Ionothermal Synthesis of High-Voltage Alluaudite ... - ACS Publications

Research Article www.acsami.org

Ionothermal Synthesis of High-Voltage Alluaudite Na2+2xFe2‑x(SO4)3 Sodium Insertion Compound: Structural, Electronic, and Magnetic Insights Debasmita Dwibedi,† Chris D. Ling,‡ Rafael B. Araujo,§ Sudip Chakraborty,§ Shanmughasundaram Duraisamy,∥ Nookala Munichandraiah,∥ Rajeev Ahuja,§ and Prabeer Barpanda*,† †

Faraday Materials Laboratory, Materials Research Center, Indian Institute of Science, C.V. Raman Avenue, Bangalore 560012, India School of Chemistry, The University of Sydney, Building F11, Sydney, NSW 2006, Australia § Condensed Matter Theory Group, Department of Physics and Astronomy, Uppsala University, Box 516, 75120 Uppsala, Sweden ∥ Inorganic and Physical Chemistry, Indian Institute of Science, C.V. Raman Avenue, Bangalore, 560012, India ‡

S Supporting Information *

ABSTRACT: Exploring future cathode materials for sodium-ion batteries, alluaudite class of Na2FeII2(SO4)3 has been recently unveiled as a 3.8 V positive insertion candidate (Barpanda et al. Nat. Commun. 2014, 5, 4358). It forms an Fe-based polyanionic compound delivering the highest Fe-redox potential along with excellent rate kinetics and reversibility. However, like all known SO4-based insertion materials, its synthesis is cumbersome that warrants careful processing avoiding any aqueous exposure. Here, an alternate low temperature ionothermal synthesis has been described to produce the alluaudite Na2+2xFeII2‑x(SO4)3. It marks the first demonstration of solvothermal synthesis of alluaudite Na2+2xMII2‑x(SO4)3 (M = 3d metals) family of cathodes. Unlike classical solidstate route, this solvothermal route favors sustainable synthesis of homogeneous nanostructured alluaudite products at only 300 °C, the lowest temperature value until date. The current work reports the synthetic aspects of pristine and modified ionothermal synthesis of Na2+2xFeII2‑x(SO4)3 having tunable size (300 nm ∼5 μm) and morphology. It shows antiferromagnetic ordering below 12 K. A reversible capacity in excess of 80 mAh/g was obtained with good rate kinetics and cycling stability over 50 cycles. Using a synergistic approach combining experimental and ab initio DFT analysis, the structural, magnetic, electronic, and electrochemical properties and the structural limitation to extract full capacity have been described. KEYWORDS: sodium-ion battery, alluaudite, Na2Fe2(SO4)3, ionothermal synthesis, DFT

1. INTRODUCTION The 21st century world has witnessed an exponential surge in myriads of portable consumer electronics, (plug-in) hybrid electric vehicles, and increasing usage of small-to-large scale of power grid-storage. The growing use of these wireless devices/ automobiles and their wide-scale dissemination across the economic strata of global population has given rise to an energy-demanding society.1−3 “Energy triangle”, i.e., generation, storage, and delivery, has become a key issue, where the electrochemical power sources in general and secondary batteries in particular have become indispensable part of modern life-style. Since 1990s, the successful commercialization of Li-ion batteries led to a multibillion dollar market.4 Lithiumion batteries stand unbeaten for applications constrained by volume/weight restriction, i.e., demanding the highest volumetric/gravimetric energy density. But, it has led to concern over the paucity of Li-rich resources, high cost, and sustainability of lithium-ion based technologies.5 While Li-ion batteries are indispensable for energy-dense applications, © 2016 American Chemical Society

sodium chemistry offers avenues to develop economic and sustainable batteries for applications not restricted by weight/ volume such as power grid storage and back-up batteries connected to renewable energy generators (e.g., solar cell, wind mills, etc.). The abundance, uniform geographic distribution, and low cost of sodium resources have generated renewed interest in developing sodium-ion batteries parallel to Li-ion technologies to cater to low-cost large-scale applications.6−8 In this scenario, various oxide ceramics have been receiving newfound interest in recent years. Though LixCoO2 and NaxCoO2 were discovered around 1980s, while LixCoO2 went into commercialization, the NaxCoO2 system took a back seat. After an almost two decade long hibernation, electrochemists have recently put unprecedented focus on suites of oxide cathodes (such as NaxCoO2, NaxFeO2, Nax(Fe1/2Mn1/2)O2, Received: November 22, 2015 Accepted: March 2, 2016 Published: March 2, 2016 6982

DOI: 10.1021/acsami.5b11302 ACS Appl. Mater. Interfaces 2016, 8, 6982−6991

Research Article

ACS Applied Materials & Interfaces NaxVO2, etc.) for Na-ion technology.6,9 These layered compounds, adopting broadly O3- or P2-type frameworks, offer scalable synthesis, decent chemical/thermal stability, and reversible capacity of 100−200 mAh/g. Parallel to the oxide chemistry, suites of polyanion insertion materials have been discovered with credible electrochemical performance e.g. NaxFePO4, NaxMSO4F, Na 2‑xFePO4F, Na2‑xMP2O7, and Na4M3(PO4)2(P2O7) [M = 3d metals].10−14 Unlike the oxides, these polyanionic cathodes inherently suffer from lower theoretical capacity stemming from their higher molecular weight, thus decreasing the net energy density. While their capacity is moderate, polyanionic systems offer rich structural diversity to design and tune the redox potential. An ideal polyanionic candidate should offer very high voltage (approaching 4 V) to deliver energy density even with moderate capacity. In this spirit, Yamada group have recently discovered a novel alluaudite Na2Fe2(SO4)3 cathode with an unprecedented 3.8 V (Fe3+/Fe2+) redox activity leading to comparable energy density with existing oxide cathodes.15,16 This earth-abundant and economic Na−Fe−S-O quaternary cathode offers unique combination of high voltage operation (approaching 4 V range of Li-ion batteries), very fast kinetics, and long cycle life. Nevertheless, similar to all known SO4-based cathodes,17,18 it poses two major challenges for materials synthesis: it suffers from dissolution in aqueous media (water) and it decomposes above 450 °C to form various decomposition products like Na2SO4, α-/β- FeSO4, Fe2O3/Fe3O4, Na6Fe(SO4)4, and SOx gases. Thus, it requires low-temperature nonaqueous synthesis. It makes the synthesis options limited and cumbersome. As an alternate scalable synthesis of this alluaudite-framework cathode, we have employed ionothermal synthesis using roomtemperature ionic liquid as reacting media. Using pristine and diol/H2O mixed ionic liquids as media as well as template, tunability in the size and morphology of final product “Na2.4Fe1.8(SO4)3” has been realized. It is the first example of solvothermal synthesis of SO4-based alluaudite cathodes that can deliver the target compound at 300 °C, the lowest reported temperature, with homogeneous nanoscale morphology. The magnetic and electronic properties of this novel alluaudite Febased cathode have been investigated for the first time combining both experimental and first-principle DFT analyses. While paramagnetic in nature at ambient temperature, it undergoes a long-range antiferromagnetic ordering below 12 K. The as-synthesized alluaudite cathode delivers reversible capacity approaching 80% of theoretical capacity with characteristic high Fe3+/Fe2+ redox potential. The origin of high-voltage (3.8 V) redox activity and structural limitation in extracting all constituent sodium species so as to get full theoretical capacity has been described using structure− property correlation. It delivers insights to understand the structural limitation, to optimize the Fe-based alluaudite cathode and to design new members of alluaudite family of cathode materials.

4, 1, 0) precursors are commercially available. During solid-state/ solvothermal synthesis, FeSO4·nH2O undergoes progressive dehydration followed by reaction with other precursors to form the desired product phase. When FeSO4·7H2O or FeSO4·4H2O was used with Na2SO4, the FeSO4·nH2O precursor solely underwent dehydration leading to a mixture of Na2SO4 and FeSO4. However, when FeSO4· H2O monohydrate precursor was used, it simultaneous released a single H2O molecule and reacted with Na2SO4 to form the final Na2.4Fe1.8(SO4)3 alluaudite phase. Although various FeSO4·nH2O (n = 7, 4, 1, 0) precursors can be used for ionothermal synthesis, successful product can be obtained only in case of FeSO4·nH2O (n = 1). A 3:4 molar mixture of Na2SO4 (Merck, 99.5%) and FeSO4.H2O was handmixed for 5 min before being transferred to a Teflon-lined steel autoclave containing 5 cm3 of EMI-TFSI ionic liquid. The mixture was stirred for 10 min to form a uniform suspension and was kept undisturbed for 5 min. This enclosed Teflon-lined steel reactor was placed inside a muffle furnace and was heated at 300 °C (heating rate = 2−5 °C/min) for 20−24 h in air atmosphere. Post cooling to ambient condition; the reaction product was recuperated from ionic liquid by centrifugation using high purity acetone as washing medium. The mixture of ionic liquid and acetone can be used to recuperate ionic liquid for further use. The recovered product was dried in hot air oven at 60 °C to obtain the product in greyish powder form. In addition, modified ionothermal synthesis was conducted by using various mixtures of EMI-TFSI and ethane-1,2-diol (ethylene glycol EG)/H2O as reacting media following the same annealing protocol. The overall ionothermal synthesis can be expressed as 3Na 2SO4 + 4FeSO4 · H 2O → 2“Na 2.4Fe1.8(SO4 )3 ” Alternately, Na2+2xFe2‑x(SO4)3 was made by conventional solid-state route15,16 using a 3:4 molar mixture of Na2SO4 and homemade FeSO4· H2O. An intimate precursor mixture was prepared by planetary ball milling for 6 h (at 350 rpm) using a Cr-hardened stainless steel (CrSS) milling media and container. The precursor mixture was pressed into pellets and was annealed at 350 °C for 12 h in an Ar-filled tubular furnace to obtain the final product. 2.2. Structural and Physical Characterization. Powder X-ray diffraction patterns were acquired with a PANalytical X’Pert Pro diffractometer having a Cu Kα source (λ1 = 1.5405 Å, λ2 = 1.5443 Å) operating at 40 kV/30 mA. Powder patterns were collected in the 2θ range of 5−90° (step size of 0.026°.s−1). Rietveld refinement19 of the XRD pattern was conducted using the GSAS program20 (with the EXPGUI front-end) and the crystal structure was illustrated with the VESTA program.21 The morphology of ionothermally obtained “Na2.4Fe1.8(SO4)3” powder (sprinkled on carbon tape) was observed with an FEI Inspect F50 scanning electron microscope operating at 5−20 kV. Transmission electron microcopy was performed with an FEI Tecnai F30 STwin unit (operating at 200 kV) by adding few drops of powder sample soaked in acetone on a copper grid. The elemental ratio was probed with EDX analysis as well as by inductively coupled plasma atomic emission spectroscopy (ICP) with a PerkinElmer Optima 8300 spectrometer. FT-Infrared spectrum of powder sample (in KBr pellet) was acquired by an Agilent Cary 660-FTIR spectrometer in the wavenumber range of 400−4000 cm−1 (number of cycles = 4). The thermal analysis (TG-DSC) was conducted using Mettler STA equipment or TA Q-50 unit in the temperature range of 25−400 °C (with Ar flow). 2.3. Magnetic Susceptibility Analysis. The magnetic susceptibility measurements were conducted with a Quantum Design physical property measurement system (PPMS) equipped with a vibrating sample magnetometer (VSM) option. The magnetization was recorded in zero field cooled (ZFC) and field cooled (FC) modes with varied applied external magnetic field (H) in the temperature range 2−300 K. Magnetization as a function of field was measured at 2 K over 6 segments to ±10000 Oe. 2.4. Computational Methodology. Here, we have carried out density functional theory (DFT) based ab initio calculations by employing the Projector Augmented Wave (PAW) method as implemented in the Vienna ab initio simulation package

2. EXPERIMENTAL SECTION 2.1. Material Synthesis. Na2+2xFe2‑x(SO4)3 phase was prepared by ionothermal synthesis using 1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl) imide (EMI-TFSI, Reinste Nanoventure Inc.) ionic liquid as the reacting media. Unlike solid-state synthesis warranting anhydrous FeSO4, we used FeSO4·H2O monohydrate precursor that was in-house prepared by annealing commercial FeSO4· 7H2O (Sigma-Aldrich, 99%) at 150 °C for 2 h in a tubular furnace under steady Ar-flow. As FeSO4 forms thermodynamically stable phases with different degree of hydration, various FeSO4·nH2O (n = 7, 6983


Research Article

ACS Applied Materials & Interfaces

alter the final size and morphology of product materials. This soft chemistry route can deliver (i) particle with preferential orientation, (ii) controlled nucleation and grain growth, and (iii) stabilization of some intermediate metastable polymorphs. Finally, although ionic liquids are relatively expensive, it can be recuperated and reused several time to offer low cost synthesis. These features make ionothermal synthesis versatile in material synthesis. In the present work, as a proof-of-concept study, 1-ethyl-3methylimidazolium bis(trifluoromethanesulfonyl) imide (EMITFSI) ionic liquid was employed. Though myriads of ionic liquids are known, EMI-TFSI was chosen as the reacting media owing to its ready availability and thermal stability up to 320 °C, which is essential for the synthesis protocol at 300 °C. Guided by previous work on fluorosulfate chemistry,11,18,34,35 FeSO4.H2O monohydrate precursor was chosen owing to (i) its easy reactivity with other precursors, (ii) in situ formation of anhydrous FeSO4 upon structural H2O removal at high temperature, and (iii) controlled rate of reaction due to slower removal of structural water embedded in hydrophobic ionic liquid media. Ionothermal synthesis at 300 °C led to the formation of desired alluaudite phase in 24 h as verified by the XRD data (Figure 1). As the precursor solubility and degree of

(VASP).22,23 The calculations have been undertaken by using the spinpolarized formalism in the Perdew, Burke, and Ernzerhof (PBE) parametrization for the exchange and correlation functional.24 We have employed GGA + U approach of Dudarev et al.25 to partially overcome the self-interaction error of the standard DFT and also to take care the strong correlation character of the 3d electrons in the transition metal ions. At this level of theory, the U and J parameters are not considered separately. In fact, an effective Ueff = U − J is the quantity to be considered and now on we will refer U as the Ueff. The U parameter for Fe assumes values of 4.0, taken from ref 26. An energy cutoff of 600 eV was employed together with a gamma-centered kpoint of 4 × 4 × 6. The magnetic configuration of Na2Fe2(SO4)3 are also examined and the antiferromagnetic state becomes the more stable configuration. Net average charge was also computed using the Bader analysis method.27,28 2.5. Electrochemical Characterization. An intimate slurry of 80 wt % active material “Na2.4Fe1.8(SO4)3”, 10 wt % acetylene carbon black and 10 wt % of polyvinylidene fluoride (PVdF) binder was prepared in minimal amount of NMP (N-methyl pyrrolidone) solvent. This slurry was painted on Al current collector disks (⌀ = 10 mm). These painted electrodes were dried at 120 °C in vacuum to get rid of NMP solvent. The final cathode loading was around 5 mg.cm−2. For galvanostatic charge−discharge measurement, swagelok-type half-cells were assembled with the painted electrode disks as working (+ve) electrode, Na metal foil as counter (-ve) electrode and two sheets of glass fiber acting as separators soaked with 1 M NaClO4 dissolved in propylene carbonate (PC) acting as electrolyte. The swageloks were assembled inside an Ar-filled Mbraun GmbH glovebox to avoid any moisture contamination and were subjected to galvanostatic cycling (at the rate of C/20) with a Bio-Logic battery tester in the voltage range of 2.0−4.5 V (at 25 °C).

3. RESULTS AND DISCUSSION Ionothermal Synthesis and Structure of “Na2.4Fe1.8(SO4)3”. PO43−-based compounds have long dominated polyanionic insertion materials, as they provide robust thermal/chemical stability in addition to good electrochemical properties. While using more electronegative SO42− increases the ionicity between M−O bonding thereby enhancing the redox potential (inductive effect29), the presence of ionic bonding imparts lower lattice energy (or enthalpy of formation, ΔHf). It makes SO4-based cathodes prone to thermal decomposition and moisture sensitive, thus excluding conventional high-temperature ceramic route and/or aqueous syntheses. Originally, alluaudite sulfates have been discovered by low-temperature (T ≈ 350 °C) solid-state route.15,30 In search for alternative sustainable synthesis, ionothermal route involving moderate heat treatment can be employed to synthesize variety of cathodes with tunable particle size and morphology.18,31−33 With flexible combination of cations and anions, suites of ionic liquids can be designed to act as reacting media with desirable solvating nature, minimal volatility and chemical/ thermal stability up to 350 °C. Hence, it can be employed for nonaqueous solvothermal synthesis of moisture sensitive inorganic materials without needing any high-temperature calcination step. The hydrophobic ionic liquids can prevent oxidation of 3d-transition metal species so it does not need steady flow of expensive inert gas (like Ar) during synthesis. In addition, ionic liquid can form a thin protective coating on electrode product phase to avoid any moisture attack during storage/handling. Unlike the hydrothermal method, ionothermal route involves zero vapor pressure ensuring safe synthesis protocol. Furthermore, the polarity, solubility, viscosity, hydrophobicity and melting point of ionic liquid can be modified by addition of aqueous/polymeric additives that can

Figure 1. Rietveld refinement fitting of XRD diffraction pattern of ionothermally synthesized “Na2.4Fe1.8(SO4)3” alluaudite phase showing the experimental data points (red), calculated pattern (black), their difference (blue) and Bragg diffraction positions (black tick marks). (Inset) Structural illustration of alluaudite structure possessing large cavities with Na atoms (blue) occupancy having FeO6 octahedra (green) and SO4 tetrahedra (yellow).

nucleation/grain growth is moderate in ionic liquid media, it warrants longer annealing duration (ca. 24 h) to complete the reaction. Further annealing leads to progressive grain growth. The viscosity of ionic liquid media can reduce drastically over 200 °C facilitating precursor solubility and product nucleation. Hence, it can realize the low-temperature synthesis of alluaudite phase at only 300 °C, benchmarking the lowest synthesis temperature reported so far. The ionothermally obtained “Na2.4Fe1.8(SO4)3” assumes a monoclinic framework (s.g. C2/ c, #15), built from edge-sharing FeO6−FeO6 [or Fe2O10] dimers, abridged by corner-sharing SO4 tetrahedra to form a three-dimensional framework. The constituent Na atoms occupy three distinct crystallographic sites (Figure 1, inset). Further, the nature of reacting media (i.e., ionic liquid) can be modified by addition of either distilled water (H2O) and/or polymeric additives (e.g., ethylene glycol). 34 Modified ionothermal synthesis was conducted employing the same reaction protocol by using 50:50 (v/v) mixture of (a) EMITFSI and ethane-1,2-diol (C2H5OH) and (b) EMI-TFSI and 6984


Research Article


Figure 2. (Left) Comparative X-ray diffraction patterns of “Na2.4Fe1.8(SO4)3” end members prepared by using pristine ionic liquid (red), 50:50 v/v mixture of ionic liquid and ethylene glycol (blue) and 50:50 v/v mixture of ionic liquid and H2O (black). (Right) Presence of H2O develops sharp peak i.e. larger particles while glycol develops broader peaks indicating smaller particles vis-à-vis pristine ionic liquid.

Figure 3. Morphology of pristine and modified ionothermally prepared sodium alluaudite materials. Representative SEM images of (a,b) uniform particles obtained using solely EMI-TFSI, (c,d) platelete morphology developed using mixture of EMI-TFSI and glycol and (e,f) large rod-shaped particles formation in the presence of EMI-TFSI and H2O mixture. (g) TEM image showing spherical nanoscale size (∼100 nm) and (h) HRTEM image of the atomic planes and selected area diffraction pattern (inset).

H2O. Unlike EMI-TFSI, both C2H5OH (TB = 197.3 °C) and H2O (TB = 100 °C) develop high vapor pressure inside autoclave upon annealing at 300 °C triggering faster nucleation and Ostwald ripening. In all cases, alluaudite “Na2.4Fe1.8(SO4)3” target products were obtained as shown in Figure 2. This particular composition, which is thermodynamically stable, was determined by Rietveld refinement of XRD data. The presence of H2O develops very sharp XRD peaks owing to extensive grain growth in the presence of high vapor pressure. In contrast, ethane-1,2-diol acts as chelating agent developing anisotropic grain growth, thus having smaller particles as evident from XRD peak broadening. It hints at possible synthesis of alluaudite sulfates solely by using glycol media to replace the expensive ionic liquid media. A detail report on glycol-based preparation of alluaudite compounds will be reported in future. The effect of reacting media was later observed by electron microscopy (Figure 3). The use of EMI-TFSI ionic liquid media develops homogeneous uniform elliptical particles with minimal aggregation (Figure 3a,b). These particles are made of smaller (∼100 nm) particles when observed under TEM

(Figure 3g). When ethane-1,2-diol is mixed to EMI-TFSI, the modified media favors chelation and anisotropic grain growth to form micrometric platelet morphology with nanometric thickness (Figure 3c,d). Similarly, the addition of H2O to EMITFSI generates high vapor pressure and grain coarsening to form large micrometric rod/needle-shaped particles (Figure 3e,f). The anisotropy of final product is less pronounced when the additive is H2O vis-à-vis ethane-1,2-diol. As summarized in Figure 3, it is possible to obtain alluaudite sulfate compound with tunable size and particle morphology. Rich diversity in tunability can be realized by simply changing the nature of reacting media that act as templating agents, e.g., (i) using ionic liquids/diols of different molecular weight, (ii) using linear vs branched ionic liquid/diols, (iii) using different terminal groups (−OH, −Cl, −CN, −CH3 etc.) and (iv) using hydrophobic surfactant to form nanodroplets or micelles etc. As alluaudite sulfates offer large tunnels for multidimensional Na+-diffusion, it can potentially deliver efficient electrochemical (de)insertion of Na+ even with large micrometric particles. EDS elemental analyses led to the verification of (i) uniform distribution of all constituent elements and (ii) a Na-rich and Fe-deficient 6985


Research Article


vibration of SO4 units [symmetric stretching ν1 ≈ 983 cm−1 and assymmetric stretching ν3 ≈ 1100 cm−1] and one from bending vibration of SO4 units [assymmetric bending ν4 ≈ 620 cm−1]. FTIR study confirms the formation of SO4-rich product without any moisture attack. The alluaudite material was found to be thermally stable up to 450 °C. Further heating led to aggressive weight loss indicating the release of SOx gaseous biproducts (Supporting Information Figure S1). Even though this low decomposition temperature is typical for SO4-based insertion materials, it offers no problem for safe practical application of battery materials. It is worth mentioning the target compound was found to have minor amount (2−4 wt %) of Fe3+ impurities as detected by Mössbauer spectrum (not shown here). It may be due to the formation of Fe2O3/ Fe3O4 impurities from some leached out Fe species from the alluaudite phase so as to maintain the thermodynamically stable Na-rich and Fe-poor composition. Magnetic Properties of Na2.4Fe1.8(SO4)3. Figure 6 (top) shows the temperature dependence of the molar magnetic susceptibility χ between 300 and 2 K for “Na2.4Fe1.8(SO4)3” in a field cooled (FC) sequence and a zero-field cooled (ZFC) sequence, using an applied field of H = 1000 Oe. An antiferromagnetic (AFM) transition clearly occurs at TN ≈ 12.0 K. The inset to Figure 6 highlighting the transition region shows very little divergence between the FC and ZFC measurements. Figure 6 also shows the inverse molar magnetic susceptibility χ−1, and a fit to the linear paramagnetic region 50 ≤ T ≤ 250 K using the Curie−Weiss law χ(T) = C/(T − ΘCW). This fit yields a Curie constant C = 3.371(6) emu K/mol which corresponds to an effective moment μeff = 5.20 μB/Fe, somewhat higher than the theoretical value of 4.89 for highspin Fe2+(d6, t2g4eg2, S = 2), but not unusual for Fe oxides with imperfectly quenched orbital contributions. For example, values of 5.29 and 5.20 μB were found for Li2FeP2O736 and LiFePO437 respectively, both with the same Fe2+ oxidation state and similar magnetic exchange pathways. The extracted Curie−Weiss constant ΘCW = −12.55(2) K is very close to the susceptibility downturn at the Néel temperature TN = 12.0 K, indicating that the sample suffers from no significant magnetic frustration. Figure 6 (bottom) depicts the magnetic field dependence of magnetization M(H) at 2 K. No hysteresis is observed that would suggest the presence of an FM component. Overall, magnetic susceptibility measurements support the onset of long-range antiferromagnetic (AFM) ordering in “Na2.4Fe1.8(SO4)3” at TN = 12.0 K, with no evidence for a significant ferromagnetic (FM) component or frustration. Crystal Structure Evolution of Na2Fe2(SO4)3 under Desodiation Process. In order to envisage the electronic structure of alluaudite Na2Fe2(SO4)3 based on DFT based simulation, we have considered the stable monoclinic phase with space group P21/c of the compound as reported in earlier work.15 It is worth mentioning the possibility of another symmetry C2/c that the structure can exhibit. However, the compound with P21/c space group has been found to be most stable one with minimum energy configuration. The energetic difference between the two symmetries at ground state has been determined as 0.7 meV, which enables us to consider throughout this study Na2Fe2(SO4)3 polyanion structure with P21/c space-group. Gibbs free energy at room temperature induces the configurational entropy, which is the prime reason behind the coexistence of both the structural symmetry in experimental environment. The DFT simulation reveals the

environment proving the off-stoichiometric composition in the end product (Figure 4). It was further verified by ICP analysis

Figure 4. Elemental analysis of ionothermally made “Na2.4Fe1.8(SO4)3” by energy dispersive spectroscopy (EDS) showing uniform distribution of constituent elements and relative Na-rich and Fe-deficient environment.

leading to a Na/Fe ratio of 1.28 (close to the exact ratio of 1.33). Although off-stoichiometric composition is thermodynamically stable, yet unknown stoichiometric alluaudite [Na2Fe2(SO4)3] can be a better cathode. The local structural coordination was later probed with infrared spectroscopy as shown in Figure 5. First, despite using

Figure 5. FT-Infrared spectrum of solution ionic-liquid made “Na2.4Fe1.8(SO4)3” product phase depicting various kinds of vibration bands arising from constituent SO4 tetrahedra. The absence of any water species (as indicated by no OH: 3600 cm−1 band) is clearly observed.

the monohydrate precursor, there is no residual H2O or surface adsorbed moisture as confirmed by the absence of any peak ∼3500 cm−1 (related to symmetric/assymmetric stretching of OH− species). Overall, low wavenumber signature bands stemming from SO42− species were observed. Three different kinds of infrared bands were captured: two from stretching 6986


Research Article


Figure 6. (Top) Temperature dependence of magnetic susceptibility (black circles) and inverse magnetic susceptibility (red circles) for “Na2.4Fe1.8(SO4)3” in an applied field of 1000 Oe. The black dashed line is a Curie−Weiss fit to the inverse susceptibility data between 50 and 250 K. The inset highlights the susceptibility close to TN. (Bottom) Field dependence of magnetization for “Na2.4Fe1.8(SO4)3” at 2 K.

Table 1. Symmetry of the Unit Cell, Magnetic Moment of Fe Ions, the Optimized Cell Parameters and Volume of the Unit Cell compound

symmetry

Na2Fe2(SO4)3

P21/c C2/c P21/c C2/c P21/c C2/c P21/c

NaFe2(SO4)3 Fe2(SO4)3

μFe (μB)

a(Å)

b(Å)

c(Å)

α(deg)

β(deg)

γ(deg)

V(Å3)

3.78 3.77

11.43 12.68 11.46 12.65 11.40 12.58 11.07

12.82 12.78 12.77 12.77 12.61 12.75 12.59

6.56 6.61 6.51 6.51 6.46 6.42 6.36

90 90 90 90 90 90 90

95.81 114.76 95.27 115.53 94.65 115.35 96.59

90 90 90 90 90 90 90

957.59 974.00 949.74 949.86 926.05 931.40 881.68

5.2 4.30/3.78 4.06 4.27

relatively small amount and this ensures further the validity of our theoretical model. We have investigated theoretically the crystal structure evolution of alluadite Na2Fe2(SO4)3 in the reaction process of desodiation. There are mainly two possibilities that have been considered regarding the desodiation process. The removal of one sodium atom per formula unit from the full sodiated

lattice parameters of the structure as 2% discrepancy with respect to the experimental outcome. Table 1 enlists the computed lattice parameters along with the unit cell volume and the magnetic moment of Na2Fe2(SO4)3 for both P21/c and C2/c symmetry. A good agreement is observed between the theory and experiment using GGA exchange correlation functional, which overestimates the lattice parameters in a 6987


Research Article


such as the sulfur atoms contributing in the energy range of −8 to −6 eV as shown in Figure 7. These contributions are

material forming NaFe2(SO4)3 is the first mechanism, while the second one is considering the possibility of complete removal of sodium ions to form Fe2(SO4)3. In the half-desodiated case NaFe2(SO4)3, the total energy difference between the P21/c and C2/c symmetries increased 0.1 eV more in comparison to the full sodiated structure. This leads to the clearer confirmation of most possible symmetry of NaFe2(SO4)3 as P21/c space-group. Moreover, the phonon dispersion calculations confirm the fact of dynamical stability of NaFe2(SO4)3 having P21/c symmetry with rare negative vibrational frequencies in phonon spectrum than C2/c symmetry. There is 3.5% volume shrinkage happened corresponding to removal of half sodium atoms in the cell reaction. The reduction in the Fe−O octahedral bond length corresponding to the oxidation process from Fe−II to Fe−III state as depicted in Table 2 is the prime reason behind this Table 2. Average Bond Distance of the FeO6 Octahedral and SO4 Tetrahedral and the Distortion Parameter Computed by 1 12 |l − l | D = 12 ∑1 i l av where li is the Bond Length from the ith av

par Fe−O and lav is the Average Bond Length of the Octahedral or Tetrahedral compound Na2Fe2(SO4)3

NaFe2(SO4)3

Fe2(SO4)3

Fe1−O Fe2−O S−O Fe1−O Fe2−O S−O Fe1−O Fe2−O S−O

Fe oxidation state

bond distance (Å)

FeO6 and SO4 distortion coefficient

+2 +2

2.17 2.25 1.49 2.16 2.06 1.49 2.05 2.05 1.49

0.022 0.019 0.009 0.017 0.030 0.011 0.044 0.044 0.023

+2 +3 +3 +3

volume decrement that leads to a single-phase transformation in the case of removal of half sodium atoms on the contrary to the insertion of half sodium atoms. The full desodiated system has been achieved upon removal of all sodium atoms from the crystal structure of NaFe2(SO4)3. We carried out the simulation for both symmetries, but for the C2/c symmetry the electronic self-consistent cycles do not completely converge itself. In case of P21/c symmetry, the total energy of the system has been converged through the electronic self-consistent with 8% of decrement of unit cell volume in comparison with the full sodiated case. The computed distortion parameter for FeO6 octahedral has attained a value of 0.044 upon desodiation as shown in Table 2. The nonuniform structural distortion makes Fe 2 (SO 4 ) 3 structure unstable, which has also been confirmed from the existence of large imaginary vibration frequencies in corresponding phonon spectrum. It means that this system is not dynamically stable and it is very likely that the full desodiation process would occur combined with a phase transition leading to irreversible sodium insertion process. (It means that this system is not dynamically stable and it is very likely that the full desodiation is an irreversible process.) Electronic Structure Analysis of Na2Fe2(SO4)3, NaFe2(SO4)3, and Fe2(SO4)3. The projected density of states (pDOS) based on DFT formalism have been computed for Na2Fe2(SO4)3, NaFe2(SO4)3, and Fe2(SO4)3 structures. The results for the full sodiated polyanion show the general trends

Figure 7. Projected density of states of Na2Fe2(SO4)3, NaFe2(SO4)3, and Fe2(SO4)3. Here, gray is representing the total density of states, red the contribution coming from Fe ions, blue the contribution coming from S atoms, and green the contribution coming from O atoms. Dashed line is representing the Fermi level.

corresponding to the hybridization of S and O atoms that forms strong covalent bonds. Moreover, in the vicinity of Fermi level and in bottom of the conduction band the main contribution comes from d states of the iron centers. The band gap value of this polyanion has been determined as 3.58 eV i.e. showing insulator behavior. With the desodiation reaction, the band gap value changes to 1.08 and 1.91 eV for NaFe2(SO4)3 and Fe2(SO4)3 compounds, respectively. 6988


Research Article


Figure 8. Electrochemical study of ionothermally prepared “Na2.4Fe1.8(SO4)3” alluaudite cathode. (a) Galvanostatic voltage-capacity profiles (at a rate of C/20, 25 °C) showing the multistep voltage profiles, (b) discharge capacity as a function of cycling rates (at 25 °C), and (c) cycling stability and Coulombic efficiency up to 50 cycles at C/20 rate.

potential reaches a value of 3.88 V for x = 1, which is in agreement with the experimental outcome, whereas for the full desodiation reaction, the predicted redox potential value changes to 4.25 V. Electrochemical Properties of Ionothermally Made “Na2.4Fe1.8(SO4)3”. Following the ab initio DFT analysis of Na+ (de)insertion behavior, the electrochemical activity of ionothermally prepared “Na2.4Fe1.8(SO4)3” was tested in Na half-cell coin-type architecture. Despite any further cathode optimization like carbon (nano)painting and/or particle downsizing, reversible Na+ (de)insertion was observed (Figure 8). Galvanostatic voltage-capacity profile reveals an ambient condition discharge capacity of 80 mAh/g for the initial cycles at C/20 rate, which is 80% of calculated one-electron theoretical capacity (QTh = 100 mAh/g). The alluaudite samples prepared by modified ionothermal synthesis delivered lower discharge capacity owing to larger/anisotropic particles (i.e., 60 mAh/g for ionic liquid-water media and 50 mAh/g for ionic liquid-diol media). Upon cycling, there is a visible drop in charge profile from the first cycle to the subsequent ones (Figure 8a) that indicates the occurrence of an irreversible structural transition during the first oxidation segment, similar to the case of Li2FeP2O7 and Li2FeSiO4.38 There is a gradual sloping voltage profile throughout the redox process that may arise from single-phase (solid solution) redox mechanism. A detail study of the underlying redox behavior using synchrotron XRD and23Na solid-state NMR analyses is under investigation. The voltage profile reveals the presence of multiple steps from second cycle onward (∼3.4, 3.8, and 4.2 V during charging and

In addition to the band gap change, the variation of the total magnetization per Fe center with the removal of Na ions has also been observed. The value changes from 3.78 μB in all Fe centers in Na2Fe2(SO4)3 compound to 4.30 μB in half of Fe centers in NaFe2(SO4)3 compound and 4.27 μB in all Fe centers in case of completely desodiated Fe2(SO4)3. This result indicates that iron centers appears as the main responsible factors for the charge rearrangement. In order to confirm this hypothesis, the Bader charge analysis has been undertaken in the case of Na2Fe2(SO4)3, NaFe2(SO4)3 and Fe2(SO4)3. For the full sodiated system the net charge amounts an average of +0.85, +1.5, +3.85, and −1.30 for Na, Fe, S and O atoms. If one considers the total ionic bond picture, oxygen atoms should assume values of −2. However, the hybridization of S states and O states showed in the pDOS leads to an effective charge of −1.30 for the O atoms. With the removal of one sodium atom per formula unit half of Fe centers undergoes a change from +1.5 to +1.85 and for the full desodiation process, all Fe centers assume net charge of +1.85 while the other atoms remains with the same value. This results lead to the confirmation of oxidation reaction occurring in Fe centers. The average intercalation potential vs Na/Na+ has been determined using the following equation: V=−

E(Na 2Fe2(SO4 )3 ) − (2 − x)E(Nabcc) − E(NaxFe2(SO4 )3 (2 − x)

where E is the total energy of the corresponding compound. The value of x is 1 in case of half desodiation and 2 in case of full desodiation process. The computed value of the redox 6989



■

3.8 and 3.4 V during discharging). With majority of Na+ (de)insertion process occurring in the voltage window of 3.6−3.8 V, the net redox potential is located at 3.7 V (vs Na/ Na+). A keen observation reveals a visible charge plateau at 4.2 V that does not appear during discharging segment. It can be correlated to large diffusion barrier at related desodiated composition. This kind of behavior has been observed in other sodium insertion cathodes like Na 2 FeP 2 O 7 and Na4Fe3(PO4)2(P2O7).13,14 The application of carbon coating can make this high-voltage plateau reversible. The experimental observation of limited capacity (80 mAh/g) corroborates with the DFT prediction of limited partial desodiation owing to structural instability. Also, the experimental redox potential value (ca. 3.7 V) is close to the DFT predicted value of 3.88 V. Further, the rate kinetics of the ionothermally obtained alluaudite product was measured (Figure 8b). It retains the multistep sloping characteristics delivering 75% and 50% of initial capacity at C/2 rate and 2 C rate, respectively. This efficient electrochemical activity can be rooted to the open 3dimensional framework of alluaudite. Although, these capacity values are less in comparison to solid-state synthesized alluaudite products,15 further cathode optimization focusing on uniform carbon coating can improve the electrochemical properties of ionothermally made compound. Finally, the cycling stability was tested up to 50 cycles retaining excellent Coulombic efficiency and ∼85% of initial capacity after 50 cycles (Figure 8c). This Fe-based alluaudite marks very high redox potential (vs Na) leading to formidable energy density comparable to many known insertion materials. With the highvoltage activity and cycling stability, this low-cost sulfate based Na-rich off-stoichiometric alluaudite insertion material can offer commercial promise in near future.39

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +91-80-2293-2873. Fax: +91-80-2360-7316. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank Department of Science and Technology (DST), Govt. of India for financial support under the aegis of Solar Energy Research Initiative (SERI) programme (DST/ TMC/SERI/FR/169). D.D. thanks Ministry of Human Resource Development (MHRD) for financial support. C.D.L. acknowledges the financial support from the Australian Research Council − Discovery Projects. R.B.A., S.C. and R.A. would like to acknowledge the Erasmus Mundus for a doctoral fellowship, Carl Tryggers Stiftelse for Vetenskaplig Forskning (CTS), Swedish Research Council (VR), STandUP for financial support. SNIC, HPC2N, and UPPMAX are acknowledged for providing computing time.

■

REFERENCES

(1) Goodenough, J. B.; Park, K.-S. The Li-ion Rechargeable Battery: A Perspective. J. Am. Chem. Soc. 2013, 135, 1167−1176. (2) Smalley, R. E. Future Global Energy Prosperity: The Terawatt Challenge. MRS Bull. 2005, 30, 412−417. (3) Dunn, B.; Kamath, H.; Tarascon, J. M. Electrical Energy Storage for the Grid: A Battery of Choices. Science 2011, 334, 928−935. (4) Whittingham, M. S. Lithium Batteries and Cathode Materials. Chem. Rev. 2004, 104, 4271−4301. (5) Tarascon, J. M. Is Lithium the New Gold? Nat. Chem. 2010, 2, 510. (6) Yabuuchi, N.; Kubota, K.; Dahbi, M.; Komaba, S. Research Development on Sodium-Ion Batteries. Chem. Rev. 2014, 114, 11636− 11682. (7) Slater, M. D.; Kim, D.; Lee, E.; Johnson, C. S. Sodium-Ion Batteries. Adv. Funct. Mater. 2013, 23, 947−958. (8) Ellis, B. L.; Nazar, L. F. Sodium and Sodium-Ion Energy Storage Batteries. Curr. Opin. Solid State Mater. Sci. 2012, 16, 168−177. (9) Yabuuchi, N.; Kajiyama, M.; Iwatate, J.; Nishikawa, H.; Hitomi, S.; Okuyama, R.; Usui, R.; Yamada, Y.; Komaba, S. P2-type Nax [Fe 1/2Mn1/2 ]O2 Made from Earth-abundant Elements for Rechargeable Na Batteries. Nat. Mater. 2012, 11, 512−517. (10) Moreau, P.; Guyomard, D.; Gaubicher, J.; Boucher, F. Structure and Stability of Sodium Intercalated Phases in Olivive FePO4. Chem. Mater. 2010, 22, 4126−4128. (11) Barpanda, P.; Chotard, J. N.; Recham, N.; Delacourt, C.; Ati, M.; Dupont, L.; Armand, M.; Tarascon, J. M. Structural, Transport, and Electrochemical Investigation of Novel AMSO4F (A = Na, Li; M = Fe, Co, Ni, Mn) Metal Fluorosulphates Prepared Using Low Temperature Synthesis Routes. Inorg. Chem. 2010, 49, 7401−7413. (12) Ellis, B. L.; Makahnouk, W. R. M.; Makimura, Y.; Toghill, K.; Nazar, L. F. A Multifunctional 3.5 V Iron-based Phosphate Cathode for Rechargeable Batteries. Nat. Mater. 2007, 6, 749−753. (13) Barpanda, P.; Ye, T.; Nishimura, S.; Chung, S. C.; Yamada, Y.; Okubo, M.; Zhou, H. S.; Yamada, A. Sodium Iron Pyrophosphate: A Novel 3.0 V Iron-based Cathode for Sodium-Ion Batteries. Electrochem. Commun. 2012, 24, 116−119. (14) Kim, H.; Park, I.; Seo, D. H.; Lee, S.; Kim, S. W.; Kwon, W. J.; Park, Y. U.; Kim, C. S.; Jeon, S.; Kang, K. New Iron-Based MixedPolyanion Cathodes for Lithium and Sodium Rechargeable Batteries: Combined First Principle Calculations and Experimental Study. J. Am. Chem. Soc. 2012, 134, 10369−10372. (15) Barpanda, P.; Oyama, G.; Nishimura, S.; Chung, S. C.; Yamada, A. A 3.8 V Earth-Abundant Sodium Battery Electrode. Nat. Commun. 2014, 5, 4358.

4. CONCLUSIONS In summary, alluaudite “Na2.4Fe1.8(SO4)3” has been synthesized via direct and modified ionothermal synthesis using room temperature ionic liquid and diol/H2O as reacting media. Benchmarking the first solvothermal synthesis of “Na2.4Fe1.8(SO4)3” sulfate alluaudite, it demonstrates the feasibility of low-temperature energy-savvy production of desired cathode at 300 °C with tunable particle size and morphology. DFT study revealed that “Na2.4Fe1.8(SO4)3” alluaudite material (i) is kinetically limited (poor conductivity) and (ii) can not realize complete desodiation owing to structural instability. It depicts low-temperature antiferromagnetic ordering at TN = 12 K. The ionothermaly obtained alluaudite exhibited high-voltage (ca. 3.7 V) redox activity delivering a reversible capacity ∼80 mAh/g. Ionothermal synthesis can be easily extended to other members of alluaudite family [Na2+2xFe2‑x(SO4)3, M = 3d metals]. This attests the versatility of ionothermal synthesis of Na-battery insertion materials.

■

Research Article

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.5b11302. TGA curve showing the thermal stability of ionothermally prepared Na2.4Fe1.8(SO4)3 (PDF) 6990


Research Article


(37) Liang, G.; Park, K.; Li, J.; Benson, R. E.; Vaknin, D.; Markert, J. T.; Croft, M. C. Anisotropy in Magnetic Properties and Electronic Structure of Single- Crystal LiFePO4. Phys. Rev. B: Condens. Matter Mater. Phys. 2008, 77, 064414−064425. (38) Shimizu, D.; Nishimura, S.; Barpanda, P.; Yamada, A. Electrochemical Redox Mechanism in 3.5 V Li2‑xFeP2O7 (0 < x < 1) Pyrophosphate Cathode. Chem. Mater. 2012, 24, 2598−2603. (39) Barpanda, P. Pursuit of Sustainable Iron-based Sodium Battery Cathodes: Two Case Studies. Chem. Mater. 2016, 28, 1006−1011.

(16) Ming, J.; Barpanda, P.; Nishimura, S.; Okubo, M.; Yamada, A. An Alluaudite Na2+2xFe2‑x(SO4)3 (x = 0.2) Derivative Phase as Insertion Host for Lithium Battery. Electrochem. Commun. 2015, 51, 19−22. (17) Barpanda, P. Sulfate Chemistry for High-Voltage Insertion Materials: Synthetic, Structural and Electrochemical Insights. Isr. J. Chem. 2015, 55, 537−557. (18) Tarascon, J. M.; Recham, N.; Armand, M.; Chotard, J. N.; Barpanda, P.; Walker, W.; Dupont, L. Hunting for Better Li-Based Electrode Materials via Low Temperature Inorganic Synthesis. Chem. Mater. 2010, 22, 724−739. (19) Rietveld, H. M. A Profile Refinement Method for Nuclear and Magnetic Structures. J. Appl. Crystallogr. 1969, 2, 65−71. (20) Larson, A. C.; Von Dreele, R. B. General Structure Analysis System (GSAS); Los Alamos National Laboratory Report, LAUR 86− 748; Los Alamos National Laboratory: Los Alamos, NM, 1994. (21) Momma, K.; Izumi, F. VESTA 3 for Three-Dimensional Visualization of Crystal, Volumetric and Morphology Data. J. Appl. Crystallogr. 2011, 44, 1272−1276. (22) Kresse, G.; Hafner, J. Ab initio Molecular Dynamics for Liquid Metals. Phys. Rev. B: Condens. Matter Mater. Phys. 1993, 47, 558−561. (23) Kresse, G.; Hafner, J. Efficient Iterative Schemes for Ab Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186. (24) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (25) Dudarev, S. L.; Botton, G. A.; Savrasov, S. Y.; Humphreys, C. J.; Sutton, A. P. Electron-Energy-Loss Spectra and Structural Stability of Nickel Oxide: An LSDA + U Study. Phys. Rev. B: Condens. Matter Mater. Phys. 1998, 57, 1505−1509. (26) Mueller, T.; Hautier, G.; Jain, A.; Ceder, G. Evaluation of Tavorite-Structured Cathode Materials for Lithium-Ion Batteries Using High-Throughput Computing. Chem. Mater. 2011, 23, 3854− 3862. (27) Bader, R. F. W. Atoms in Molecules: A Quantum Theory; Oxford University Press: New York, 1990. (28) Tang, W.; Sanville, E.; Henkelman, G. A Grid-Based Bader Analysis Algorithm Without Lattice Bias. J. Phys.: Condens. Matter 2009, 21, 084204. (29) Manthiram, A.; Goodenough, J. B. Lithium Insertion Into Fe2(SO4)3 Frameworks. J. Power Sources 1989, 26, 403−408. (30) Dwibedi, D.; Araujo, R. B.; Chakraborty, S.; Shanbogh, P. P.; Sundaram, N. G.; Ahuja, R.; Barpanda, P. Na2.44Mn1.79(SO4)3: A New member of the Alluaudite Family of Insertion Compounds for Sodium Ion Batteries. J. Mater. Chem. A 2015, 3, 18564−18571. (31) Parnham, E. R.; Morris, R. E. Ionothermal Synthesis of Zeolites, Metal-Organic Frameworks, and Inorganic-Organic Hybrids. Acc. Chem. Res. 2007, 40, 1005−1013. (32) Recham, N.; Dupont, L.; Courty, M.; Djellab, K.; Larcher, D.; Armand, M.; Tarascon, J. M. Ionothermal Synthesis of Tailor-Made LiFePO4 Powders for Li-ion Battery Applications. Chem. Mater. 2009, 21, 1096−1107. (33) Barpanda, P.; Djellab, K.; Recham, N.; Armand, M.; Tarascon, J. M. Direct and Modified Ionothermal Synthesis of LiMnPO4 with Tunable Morphology for Rechargeable Li-ion Batteries. J. Mater. Chem. 2011, 21, 10143−10152. (34) Barpanda, P.; Recham, N.; Chotard, J. N.; Djellab, K.; Walker, W.; Armand, M.; Tarascon, J. M. Structural and Electrochemical Properties of Novel Mixed Li(Fe1‑xMx)SO4F (M = Co, Ni, Mn) Phases fabricated by Low Temperature Ionothermal Synthesis. J. Mater. Chem. 2010, 20, 1659−1668. (35) Barpanda, P.; Chotard, J. N.; Delacourt, C.; Reynaud, M.; Filinchuk, Y.; Armand, M.; Deschamps, M.; Tarascon, J. M. LiZnSO4F made in an Ionic Liquid: A ceramic Electrolyte Composite for SolidState Lithium Batteries. Angew. Chem., Int. Ed. 2011, 50, 2526−2531. (36) Barpanda, P.; Rousse, G.; Ye, T.; Ling, C. D.; Mohamed, Z.; Klein, Y.; Yamada, A. Neutron Diffraction Study of the Li-ion Battery Cathode Li2FeP2O7. Inorg. Chem. 2013, 52, 3334−3341. 6991


Ionothermal Synthesis of High-Voltage Alluaudite ... - ACS Publications

Recommend Documents