Competition between metal dissolution and gas release in Li-rich

Subscriber access provided by University of Sunderland

Article

Competition between metal dissolution and gas release in Li-rich LiRuIr O model compounds showing anionic redox 3

y

1-y

4

Quentin Jacquet, Antonella Iadecola, Matthieu Saubanère, Louis Lemarquis, Erik J. Berg, Daniel Alves Dalla Corte, Gwenaëlle Rousse, Marie-Liesse Doublet, and Jean-Marie Tarascon Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/acs.chemmater.8b02955 • Publication Date (Web): 19 Oct 2018 Downloaded from http://pubs.acs.org on October 22, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemistry of Materials

Competition between metal dissolution and gas release in Li-rich Li3RuyIr1-yO4 model compounds showing anionic redox Quentin Jacqueta,b,c, Antonella Iadecolac, Matthieu Saubanèrec,d, Louis Lemarquisa, Erik J. Berge, Daniel Alves Dalla Cortea,c, Gwenaëlle Roussea,b,c, Marie-Liesse Doubletc,d and Jean-Marie Tarascona,b,c

a

Collège de France, Chaire de Chimie du Solide et de l’Energie, UMR 8260, 11 place Marcelin Berthelot, 75231 Paris CEDEX 05, France b c

Sorbonne Université, 4 place Jussieu, F-75005 Paris, France

Réseau sur le Stockage Electrochimique de l’Energie (RS2E), FR CNRS 3459 – France

d Institut

Charles Gerhardt, CNRS UMR 5253, Université Montpellier, Place E. Bataillon, 34 095 Montpellier, France

e Electrochemistry

Laboratory, Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland

1 ACS Paragon Plus Environment

Chemistry of Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 26

Abstract Li-ion batteries have been the object of tremendous progress over the last decade; however their energy density should still be increased to power electric vehicles. Following this aim, the energy density of the cathode material can be drastically boosted by making use of anionic redox although it often comes along with material degradation. In this study, through a detailed analysis of the charge compensation mechanism of Li3RuO4 by OEMS, XAS and UV spectroscopy, we unveiled a new degradation mechanism for cathode material showing anionic redox, namely the dissolution of Ru forming RuO4/RuO4- species with limited gas release from the material. We show that this dissolution can be effectively tackled substituting Ru by Ir. However, such a strategy leads to a massive increase of O2 gas release at the end of charge. DFT calculations prove that the relative stability of the end members RuO4 and IrO4 versus oxygen release is at the origin of this competition between metal dissolution and gas release.

TOC :

RuyIr1-yO4 Gas release RuO4

LixRuyIr1-yO4

Dissolution Reversible anionic redox

0

x intermediate

Li3RuyIr1-yO4 Li3RuO4 3 x in LiXMO4


Page 3 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION : Li – ion batteries are widely used in cellphones and hybrid vehicles due to their tremendous energy density compared to other electrochemical energy storage systems. However, for electric vehicles, their performances should still be increased to meet the consumers demand in terms of autonomy. The energy density of Li-ion batteries being mainly governed by the cathode material, this calls for the search of new high capacity positive electrodes 1. Discovered in 2002, Li-rich materials of formula Li1.2Ni0.13Mn0.54Co0.13O2 opened a fertile ground as they show capacities exceeding 300 mAh/g, nearly two times greater than commercial LiCoO2 2,3,4. A few years later (2013), the origin of such extra capacity was demonstrated to be nested in the electrochemical activity of the oxygen network offering a new direction for creating advanced high capacity electrodes5,6,7. Fundamental insights concerning this anionic redox activity were obtained by theoretical studies which revealed the key role played by oxygen non-bonding states in the redox activity of the electrodes8,9. They provided some guidance to further increase the anionic redox activity; that is increasing, the number of oxygen lone-pairs by increasing the O/M ratio and the Li content in the chemical composition, M being the transition metal 10. These findings have led to the study of various Li-rich compounds such as Li5FeO411, Li3NbO4 derivatives12,13 or Li3IrO414, for which high capacity can be achieved but at the expense of the materials stability. Indeed, besides the intrinsic drawbacks pertaining to Li-rich materials (sluggish kinetics, voltage fade15,16) it was recently shown that pushing the anionic redox to its limit by increasing the O/M ratio can lead to massive O2 release (Li3IrO4) or phase decomposition (Li5FeO4) upon oxidation. To combat these issues so as to design practical Li-rich materials with a mastered anionic redox activity, a greater understanding of the degradation mechanisms upon oxidation is sorely needed. Some clues already exist. For instance, studies on Li3IrO4 have shown a stable and reversible anionic activity until the removal of 2 Li, while further delithiation induces massive O2 release leading to an IrO3 phase at the end of charge. In contrast, XPS studies on Li3RuO4 electrode revealed an anionic activity upon charging while limited gas release is observed but with issues related to metal dissolution17. Therefore, moving from Ir to Ru in the Li3MO4 rocksalt series seems to have important effects on the degradation mechanism of the materials. Here, we embark into the understanding of the difference between these two model compounds so as to identify the key parameters governing the material degradation mechanisms upon activation of the anionic redox process. First, we perform a detailed analysis of the charge compensation and degradation mechanism of Li3RuO4. We used operando X-ray Absorption Spectroscopy (XAS) and Online Electrochemical Mass Spectrometry (OEMS) to unveil the electrochemical activity of oxygen with very little gas release (mostly CO2) and show direct evidence of Ru dissolution as 3 ACS Paragon Plus Environment


Page 4 of 26

characterized via operando Ultraviolet-Visible (UV) spectroscopy. In light of the difference between the degradation mechanisms of Li3RuO4 and Li3IrO4, we explore the Li3RuyIr1-yO4 series for which we show that a small Ir substitution in Li3RuO4 (y = 0.75) is sufficient to fully prevent Ru dissolution while triggering massive gas release. The relative stability of RuO4/IrO4 are then investigated through Density Functional Theory (DFT) calculations showing that the former is stable at standard pressure and temperature conditions while the latter is prone to O2 gas release. The paper is organized as follow: after describing the synthesis of Ru/Ir samples, we focus on the charge compensation and degradation mechanism unveiled by in situ UV spectroscopy, OEMS and operando XAS spectroscopy results for Li3RuO4 prior to present the structural/electrochemical results obtained for the Li3RuyIr1-yO4 series and discuss our results based on DFT calculations. Experimental part Synthesis Classical solid state synthesis from the metallic/oxide precursors, Ir black (Alfa Aesar 99.9%) and RuO2 (Alfa Aesar, 99.9%) and with 10% excess Li2CO3 (Sigma Aldrich, 99%), was employed to prepare the Li3RuyIr1-yO4 compounds with y = 1, 0.75, 0.5, 0.25, 0. These compositions were chosen so as to get a representative view of the phase diagram. The reactants were mixed using mortar and pestle before firing the mixture for 24 h at 900°C in air. Crystal structure analysis X-ray diffraction measurements (XRD) were performed using a BRUKER D8 Advance diffractometer with Cu Kα radiation (λKα1=1.54056 Å, λKα2=1.54439 Å) and a Lynxeye XE detector. In situ X-ray powder diffraction (XRD) measurements were done using a home-made airtight electrochemical cell equipped with a Be window18. Rietveld refinements were performed using the FullProf program19. Electrochemical testing All electrochemical characterizations were performed in Swagelok-type cells. The positive electrode material was grinded in SPEX-type ball mill for 15 min with 20% carbon SP before testing. Li metal was used as anode. Whatmann GF/D borosilicate glass fibers sheets were used as separators and the electrolyte was LP100 – 1/1/3 EC/PC/DMC with 1M LiPF6. Cells were assembled in Ar-filled glovebox. Galvanostatic cycling was performed at C/10 (meaning 1 Li is extracted in 10 h) with different cutoff voltages depending on the composition of the active material, from 4.2 V to 4.6 V vs Li+/Li. Ex situ samples were washed with DMC inside an Ar-filled glovebox before being dried under vacuum. In situ gas analysis was performed into a special homemade cell20 using the same electrode material, electrolyte and negative electrode as for the standard electrochemical tests.


Page 5 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Online Electrochemical Mass Spectrometry (OEMS). Detailed schematics of the OEMS setup are provided elsewhere21. Briefly, the OEMS operates with a quadrupole mass spectrometer (QMS 200, Pfeiffer, Germany) for partial pressure measurements and a pressure transducer (PAA-33X, Keller Druck AG, Switzerland) for total cell pressure recordings, temperature, and internal volume determination. Stainless steel gas pipes and Swagelok fittings (3 mm compression tube fittings, Swagelok, OH, USA) are employed to connect a homemade OEMS cell, a set of solenoid valves (2-way magnetic valve, Series 99, silverplated nickel seal, Parker, USA) and a scroll pump (nXDS15i, EDWARDS GmbH, UK) for efficient flushing. The magnetic valves are electronically controlled with a Solid State Relay Module (NI 9485 measurement System, National Instruments, TX, USA) connected to a computer with a LabView Software (NI Labview 2013, National Instruments, TX, USA). For partial pressure and gas evolution rate analysis 0.8 mL of gas are extracted from the headspace (~ 4 mL) of the cell and replaced by pure Ar (quality 5.0). Calibration gas bottles were utilized to relate the MS ion-current signals at m/z = 32 and 44 to known concentrations of O2 and CO2 (1000 ppm of O2 and CO2 in Ar, respectively), before and after the measurement. UV visible spectroscopy. In situ UV-visible spectroscopy was performed in air tight electrochemical cell (UVC cell, SPHERE-ENERGY, France). PTFE self-standing electrode of Li3RuO4 (72%wt), carbon SP (18%wt) and PTFE (10%wt) coated on stainless steel mesh are used as positive while the negative electrode is Li foil coated on another stainless steel mesh. 2mL of LP100 is used as electrode without any separator. UV spectrometer is a Mettler Toledo UV5bio. Operando X-ray absorption spectroscopy (XAS). Operando XAS measurements at the Ru K-edge were performed in transmission mode at the ROCK beamline22 of synchrotron SOLEIL (France). A Si (220) channel-cut quick-XAS monochromator with an energy resolution of 2 eV at 22 keV was used. The intensity of the monochromatic X-ray beam was measured by three consecutive ionization detectors. The in situ electrochemical cell18 was placed between the first and the second ionization chambers. Selfstanding PTFE films of the active material were used and cycled at C/4. The cell was started on reduction down to 1.4 V, then charged up to 4.2 V corresponding to Li0.3RuO4 composition before discharging to 1 V. Successive spectra were collected at a rate of 2 Hz and averaged out over periods of 5 minutes. The energy calibration was established with simultaneous absorption measurements on RuO2 (E = 22135 eV) placed between the second and the third ionization chamber. The data was treated using the Demeter package for energy calibration and normalization23. The normalized spectra were then globally analyzed with Principal Component Analysis (PCA)24 in order to individuate the orthogonal components able to describe the whole evolution during cycling. The number of principal components was then used as basis for Multivariate Curve Resolution-Alternating Least Squares (MCRALS)25 analysis. Finally, the reconstructed components were fitted using the Artemis software. Fourier transforms of k3 weighted EXAFS oscillations were carried out in k-range from 5.3 Å−1 to 14.5 Å−1. Fitting was performed in R-



Page 6 of 26

range from 1.0 to 3.2 Å. EXAFS amplitudes and phase-shifts were calculated by FEFF7 with a starting model derived from Li3RuO4. DFT calculations. Spin-polarized density functional theory (DFT) calculations were performed using the plane-wave density functional theory VASP (Vienna ab initio simulation package) code26,27 within the generalized gradient approximation of Perdew−Burke−Ernzerhof (PBE) to describe electron exchange and correlation28. The rotationally invariant Dudarev method (DFT + U)29 was used to correct the self-interaction error of conventional DFT for correlated d-electrons. Several configurations were computed for the delithiated compounds. The crystal structures, electronic and electrochemical properties of the LixMO4 phases (M = Ir, Ru) were performed for different Ueff = U − J = 0, 2 and 4 eV.

Li3RuyIr1-yO4: synthesis and characterization The synthesis of Li3RuyIr1-yO4 for y = 1, 0.75, 0.5, 0.25 and 0 was successfully carried out through solid state method. The XRD patterns of the obtained compounds are presented in Figure 1. Li3RuO4 can be described in the P 2/a space group30 with a volume per formula unit of 71.79 Å3 while Li3IrO4 has been recently reported to crystallize in R -3 m with a volume per formula unit of 72.77 Å3 14. Both are layered rocksalt derivatives composed of LiO6 and (M0.5Li0.5)O6 layers (M being Ru or Ir) with the only difference being nested in the way the Li is arranged with respect to M within the (M0.5Li0.5)O6 layers. Indeed, while in Li3RuO4, Li and Ru are ordered in zig-zag chains, and in Li3IrO4, Li and Ir are randomly distributed within the layers. Regarding the intermediate compositions, Li3Ru0.25Ir0.75O4 and Li3Ru0.5Ir0.5O4 show the same pattern as Li3IrO4 while Li3Ru0.75Ir0.25O4 pattern is similar to the one of Li3RuO4.

Accordingly,

Li3Ru0.75Ir0.25O4 was satisfactorily refined using the structural model of Li3RuO4 (P 2/a) while introducing a slight Li/M cation mixing in the (M0.5Li0.5)O6 layers (Table SI1). Rietveld refinement of Li3Ru0.25Ir0.75O4 and Li3Ru0.5Ir0.5O4 using Li3IrO4 (R -3 m) structural model led to a decent fit, which could however be improved using the structural model of Li3RuO4 with a total disordering between Li and M in the transition metal layer as presented in Figure SI1 and Tables SI2 and SI3. Across the series, the volume per formula unit increases from 71.97 Å3 to 72.24 Å3 and 72.55 Å3 for y = 0.75, 0.5 and 0.25, respectively, which is fully consistent with the slightly larger ionic radii of Ir (0.57 Å) compared to Ru (0.56 Å)31. Overall, through the Li3RuyIr1-yO4 phase diagram, a phase transition from a perfectly ordered layered structure to a disordered layered structure is unveiled. Since high valence ruthenate or iridate compounds have received a growing interest for their magnetic properties, zero field cooling magnetic measurements of the 6 ACS Paragon Plus Environment

Page 7 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


synthetized materials were performed. All data were fitted using a Curie-Weiss law with the resulting constants presented in Figure SI2. Li3RuO4 shows an antiferromagnetic transition at 75 K with an effective moment of 3.62 μB, both being in agreement with previously reported data 32,33. In the Ir-doped Li3RuyIr1-yO4 phases, the effective magnetic moment drops continuously from 3.36 to 2.64, 1.69 and 0.77 μB for y = 0.75, 0.5, 0.25 and 0, respectively. The antiferromagnetic transition temperature also shifts to lower values, i.e. 10 K and 4 K for Li3Ru0.75Ir0.25O4 and Li3Ru0.5Ir0.5O4 and disappears completely for Li3Ru0.25Ir0.75O4 and Li3IrO4. The disappearance of the ordering temperature is associated to the increase of the cationic disordering while the decrease of the effective moment is consistent with the addition of low spin Ir5+, as previously reported 34. Degradation and charge compensation mechanism in Li3RuO4. Electrochemical properties of Li3RuO4 as cathode material for Li – ion batteries have been reported previously by our group 17. It was found that nearly three lithium atoms per formula unit can be extracted from the material below 4 V in a single process with the subsequent discharge showing a sloppy profile during the “reinsertion” of 2.5 Li down to 1.5 V. The “S” shape is preserved on subsequent cycles while the capacity rapidly decays. Ruthenium dissolution problems were mentioned as a possible origin of this decay but not further explored. Moreover, Li3RuO4 was reported to reversibly uptake 1.5 Li at a voltage of 1.5 V, leading to Li4.5RuO4. Through in situ XRD and ex situ XPS, electrochemically-driven amorphization of Li3RuO4 together with the participation of oxygen atoms in the charge compensation mechanism during charge is observed. In this study, we provide a detailed analysis of the degradation and the charge compensation mechanism by i) studying the dissolution of Ru, ii) analyzing the chemical composition of the gas released at high state of charge, and iii) probing the role of Ru in the charge compensation mechanism. 1) Degradation mechanism of Li3RuO4 : a) Ru dissolution and the instability of the RuO4-/RuO4 dissolved species To assess Ru dissolution, we performed SEM-EDX measurements on the Li counter electrode after the first electrochemical oxidation of Li3RuO4. Results unveil the presence of Ru, as it can be seen Figure SI3, suggesting the dissolution of some Ru ions/species in the electrolyte upon charge. To quantify the amount of dissolution, we measured the active material mass loss during charge as shown Figure 2 d. We found that dissolution starts after the removal of the 1st Li and reaches 35% after the removal of three lithium atoms from Li3RuO4. Accordingly, the electrochemical data are systematically renormalized based on the active material loss. Characterization of the soluble ruthenium species formed in the electrolyte 7 ACS Paragon Plus Environment


Page 8 of 26

was performed via operando UV absorption spectroscopy was performed and the results are presented in Figure 2 a-b-c. No change is observed in the UV spectra during the extraction of the first Li from Li3RuO4, consistent with the mass loss data, while the absorbed intensity grows rapidly upon further delithiation from Li2RuO4 to Li0.3RuO4. Two observations can be made on the UV spectra. First, high absorbance in ultra violet region decreasing as the wavelength increases is observed. Such signal is typically due to light scattering caused by the presence of particles in the electrolyte. Such particles can also be spotted visually in the cell (see Figure SI 4. Second, broad peaks can be observed at 310 nm and 360 nm and at 340 nm and 500 nm, the two former being ascribed to RuO4-/RuO4 and the two latter to RuO42- 35. Therefore, from operando UV spectroscopy, we could determine the formation of RuO4/RuO4soluble species in the electrolyte, together with the appearance of particles and RuO42-. Note that similar observations were done in LP100 – LP30 or PC containing 1M of LiClO4. Being powerful oxidizing agents, RuO4 and RuO4- might chemically react with the electrolyte 36. To assess this hypothesis, we added LP100 onto KRuO4 (K+, RuO4-) powder and monitored the evolution of the obtained solution by UV spectroscopy, as presented in Figure SI 5-a. The formation of RuO4- species is initially observed together with the appearance of particles. Then, the UV signal associated to RuO4- disappears while the amount of particles precipitating in the solution increases (Figure SI 5-b). Similar observations were made with RuO4. To assess if the chemical reaction of RuO4-/RuO4 with the electrolyte can lead to gas evolution, mass spectrometry measurements were performed on an argon filled reaction tube containing KRuO4 and LP100 and it revealed the presence of CO2 (see Figure SI 5-c). Altogether, these results demonstrate the partial decomposition of the Li2RuO4, upon further oxidation, in RuO4/RuO4- species that could further chemically react with the carbonate constituents of the LP100 electrolyte to form insoluble amorphous RuOx particles and CO2 gas (see scheme in Figure 2 –e).

b) Gas formation in the cell, study of its composition and origin. To further explore the electrochemical-driven gas release during the oxidation process of Li3RuO4 and upon cycling, OEMS experiments were performed (see Figure 3-a). Gas evolution starts after the removal of 0.5 Li with the formation of CO2 and progresses constantly until Li0.3RuO4, leading to a cumulated release of 0.2 moles of CO2 per mole of Li3RuO4. Pushing the oxidation process beyond Li0.3RuO4 results in drastic increase of the CO2 release which comes along with the production of O2 gas but in much smaller amount (20 times less), hence the need to limit our cycling between Li3RuO4 and 8 ACS Paragon Plus Environment

Page 9 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Li0.3RuO4. Pressure cell measurements carried out in this composition range confirm a relatively low gas evolution during the first cycle (Figure 3b). To account for this gas release three contributions should be considered: i) surface Li2CO3 decomposition at the beginning of charge, ii) reaction of RuO4-/RuO4 with LP100 after the removal of the first Li and iii) surface reactivity of charged LixRuO4. Assessing the exact contribution of each of these mechanisms is not trivial; however the second process is believed to be the most important source of gas release, keeping in mind that 35% of the electrode dissolves into RuO4/RuO4. 2) Role of Ru in the charge compensation mechanism. Although XPS data suggested anionic redox to be responsible for the charge compensation mechanism in Li3RuO4, no analyses were done to assess the evolution of the Ru oxidation state upon cycling. We thus decided to perform operando XAS measurements at Ru K-edge. The results are shown in Figure 4 for a cell initially started on discharge. During Li intercalation in Li3RuO4 (process 1 on the curve), there is a clear shift of the edge position towards lower energy indicative of the Ru reduction from 5+ to 3.5+ since 1.5 Li can be intercalated into Li3RuO4. This process is fully reversible since 1.5 Li can be removed from Li4.5RuO4 to restore Li3RuO4 (process 2). In contrast, when Li3RuO4 is oxidized (process 3) the shift of Ru K-edge XAS spectra is small and shows the appearance of a pre-edge peak, indicative of a non-centrosymmetric distortion around Ru. For the discharge processes 4 and 5, the pre-edge peak disappears and the edge position shifts continuously to lower energy, suggesting the reduction of Ru. Using the first reduction process as our reference and assuming a linear relationship between the edge energy and the Ru oxidation state, one can extrapolate the oxidation state of Ru to be Ru2.8+ at the end of discharge (see Figure SI6). To better understand the charge compensation mechanism during the electrochemical cycling, Principal Component Analysis (PCA) combined to Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) has been carried out on all XAS spectra (Figure SI7 and SI8). We find that seven components are necessary to describe the complex evolution of the spectra, Figure 5 a-b. The XANES spectra and the Fourier Transforms of the EXAFS oscillations of the principal components are reported in supplementary information (Figure SI 9 and Figure SI 10 respectively). The Ru-O and Ru-Ru distances obtained from the analysis of the EXAFS oscillations are reported in Figure 5-c) and summarized in Table SI 4. Each process, namely the reduction, the charge and the discharge will be described separately for clarity reasons. Three components are needed to describe the spectral evolution during the reduction process: the 1st component (C1) corresponds to the pristine material, the 2nd component (C2) 9 ACS Paragon Plus Environment


Page 10 of 26

corresponds to Ru4+ and the 3rd component (C3) corresponds to Ru3.5+. In agreement with the Ru reduction, the average Ru-O bond lengths increase from 1.98 Å to 2.02 Å and 2.06 Å for C1, C2 and C3, respectively. Concerning the Ru – Ru distances, the appearance of short Ru – Ru distance of 2.7 Å is observed during the reduction. Such Ru – Ru dimerization has been observed in other compounds with Ru4+ 37. This process is fully reversible since the oxidation from Li4.5RuO4 to Li3RuO4 follows the reverse transformation. The oxidation of Li3RuO4 can be described with only two components: the pristine (C1) and fully charged component (C4) which concentration reaches 100% at the end of charge. At this point, one should note that the dissolved RuO4/RuO4- species have little influence on the spectra since their contribution could not be spotted through the PCA analysis. This is attributed to the dispersion of the soluble species inside the cell, diluting their concentration in the beam. The Ru oxidation state of C4 corresponds to Ru5.4+, estimated using K2RuO4 (Ru6+) as a reference and assuming a linear relationship between the edge position and the metal oxidation state

38.

In C4, Ru adopts a highly distorted

environment composed of 1 short, 3 medium and 2 long Ru – O distances of 1.72 Å, 1.91 Å and 2.07 Å, respectively. The distortion of the RuO6 local environment is confirmed by the presence of the pre-edge peak in the XANES region of the spectra. For the reduction part, four components are needed: C4, C5, C6 and C7 which corresponds to Ru5.4+, Ru4.7+, Ru3.6+ and Ru2.8+, respectively. In terms of Ru – O distances, the average distance in C5, C6 and C7 are 1.93 Å, 2.03 Å and 2.07 Å, respectively, consistent with their respective oxidation states. Again, short Ru – Ru distances of 2.6 Å are observed in these components. Overall, operando Ru K-edge XAS spectra on Li3RuO4 confirm the cationic redox, Ru5+  Ru3.5+ for Li insertion into Li3RuO4. Regarding the charge of Li3RuO4, we have shown that while oxygen oxidation is responsible for the charge compensation in the bulk of the material, the formation of RuO4-/RuO4 species having Ru in its 7+/8+ oxidation state clearly shows that Ru oxidation is also active. During discharge both oxygen and ruthenium are involved in the electrochemical processes. At this point, we’ll try to count electrons and quantify the importance of cationic and anionic redox in the charge compensation mechanism. During charge, 2.7 Li+ are removed from Li3RuO4, corresponding to the extraction of 2.7 e-. As detailed in SI, 1.1 e- can be attributed to cationic redox, which by simple difference gives 1.6 e- for anionic redox. Turning to the subsequent discharge, Li0.3RuO4 can uptake 4.2 Li+ and equivalently 4.2e-. 2.6 e- can be attributed to the reduction of Ru5.4+  Ru2.8+ while the remaining electrons, 1.6 e-, arise from the oxygen reduction. Also not fully accurate, this calculation strongly suggests a decent reversibility for the oxygen activity in this material, despite its partial dissolution. 10 ACS Paragon Plus Environment

Page 11 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Li3RuO4 bears some similarity with Li3IrO4 with however a striking difference being that pushing the anionic redox to its limits triggers metal dissolution rather than gas formation. To get more insights about the differences between these two systems, the electrochemical properties of Li3RuyIr1-yO4 are next explored. 3) Understanding the difference between Li3RuO4 and Li3IrO4 through the study of Li3RuyIr1-yO4 solid solution. Electrochemical properties for the 1st cycle of the Li3RuyIr1-yO4 members are presented in Figure 6 a). All show similar behaviors irrespective of y: nearly 3 Li can be extracted up to 4.8 V, while 2 Li can be re-inserted down to 2 V. Gas pressure measurements during cycling (Figure 6 b) show that while Li3RuO4 gas evolution is rather small during delithiation, it increases dramatically up to 6x10-3 bar/mg at the end of charge once Ir is added to the chemical composition. Converting pressure to moles, it corresponds to a molar ratio between the gas and the active material of 50%. Interestingly, the onset of this gas release is correlated to the y ratio through the Li content (x) in LixRuyIr1-yO4 as it occurs at x = 2.4, 1.8, 1.4 and 1 for y = 0.75, 0.5, 0.25 and 1, respectively. In short, increasing the composition in Ir postpones the value of x at which gas release occurs. As for Li3IrO4, if we assume all O atoms from CO2/O2 come from the material, one can deduce, independently of y, a 25% oxygen loss from the material leading to a RuyIr1-yO3 composition at the very end of charge. To check the impact of gas release on the material structure, operando XRD was conducted on each composition and the results are presented in Figure 7 for Li3Ru0.5Ir0.5O4 and in Figure SI 11 and SI 12 for Li3Ru0.75Ir0.25O4 and Li3Ru0.25Ir0.75O4, respectively. Upon oxidation of Li3Ru0.5Ir0.5O4, the growth of a new set of peaks is observed as emphasized in the inset in Figure 7-a). They can be ascribed to a new Li poor phase coexisting with the pristine one and having a similar structure with however larger cell parameters, a = 2.965 Å and c = 14.525 Å compared to a = 2.945 Å and c = 14.425 Å (described in R-3m space group), and a volume per formula unit of 73.81 Å3 compared to 72.24 Å3. The growth of the cell parameters is likely due to the increase of the electrostatic repulsions between the oxygen layers which are less efficiently screened upon Li extraction. After x = 1.6, further oxidation leads to a drastic broadening and shift of the peaks corresponding to the new phase suggesting its amorphization. For Li3Ru0.25Ir0.75O4, a biphasic process between the pristine structure and a new phase with larger cell parameters and volume per unit formula, i.e. 72.55 Å3 and 74.53 Å3, respectively, is also observed from the beginning of charge up to x = 1.3, followed by the amorphization of the new phase upon further oxidation. For Li3Ru0.75Ir0.25O4, the loss of crystallinity is also observed at the beginning of charge. In light of these results, it can be concluded that the onset of gas release upon oxidation from the Li3RuyIr1-yO4 members is associated with the amorphization of an intermediate Li-poor 11 ACS Paragon Plus Environment


Page 12 of 26

phase. Lastly, and quite interestingly, operando UV experiments carried out on the various members of the Li3RuyIr1-yO4 series (the data is only shown for Li3Ru0.5Ir0.5O4 in Figure 8) revealed no metal dissolution whatever is the Ir content, demonstrating the benefit of this substitution for preventing electrode dissolution. 4) Rationalization of the degradation mechanism through DFT calculations First-principles DFT calculations were performed to address the origin of the decomposition of LixRuO4 which shows limited O2 release and does not stabilize RuO3 at the end of charge compared to LixIrO4. The phase stability diagrams are computed for LixRuO4 and LixIrO4 (see Figure 9). Note that the phase stability diagram of Li3IrO4 was reported in ref 14 and reused here for sake of comparison. It can be seen that both electrodes should undergo an equivalent structural phase transition at x = 1 for Ru and x = 0 for Ir. The transition converts the metal oxide framework made of condensed chains of edge-shared MO6 octahedra into a more opened framework consisting in fully disconnected MO4 tetrahedra. It is done through modification of the oxygen lattice without impacts on the cationic framework. Note that this structural type is equivalent to the one observed for strongly covalent transition metal phosphides, LixMP4, for which anionic redox was previously demonstrated

39.

In the case of Li3IrO4, the reaction

enthalpy computed for the O2 release mechanism is negative in LiIrO4 and positive in IrO4. Such findings show that i) the phase transition from Oh (octahedral) to Td (tetrahedral) stabilizes the O-network towards oxygen release but that ii) LiIrO4 will degas and IrO4 will never form. One should note that these calculations are performed with an ordered Li/Ir configuration, however in disordered systems, Li/Ir mixing creates statistically Ir-poor O environments which will degas at x > 1 in LixIrO4. For the ruthenium phase however, the Oh to Td transition takes place at lower Li content hence preserving LixRuO4 from oxygen release. We believe such a difference mainly arises from the number of electrons per metal, Ru having one electron less in its d-shell (Ru 4d8 compared to Ir 5d9). At last, the LixRuTdO4 opened framework is predicted to be stable vs. O2 release down to x = 0, however it might behave as a salt for which the (RuTdO4)- ionic species are prone to dissolution in the electrolyte. These results correlate perfectly with the experimental data since tetrahedral RuO4- species are indeed detected, proving the structural transition mentioned above is possible in this system. Hence, the winning strategy that allowed us to bypass the oxygen instability in the Ru-based electrode is negated by chemical dissolution, as schematized Figure 10. Conclusions


Page 13 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Through this study we unveiled the participation of ruthenium and oxygen in the charge compensation mechanism of Li3RuO4 via complementary XRD, OEMS and XAS measurements. We proved that the oxygen redox process is active both in oxidation and in reduction with limited gas release but with substantial Ru dissolution via the formation of soluble RuO4 or RuO4- species that react with the electrolyte to liberate CO2, hence leading to irreversible capacity. Such a degradation mechanism was never reported before and is totally different from the usual gas release observed in other material showing anionic redox, Li3IrO4 for instance. Therefore, the degradation mechanism of members among the series Li3RuyIr1-yO4 was studied. Strikingly, we show that metal dissolution is inhibited as soon as Ir is added to the composition while drastic increase of gas evolution is observed as the result of an intermediate phase that does form and becomes amorphous under harsh oxidizing conditions. DFT calculations rationalize the difference between Ru/Ir based materials as being nested in the stability of the end of charge compound, MO4. On one hand, RuO4 is stable versus gas release thanks to a phase transition from a phase having RuO6 octahedral (Oh) to RuO4 tetrahedral (Td) occurring at x = 1 which stabilizes the O-network. The RuO4-/RuO4 species are unfortunately soluble in current electrolytes therefore leading to a progressive dissolution of the electrode upon cycling. On the other hand, the Oh to Td phase transition is postponed to x = 0 for the Ir-based electrode, therefore preventing from electrode dissolution since LixIrO4 is unstable towards O2 formation, hence leading to gas evolution during oxidation. Of great interest as follow-up of this study, is the outcome of the dissolution process in solid state batteries, as it is being investigated. Whatever, we hope such finding provides useful guidelines for designing new practical cathode material for Li-ion batteries showing anionic redox.

Acknowledgments: The authors would like to thank A. Perez, A. Grimaud, W. Yin, M. Sathiya for fruitful discussion. The authors also thank V. Briois and S. Belin for helpful discussions on XAS analysis and synchrotron SOLEIL (France) for providing beamtime at the ROCK beamline (This work was supported by a public grant overseen by the French National Research Agency (ANR) as part of the “Investissements d’Avenir” program (reference : ANR10-EQPX45), proposal #20160095). Q. J. thanks the ANR “Deli-Redox” for PhD funding. J.-M.T. acknowledges funding from the European Research Council (ERC) (FP/2014)/ERC GrantProject 670116-ARPEMA.



Page 14 of 26

Author information: Corresponding author: [email protected] Supporting information: Crystallographic data, additional Rietveld refinements, EXAFS analysis. Author Contribution: Q.J. carried out the synthesis, Q.J. and L.L. conducted the electrochemical measurements, Q.J. and G.R. did the diffraction experiments and analysis, Q.J. and A.I. conducted the XAS measurements and analysis, M.S. and M.-L.D. performed the DFT study, E.J.B. conducted the OEMS experiments, D.A.D.C. performed the SEM, Q.J. and D.A.D.C. conducted to UV and MS measurements, Q.J., G.R., M.-L.D. and J.M.T. wrote the manuscript, and all authors discussed the experiments and edited the manuscript.

Bibliography (1) (2) (3) (4) (5) (6) (7)

Tarascon, J.-M.; Armand, M. Issues and Challenges Facing Rechargeable Lithium Batteries. Nature 2001, 414, 359–367. Lu, Z.; Dahn, J. R. Understanding the Anomalous Capacity of Li/LiNixLi1/3−2x/3Mn2/3−x/3O2 Cells Using In Situ X-Ray Diffraction and Electrochemical Studies. J. Electrochem. Soc. 2002, 149, A815-A822. Kim, J.-S.; Johnson, C. S.; Vaughey, J. T.; Thackeray, M. M.; Hackney, S. A.; Yoon, W.; Grey, C. P. Electrochemical and Structural Properties of x Li2M‘O 3 ·(1− x )LiMn0.5Ni0.5O 2 Electrodes for Lithium Batteries (M‘ = Ti, Mn, Zr; 0 ≤ x ⩽ 0.3). Chem. Mater. 2004, 16, 1996–2006. Mizushima, K.; Jones, P. C.; Wiseman, P. J.; Goodenough, J. B. LixCoO2 (0< X

Competition between metal dissolution and gas release in Li-rich

Recommend Documents