Reversible intercalation of aluminium into vanadium pentoxide xerogel for aqueous rechargeable batteries

Abstract

Rechargeable batteries based on the intercalation of aluminium ions may be competitive against lithium-ion batteries, but their development and comprehension are full of difficulties. The charge/discharge processes are particularly complex in aqueous electrolyte solutions. The electrochemical behaviour of orthorhombic V₂O₅, obtained from xerogel, in an aluminium cell is studied here by using electrochemical cycling, impedance spectroscopy, XRD and XPS results. After electrochemical intercalation of aluminium, the resulting (Al³⁺)_x/3[(V⁴⁺)_x,(V⁵⁺)_2−x]O₅·nH₂O is XRD-amorphous at approximately x = 0.5. The reversible capacity is ca. 120 mA h g⁻¹ (equivalent to Al_0.27V₂O₅). The loss of crystallinity induced by the electrochemical intercalation enhances the chemical exchange between the electrode material and the electrolyte solution. In the presence of acidic water solution, besides the faradic electrochemical process driven by the electrical current, aluminium, protons and water also can be intercalated into V₂O₅ by chemical reactions or ion exchange.

---

Introduction

Lithium-ion batteries have been successfully commercialized since the 1990s. At present, batteries based on aluminium may be advantageous against lithium batteries in terms of economy, sustainability and energy density.¹ Although, lithium and sodium intercalation reactions have been extensively studied over decades, the study of the insertion of multivalent cations has received less attention and it is still in its infancy. Several of the most promising materials proposed for reversible intercalation of aluminium ions are titanium dioxides^2–4 and vanadium oxides.^5–8

Orthorhombic vanadium pentoxide (V₂O₅) possesses a layered structure built up from VO₅ square pyramids sharing edges and corners, with V₂O₅ layers held together via weak vanadium–oxygen interaction.⁹ The layered character of V₂O₅ and the high redox potential of V⁵⁺ make it an interesting and versatile host material for many guest species and for batteries based on intercalation of ions such as lithium, sodium, magnesium and aluminium. It is worth to note that xero-V₂O₅ adsorbs water at the oxide–air interface and that, according to the layer structure of V₂O₅, water absorption may be described as an intercalation process.¹⁰ The xerogel (xero-V₂O₅) consists of V₂O₅ bilayers with water between them and it is a Brønsted acid. Thus, the intercalated water can play an important role in the electrochemistry of V₂O₅.^11–13 In addition, vanadium pentoxide gels are very sensitive to some reduction of vanadium ions (weak mixed V⁵⁺–V⁴⁺ character) and its content in V⁴⁺ is related to the water content (dehydratation leads to some reduction of vanadium). In non-aqueous lithium cell, vanadium pentoxide exchanges its residual water with the organic solvent and probably the discrepancies in the electrochemical behaviour reported by different authors may be due to different water content in their samples.^10–15 According to the literature,^14,15 the xerogel of V₂O₅ can contain protonated water (H₃O⁺) in the interlayer space and the layers of vanadium oxide are negatively charged (V₂O₅^m−) with about 0.2 charge per vanadium atom. This fact is due to the acidity of V⁵⁺ which drives to hydrolysis. Thus, the xero-V₂O₅ is a protonic conductor and presents ionic exchange properties between the H₃O⁺ cations and various charged species (Mⁿ⁺), such as Al³⁺.¹⁴ By immersion of amorphous or poorly crystallized xero-V₂O₅ in an aqueous solution of cations, the negative charge of the V₂O₅^m− layer remains fixed while the H₃O⁺ ions in the interlayer space are exchanged with the outer cations of the solution. By thermal annealing of the M-intercalated xero-V₂O₅ materials (M_xⁿ⁺, (V₂O₅)^m−, yH₂O), vanadium bronzes are formed (M_xV₂O₅) in which the initial vanadium is partially reduced (about 600 °C for M = Al).¹⁵

Surprisingly, in contrast to lithium and magnesium batteries, the role of water molecules in the aluminium intercalation into V₂O₅ has received little attention. Within the field of post-lithium batteries, the intercalation of aluminium into TiO₂ for batteries using aqueous electrolyte solutions has been explored.^2,3 However, to the best of our knowledge, the intercalation of aluminium into vanadium pentoxide for batteries using aqueous electrolyte has not been reported yet; irrespectively of the reports using electrolytes based on ionic liquids.^16–18

Experimental

Synthesis of V₂O₅

The V₂O₅ sample was prepared by using the ion exchange method. A 0.5 M solution of sodium metavanadate in water was eluted through a proton exchange resin (Downing). The resulting hydrogel was aged and dried in an air atmosphere for two weeks. Amorphous xerogel of vanadium pentoxide is obtained at this stage. Then, the xerogel (xero-V₂O₅) is manually milled, further dried at 120 °C and crystallized at 300 °C for three hours in air atmosphere.

Characterization of materials

X-ray diffraction (XRD) patterns were recorded in a Bruker D8 instrument with CuKα radiation and reflection geometry. For recording XRD of the recuperated electrodes, the samples were protected against reaction with air by covering with kapton tape.

For the images of electron microscopy, TEM (JEM 1400) and FE-SEM (JSM 7800F) were used.

Thermogravimetric analysis (TGA) was carried out in Shimadzu instrument at 10 °C min⁻¹ of heating rate in air atmosphere.

The chemical state and composition were analyzed by X-ray photoelectron (XP) spectra that were recorded using a SPECS Phobios 150MCD and Mg Kα source. Before recoding the XPS of the electrode recuperated from the electrochemical cell, the electrode material was washed with water to remove the remaining electrolyte solution and then dried under vacuum.

Electrochemical tests

All the electrochemical experiments were performed in a VMP Biologic instrument. The electrochemical cells were of Swagelok-type in T-configuration and, in order to prevent corrosion, the use of metals as current collectors was avoided. The active material (xero-V₂O₅ annealed at 300 °C, 80%) was mixed with polytetrafluoroethylene (PTFE, 10%) binder and carbon black (conductive additive, 10%), and the resulting mixture was spread on carbon paper (graphite, Goodfellow). The electrolyte solution was 1 M AlCl₃ in deionized water which was previously outgassed by flowing Ar. Several discs of Whatman glass fiber filter were used like separator. The counter electrode was glassy carbon. Mercury/mercurous sulfate electrode Hg/Hg₂SO₄, K₂SO₄(sat'd) was used as reference electrode (MSE).

Results and discussion

Microstructure

The V₂O₅ xerogel dried at 120 °C is XRD-amorphous (Fig. 1a). After annealing at 300 °C it shows a XRD-pattern (Fig. 1b) that is ascribed to orthorhombic vanadium pentoxide (layered structure), with space group Pmmn. The resulting lattice parameters are a = 11.521(2) Å, b = 3.563(1) Å and c = 4.3719(6) Å. The relative intensity of the reflection (001) is in good agreement with the layered character and with preferred orientation of the particles. The average crystallite size obtained by applying the Scherrer equation (L = λβ⁻¹ cos θ⁻¹) to the broadening (β) of the Bragg reflections is L = 45(4) nm.

Fig. 1 XRD pattern of xero-V₂O₅ dried at 120 (a) and 300 °C (b). The Miller indexes are given.

The morphology of the particles was examined by SEM and TEM (Fig. 2). In the SEM results the particles of the sample dried at 120 °C have the appearance of crumpled layers, and after annealing at 300 °C the layered particles become less crumpled and more flat. In the TEM micrographs (Fig. 2E and F), the particles show layered character, with around 30–250 nm of diameter.

Fig. 2 SEM (A–D) and TEM (E and F) micrographs of xero-V₂O₅ heated to 120 (A and B) and 300 °C (C–F).

Thermogravimetric analysis

The water content of V₂O₅ can greatly influence on its electrochemical cycling and capacity in batteries.^11,12 Water may be detrimental for non-aqueous batteries, but a small amount of water can help to the intercalation of the cation into vanadium oxide. It is possible to obtain the anhydrous form of V₂O₅ whether the sample is heated for enough time at 250 °C or higher temperature, although the water can be reabsorbed. Fig. 3 shows to the TGA results of a xero-V₂O₅ sample annealed at 300 °C and stored in air atmosphere for two months. The xero-V₂O₅ sample contains a small amount of water (V₂O₅·nH₂O). Firstly, weakly bonded or adsorbed water is released at ca. 100 °C. Then, more strongly water is lost at 250–300 °C. If one takes into account all the water content found in the TGA up in all the range of temperature from 30 to 300 °C, then it results that n = 0.29 in V₂O₅·nH₂O. If the water loss at lower temperature is considered to be free or weakly adsorbed water, the content of more strongly bonded water is n = 0.16 (from 100 to 300 °C). According to the resulting spacing of the basal plane obtained from the XRD pattern and the literature,¹⁹ the water molecules are not placed in the interlayer spacing. Even after annealing at 600 °C, the XRD pattern of vanadium oxide remains nearly unchanged (not shown), confirming that there was not a significant amount of water in the interlayer space. These results show that xero-V₂O₅ has tendency to adsorb limited amounts of water from the atmosphere.

Fig. 3 TGA curve of xero-V₂O₅ (heated to 300 °C two months before recording the TGA experiment). Firstly, the sample was heated at a rate of 10 °C min⁻¹ up to 300 °C, and then the temperature was hold at 300 °C for one hour.

Electrochemical performance

The electrochemical behavior of the as-prepared V₂O₅ sample was examined using cyclic voltammetry (Fig. 4) and galvanostatic cycling (Fig. 5). In the cyclic voltammogram (CV) of V₂O₅ in aluminium cell a pair of redox peaks can be observed at −0.30 and −0.03 V (vs. MSE) in the cathodic and anodic sweep, respectively (Fig. 4A). These peaks agree well with a reversible faradic process. Corrosion and electrolyte decomposition are not taking place in the voltage range in which V₂O₅ is employed.

Fig. 4 Cyclic voltammetry results. (A) Comparison of cyclic voltammograms (CV) of V₂O₅ (working electrode) and blank experiment (no V₂O₅) using different solutions and at scanning rate of 10 mV s⁻¹. (B) Capacity as a function of cycle number at different scan rates in 1 M AlCl₃(aq.) solution (pH 2.3). (C) CV of V₂O₅ in 1 M AlCl₃(aq.) at 0.5 mV s⁻¹.

Fig. 5 Galvanostatic experiments. (A) A typical voltage–capacity curve of V₂O₅ in aluminium aqueous cell at 60 mA g⁻¹ of current density. (B) Specific capacity as a function of cycle number at different current densities.

As a blank test, the CV obtained in the absence of active material (the working electrode is a carbon foil) shows no redox activity in the voltage range between −1.3 and +0.5 V. In comparison with anatase,² the voltage of V₂O₅ is higher, and this fact can involve higher energy density and that H₂ is not evolved during the reduction process. For the sake of comparison, aqueous solutions of AlCl₃ an HCl were used as electrolytes.

If one looks at the cyclic voltammetry results of V₂O₅ in hydrochloric solution (Fig. 4A), it is evident that the protons can play an important role. In solution of aluminium-free hydrochloric acid at pH 1.3 the current becomes more intense than at pH 2.3, strongly suggesting that more protons are inserted at more acidic pH.

The electrochemical intercalation in aqueous solution of protons yielding H_xV₂O₅ (0 < x < 0.5) and the insertion of lithium to form Li_xV₂O₅ (0 < x < 1.0) were previously reported by other authors.²⁰ The same authors reported that larger cations are not intercalated into V₂O₅. Taking into account that the crystal ionic radius of Al³⁺ (0.68 Å) is smaller than Li⁺ (0.90 Å), it seems reasonable to believe that aluminium can be intercalated. In this sense, other authors found minimum activation energy for aluminium insertion (by exchange in water solution) into xero-V₂O₅.²¹ Thus, in Fig. 4A, for a pH value of 2.3, the CV curve is modified in the presence of aluminium ions in comparison with aluminium-free solution at the same pH, indicating that aluminium insertion is preferential to proton insertion. The redox peaks in 1 M AlCl₃(aq.) are resulted from aluminium insertion. In Fig. 4B, it is observed that the capacity is higher at lower scan rates, and the capacity retention is good. In the CV at low rate the peak intensity is smaller (Fig. 4C), as expected for a diffusion-controlled process.

In the galvanostatic experiment, the voltage–capacity curve shows a reversible pseudoplateau at ca. −0.15 V in the discharge process and at ca. 0.0 V in the charge process (Fig. 5A). The origin of this pseudoplateau can be the intercalation process of hydrated aluminium and the coexistence of crystalline and amorphous phase as it is further discussed below in the light of XRD and XPS results. The reversible capacity is substantially lower at higher rates (ca. 20 mA h g⁻¹ at 200 mA g⁻¹ of current density) than at low rate (in the order of 120 mA h g⁻¹ at 60 mA g⁻¹) (Fig. 5B), suggesting that the reaction is controlled by diffusion.

Hydrated aluminium ions are intercalated into the interlayer space of V₂O₅ and V⁵⁺ ions are reduced to V⁴⁺ according to the next reaction:

x/3Al(H₂O)_n³⁺ + (V⁵⁺)₂O₅ + xe⁻ → (Al³⁺)_x/3[(V⁴⁺)_x,(V⁵⁺)_2−x]O₅·nH₂O (1)

Water molecules can shield the charge of the aluminium ion facilitating its insertion/deinsertion into vanadium oxide. Thus, it is expected that water molecules are co-intercalated simultaneously to aluminium-intercalation. On the other hand, it is known that in the interlayer space of V₂O₅ in contact with water, protonated water (H₃O⁺) and Al³⁺ can be chemically exchanged.²² In the same sense, other authors found that V₂O₅ can be doped with aluminium ions (Al_xV₂O₅·nH₂O) using soft chemistry.^15,23 Thus, chemical exchange can overlap with the electrochemical process that is driven by the flow of electrons through the external electrical circuit during the discharge/charge. There is a competition between chemical exchange and electrochemical insertion/deinsertion of aluminium ions and water molecules (or H₃O⁺). Most probably, some protons are also intercalated during the discharge process. Then, aluminium ions are also chemically inserted into V₂O₅ by cationic exchange with protons in competition to the electrochemical deinsertion during the charge process, though the ion exchange process is slower than the faradic reaction. It involves that the aluminium content in Al_xV₂O₅, determined from charge passed through the cell, would be overestimated. Therefore, the apparent discharge capacity could be slightly lower than the charge capacity, particularly at low kinetics (60 mA g⁻¹ in Fig. 5B).

For the sake of comparison, the electrochemical intercalation of lithium into V₂O₅ was explored. The results agree well with the literature and hence the capacity fade upon cycling in Fig. 5 and 6 is due to the dissolution of vanadium compounds.^24–26 A discharge capacity around 130–250 mA h g⁻¹ and a charge capacity of only ca. 70 mA h g⁻¹ is observed in (Fig. 6). The operating voltage is lower in the lithium electrolyte than in aluminium one. In contrast to aluminium, lithium is intercalated in non-hydrated form,²⁰ only a fraction of lithium can be deintercalated and the charge capacity is substantially smaller than the discharge capacity (Fig. 6). Some amount of lithium is irreversibly trapped in the framework of V₂O₅, and the charge of Li⁺ helps to shield the repulsions between the oxygens in the framework of V₂O₅. It is worth to note that, in comparison with other ions, such as Li⁺ and Na⁺, Al³⁺ is a smaller ion with a higher hydration number. It is expected that the chemical exchange between the solution and intercalated vanadium oxide is enhanced for aluminium, and that water (or protonated water) helps to shield the repulsions between the oxygens in the framework of Al_xV₂O₅.

Fig. 6 Electrochemistry of V₂O₅ in lithium cell: (A) CV at 10 mV s⁻¹ of scan rate and (B) galvanostatic cycling at 20 mA g⁻¹ of current density. Electrolyte: 1 M LiCl in water.

Ex situ XRD

According to the ex situ XRD patterns (Fig. 7) of the electrodes recorded at different states of discharge, the structure of V₂O₅ is initially preserved during the beginning of the discharge process, the position of the Bragg peaks remain nearly unchanged and become more broadened. New Bragg peaks are not observed. It is worth to note that the interlayer spacing in V₂O₅ is much larger that the ionic size of Al³⁺. With further discharge depth, the oxide irreversibly becomes XRD-amorphous at the nominal composition Al_0.5V₂O₅ (equivalent to 221 mA h g⁻¹). The amorphous state of Al-intercalated vanadium oxide is strongly indicative of the hydration of the interlayer spacing.¹⁰ In other words, aluminium and water molecules are co-intercalated and the resulting hydrated vanadium oxide (Al³⁺)_x/3[(V⁴⁺)_x,(V⁵⁺)_2−x]O₅·nH₂O is XRD-amorphous. Another possible explanation would be that the chemical exchange between intercalated aluminium and protons from the solution in the form of H₃O⁺ provokes the amorphization of the vanadium oxide electrode, but it could not explain alone the high capacity of the charge process. In contrast to the insertion of hydrated aluminium, according to the literature the insertions of lithium and protons preserve the crystalline character of V₂O₅.²⁰

Fig. 7 Ex situ XRD patterns of (a) raw V₂O₅ and electrodes recuperated from aluminium cells at selected states of discharge (b–e) and charge (f) at slow current density. The (002) reflection of hexagonal graphite (substrate of the electrode active material) is also observed at ca. 26.4°.

Chemical and electrochemical intercalation

The redox potential of V⁵⁺ is quite high (E°(VO₂⁺/VO²⁺) = +1.0 V vs. SHE), but this potential in principle would not be enough to decompose water (E°(O₂/H₂O) = +1.229 V vs. SHE). In contrast to crystalline V₂O₅, it is known that amorphous or poorly crystallized xero-V₂O₅ immersed into an aqueous solution containing certain cations can intercalate these cations by easy exchange with the H₃O⁺ ion.¹⁵ We tested the stability of crystalline V₂O₅ in aqueous solution containing Al³⁺ ions. After immersion of V₂O₅ in 1 M AlCl₃(aq.) solution during two hours, a slight decrease of the pH value (ca. 0.15 units) and formation of tiny bubbles (more probably oxygen) were observed. In addition, the crystalline structure of vanadium oxide is preserved (see XRD in ESI†) and no relevant change of the unit cell parameters was observed. The solution remained colourless. These results suggest that, although V₂O₅ is not soluble, it may spontaneously suffers a very limited intercalation, proton/aluminium exchange and redox reaction in which vanadium oxide would be slightly reduced and O₂ is released:

V₂O₅ + xAl(H₂O)_n³⁺ + 3x/2H₂O → Al_xV₂O₅·nH₂O + 3x/4O₂ + 3xH⁺ (2)

The positive charge of the chemically intercalated aluminium ion balances the negative charge left in V₂O₅^3x− by the electrons released by oxygen.

In the electrochemical cell, the loss of crystallinity that is induced by the electrochemical intercalation of aluminium (reaction (1)) can enhance the extent of the chemical exchange in XRD-amorphous (Al³⁺)_x/3[(V⁴⁺)_x,(V⁵⁺)_2−x]O₅·nH₂O (reaction (2)). The reaction (2) also can explain the apparently higher charge capacity observed in some cycles of the galvanostatic experiments. In the same sense, the reduction of V⁵⁺ and formation of O₂ was reported by Znaidi et al.¹⁴ In addition, Schöllhorn et al. reported the proton-exchange reaction in vanadium bronzes.²²

X-ray photoelectron spectroscopy

To further understand the mechanism of aluminium intercalation into V₂O₅, the X-ray photoelectron spectra of the samples are shown in Fig. 8. For the raw V₂O₅ sample the main XP lines of O 1s (Fig. 8Aa) and V 2p_3/2 (Fig. 8Ba) are placed at the binding energies lines of 530.5 (O²⁻) and 517.5 eV (V⁵⁺), respectively. These values agree very well with the reported data for V₂O₅.²⁷ A minor component at 516.1 eV is ascribed to traces of V⁴⁺. These V⁴⁺ ions can increase the electronic conductivity. There are two additional O 1s components with very low intensity and placed at ca. 532 and 534 eV that are ascribed to more covalently bonded oxygen in defects and contamination on the surface of V₂O₅.^27,28

Fig. 8 X-ray photoelectron spectra of (A) O 1s core level and (B) V 2p_3/2, (C) Al 2p and V 3s core levels for (a) as prepared V₂O₅, (b) discharged electrode (end of the first discharge at −0.3 V vs. MSE) and (c) electrode after the first discharge–charge cycle (+0.15 V vs. MSE).

After reaction with aluminium, the V 2p line is shifted to lower binding energy as due to the reduction of V⁵⁺ to V⁴⁺ (Fig. 8Bb), and the O 1s line is shifted to higher energy (Fig. 8Ab). The charge process leads to the re-oxidation of vanadium (Fig. 8Bb). The broadening of XPS in the region of O 1s core peak in the charged electrode (Fig. 8Ac) is in good agreement with a more disordered structure.

In the discharged electrode, the Al 2p line is observed at 74.6 eV (Fig. 8C). The V 3s core hole lines have a binding energy of ca. 70.3 eV and a small satellite at ca. 4–8 eV higher binding energy (Fig. 8C). It has been suggested that the satellite structure of V 3s is caused by core hole screening.²⁹ The satellite observed in raw V₂O₅ (Fig. 8Ca) is in good agreement with the presence of traces of V⁴⁺. According to the literature, in vanadium oxides there is exchange coupling of the 3s electrons and the valence shell (3d electrons), but this exchange does not directly reflect the 3d occupancy.²⁹ In the discharged electrode (Fig. 8Cb), more probably the V 3s signal is too small to be observed.

The Auger transitions are significant to reveal the reversible change of the oxidation state of vanadium (Fig. 8D). The V(L₃M_4,5M_4,5) transition is mixed with the O(KL_2,3L_2,3) transition, making this Auger peak more intense.²⁷ The V(L₃M_4,5M_4,5) transition is shifted to higher energies for the discharged electrode that contains V⁴⁺ (3d¹). The peak of the V (L₃M_2,3M_4,5) Auger transition is due not only to the vanadium transition, but also to interatomic transitions between vanadium and oxygen, and such peak has lower intensity for the discharged electrode (Fig. 8Db) as expected according to the literature.²⁷ In V₂O₅ (V⁵⁺, 3d⁰), the greater covalency leads to a higher density of electrons (coming from oxide ion) around vanadium ions as compared with Al_xV₂O₅.

From the quantitative analysis of the XP spectra, the chemical compositions of the discharged and charged electrodes would be Al₅V₂O₅x5.6H₂O and Al_0.52V₂O₅, respectively. These formulas must be regarded with a lot of caution because XPS is a surface technique and the electrode surface can be contaminated with traces of electrolyte, but the results strongly suggest that hydrated aluminium is reversibly intercalated into vanadium pentoxide in good agreement with reaction (1).

Impedance spectra

The electrochemical impedance spectra (Fig. 9) are in good agreement with the diffusion of aluminium ions into the framework of V₂O₅. The spectrum recorded on the freshly assembled cell, at −0.05 V of open circuit voltage (OCV), shows a linear behaviour with an angle of ca. 45°. Is is characteristic of a system under diffusion control since the kinetics of the charge transfer at the electrode–electrolyte interface is faster than the diffusion of the Al³⁺ ions in V₂O₅. After the first discharge at −0.2 V, the resulting spectrum still shows a linear behaviour. After 10 cycles, a depressed semicircle is observed at high–medium frequencies, which is ascribed to charge transfer resistance and double-layer capacitance. The detailed view of the high–medium frequencies region is shown in Fig. 9B. The linear part at medium–low frequencies is ascribed to Warburg impedance associated with diffusion of aluminium in V₂O₅. The small size of the semicircle shows that the charge transfer resistance is low, and according to the fitting using the equivalent circuit shown as an inset in Fig. 9B, the resistance values is R₂ = 18 Ω.

Fig. 9 (A) Impedance spectra (Nyquist plots) of V₂O₅ electrode measured at OCV and at the end of the first discharge (−0.3 V vs. MSE) and after ten cycles. (B) A detailed view of the high–medium frequencies region after ten cycles and equivalent circuit. W = Warburg impedance. R = resistance. Q = constant phase element.

Conclusions

Vanadium pentoxide has been used as an active electrode material in rechargeable aluminium battery based on aqueous electrolyte. These batteries can be environmentally friendly and more sustainable in comparison with lithium batteries. The intercalation/deintercalation process of aluminium is complex. During the discharge process, vanadium is reduced and the resulting product becomes XRD-amorphous at Al_0.5V₂O₅, in good agreement with bulk intercalation. As due to the high charge-to-size ratio of Al³⁺ and the chemical exchange properties, the electrochemical (faradic) intercalation of aluminium into V₂O₅ is parallel to water intercalation, and some protons also can be intercalated. During the charge (oxidation) process the chemical intercalation and exchange of protons and aluminium can compete against the faradic deintercalation process. These features should be taken into account for the future development of batteries based on intercalation of multivalent ions. Perhaps, further optimization of the electrolyte composition, the selection of another binder and the formation of composite materials may improve the coulombic efficiency and cycling stability.

Acknowledgements

The authors would like to greatly acknowledge MINECO, FEDER (MAT2014-56470-R). The instruments for XPS, SEM and TEM were provided by SCAI (UCO Central Service for Research Support). We also thank Institute of Fine Chemistry and Nanochemistry and European Research Institute ERI-ALISTORE.

References

1. M. C. Lin, M. Gong, B. Lu, Y. Wu, D. Y. Wang, M. Guan, M. Angel, C. Chen, J. Yang, B. J. Hwang and H. Dai, Nature, 2015, 520, 324; V. Rani, V. Kanakaiah, T. Dadmal, M. S. Rao and S. Bhavanarushi, J. Electrochem. Soc., 2013, 160, A1781; R. Mori, RSC Adv., 2013, 3, 11547; R. Mori, RSC Adv., 2014, 4, 1982.
2. S. Liu, J. J. Hu, N. F. Yan, G. L. Pan, G. R. Li and X. P. Gao, Energy Environ. Sci., 2012, 5, 9743.
3. Y. J. He, J. F. Peng, W. Chu, Y. Z. Li and D. G. Tong, J. Mater. Chem. A, 2014, 2, 1721.
4. S. Sang, Y. Liu, W. Zhong, K. Liu, H. Liu and Q. Wu, Electrochim. Acta, 2016, 187, 92.
5. H. Wang, Y. Bai, S. Chen, X. Luo, C. Wu, F. Wu, J. Lu and K. Amine, ACS Appl. Mater. Interfaces, 2015, 7, 80.
6. N. Jayaprakash, S. K. Das and L. A. Archer, Chem. Commun., 2011, 47, 12610.
7. W. Wang, B. Jiang, W. Xiong, H. Sun, Z. Lin, L. Hu, J. Tu, J. Hou, H. Zhu and S. Jiao, Sci. Rep., 2013, 3, 3383.
8. D. B. Le, S. Passerini, F. Coustier, J. Guo, T. Soderstrom, B. B. Owens and W. H. Smyrl, Chem. Mater., 1998, 10, 682.
9. R. Enjalbert and J. Galy, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1986, 42, 1467.
10. J. Livage, Chem. Mater., 1991, 3, 578; H. H. Kristoffersen and H. Metiu, J. Phys. Chem. C, 2016, 120, 3986.
11. J. Wang, C. J. Curtis, D. L. Schulz and J. G. Zhang, J. Electrochem. Soc., 2004, 151, A1.
12. Y. Wang, H. Shang, T. Chou and G. Cao, J. Phys. Chem. B, 2005, 109, 11361.
13. M. Vujković, I. Pašti, I. Stojković Simatovića, B. Šljukić, M. Milenković and S. Mentus, Solid State Ion., 2015, 176, 130.
14. L. Znaidi, N. Baffier and M. Huber, Mater. Res. Bull., 1989, 24, 1501.
15. N. Baffier, L. Znaidi and M. Huber, Mater. Res. Bull., 1990, 25, 705.
16. N. Jayaprakash, S. K. Das and L. A. Archer, Chem. Commun., 2011, 47, 12610.
17. L. D. Reed and E. Menke, J. Electrochem. Soc., 2013, 160, A915.
18. M. Chiku, H. Takeda, S. Matsumura, E. Higuchi and H. Inoue, ACS Appl. Mater. Interfaces, 2015, 7, 24385.
19. P. Aldebert, N. Baffier, N. Gharbi and J. Livage, Mater. Res. Bull., 1981, 16, 669.
20. P. G. Dickens, A. M. Chippindale and S. J. Hibble, Solid State lon., 1989, 34, 79.
21. J. C. Badot and N. Baffier, J. Mater. Chem., 1992, 2, 1167.
22. R. Schöllhorn, F. Klein-Reesink and R. Reimold, J. Chem. Soc., Chem. Commun., 1979, 9, 398.
23. K. Zhu, H. Qiu, Y. Zhang, D. Zhang, G. Chen and Y. Wei, ChemSusChem, 2015, 8, 1017.
24. I. Stojkovic, N. Cvjeticanin, I. Pašti, M. Mitric and S. Mentus, Electrochem. Commun., 2009, 11, 1512.
25. M. Vujković, I. Pašti, I. Stojković Simatović, B. Šljukić, M. Milenković and S. Mentus, Electrochim. Acta, 2015, 176, 130.
26. H. Wang, Y. Zeng, K. Huang, S. Liu and L. Chen, Electrochim. Acta, 2007, 52, 5102.
27. Q. H. Wu, A. Thißen and W. Jaegermann, Appl. Surf. Sci., 2005, 252, 1801; G. Silversmit, D. Depla, H. Poelman, G. B. Marin and R. De Gryse, Surf. Sci., 2006, 600, 3512; A. Benayad, H. Martinez, A. Gies, B. Pecquenard, A. Levasseur and D. Gonbeau, J. Electron Spectrosc. Relat. Phenom., 2006, 150, 1; A. Meisel, K. H. Hallmeier, R. Szargan, J. Muller and W. Schneider, Phys. Scr., 1990, 41, 513; J. Kasperkiewicz, J. A. Kovacich and D. Lichtman, J. Electron Spectrosc. Relat. Phenom., 1983, 32, 123; P. Mezentzeff, Y. Lifshitz and J. W. Rabalais, Nucl. Instrum. Methods Phys. Res., Sect. B, 1990, 44, 296.
28. A. Thissen, D. Ensling, F. J. Fernández-Madrigal, W. Jaegermann, R. Alcántara, P. Lavela and J. L. Tirado, Chem. Mater., 2005, 17, 5202.
29. P. Krüger, M. Taguchi, J. C. Parlebas and A. Kotani, Phys. Rev. B: Condens. Matter Mater. Phys., 1997, 55, 16466; R. Zimmermann, R. Claessen, F. Reinert, P. Steiner and S. Hüfner, J. Phys.: Condens. Matter, 1998, 10, 5697.

---

Footnote

† Electronic supplementary information (ESI) available: XRD results. See DOI: 10.1039/c6ra11030d

---