The elemental maps (Fig. 2e–k) show the uniform distribution of C, F, P, O, V, and Na through the whole nanocube. Furthermore, the N₂ adsorption/desorption isotherms of Na₃V₂(PO₄)₂F₃–SWCNT-2 show typical type-IV isotherms in Fig. S3.† The specific surface area of Na₃V₂(PO₄)₂F₃–SWCNT-2 is calculated to be 69.3 m² g⁻¹ based on the BET method. The pore size distribution derived via the BJH method is in the range of 7–22 nm, as shown in the inset of Fig. S3.†

To evaluate the electrochemical properties of Na₃V₂(PO₄)₂F₃–SWCNT, a series of half cells (Na₃V₂(PO₄)₂F₃–SWCNT-2//Na) have been tested. All the capacities in this work are calculated based on the mass of the Na₃V₂(PO₄)₂F₃–SWCNT-2 composites, although the SWCNTs have a limited capacity (ca. 12 mA h g⁻¹) (Fig. S4†). Fig. 3a shows the CV results at different scan rates from 0.1 to 0.4 mV s⁻¹ in the potential range between 2.5 and 4.3 V vs. Na⁺/Na. There are two cathodic peaks (located at 3.54 and 3.95 V of 0.1 mV s⁻¹) and two anodic peaks (near 4.08 and 3.7 V of 0.1 mV s⁻¹), corresponding to the redox reaction of V³⁺/V⁴⁺ with extraction/insertion of two sodium ions. As the scan rate increases, the height and the area of the CV curve increase due to the constant capacity of the electrode. The peak current ratio of the anodic and cathodic peak appears to be unity, i.e., |i_pc/i_pa| = 1, indicating the excellent reversibility of the sodium insertion and extraction. The profiles between i_p and ν^1/2 based on the data in Fig. 3a are shown in Fig. 3b. The excellent linear fitting results of i_pvs. ν^1/2 confirm the diffusion-controlled behavior of the electrode reactions in Na₃V₂(PO₄)₂F₃–SWCNT-2.^48,49 The sodium ion diffusion coefficient can be calculated from the Randles–Sevcik formula:^50,51

where i_p is the peak current of the anodic peaks for the sodium ions at the Na₁ or Na₂ site, F is the Faraday constant (96 485 C mol⁻¹), R is the molar gas constant (8.314 J K⁻¹ mol⁻¹), T is the temperature in our experiment (298 K), n is the number of electrons per molecule taking part in the electron transfer reaction (n = 2), A is the electrode area (here the geometric area of the electrode, 0.785 × 10⁻⁴ m², is used for simplicity⁴⁸), C_Na⁺ is the concentration of Na⁺ in the electrode (9.1 mol L⁻¹), D_Na⁺ is the diffusion coefficient, and ν is the scan rate referring to peaks O1, R1, O2, and R2, respectively. Therefore, the apparent Na⁺ diffusion coefficient (D_Na⁺) can be calculated to be 2.52 × 10⁻¹², 2.40 × 10⁻¹², 2.48 × 10⁻¹², and 2.86 × 10⁻¹² cm² s⁻¹, respectively, according to eqn (2). It is worth noting that the presently calculated D_Na⁺ values of Na₃V₂(PO₄)₂F₃–SWCNT are higher than those of other polyanion type phosphate cathodes for SIBs,^32,52,53 and are significantly higher than those of LiFePO₄ measured by Prosini et al. (10⁻¹⁶ to 10⁻¹⁴ cm² s⁻¹)⁵⁴ and Franger et al. (10⁻¹⁴ to 10⁻¹³ cm² s⁻¹).⁵⁵ This high D_Na⁺ indicates the excellent diffusivity of the NASICON-type structure of Na₃V₂(PO₄)₂F₃, allowing Na⁺ ion transport in an open 3D framework and resulting in a high ion diffusion capability.

Fig. 3 The electrochemical behaviour of the Na₃V₂(PO₄)₂F₃–SWCNT electrodes: (a) CV curves at various scan rates. (b) The linear relationship between the peak current (i_p) and square root of the scan rate (ν^1/2). (c) The first charge/discharge profile at a current density of 0.5C. (d) The corresponding cycling performance tested in the voltage range of 2.5–4.3 V vs. Na⁺/Na. (e) The rate performance of Na₃V₂(PO₄)₂F₃–SWCNT-2. (f) The cycle performance of Na₃V₂(PO₄)₂F₃–SWCNT-2 at 10C.

To clarify the Na intercalation/extraction processes of Na₃V₂(PO₄)₂F₃–SWCNT as a cathode material for SIBs, the galvanostatic charge/discharge curves and cycle life curves are investigated. Fig. 3c indicates the typical charge/discharge profiles of the Na₃V₂(PO₄)₂F₃, Na₃V₂(PO₄)₂F₃–SWCNT-1, Na₃V₂(PO₄)₂F₃–SWCNT-2, and Na₃V₂(PO₄)₂F₃–SWCNT-3 electrodes at a current rate of 0.5C (1C = 128 mA g⁻¹) between 2.5 and 4.3 V vs. Na⁺/Na in the first cycle. The cells exhibit charge/discharge capacities of 97.4/87.6, 135.9/92.4, 160.6/116.7 and 83.5/49.8 mA h g⁻¹, respectively. It can be concluded that the specific capacity of the materials composited with a certain amount of SWCNTs is greatly improved compared with that of the Na₃V₂(PO₄)₂F₃ electrode under the same conditions. Note that the discharge capacity of Na₃V₂(PO₄)₂F₃–SWCNT-2 reached 117 mA h g⁻¹, which is very close to the theoretical capacity of Na₃V₂(PO₄)₂F₃ (128 mA h g⁻¹), indicating a full use of the material for the Na⁺-storage reaction. The cycling performance of the Na₃V₂(PO₄)₂F₃–SWCNT serial samples is given in Fig. 3d. It is found that Na₃V₂(PO₄)₂F₃–SWCNT-2 has the highest specific capacity and good cycling stability. At the 100th cycle, the capacity can be maintained at 111.4 mA h g⁻¹ with a capacity retention of 92.4%, demonstrating the good stability of this material. Therefore, the following electrochemical properties are studied using Na₃V₂(PO₄)₂F₃–SWCNT-2. In addition, the rate performances are further investigated at different current rates ranging from 0.2 to 10C. As shown in Fig. 3e, Na₃V₂(PO₄)₂F₃–SWCNT-2 delivers reversible capacities of 117.7, 114.3, 110.2, 109.1, 104.7, 100.7 and 120 mA h g⁻¹ at current densities of 0.2, 0.5, 1, 2, 5, and 10C, respectively. In particular, when the current density returns to 0.2C after 30 cycles, the specific capacity nearly recovers to the same extent as observed in the initial five cycles, indicating an excellent reversibility even at high-rate charge/discharge. The cycle stability of Na₃V₂(PO₄)₂F₃–SWCNT-2 is also recorded for 500 cycles and shown in Fig. 3f with a capacity retention of 96% at 10C. Additionally, when the loading mass of Na₃V₂(PO₄)₂F₃–SWCNT-2 is more than 1 mg cm⁻², the electrochemical performance is still comparable with that of lower loading mass (∼0.3 mg cm⁻²). The excellent capability of the Na₃V₂(PO₄)₂F₃–SWCNT-2 composites can be attributed to their 3D framework with large tunnels as well as nanoscale morphology, which enables rapid sodium-ion extraction/insertion.^46,49

It has been reported that Na₃V₂(PO₄)₂F₃ only allows one Na⁺ insertion/extraction in aqueous solution.^22,56–58 In theory, when the number of sodium ions that can be stored doubles, the full battery's energy density can be increased by 2.67 times (based on the mass of cathode). Therefore, it is especially important to make full use of the high potential platform in aqueous electrolyte for the development of EES. Fig. 4 shows the electrochemical properties of the Na₃V₂(PO₄)₂F₃–SWCNT-2 composites in 17 m NaClO₄ aq. The CV plot for the Na₃V₂(PO₄)₂F₃–SWCNT composites is shown in Fig. 4a. Two pairs of charge/discharge plateaus appearing at 1.21/1.1 V and 0.82/0.71 V vs. Ag/AgCl are clearly observed in the first cycle, which are reversible Na⁺ ion extraction and insertion from/into the structure. Moreover, the oxygen evolution potential which is located at 1.38 V vs. Ag/AgCl is completely separated from the redox potential, which implies the possibility of high potential in 17 m NaClO₄ aq. The charge/discharge voltage profiles for different cycles are shown in Fig. 4b. The initial charge and discharge capacities are 92.4 and 81.3 mA h g⁻¹, corresponding to the first coulombic efficiency of 87.98%. This low coulombic efficiency is attributed to the formation of a solid-state interface (SEI) film caused by the irreversible interfacial storage. In addition, after 60 cycles, the capacity decreases to 58 mA h g⁻¹ with a retention of 71.3% of the capacity in the first cycle (Fig. S5†). The XRD patterns (Fig. S6†) and SEM images (Fig. S7a and b†) for the pristine Na₃V₂(PO₄)₂F₃–SWCNT-2 electrode before cycling and the Na₃V₂(PO₄)₂F₃–SWCNT-2 electrode after 60 cycles show the structural degradation of Na₃V₂(PO₄)₂F₃–SWCNT. Details are discussed in the ESI.† The redox reaction in the cathode and anode, and the formation of the SEI in the cathode can be verified by XPS.

Fig. 4 (a) CVs obtained at 2 mV s⁻¹, and (b) charge/discharge curves at 1C (1C = 128 mA g⁻¹) for the Na₃V₂(PO₄)₂F₃–SWCNT-2 composite in 17 m NaClO₄ aq.

Fig. S8† shows the XPS spectra of the V element of the Na₃V₂(PO₄)₂F₃–SWCNT electrode and the Ti element of the NaTi₂(PO₄)₃–MWCNT electrode at selected pristine, charged and discharged states during the initial cycle. These valence changes indicate a normal redox process. Details are discussed in the ESI.† It's generally known that the SEI film is mainly composed of C and F elements. To investigate the surface change of Na₃V₂(PO₄)₂F₃–SWCNT after cycling, comparisons of the XPS spectra of C_1s and F_1s at pristine, fully charged 1.2 V vs. Ag/AgCl and fully discharged 0.2 V vs. Ag/AgCl states are drawn. As shown in Fig. 5a, the C_1s spectra show a new peak at 288.3 eV for the fully charged and discharged electrodes compared with the that for the pristine (electrolyte-soaked Na₃V₂(PO₄)₂F₃–SWCNT) electrode, which can be attributed to the O–C=O functionality of the SEI film.⁵⁸ Similarly, in comparison with that of the pristine electrode, the new signal occurring at 684.2 eV in the F_1s spectra (Fig. 5b) of the cycled Na₃V₂(PO₄)₂F₃–SWCNT electrode originates from NaF due to the formation of the SEI film.⁵⁹ Therefore, the main composition of the SEI film on the cathode may be Na₂CO₃ and NaF. The SEI film on the cathode favors the suppression of cathode dissolution and reduction of water splitting, which occurs in aqueous electrolyte, according to the theory proposed by Chen,⁶⁰ Wang,⁶¹ and Wang and Jiang.⁵⁸

Fig. 5 XPS spectra of (a) C_1s and (b) F_1s, of the Na₃V₂(PO₄)₂F₃–SWCNT electrodes at different charge–discharge stages in 17 m NaClO₄ aq.

To look into the functioning of Na₃V₂(PO₄)₂F₃–SWCNT in aqueous electrolyte during charging and discharging, we performed ex situ XRD at various states of charge and discharge in the first cycle in 17 m NaClO₄ aq. (Fig. S9a and b†). All these patterns present almost coincident characteristic peaks. As shown in the enlarged patterns within the 2θ ranges of 27–34°, 39–46°, 48–51° and 56–60° (Fig. S9c–f†), when the Na₃V₂(PO₄)₂F₃–SWCNT electrodes are charged to 0.8 V, the characteristic peaks belonging to the (220), (222), (400), (333), and (440) planes shift to higher angles, corresponding to the transformation from Na₃V₂(PO₄)₂F₃ to Na₂V₂(PO₄)₂F₃ with the extraction of Na⁺ at fully occupied positions. As the voltage reaches 1.2 V vs. Ag/AgCl, the positions of these peaks change even more, indicating that another Na⁺ is extracted from Na₂V₂(PO₄)₂F₃ to give NaV₂(PO₄)₂F₃.^47,49,62 The locations of diffraction peaks can be fully recovered by discharging to 0.8 V and further to 0.2 V. These results demonstrate that the electrochemical behavior of the Na₃V₂(PO₄)₂F₃–SWCNT electrodes is a reversible process of Na⁺ extraction/insertion (Na₃V₂(PO₄)₂F₃ ↔ NaV₂(PO₄)₂F₃ + 2Na⁺ + 2e⁻), which is consistent with the mechanism in non-aqueous electrolyte.^47,49 Moreover, a full battery is assembled for the convenience of the in situ Raman test on the Na₃V₂(PO₄)₂F₃–SWCNT electrode. The full cell uses NaTi₂(PO₄)₃–SWCNT as the anode, which can accommodate 2 Na⁺ per molecular unit.⁶³ The performance characterization is shown in Fig. S10† and is also discussed in detail in the ESI.†Fig. 6a shows the in situ Raman results between 0.6 and 2.1 V. The corresponding charge/discharge profiles are shown in Fig. 6b. At open circuit voltage, one major distinct peak in the Raman spectrum is observed at 941.3 cm⁻¹ (P–O symmetrical stretching vibration of PO₄) for Na₃V₂(PO₄)₂F₃–SWCNT, and four major peaks for the SWCNTs are observed at 222 (RBM), 1330 (D band), 1586 (G band), and 2641 cm⁻¹ (2D band). As the voltage increases, the band at 941.3 cm⁻¹ weakens quickly, which can be attributed to the effect of changes in local symmetry on the ν(PO₆) vibration mode. As sodium ions are deintercalated from the bulk, the V electronic configurations change to balance the charge. Because each phosphate ion is surrounded by four vanadium ions, the removal of Na⁺ and the associated oxidation of V³⁺ ion to V⁴⁺ will affect the four adjacent phosphate groups.⁶⁴ In addition, the intensity of each Raman characteristic peak of SWCNTs decreases, which can be attributed to some Na⁺ being captured by the SWCNTs (poor but existing capacity, ca. 25 mA h g⁻¹ Fig. S11†), which leads to a modification of the nanotube electronic structure and causes changes (reduction in intensity) in the resonance Raman signal.^65,66 The in situ Raman spectra of the Na₃V₂(PO₄)₂F₃–SWCNT electrode in the discharge process show a recovery along the original path. This excellent electrochemical reversibility in the range of 0.6–2.1 V makes this material a highly rechargeable positive electrode for Na₃V₂(PO₄)₂F₃–SWCNT//NaTi₂(PO₄)₃–MWCNT batteries.

Fig. 6 (a) In situ Raman spectra collected for the Na₃V₂(PO₄)₂F₃–SWCNT electrode in the Na₃V₂(PO₄)₂F₃–SWCNT//NaTi₂(PO₄)₃–MWCNT aqueous full-cell. (b) The corresponding galvanostatic charge/discharge curve. (c) Partial enlargement of images of the in situ Raman spectra. The in situ Raman spectra are collected during the 1st cycle at a 1C rate between 0.6 and 2.1 V with a sampling interval of 3 min.

Based on the above mentioned results, we investigate the electrochemical performance of the Na₃V₂(PO₄)₂F₃–SWCNT//NaTi₂(PO₄)₃–MWCNT aqueous full-cell. All the specific capacities of the full cell are calculated based on the mass of Na₃V₂(PO₄)₂F₃–SWCNT-2. Fig. 7a gives the CVs and the charge/discharge performances of the Na₃V₂(PO₄)₂F₃–SWCNT//NaTi₂(PO₄)₃–MWCNT aqueous full-cell in a voltage range of 0.6 V to 2.1 V. Three obvious pairs of symmetric redox bands at 1.06/0.96, 1.52/1.46, and 1.94/1.89 V can be observed. They are attributed to extraction/insertion reactions of the sodium ions.^67,68 The voltage for the Na₃V₂(PO₄)₂F₃–MWCNT//NaTi₂(PO₄)₃–MWCNT full-cell is 1.92 V, which is one of the highest voltage obtained by using intercalation anode and cathode materials in aqueous electrolyte for ARSBs. EIS measurements of the full cell are also performed (as shown in Fig. S12†). The EIS curve is composed of a semicircle in the high-frequency region and a sloping line in the low-frequency region. The low impedance of 4.23 Ω benefits from the high conductivity of this aqueous electrolyte. The charge/discharge voltage profiles for each cycle are shown in Fig. 7b. A discharge capacity of 75.2 mA h g⁻¹ is achieved at a 1C rate (128 mA g⁻¹), with an excellent coulombic efficiency of 99%. Although the capacitance retention (ca. 74% for the 20th cycle), depicted in the inset of Fig. 7b, is not very superior due to the unwanted decomposition of water and the variation of the structure of the electrode material during ion intercalation/de-intercalation,⁶⁹ the significance of this work is that it at least verified the possibility of such a high voltage. In our opinion, we believe that with the modification of Na₃V₂(PO₄)₂F₃ and the electrolyte, there will be remarkably excellent performance.

Fig. 7 (a) CVs obtained at 1 mV s⁻¹ and (b) charge/discharge curves at 1C (1C = 128 mA h g⁻¹) for the Na₃V₂(PO₄)₂F₃–SWCNT//NaTi₂(PO₄)₃–MWCNT aqueous full-cell in 17 m NaClO₄ aq. (the inset is the cycle performance).

Conclusions

We have synthesized Na₃V₂(PO₄)₂F₃–SWCNT by a low temperature solvothermal method, and the Na₃V₂(PO₄)₂F₃–SWCNT cathode in a sodium ion battery exhibits a specific capacity of 117 mA h g⁻¹ with a voltage plateau of 4.1 V. When measured at 0.5C, Na₃V₂(PO₄)₂F₃–SWCNT could produce 114 mA h g⁻¹ and retain 92.4% after 100 cycles. In addition, the Na₃V₂(PO₄)₂F₃–SWCNT//NaTi₂(PO₄)₃–MWCNT aqueous full-cell can deliver a high voltage of 1.92 V, and an average voltage of 1.7 V, achieving a high energy density of 150 W h kg⁻¹ in 17 m NaClO₄ aq. These encouraging results may accelerate the development of ARSBs with a high concentration electrolyte, high potential cathode, and low potential anode.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Key R&D Program of China (2016YFB0901500, and 2016YFB0101201), the National Natural Science Foundation of China (51771094), the Ministry of Education of China (B12015 and IRT13R30) and the Fundamental Research Funds for the Central Universities.

Notes and references

1. D. Kundu, E. Talaie, V. Duffort and L. F. Nazar, Angew. Chem., Int. Ed., 2015, 54, 3431.
2. L. M. Suo, Y. S. Hu, H. Li, M. Armand and L. Q. Chen, Nat. Commun., 2013, 4, 2513.
3. J. B. Goodenough and K. S. Park, J. Am. Chem. Soc., 2013, 135, 1167.
4. W. Luo, F. Shen, C. Bommier, H. L. Zhu, X. L. Ji and L. B. Hu, Acc. Chem. Res., 2016, 49, 231.
5. Y. Y. Lu, P. F. Zhou, K. X. Lei, Q. Zhao, Z. L. Tao and J. Chen, Adv. Energy Mater., 2017, 7, 1601973.
6. X. D. Xiang, K. Zhang and J. Chen, Adv. Mater., 2015, 27, 5343.
7. S. Liu, L. Y. Shao, X. J. Zhang, M. Zhou, Z. L. Tao and J. Chen, J. Alloys Compd., 2018, 754, 147.
8. L. Y. Shao, Q. Zhao and J. Chen, Chem. Commun., 2017, 53, 435.
9. V. Palomares, P. Serras, I. Villaluenga, K. B. Hueso, J. Carretero-González and T. Rojo, Energy Environ. Sci., 2012, 5, 5884.
10. J. F. Mao, C. Luo, T. Gao, X. L. Fan and C. S. Wang, J. Mater. Chem. A, 2015, 3, 10378.
11. N. Yabuuchi, K. Kubota, M. Dahbi and S. Komaba, Chem. Rev., 2014, 114, 11636.
12. Y. Y. Huang, X. Lia, J. S. Wang, L. Miao, C. Li, J. T. Han and Y. H. Huang, Energy Storage Materials, 2018, 15, 108.
13. Y. Y. Huang, Y. H. Zheng, X. Li, F. Adams, W. Luo, Y. H. Huang and L. B. Hu, ACS Energy Lett., 2018, 3, 1604.
14. J. Zhao, Y. Gao, Q. Liu, X. Meng, N. Chen, C. Z. Wang, F. Du and G. Chen, Chem.–Eur. J., 2018, 24, 2913.
15. W. X. Song and S. Q. Liu, Solid State Sci., 2013, 15, 1.
16. B. L. Ellis, K. T. Lee and L. F. Nazar, Chem. Mater., 2010, 22, 691.
17. Z. Li, D. Young, K. Xiang, W. C. Carter and Y. M. Chiang, Adv. Energy Mater., 2013, 3, 290.
18. Z. Chang, Y. Q. Yang, M. X. Li, X. W. Wang and Y. P. Wu, J. Mater. Chem. A, 2014, 2, 10739.
19. Y. S. Wang, J. Liu, B. Lee, R. M. Qiao, Z. Z. Yang, S. Y. Xu, X. Q. Yu, L. Gu, Y. S. Hu, W. L. Yang, K. Kang, H. Li, X. Q. Yang, L. Q. Chen and X. J. Huang, Nat. Commun., 2015, 6, 6401.
20. Y. S. Wang, L. Q. Mu, J. Liu, Z. Z. Yang, X. Q. Yu, L. Gu, Y. S. Hu, H. Li, X. Q. Yang, L. Q. Chen and X. J. Huang, Adv. Energy Mater., 2015, 5, 1501005.
21. L. M. Suo, O. Borodin, Y. S. Wang, X. H. Rong, W. Sun, X. L. Fan, S. Y. Xu, M. A. Schroeder, A. V. Cresce, F. Wang, C. Y. Yang, Y. S. Hu, K. Xu and C. S. Wang, Adv. Energy Mater., 2017, 7, 1701189.
22. H. W. Long, W. Zeng, H. Wang, M. M. Qian, Y. H. Liang and Z. C. Wang, Adv. Sci., 2018, 1700634.
23. Z. W. Guo, Y. Zhao, Y. X. Ding, Y. Y. Xia, H. S. Peng and Y. G. Wang, Chem, 2017, 3, 348.
24. H. Qin, Z. P. Song, H. Zhan and Y. H. Zhou, J. Power Sources, 2014, 249, 367.
25. H. C. Gao and J. B. Goodenough, Angew. Chem., Int. Ed., 2016, 128, 12960.
26. X. Y. Wu, M. Y. Sun, Y. F. Shen, J. F. Qian, Y. L. Cao, X. P. Ai and H. X. Yang, ChemElectroChem, 2014, 7, 407.
27. P. R. Kumar, Y. H. Jung, J. E. Wang and D. K. Kim, J. Power Sources, 2016, 324, 421.
28. Q. Zhang, C. Y. Liao, T. Y. Zhai and H. Q. Li, Electrochim. Acta, 2016, 196, 470.
29. K. Nakamoto, R. Sakamoto, M. Ito, A. Kitajou and S. Okada, Electrochemistry, 2017, 85, 179.
30. Z. G. Hou, X. Q. Zhang, X. N. Li, Y. C. Zhu, J. W. Liang and Y. T. Qian, J. Mater. Chem. A, 2017, 5, 730.
31. Y. Yao, L. Zhang, Y. Gao, G. Chen, C. Z. Wang and F. Du, RSC Adv., 2018, 8, 2958.
32. W. X. Song, X. B. Ji, Z. P. Wu, Y. C. Yang, Z. Zhou, F. Q. Li and C. E. Banks, J. Power Sources, 2014, 256, 258.
33. X. Q. Xie, M.-Q. Zhao, B. Anasori, K. Maleski, C. E. Ren, J. W. Li, B. W. Byles, E. Pomerantseva, G. X. Wang and Y. Gogotsi, Nano Energy, 2016, 26, 513.
34. G. L. Polles, J. J. Videau, R. Olazcuaga and M. Couzi, J. Solid State Chem., 1996, 127, 341.
35. Z. L. Jian, W. Z. Han, X. Lu, H. X. Yang, Y. S. Hu, J. Zhou, Z. B. Zhou, J. Q. Li, W. Chen, D. F. Chen and L. Q. Chen, Adv. Energy Mater., 2013, 3, 156.
36. P. R. Kumar, Y. H. Jung and D. K. Kim, J. Solid State Electrochem., 2017, 21, 223.
37. Y. Zhang, S. R. Guo and H. Y. Xu, J. Mater. Chem. A, 2018, 6, 4525.
38. L. Li, Y. L. Xu, X. F. Sun, S. N. He and L. Li, Chem. Eng. J., 2018, 331, 712.
39. J. Z. Guo, P. F. Wang, X. L. Wu, X. H. Zhang, Q. Y. Yan, H. Chen, J. P. Zhang and Y. G. Guo, Adv. Mater., 2017, 29, 1701968.
40. F. Sauvage, E. Quarez, J. M. Tarascon and E. Baudrin, Solid State Sci., 2006, 8, 1215.
41. Y. R. Qi, L. Q. Mu, J. M. Zhao, Y. S. Hu, H. Z. Liu and S. Dai, Angew. Chem., Int. Ed., 2015, 54, 9911.
42. S. Yu, X. Wang, R. Zhang, T. Yang, Y. Ai, T. Wen, W. Huang, T. Hayat, A. Alsaedi and X. Wang, Sci. Rep., 2017, 7, 39625.
43. J. M. Shim, D. G. Kim, H. J. Kim, J. H. Lee, J. H. Baika and J. C. Lee, J. Mater. Chem. A, 2014, 2, 13873.
44. L. Shi, X. X. Fu, C. Y. Fan, S. Q. Yu, G. D. Qian and Z. Y. Wang, RSC Adv., 2015, 5, 85179.
45. R. A. Shakoor, D. H. Seo, H. Kim, Y. U. Park, J. Kim, S. W. Kim, H. Gwon, S. Leec and K. Kang, J. Mater. Chem., 2012, 22, 20535.
46. C. B. Zhu, C. Wu, C. C. Chen, P. Kopold, P. A. van Aken, J. Maier and Y. Yu, Chem. Mater., 2017, 29, 5207.
47. W. X. Song, X. Y. Cao, Z. P. Wu, J. Chen, Y. R. Zhu, H. S. Hou, Q. Lan and X. B. Ji, Langmuir, 2014, 30, 12438.
48. H. M. Yi, M. X. Ling, W. B. Xu, X. F. Li, Q. Zheng and H. M. Zhang, Nano Energy, 2018, 47, 340.
49. Y. S. Cai, X. X. Cao, Z. G. Luo, G. Z. Fang, F. Liu, J. Zhou, A. Q. Pan and S. Q. Liang, Adv. Sci., 2018, 1800680.
50. Y. N. Xu, Q. L. Wei, C. Xu, Q. D. Li, Q. Y. An, P. F. Zhang, J. Z. Sheng, L. Zhou and L. Q. Mai, Adv. Energy Mater., 2016, 6, 1600389.
51. D. Wang, X. Bie, Q. Fu, D. Dixon, N. Bramnik, Y. S. Hu, F. Fauth, Y. Wei, H. Ehrenberg, G. Chen and F. Du, Nat. Commun., 2017, 8, 15888.
52. Y. B. Niu, M. W. Xu, Y. Zhang, J. Han, Y. Wang and C. M. Li, RSC Adv., 2016, 6, 45605.
53. Q. Y. Wang, B. D. Zhao, S. Zhang, X. H. Gao and C. Deng, J. Mater. Chem. A, 2015, 3, 7732.
54. P. P. Prosini, M. Lisi, D. Zane and M. Pasquali, Solid State Ionics, 2002, 148, 45.
55. S. Franger, F. L. Cras, C. Bourbon and H. Rouault, Electrochem. Solid-State Lett., 2002, 5, A231.
56. P. R. Kumar, Y. H. Jung, B. Moorthy and D. K. Kim, J. Electrochem. Soc., 2016, 163, A1484.
57. P. R. Kumar, Y. H. Jung, C. H. Lim and D. K. Kim, J. Mater. Chem. A, 2015, 3, 6271.
58. W. Lia, K. L. Wang, S. J. Cheng and K. Jiang, Energy Storage Materials, 2018, 15, 14.
59. M. A. Muñoz-Marquez, M. Zarrabeitia, E. Castillo-Martínez, A. Eguía-Barrio, T. Rojo and M. Casas-Cabanas, ACS Appl. Mater. Interfaces, 2015, 7, 7801.
60. N. Zhang, F. Y. Cheng, J. X. Liu, L. B. Wang, X. H. Long, X. S. Liu, F. J. Li and J. Chen, Nat. Commun., 2017, 8, 405.
61. F. Wang, Y. X. Lin, L. M. Suo, X. L. Fan, T. Gao, C. Y. Yang, F. D. Han, Y. Qi, K. Xu and C. S. Wang, Energy Environ. Sci., 2016, 9, 3666.
62. M. Bianchini, F. Fauth, N. Brisset, F. Weill, E. Suard, C. Masquelier and L. Croguennec, Chem. Mater., 2015, 27, 3009.
63. S. I. I. Park, I. Gocheva, S. Okada and J. Yamaki, J. Electrochem. Soc., 2011, 158, A1067.
64. Z. G. Liu, Y. Y. Hu, M. T. Dunstan, H. Huo, X. G. Hao, H. Zou, G. M. Zhong, Y. Yang and C. P. Grey, Chem. Mater., 2014, 26, 2513.
65. Y. A. Kim, M. Kojima, H. Muramatsu, S. Umemoto, T. Watanabe, K. Yoshida, K. Sato, T. Ikeda, T. Hayashi, M. Endo, M. Terrones and M. S. Dresselhaus, Small, 2006, 2, 667.
66. M. Kalbac, Y. P. Hsieh, H. Farhat, L. Kavan, M. Hofmann, J. Kong and M. S. Dresselhaus, Nano Lett., 2010, 10, 4619.
67. C. Shen, H. Long, G. C. Wang, W. Lu, L. Shao and K. Y. Xie, J. Mater. Chem. A, 2018, 6, 6007.
68. I. L. Matts, S. Dacek, T. K. Pietrzak, R. Malik and G. Ceder, Chem. Mater., 2015, 27, 6008.
69. W. X. Song, X. B. Ji, Y. R. Zhu, H. J. Zhu, F. Lu, Y. P. Yao and C. E. Ban, ChemElectroChem, 2014, 1, 871.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details and additional results are included. See DOI: 10.1039/c8ta09194c

---