Article pubs.acs.org/cm
A2VO(SO4)2 (A = Li, Na) as Electrodes for Li-Ion and Na-Ion Batteries Meiling Sun,†,‡ Gwenael̈ le Rousse,†,‡,§ Matthieu Saubanère,§,∥ Marie-Liesse Doublet,§,∥ Daniel Dalla Corte,† and Jean-Marie Tarascon*,†,‡,§ †
UMR 8260 “Chimie du Solide et Energie”, Collège de France, 11 Place Marcelin Berthelot, 75231 Paris Cedex 05, France Sorbonne Universités - UPMC Univ Paris 06, 4 Place Jussieu, F-75005 Paris, France § Réseau sur le Stockage Electrochimique de l’Energie (RS2E), FR CNRS 3459, 33 rue Saint Leu, 80039Amiens Cedex, France ∥ ICGM Institut Charles Gerhardt - CNRS and Université Montpellier, Place Eugène Bataillon, 34095 Montpellier, France ‡
S Supporting Information *
ABSTRACT: We herein report the synthesis, crystal structure, and electrochemical performances of a new Li-based vanadium oxysulfate phase, Li2VO(SO4)2, whose structure is built on vanadyl-containing VO5 square-based pyramids and SO4 groups linked by vertices to form layers. Li2VO(SO4)2 presents a redox activity at 4.7 V vs Li+/Li0 with a reversible capacity of 50 mA·h/g, while the chemically similar but structurally different Na2VO(SO4)2 delivers a capacity of 60 mA h/g at 4.5 V vs Na+/Na0. Density functional theory calculations are also performed to rationalize our findings in terms of redox activity and phase stability.
■
INTRODUCTION Lithium ion batteries have been recognized as attractive energy storage systems not only for portable electronics but also for powering electric vehicles (EV and HEV). They are also considered for grid applications for which cost is one of the overriding figures of merit. Cost reduction calls for new positive electrode materials with high energy densities that can be made from abundant elements via low-temperature processes. Within such a context, polyanionic compounds are quite attractive and have been the subject of intense studies over the past decade, with LiFePO4 being the “stellar” material for EV applications. Traditional polyanionic systems present benefits in terms of cost and safety, but they suffer from a relatively low energy density (∼580 W·h/kg for LiFePO4 and ∼550 W·h/kg for triplite LiFeSO4F)1,2 compared to the layered oxide materials (∼750 W· h/kg).3 This is due to the dead weight penalty associated with the heavy polyanions replacing oxygens. This can be overcome by lowering the weight of the polyanionic group (e.g., replacing PO43− by BO33− to obtain LiFeBO3).4 Another route relies on making use of multielectron transfer processes with vanadium being one of the most prone 3d metals to undergo such a process, the reason why they have recently attracted great interest. LiVPO4F displays for instance a cumulative redox capacity of 312 mA·h/g parted between two plateaus occurring at 4.25 and 1.8 V and corresponding to the V4+/V3+ and V3+/V2+ redox couples, respectively.5 Equally, LiVOPO4 presents one plateau at 3.95 V on oxidation and a cascade of three plateaus (2.4, 2.2, and 2 V) upon insertion of 1 Li+ on reduction which corresponds to the V5+/V4+ and V4+/V3+ redox couples, respectively.6−10 In parallel, Pralong and co-workers reported the synthesis and structure of © 2016 American Chemical Society
another vanadyl phosphate, Li4VO(PO4)2, exhibiting a layered structure with a reversible capacity of 60 mA·h/g at 4.1 V vs Li+/ Li0 (V5+/V4+ redox couple).11 The same group also electrochemically synthesized the Li5VO(PO4)2 phase from Li4VO(PO4)2,12 with such a lithiation being accompanied by a 2D to 3D structural phase transition. At this stage, in view of reaching higher potential via inductive effect,13 an obvious extension of the previously mentioned studies consists of substituting PO43− by a more electronegative SO42− polyanion so as to obtain Li-based vanadium oxysulfates. Previous studies focused on the V4+/V3+ couple, with for example the intercalation of Li+ in β-VOSO4 that revealed an equilibrium redox potential of 2.84 V vs Li+/Li0, a pretty low value that was ascribed to a shortened vanadyl bond (i.e., showing an enhanced covalency).14 A legitimate question regards the feasibility of synthesizing Li-based vanadium(IV) oxysulfates that can present V5+/V4+ and V4+/V3+ redox activities. Herein, we report the synthesis, structure, ionic conductivity, and electrochemical performances of a new Li2VO(SO4)2 compound which shows a redox potential of 4.7 V for the V4+/V5+ redox couple. Moreover, for the sake of completion of this study and due to the existence of a reported Na2VO(SO4)2 compound15 having a chemical formula analogous to our newly identified Li2VO(SO4)2 phase, we decided to explore the structural relation between the two phases together with the electrochemical properties of the NaReceived: July 6, 2016 Revised: August 30, 2016 Published: August 30, 2016 6637
DOI: 10.1021/acs.chemmater.6b02759 Chem. Mater. 2016, 28, 6637−6643
Article
Chemistry of Materials
cation migration were obtained by fitting the ac and dc data measured according to the Arrhenius equation σ(T) = σ0 exp(−Ea/kBT) in which σ is the conductivity at the temperature T, σ0 is a pre-exponential factor, and kB is the Boltzmann constant.
based phase so far not reported. A redox activity of 4.5 V vs Na+/ Na0 and a capacity of 60 mA·h/g was obtained.
■
EXPERIMENTAL SECTION
■
Synthesis. Li2SO4 (Alfa Aesar, Ward Hill, MA; 99.7%), Na2SO4 (Alfa Aesar, 99%), and VOSO4·xH2O (Alfa Aesar, 99.9% metal basis) were used as Li-, Na-, and V-based precursors for the targeted Li2VO(SO4)2 phase. VOSO4·xH2O was first dehydrated in argon at 260 °C to prepare α-VOSO4 as illustrated in the literature.16 Stoichiometric amounts of Li2SO4 and VOSO4 were thoroughly ball-milled, pressed into a pellet, and annealed under Ar for 12 h at 400−415 °C to produce a well-crystallized sample with sharp reflections on the XRD powder pattern reminiscent of a new phase according to the following reaction:
RESULTS AND DISCUSSION Structural Characterization. The Li2VO(SO4)2 structure was examined using X-ray and neutron powder diffraction. The finely ground greenish powder was loaded in a 0.7 mm diameter capillary and the sample was measured in transmission mode (λ = 0.41417 Å). For neutrons, the sample (from the exact same batch) was put in a cylindrical vanadium container and measured on the HRPT high-resolution neutron diffractometer with a wavelength of 1.4934 Å. The neutron and X-ray diffraction patterns obtained for Li2VO(SO4)2 were first refined using the tetragonal structural models reported for the phosphate Li4VO(PO4)2 and arsenate Li4VO(AsO4)2 compounds reported by Pralong and co-workers and Aranda et al.11,27 While these two compounds crystallize in the same tetragonal unit cell, they present different space groups: P4/n and P4/ncc, respectively. However, our trials to refine Li2VO(SO4) starting from these analogs led to unsatisfactory refinements, as shown in the Supporting Information, Figure S1. The structure of Li2VO(SO4)2 was therefore solved ab initio. First, synchrotron X-ray diffraction peaks could be indexed using the Dicvol program18 in a tetragonal unit cell, with lattice parameters a = 8.81158(4) Å and c = 8.75256(6) Å. The corresponding volume (V = 679.584(6) Å3) is suitable to accommodate four formulas per unit cell, and this unit cell is metrically similar to the one reported for the phosphate and arsenate compounds mentioned above. However, the observed (hkl) reflections indicate a body-centered unit cell. At this stage, combined neutron/X-ray powder diffraction explorations were performed in different space groups using FOX software.19 SO4 groups were treated as rigid tetrahedra with S−O distances of 1.49 Å. The best structural model was obtained in space group I4cm that we confirmed by performing a combined Rietveld refinement of the neutron and synchrotron XRD patterns with the FullProf program,17 and the result is shown in Figure 1. For the latter, the peak shapes were described using Thomson-CoxHastings profile functions; all atoms and their isotropic temperature factors were freely refined, except for vanadium which is placed at (0, 0, 0) to fix the floating z-origin. Final atomic positions are reported in Table 1. An overview of the structure of Li2VO(SO4)2 is shown in Figure 2. Vanadium atoms are coordinated by one oxygen atom at a short distance of 1.58 Å (vanadyl bond) and four equatorial oxygen atoms at 2.00 Å, so as to form a square pyramidal environment commonly observed in V4+-containing compounds.28 The resulting VO5 square-based pyramids are linked to SO4 groups through vertices (Figure 2a,b) to form layers perpendicular to [001]. Note that if longer V−O distances are considered, then the structure reveals alternate short (1.58 Å) and long (2.79 Å) V−O bonds along [001], similar to those observed in oxyphosphates, LiVOPO4 and Li4VO(PO4)2 (Figure 3). Lastly, Li atoms stay in the middle of regular square-based pyramids, sharing edges to form dimers (Figure 2c). On the other hand, the structure of Na2VO(SO4)2 was checked by Rietveld refinement of the synchrotron X-ray diffraction based on the model previously published,15 as shown in Figure 4. Na2VO(SO4)2 crystallizes in an orthorhombic structure, with space group P212121 and lattice parameters a = 6.310020(13) Å, b = 6.807419(13) Å, and c = 16.69296(3) Å.
VOSO4 + Li 2SO4 → Li 2VO(SO4 )2 Regarding the synthesis of Na2VO(SO4)2, we deviated from the reported approach that consists of the dissolution of V2O5 in molten Na2S2O7 while bubbling a SO2−N2 gas mixture through the solution,15 and rather used a solid-state reaction route:
VOSO4 + Na 2SO4 → Na 2VO(SO4 )2 The experiment was conducted similar to Li2VO(SO4)2 with only Na rather than Li sulfate as the precursor. The recovered powder displayed an X-ray diffraction pattern identical to the reported pattern for Na2VO(SO4)2 in ref 15. Structural Characterization. All compounds were checked by laboratory X-ray Powder diffraction (XRD) using a Bruker D8 Advance diffractometer equipped with a copper source (λKα1 = 1.54056 Å, λKα2 = 1.54439 Å) and a LynxEye detector. Additional Synchrotron diffraction patterns (λ = 0.41417 Å) were recorded on the 11BM beamline at Argonne National Laboratory. High-resolution neutron powder diffraction was performed on the HRPT instrument at SINQ-PSI (Villigen, Switzerland) with a wavelength of λ = 1.4934 Å. All patterns were refined using the Rietveld method as implemented in the FullProf program.17 Dicvol was used for indexation18 and FOX for structure solution.19 Electrochemical Characterization. Electrochemical performances of Li2VO(SO4)2/Li, Na2VO(SO4)2/Na, and Na2VO(SO4)2/Li cells were tested using Swagelok cells which were assembled in an argon dry glovebox and cycled with a VMP system (Biologic S.A., Claix, France) operating in galvanostatic mode. Prior to use as positive electrodes, the active materials were ball-milled with 25% in mass of carbon SP for 15 min in a SPEX-8000M mixer-miller. The positive electrode was separated from the Li/Na metal disc negative electrode by a Whatman GF/D borosilicate glass fiber sheet saturated with either 1 M LiPF6 in ethylene carbonate, propylene carbonate, and dimethyl carbonate [1:1:3 (w/w/w)] (LP100) or 1 M NaPF6 in ethylene carbonate, dimethyl carbonate, and fluoroethylene carbonate [97:97:6 (w/w/w)] as electrolyte. DFT+U Calculations. Spin-polarized DFT+U (DFT, density functional theory) calculations were performed using the VASP code (Vienna Ab Initio Simulation Package)20,21 and using the rotationally invariant Dudarev method22 and the generalized gradient approximation with the PBE functional to describe electron exchange and correlation.23 Different effective Hubbard corrections were tested for the vanadium d-electrons (Ueff = 0.0, 2.0, and 4.0 eV). A plane-wave cutoff of 600 eV was used to define the basis set, with well-converged kpoint sampling for each compound. The COOPs were computed using the Lobster program developed by Dronskowski and co-workers.24−26 Ionic Conductivity Measurement. The ac conductivity was measured for temperatures ranging from 75 to 300 °C, using a Bio-Logic MTZ-35 Impedance Analyzer in a frequency range of 35 MHz to 0.01 Hz and an excitation voltage of 100 mV. The dc conductivity values were determined at the same temperatures by applying polarization voltages of 100−500 mV. The pellets were sintered at 400 °C (⌀ 10 mm, compactness 77% for both), sputtered with gold and measured between platinum blocking electrodes under argon flow. During heating, the compounds were equilibrated at a constant temperature automatically prior to the impedance measurement. Values of activation energy Ea for 6638
DOI: 10.1021/acs.chemmater.6b02759 Chem. Mater. 2016, 28, 6637−6643
Article
Chemistry of Materials
Figure 1. Combined Rietveld refinement of the synchrotron and neutron diffraction patterns of Li2VO(SO4)2 (T = 300 K). The red crosses, black continuous line, and bottom blue line represent the observed, calculated, and difference patterns, respectively. Vertical green tick bars mark the Bragg reflections.
The structure (Figure 4, atomic positions in the Supporting Information, Table S1) consists of corner-sharing VO6 square bipyramids and SO4 tetrahedra, forming a 3D interconnected framework. VO6 square bipyramids in Na2VO(SO4)2 present vanadyl bonds (VO distance of 1.60 Å) and trans V−O bonds of 2.14 Å. Note that despite both Na2VO(SO4)2 and Li2VO(SO4)2 structures rely on corner-sharing SO4 tetrahedra and VOn (n = 5 or 6) polyhedra, the relative arrangement of atoms differ and there is not any structural relationship between them. Electrochemical Characterization. The electrochemical performance of Li2VO(SO4)2 and Na2VO(SO4)2 as electrodes vs Li or Na were tested. All cells were started on oxidation and their voltage−composition curves are shown in Figure 5. With the charging rate of C/20 (1 Li in 20 h), the potential of the Li2VO(SO4)2/Li cell rapidly increases to 4 V and then progressively reaches a plateau located at 4.7 V whose amplitude corresponds to 1.1 mol of Li. Upon discharge, only ∼0.5 out of the 1.1 Li can be reinserted on a 4.6 V plateau (Figure 5a). This leads to an overall reversible capacity of 50 mA·h/g. There is a large irreversible capacity between the first charge and discharge which is followed by a slippage of the voltage−composition
Figure 2. (a) Layered view of the structure of Li 2 VO(SO 4 ) 2 perpendicular to [100] direction. VO5 square pyramids are colored in blue, SO4 groups are green, Li are shown as yellow balls. (b) View of one layer along [001], built on VO5 square pyramids (blue) sharing vertices with SO4 groups (green). (c) Li coordination environment (LiO5 square-based pyramids sharing edges, colored in yellow).
Table 1. Structural Parameters for Li2VO(SO4)2, Deduced from the Combined Rietveld Refinement of the Neutron and Synchrotron X-ray Diffraction Patterns at 300 Ka atom
Wyckoff site
x
y
z
occupancy
B (Å2)
Li V S O1 O2 O3 O4
8c 4a 8c 8c 16d 8c 4a
0.3865(6) 0 0.30544(13) 0.3902(2) 0.36268(17) 0.3454(2) 0
0.8865(6) 0 0.80544(13) 0.8902(2) 0.32610(18) 0.8454(2) 0
0.1681(8) 0 0.5250(4) 0.6377(4) 0.5522(3) 0.3704(4) 0.8200(5)
1 1 1 1 1 1 1
0.737(78) 0.497(35) 0.617(85) 1.303(12) 0.879 (73) 1.686(59) 1.267(74)
Li2VO(SO4)2: space group I4cm, a = 8.81158(4) Å, c = 8.75256(6) Å, V = 679.584(6) Å3, Z = 4, density = 2.668 g/cm3. Reliability parameters: χ2 = 3.16; Bragg R-factor = 4.81% (neutron). χ2 = 1.43; Bragg R-factor = 4.81% (synchrotron). a
6639
DOI: 10.1021/acs.chemmater.6b02759 Chem. Mater. 2016, 28, 6637−6643
Article
Chemistry of Materials
observed in the X-ray powder pattern until the voltage reaches the plateau at 4.7 V. This suggests that the initial capacity in oxidation is mainly due to electrolyte decomposition. Once the plateau is reached, a new set of XRD peaks appear and they grow upon further delithiation at the expense of the pristine phase. The peaks of the new phase can be indexed with the same tetragonal structural model as pristine Li2VO(SO4)2 but with lattice parameters a = 8.51 Å and c = 9.10 Å, that is, corresponding to a volume of 658 Å3 (vs 679 Å3 for pristine, ΔV ∼ −3.1%), as shown in Figure 6b. From exploitation of the XRD pattern, we estimate this new phase, whose exact composition in Li cannot be determined from Coulometric titration because of electrolyte decomposition, to account for 20% of the total amount of material in the fully charged state. Therefore, at the end of charge, the powder consists of a mixture of Li2VO(SO4)2 and LiVO(SO4)2 with a ∼80%/20% ratio. On discharge, peaks from the LiVO(SO4)2 phase fully vanish to give a powder pattern analogous to that of the pristine Li2VO(SO4)2 compound, hence indicating a reversible biphasic process. Moreover, starting a cell on discharge led to the appearance of a poorly reversible plateau around 2 V (Supporting Information, Figure S2), which we did not further explore. We equally tested the electrochemical activity of Na2VO(SO4)2 electrodes vs Na using 1 M NaPF6 in EC−PC (3% FEC) as electrolyte (Figure 5b) at the rate of C/30. Similar to Li2VO(SO4)2, there is a large irreversible capacity between the first charge and discharge due to electrolyte oxidation. However, from the second cycle and outward the cell delivers a reversible capacity of 60 mA·h/g, which nearly corresponds to delithiation and lithiation of 0.7 Na+. Two plateaus located at 4.5 and 3.8 V can be observed on the derivative dx/dV curve for the second cycle (Figure 5b). From in situ X-ray diffraction we could deduce that such plateaus correspond to reversible phase transformations (Supporting Information, Figure S3) that we could not exploit further because of strong peak broadening. Lastly, although these two compounds are structurally different, it was tempting to check whether we could prepare Li2VO(SO4)2 from Na2VO(SO4)2 via ion exchange. All the attempts we have tried enlisting either solution (LiCl in acetonitrile) or molten salt (LiNO3) with temperatures ranging from 80 to 250 °C have failed, leading mainly to phase decomposition. To further pursue in this direction, we explored the electrochemical behavior (Figure 5c) of a Na2VO(SO4)2 electrode vs Li+/Li0 (e.g., metallic Li) using LP100 electrolyte. Please note the feasibility to remove nearly 1 Na+ during the first oxidation like for Na2VO(SO4)2/Na cells. This removal occurs at a slightly lower potential (4.5 instead of 4.7 V) as expected due to the lower reducing potential (∼300 mV) of Na compared to that of Li. The discharge curve is, in contrast, quite different between the two cells with a cascade voltage profile for Na2VO(SO4)2/Na as compared to a smooth voltage decrease for Na2VO(SO4)2/Li. Such a different profile is indicative of the reinsertion of Li+ rather than Na+ in the “Na1.3VO(SO4)2” phase. The Na content of ∼1.3 was deduced by EDX analysis, bearing in mind once again the difficulty to exploit coulometric titration. Upon subsequent cycles, the electrode can reversibly uptake ∼0.7 (Li+/Na+) at an average voltage of 4.65 V, leading to a reversible capacity of ∼80 mA·h/g, which progressively decays upon cycling. X-ray diffraction data of Na1.3Li0.5VO(SO4)2 nearly resemble that of the mother Na2VO(SO4)2 phase with however contracted lattice parameters, indicating that the Na2VO(SO4)2 structural framework is preserved through the cycling process (Supporting Information, Figure S4). Such a finding indicates the feasibility to
Figure 3. Comparison of alternate short and long V−O bonds (units: Å) among different vanadyl oxysulfate and oxyphosphate: Li2VO(SO4)2, Na2VO(SO4)2, Li4VO(PO4)2 and three polymorphs of LiVOPO4 (space group for polymorph α, β, and α1: P1̅, Pnma, and P4/nmm).
Figure 4. Rietveld refinement of the synchrotron powder diffraction pattern of Na2VO(SO4)2 (T = 300 K). The red crosses, black continuous line, and bottom blue line represent the observed, calculated, and difference patterns, respectively. Vertical green tick bars mark the Bragg reflections. The structure is shown in the lower part; VO6 square bipyramids are colored in blue, sulfate groups are green, and yellow balls are Na atoms.
curves toward lower x in LixVO(SO4)2, suggesting copious electrolyte decomposition together with the formation of an SEI (Solid Electrolyte Interface) layer. Its growth seems to persist until 20 cycles, beyond which the charge and discharge capacities nearly superimpose. The Li extraction−insertion mechanism was further examined by in situ X-ray diffraction using a homemade stainless steel cell with an X-ray transparent beryllium window (Figure 6a). During cell charging, no noticeable changes are 6640
DOI: 10.1021/acs.chemmater.6b02759 Chem. Mater. 2016, 28, 6637−6643
Article
Chemistry of Materials
Figure 5. Voltage−composition curve for (a) Li2VO(SO4)2/Li cell, (b) Na2VO(SO4)2/Na cell, and (c) Na2VO(SO4)2/Li cell. The capacity retention is shown as the inset and derivative curve of the second cycle is shown underneath for each cell, respectively.
leading to phase segregation into 1/2Li2VO(SO4)2 + 1/2Na2VO(SO4)2 and explaining why this phase could not be observed experimentally. Electrochemically driven insertion reactions, relying on topotactic mechanisms, are a nice way to overcome such limitations and to stabilize metastable phases, as clearly indicated herein by our feasibility to prepare the Na1.3Li0.5VO(SO4)2 phase. Ionic Conductivity Measurement. Owing to the different crystal structures and electrochemical activities, we decided to measure the transport properties for Li2VO(SO4)2 and Na2VO(SO4)2. Ac impedance spectroscopies were performed to measure their conductivities. Activation energies (Figure 8) of 0.75 and 0.65 eV were obtained for the Na2VO(SO4)2 and Li2VO(SO4)2 phases, and extrapolated ac conductivities at room temperature are 2.6 × 10−10 and 4.2 × 10−10 S cm−1 for the Na and Li counterparts, respectively. For both samples, the lowfrequency tail of the impedance spectra (Figure 8, inset) suggests an ionic component to the overall conductivity. To shed some light on this issue, dc conductivity measurements were done. Activation energies of 0.33 and 0.19 eV were found for the Liand Na-based oxysulfates, respectively, with room temperature dc values nearly equal to or slightly greater than the ac ones. This indicates a predominant contribution of the electronic conductivity to the transport properties of these phases at room temperature as previously observed for other polyanionic compounds. Bond valence energy landscapes (BVEL) calculations30 conducted on these phases (Supporting Information, Figure S6) to evaluate how Li/Na ions would diffuse in these two different structures lead to nearly equal “activation energy values” (defined as the minimum energy to obtain at least one infinitely connected path in a crystallographic direction) for Na in Na2VO(SO4)2 and Li in Li2VO(SO4)2. Overall, although quite different from a structural point of view, these two compounds show transport properties quite similar to each other, which is consistent with the obtained electrochemical data.
partially substitute minute amounts of Na for Li in Na2VO(SO4)2 while preserving the same structure. DFT+U Calculations. Whatever the Li or Na phases explored herein, the study above reveals a redox activity at potentials greater than ∼4.5 V that are particularly high for the V5+/V4+ redox couple. To grasp some light on the origin of this, DFT calculations were undertaken. For both Li2VO(SO4)2 and Na2VO(SO4)2 the density of states combined with a crystal orbital overlap population (COOP) analysis show that the bonding electronic states of the vanadyl bond are far below the Fermi level and therefore are transparent to the oxidation process (Figure 7). In more detail, the f- Fukui functions (Figure 7d,e) computed for Li2VO(SO4)2 and Na2VO(SO4)2 confirm that the electronic states involved in the oxidation process arise from a metallic orbital perpendicular to the VO bond with mainly oxygen contributions of the SO4 groups, hence stressing the efficacy of the inductive effect associated with the SO42− polyanion. Moreover, the f+ Fukui functions (Figure 7d) show that the reduction process from pristine Li2VO(SO4)2 should now involve the VO bonds as the main redox center, like the βVOSO4 phase14 for which the strongly covalent VO bond governs the V4+/V3+ redox center (2.8 V). This Li insertion occurs at a much lower reduction potential of ∼2 V vs Li+/Li0 for the V4+/V3+ couple (Supporting Information, Figure S2). At this stage, we should recall that the oxidation potential of Li2VO(SO4)2 is about 0.6 V greater than that for the Li4VO(PO4)2 homologue. This voltage difference is due to a large part to the inductive effect associated with the replacement of PO43− by SO42− and to a lesser extent to the lower electrostatics of Li/Na cationic sites as a result of the higher Li/Na stoichiometry in (PO4)3−-based systems.29 Regarding the possibility of partially exchanging sodium by lithium in the Na2VO(SO4)2 polymorph, DFT calculations were also performed to investigate the stability of the NaLiVO(SO4)2 composition with respect to a proportional mixture of Li2VO(SO4)2 and Na2VO(SO4)2. All configurations tested (Supporting Information, Figure S5) reveal that this “mixed phase” in Li/Na is thermodynamically less stable than the end members, hence 6641
DOI: 10.1021/acs.chemmater.6b02759 Chem. Mater. 2016, 28, 6637−6643
Article
Chemistry of Materials
Figure 7. (top) Atom-projected density of states (DOS) and VO and V−OSO3 crystal orbital overlap population (COOP) computed for (a) Li2VO(SO4)2 and (b) Na2VO(SO4)2 phases with DFT+U (Ueff = 4 eV for V). (bottom: c) t2g-orbital splitting expected for the octahedral to pyramidal distortion showing the shape of the electronic levels involved in the oxidation and reduction processes of the A2VO(SO4)2 phases (A = Li, Na). The f- and f+ Fukui functions (d, e) computed for these two processes confirm the expected orbital shape and indicate the redox centers in the two processes, that is, the V−OSO3 ionic bonds in oxidation and the VO strongly covalent bond in reduction.
Figure 6. Behavior of a Li2VO(SO4)2/Li cell on charge. (a) In situ XRD patterns recorded behind a Be window while the cell is charged to 4.9 V (voltage−time curve on the right). The red pattern corresponds to the pristine phase, the green one refers to the fully oxidized sample, and the blue pattern on top was recorded at the subsequent discharge, indicating the reversibility of this process. (b) Refinement of the fully charged sample (4.9 V) with two phases: Li2VO(SO4)2 (red vertical tick marks) and Li1VO(SO4)2 (blue vertical tick marks) in an ∼80/20 ratio.
■
CONCLUSIONS In this paper, we reported the synthesis, structure, and electrochemical activity toward Li of a novel phase, Li2VO(SO4)2. This new compound displays a capacity of 50 mA·h/g at 4.7 V that is associated with the V4+/V5+ redox couple. Interestingly, while Li2VO(SO4)2 is structurally related to Li4VO(PO4)2, it shows a slightly lower capacity but a higher voltage, leading to similar energy density (235 vs 244 W·h/kg). DFT calculations suggest that the electronic states involved in the oxidation process are mainly due to a metallic orbital with oxygen contributions of the SO4 groups. This is in contrast to our recently reported Cu-based oxysulfate, Li2Cu2O(SO4)2, which also displays a plateau at 4.7 V vs Li+/Li0.31 In that case we demonstrated that the band involved in the oxidation process is mainly attributed to O(2p) orbitals of the oxygen atoms bridging the CuO4 square planes. Lastly, we note that, upon reduction, Li2VO(SO4)2 can uptake Li+ at a potential of ∼2 V (Supporting Information, Figure S2) through a redox process that enlists the VO bonds as the main redox center, as suggested by DFT. Such low potential was not further explored herein because of poor reversibility and mediocre kinetics. Obvious direct extensions of this work range from finding ways to extract
Figure 8. Transport properties of Li2VO(SO4)2 (red circles) and Na2VO(SO4)2 (blue squares), the filled and open circles/squares refer to ac and dc measurements, respectively; the inset shows impedance spectra (filled squares and circles) and the fit of each spectra (continuous line) of Li2VO(SO4)2 (red) and Na2VO(SO4)2 (blue) in argon at 200 °C.
greater amounts of Li+ or Na+ at high voltage for enhancing their practical energy densities, to means of mastering the two electrons V5+/V3+ process offered by this compound. This calls for innovative electrode wiring together with the development of electrolytes highly stable against oxidation.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.chemmater.6b02759. 6642
DOI: 10.1021/acs.chemmater.6b02759 Chem. Mater. 2016, 28, 6637−6643
Article
Chemistry of Materials
■
(14) Gaubicher, J.; Chabre, Y.; Angenault, J.; Lautié, A.; Quarton, M. Lithium electrochemical intercalation in β-VOSO4. J. Alloys Compd. 1997, 262−263, 34−38. (15) Fehrmann, R.; Boghosian, S.; Papatheodorou, G. N.; Nielsen, K.; Berg, R. W.; Bjerrum, N. J. Crystal structure and vibrational spectra of disodium oxo(disulfato)vanadate. Inorg. Chem. 1990, 29, 3294−3298. (16) Longo, J. M.; Arnott, R. J. Structure and magnetic properties of VOSO4. J. Solid State Chem. 1970, 1, 394−398. (17) Rodriguez-Carvajal, J. Recent advances in magnetic-structure determination by neutron powder diffraction. Phys. B 1993, 192, 55−69. (18) Boultif, A.; Louer, D. Indexing of powder diffraction patterns for low-symmetry lattices by the successive dichotomy method. J. Appl. Crystallogr. 1991, 24, 987−993. (19) Favre-Nicolin, V.; Cerny, R. FOX, free objects for crystallog̀ raphy’: a modular approach to ab initio structure determination from powder diffraction. J. Appl. Crystallogr. 2002, 35, 734−743. (20) Kresse, G.; Hafner, J. Ab initio molecular dynamics for liquid metals. Phys. Rev. B: Condens. Matter Mater. Phys. 1993, 47, 558−561. (21) Kresse, G.; Furthmüller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 1996, 6, 15−50. (22) Dudarev, S. L.; Botton, G. A.; Savrasov, S. Y.; Humphreys, C. J.; Sutton, A. P. Electron-energy-loss spectra and the structural stability of nickel oxide: An LSDA+U study. Phys. Rev. B: Condens. Matter Mater. Phys. 1998, 57, 1505−1509. (23) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (24) Maintz, S.; Deringer, V. L.; Tchougréeff, A. L.; Dronskowski, R. Analytic projection from plane-wave and PAW wavefunctions and application to chemical-bonding analysis in solids. J. Comput. Chem. 2013, 34, 2557−2567. (25) Dronskowski, R.; Bloechl, P. E. Crystal orbital Hamilton populations (COHP): energy-resolved visualization of chemical bonding in solids based on density-functional calculations. J. Phys. Chem. 1993, 97, 8617−8624. (26) Deringer, V. L.; Tchougréeff, A. L.; Dronskowski, R. Crystal orbital hamilton population (COHP) analysis as projected from planewave basis sets. J. Phys. Chem. A 2011, 115, 5461−5466. (27) Aranda, M. A. G.; Attfield, J. P.; Bruque, S.; Martinez-Lara, M. Order and disorder of vanadyl chains: crystal structures of vanadyl dihydrogen arsenate (VO(H2AsO4)2) and the lithium derivative Li4VO(AsO4)2. Inorg. Chem. 1992, 31, 1045−1049. (28) Schindler, M.; Hawthorne, F. C.; Baur, W. H. Crystal chemical aspects of vanadium: polyhedral geometries, characteristic bond valences, and polymerization of (VOn) polyhedra. Chem. Mater. 2000, 12, 1248−1259. (29) Saubanère, M.; Yahia, M. B.; Lebègue, S.; Doublet, M. L. An intuitive and efficient method for cell voltage prediction of lithium and sodium-ion batteries. Nat. Commun. 2014, 5, 5559. (30) Adams, S. From bond valence maps to energy landscapes for mobile ions in ion-conducting solids. Solid State Ionics 2006, 177, 1625− 1630. (31) Sun, M.; Rousse, G.; Abakumov, A. M.; Saubanere, M.; Doublet, M.-L.; Rodríguez-Carvajal, J.; Van Tendeloo, G.; Tarascon, J.-M. Li2Cu2O(SO4)2: a possible electrode for sustainable Li-based batteries showing a 4.7 V redox activity Li+/Li0. Chem. Mater. 2015, 27, 3077− 3087.
Additional XRD pattern, electrochemical curves, and DFT calculations (PDF) CIF file of Li2VO(SO4)2 (CIF)
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank M. Courty for TGA experiments. We thank V. Pumjakushin for help in neutron experiments and SINQ for allocated beamtime. Use of the 11-BM mail service of the APS at Argonne National Laboratory was supported by the U.S. Department of Energy under Contract No. DE-AC0206CH11357 and is greatly acknowledged. M.S. and M.-L.D acknowledge HPC resources from GENCI-CCRT/CINES (Grant cmm6691).
■
REFERENCES
(1) Kim, M.; Jung, Y.; Kang, B. High electrochemical performance of 3.9 V LiFeSO4F directly synthesized by a scalable solid- state reaction within 1 h. J. Mater. Chem. A 2015, 3, 7583−7590. (2) Wang, Y.; Wang, Y.; Hosono, E.; Wang, K.; Zhou, H. The design of a LiFePO4/Carbon nanocomposite with a core−shell structure and its synthesis by an in situ polymerization restriction method. Angew. Chem., Int. Ed. 2008, 47, 7461−7465. (3) Koga, H.; Croguennec, L.; Ménétrier, M.; Douhil, K.; Belin, S.; Bourgeois, L.; Suard, E.; Weill, F.; Delmas, C. Reversible oxygen participation to the redox processes revealed for Li1.20Mn0.54Co0.13Ni0.13O2. J. Electrochem. Soc. 2013, 160, A786−A792. (4) Yamada, A.; Iwane, N.; Harada, Y.; Nishimura, S.-i.; Koyama, Y.; Tanaka, I. Lithium iron borates as high-capacity battery electrodes. Adv. Mater. 2010, 22, 3583−3587. (5) AtebaMba, J.-M.; Masquelier, C.; Suard, E.; Croguennec, L. Synthesis and crystallographic study of homeotypic LiVPO4F and LiVPO4O. Chem. Mater. 2012, 24, 1223−1234. (6) Barker, J.; Saidi, M. Y.; Swoyer, J. L. Electrochemical properties of beta-LiVOPO4 prepared by carbothermal reduction. J. Electrochem. Soc. 2004, 151, A796−A800. (7) Bianchini, M.; Ateba-Mba, J. M.; Dagault, P.; Bogdan, E.; Carlier, D.; Suard, E.; Masquelier, C.; Croguennec, L. Multiple phases in the εVPO4O-LiVPO4O-Li2VPO4O system: a combined solid state electrochemistry and diffraction structural study. J. Mater. Chem. A 2014, 2, 10182−10192. (8) Gaubicher, J.; Le Mercier, T.; Chabre, Y.; Angenault, J.; Quarton, M. Li/β-VOPO4: A new 4 V system for lithium batteries. J. Electrochem. Soc. 1999, 146, 4375−4379. (9) He, G.; Bridges, C. A.; Manthiram, A. Crystal chemistry of electrochemically and chemically lithiated layered αI-LiVOPO4. Chem. Mater. 2015, 27, 6699−6707. (10) Saravanan, K.; Lee, H. S.; Kuezma, M.; Vittal, J. J.; Balaya, P. Hollow α-LiVOPO4 sphere cathodes for high energy Li-ion battery application. J. Mater. Chem. 2011, 21, 10042−10050. (11) Kishore, M. S.; Pralong, V.; Caignaert, V.; Varadaraju, U. V.; Raveau, B. Synthesis and electrochemical properties of a new vanadyl phosphate: Li4VO(PO4)2. Electrochem. Commun. 2006, 8, 1558−1562. (12) Satya Kishore, M.; Pralong, V.; Caignaert, V.; Malo, S.; Hebert, S.; Varadaraju, U. V.; Raveau, B. Topotactic insertion of lithium in the layered structure Li4VO(PO4)2: The tunnel structure Li5VO(PO4)2. J. Solid State Chem. 2008, 181, 976−982. (13) Padhi, A. K.; Manivannan, V.; Goodenough, J. B. Tuning the Position of the Redox Couples in Materials with NASICON Structure by Anionic Substitution. J. Electrochem. Soc. 1998, 145, 1518−1520.
■
NOTE ADDED AFTER ASAP PUBLICATION This paper was published ASAP on September 14, 2016, with an error in the Electrochemical Characterization Section. The corrected version reposted on September 15, 2016.
6643
DOI: 10.1021/acs.chemmater.6b02759 Chem. Mater. 2016, 28, 6637−6643