The Role of Manganese in Lithium- and Manganese-rich Layered

Subscriber access provided by BUFFALO STATE

Energy Conversion and Storage; Plasmonics and Optoelectronics

The Role of Manganese in Lithium- and Manganese-rich Layered Oxides Cathodes Laura Simonelli, Andrea Sorrentino, Wojciech Olszewski, Carlo Marini, Nitya Ramanan, Dominique Heinis, Angelo Mullaliu, Agnese Birrozzi, Nina Laszczynski, Marco Giorgetti, Stefano Passerini, and Dino Tonti J. Phys. Chem. Lett., Just Accepted Manuscript • DOI: 10.1021/acs.jpclett.9b01174 • Publication Date (Web): 29 May 2019 Downloaded from http://pubs.acs.org on May 30, 2019

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 30 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

The Journal of Physical Chemistry Letters

The Role of Manganese in Lithium- and Manganese-rich Layered Oxides Cathodes Dr. Laura Simonelli*, Dr. Andrea Sorrentino, Dr. Wojciech Olszewski, Dr. Carlo Marini, Dr. Nitya Ramanan, Dr. Dominique Heins, ALBA Synchrotron Light Facility, Carrer de la Llum 2-26, 08290 Cerdanyola del Vallès, Spain E-mail: [email protected] Dr. Angelo Mullaliu, Dep. of Industrial Chemistry Toso Montanari Univ. of Bologna, Viale del Risorgimento 4, 40136 Bologna, Italy Dr. Agnese Birrozzi, Dr. Nina Laszczynski, Helmholtz Inst. Ulm (HIU), Electrochemistry I Helmholtzstraẞe 11, 89081 Ulm, and Karlsruhe Inst. of Technology (KIT) PO Box 3640, 76021 Karlsruhe (Germany) Prof. Marco Giorgetti, Dep. of Industrial Chemistry Toso Montanari Univ. of Bologna, Viale del Risorgimento 4, 40136 Bologna, Italy Prof. Stefano Passerini, MEET Battery Research Centre, Inst. of Physical Chemistry Univ. of Muenster, Corrensstr. 46, 48149 Muenster, Germany; Helmholtz Inst. Ulm (HIU), Electrochemistry I Helmholtzstraẞe 11, 89081 Ulm, and Karlsruhe Inst. of Technology (KIT) PO Box 3640, 76021 Karlsruhe, Germany Dr. Dino Tonti Inst. de Ciència de Materials de Barcelona, Consejo Superior de Investigaciones Cientìficas, Campus UAB Bellaterra, Spain Keywords: Rechargeable Li-ion batteries, Li- and Mn-rich NMC cathodes, local Mn electronic properties, VOx-coating, strain effects

ABSTRACT: Lithium-rich transition-metal-oxide cathodes are among the most promising materials for the next lithium-ion-batteries generation because they operate at high voltages and deliver high capacities. However, their cycle-life remains limited and individual roles of the transition-metals are still not fully understood. By bulk-sensitive X-ray absorption and emission spectroscopy on Li[Li0.2Ni0.16Mn0.56Co0.08]O2 we inspect the behavior of Mn, generally considered inert upon the electrochemical process. During the first charge Mn appears to be redox-active showing a partial transformation from high-spin Mn4+ to Mn3+ in both high and low spin configurations, where the

ACS Paragon Plus Environment

The Journal of Physical Chemistry Letters 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

latter is expected to favor reversible cycling. The Mn redox-state along cycling continues changing in opposition to the expected charge compensation and is correlated with Ni oxidation/reduction, also spatially. The findings suggest the strain induced on the Mn-O sublattice by the Ni oxidation to trigger the Mn reduction. These results unravel the Mn role in controlling the electrochemistry of Li-rich cathodes.

The wide use of rechargeable lithium-ion batteries and the continuously growing demands of increased energy and power densities stimulate the investigation of novel high-voltage cathode materials.1, 2 Lithiated transition-metal-oxides are under intensive investigation as cathode materials for Li-ion batteries. The best performing cathodes show an ordered layered structure, which locates the Li ions in between the metal-oxygen layers.3, 4 Generally these materials offer the best cycle life when the layered structure is maintained during the delithiation/lithiation process. Among Mn, Co, and Ni, only Co3+ and Ni3+ enable 2D layered Li-based oxides. LiNiO2, however, has various drawbacks related to its crystal structure: difficulty to be synthesized, poor cycling performance, and poor thermal stability.5 As a matter of fact, LiCoO2 is the 2D layered oxide showing the best electrochemical performance and the most commonly used material in Li-ion batteries6. The


Page 2 of 30

Page 3 of 30 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


delithiation process in LiCoO2/C appears to be limited to slightly less than 0.5 Li, meaning that only about 50% of the total material capacity can be actually used, which corresponds to around 140 mAh/g. Attempts to go beyond such a limit are not reversible because of the strong repulsion between the CoO2 layers no longer sufficiently screened by the Li+ cations, finally resulting in the material degradation. To overcome this limitation, chemical substitution at the transition metal (M) sites has been explored, resulting in several M combinations offering decreased production costs, increased safety, and enhanced energy densities. In particular, when the Co3+ is partially replaced by Mn4+, Ni2+, and Li+, a new material is formed, herein called "Li-richNMC", with promising electrochemical performances.7-10 This new material combines the different beneficial effects of Ni, Co and Mn with the possibility of storing extra Li in the transition metal layers in addition to the Li present in the Van der Waals gap. Its structure can be described as the superposition of the layered rhombohedral structure with a monoclinic Li2MnO3 superlattice. It shows capacities exceeding 280 mAh/g,3 about twice that of conventional LiCoO2.6 Despite the great potential, the implementation of Li-rich NMC in practical Li-ion batteries is limited by the poor kinetics and reversibility and the large voltage decay upon cycling. To circumvent this issue, many synthesis efforts have been performed,4,

7-8, 11-21

however, understanding the problem from the structural/chemical point of view is still limited. The exploration of Li-rich materials formed by 4d rather than 3d metals suggests that to avoid the voltage decay upon cycling it is mainly necessary to control the size of the metal ionic radii to prevent the possible trapping of metal ions in interstitial tetrahedral sites.22 The relative M size depends on the atomic number, oxidation state, and spin state, and it has been proposed, ones compared with the alkaline metal size, to be the main parameter governing the structural aspects in layered oxides.3 In all Li-rich-NMC materials, the first charge involves severe structural and



chemical rearrangements that are still not fully understood. Some irreversible reactions occur as demonstrated by the disappearance of the voltage plateau at 4.4 V in the following discharge and the next cycles.13, 14 It has been proposed that, in the first charge, two reactions occur in series: (i) one involving the LiMO2 component, i.e., the Ni2+/Ni4+ and Co3+/Co4+ redox reactions, while Mn is expected to remain in the Mn4+ oxidation state, and (ii) a second one, involving the activation reaction of the Li2MnO3 phase, where manganese is not expected to change the oxidation state (Mn4+), but the oxide anion is oxidized with possible oxygen release. Furthermore, it has been shown that after the first delithiation, the following recharges are highly reversible,23-24 where the involvement of the Ni2+/Ni4+, Co3+/Co4+, Mn4+/Mn3+ redox reactions are expected on the basis of the experimentally released capacity.13 Despite all the experimental efforts, it looks that this model is not able to describe completely the charge compensation mechanism in Li-rich-NMC material, where the role of manganese, which at least during delithiation is generally foreseen to merely act as a spectator, is still controversial. In fact, some authors have reported an unexpected Mn reduction, mainly during the first charge, at the surface of the active material,25-27 associated to the formation of oxygen vacancies. Others proposed the formation of a little amount of Mn3+ in the bulk at the end of the first charge. 28-29 In this work we directly access the Mn electronic and magnetic properties evolution as a function of charge state in Li[Li0.2Ni0.16Mn0.56Co0.08]O2, a cobalt-poor and lithium-rich-NMC material which shows promising long term cycling and rate performances.23 We also investigate the influence of a VOx-coating on the electronic properties of the host compound during cycling. By mean of multiple synchrotron bulk-sensitive x-ray techniques, we finally reveal the Mn oxidation state evolution, surprisingly constantly opposite to the main charge compensation mechanism. The detected partial reduction of Mn during charge indicates a role of Mn in the charge


Page 4 of 30

Page 5 of 30 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


compensation mechanism and confirms as well the recent proposition that oxygen is oxidizing during charge, 28-30 in an even higher extent to compensate as well the Mn reduction. Moreover, the data suggest that the strain induced on the Mn-O network by the strong Ni size variation during its oxidation is triggering the Mn reduction at the nanoscale. Finally we report important and novel insides on the Mn spin state, which combination with the Mn oxidation state looks to be key in defining the Mn size and then in controlling the structural properties, i.e. the cathode performances. Mn and O K-edge X-ray absorption near edge structure (XANES) collected on the uncoated and VOx-coated samples at different states of charge are depicted in Fig. 1. The voltage profile upon the

first

charge/discharge

cycle

at

C/10

of

uncoated

and

VOx-coated

Li[Li0.2Ni0.16Mn0.56Co0.08]O2, is reproduced in Fig. 1g, where the charge points investigated in the present study are indicated, following the labeling of Ref. 13 (P01, P03, P04, P05, and P08). The charge capacity is higher for VOx-coated sample. It have been proposed that, upon cell operation, the coating delays the transformation of the active layer material into the spinel phase, occurring at low voltages, thus delaying its typical capacity fade upon cycling, and that contribute to the overall material capacity.14 Fig. S1 shows the voltage profile of the uncoated system, where also the charge points P12, and P16 have been indicated. Upon charge, from Fig. 1a,b it is possible to observe a partially irreversible increase of intensity of the whole Mn K-edge pre-edge region that most likely corresponds to increased deformations of the MnO6 octahedra and the consequent increase in the Mn 3d-4p orbital hybridizations.31 On the Mn K-edge absorption spectra, the energy position of the main- (Mn 1s → 4p) and pre-edge (Mn 1s → 3d / 4p) features, sensitive to changes in shielding of the nuclear charge provided by the valence electrons, may be used for the determination of the Mn valence. Unfortunately it is



not always straightforward, because these spectral features also depend on the local magnetic properties, site symmetry and nature of bonding with surrounding ligands (Fig. S6). Moreover, considering that the Li2MnO3-like component is expected to be fully converted only at the end of the first charge,32 the Mn site of the pristine electrode most likely corresponds to two components, the monoclinic Li2MnO3-like superlattice and the layered LiMn0.4Ni0.4Co0.2O2. In this complex scenario, complementary information can be extracted from the O K-edge XANES (Fig. 1c,d). The O K-edge absorption spectra have been extracted from full field transmission soft X-ray microscopy (TXM) images collected varying the energy across the O K-edge. The O K-pre-edge peak region (around 530 eV) corresponds to the transition from O 1s to hybridized states of M (M = Mn, Ni, and Co) 3d and O 2p orbitals. Similarly to the Mn, the O absorption pre-peak shows a not fully reversible trend, confirming the occurrence of some irreversible reactions along the first charge/discharge. Instead, the broad band above 534 eV corresponds to the hybridized M 4p and O 2p orbitals, which is expected to be sensitive to the oxygen oxidation state and the metal-ligand bond distance.33-34 Upon cycling, a reversible evolution of the broad band above 534 eV in the O K-edge spectra is detected. Within the ionic picture, this trend is compatible with both, the O2-/O- redox reactions, which have been recently proposed to take place during charge,28-30 and the shortening of Ni-O and Co-O bonds upon Ni and Co oxidation.35 Representative element maps obtained composing images at different energies are reported in Fig. 1e,f. Mn spatial distribution is reported in green, oxygen in red and vanadium (for the VOxcoated samples) in blue. The yellow color results from the superimposition of green and red. The appearance of systematic reddish borders on the imaged sample particles suggests a higher O to Mn ratio in these regions. Additional oxygen on the particle edges is evident on both


Page 6 of 30

Page 7 of 30 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


uncoated and VOx-coated samples because of a native precipitate, or the formations of a shell of solid electrolyte interphase (SEI) after exposition to the electrolyte, in agreement with previous work.36 This is confirmed by the spectral differences at the O K-edge pre-edge regions, bulk respect to the border of isolated particles (Fig. S15), which resulted smaller for the VOx-coating, supporting some hindered SEI formation from reaction with the electrolyte, much more significant with the uncoated samples. The difference of bulk and border spectra integral around the pre-edge region as a function of the charge state (Fig. S15g) strongly suggests formation of compounds with the oxygen expected to be released at the end of the voltage plateau in the uncoated material, but which seems suppressed by the VOx-coating.14 In the following we continue to focus on the pre-edge absorption features which are less influenced by the chemical environment being more directly connected to the M valence and magnetic state. In Fig. 1h-k and S2 we compare the O and Mn pre-edge peak features. The Mn K-edge spectra (black) have been shifted in energy for comparison. A common double peak structure is evident, which corresponds to the Mn4+ phase, as expected for the investigated systems.13,

37-38

We labelled as (II) and (IV) the two features composing the Mn4+ spectral

contribution. In the O pre-edge region more features can be appreciated. Feature (I) corresponds to the O 1s → Ni4+ 3d hybridized with O 2p transition,39-41 which is expected in the charge states in agreement with Ni K-edge absorption investigation.13 The presence of Mn3+ can be hypothesized in the cycled samples, 37 where features (III) and (V) become visible, despite the possible contributions from oxidized O species that may also overlap in this energy range, in agreement with the Mn L-edge spectra reported in Fig. S7. To disentangle the contribution of the different components from the two sets of collected spectra, the same 5 Gaussian model fit has been used to fit both the Mn and O K-edge pre-edge



peaks (details in supplementary Fig. S3, S4, S5 and Table S1). The empirically determined width values reproduce well the different energy experimental resolutions. The energy differences among (II), (III), and (IV) components are similar for the two sets of spectra as a function of charge, confirming the main Mn character of the corresponding excitations in the O K-edge spectra. The behavior of the individual Mn K-edge pre-edge peak features intensity is similar, and can be represented by the evolution of the total integrated intensity. It grows as a function of charge to decrease again upon discharge without, however, reaching the original value (insets in Fig. 1a,b). The uncoated and VOx-coated materials present a similar behavior, with, except for the pristine state (P01), systematically lower values for the VOx-coated samples. Delithiation (charge, P01 → P05) seems to favor the Mn 3d-4p hybridization, in a partially reversible way (P08 ≠ P01).31 The strains induced by the VOx-coating could potentially affect the Mn 3d-4p hybridization. From Mn K-edge extended X-ray absorption fine structure (EXAFS) data it results that the strains induced by the VOx-coating strongly affect the Mn local structure in the pristine compound, but it relaxes during the first charge cycle, being negligible in the charged state (Fig. S9). The detected Mn 3d-4p hybridization reduction by VOx-coating appears to not correlate to structural strains. The O pre-edge peak feature integrated intensities as a function of charge evolve differently, reflecting the main effect of the M 3d - O 2p hybridization. Being the Ni4+ contribution (I) isolated from the Mn one, the O K-edge data set allows to follow directly the Ni4+ formation. In Fig. 2a the evolution as a function of the charge state of the integrated intensity of feature (I) is reported. It appears that Ni2+ is partially irreversibly converted into Ni4+ during the charge. The


Page 8 of 30

Page 9 of 30 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


Ni oxidation results almost completed at the beginning of the voltage plateau for the uncoated sample, while it spans over the whole plateau for the coated one. Assuming features (II) and (III) to correspond principally to Mn4+ and Mn3+ phases,37 the relative intensity of the feature (II) is related to the Mn4+ fraction of Mn. In Fig. 2b the relative intensity of the feature (II) as a function of the charge state for the two sets of data shows similar trends, with the results obtained from the O absorption data being slightly affected by the presence of contributions coming from Ni and Co and by total absorption effects. Differences in the orbital hybridizations and density of states of the two Mn oxidation species are most likely introducing systematic errors which permit only qualitative considerations. The parallel evolution of the relative fractions of the (II) and (III) components at the Mn and O K-edge demonstrates that, during the first charge, Mn4+ is in average partially reduced to Mn3+, even if an overall oxidation process is taking place, as supported by the Mn L-edge spectra reported in Fig. S7. In P01 Mn appears to be in the pure Mn4+ phase, which is continuously, although partially, reduced upon charging, with a faster trend for the uncoated material. Upon the following discharge (reduction process) the produced Mn3+ is partially re-oxidized, which agrees with some previous reports.2527

In the following cycles, which seem to be reversible within the error bar (P12 and P16

compared with P05), the relative Mn4+ fraction reduces again in favor of the Mn3+ one during the charges, contrarily to the main charge compensation process and to what assumed in the past. The Mn evolution, constantly opposite to the main charge compensation mechanism during lithium extraction and insertion, can be considered as a collateral effect occurring in parallel to Ni, Co and O oxidation and subsequent reduction. Interestingly, by comparing the evolution as a function of the charge state of the spectra collected over uncoated and VOx-coated samples, the Ni2+/Ni4+ redox reaction (Fig. 2a) seems correlated



to the observed Mn reduction (Fig. 2b). Fig. 2c reports the relative intensity of the feature (II) as a function of the integrated intensity of feature (I). A linear tendency is clearly visible, evidencing the simultaneity of the two effects, whose overall evolution is different in uncoated and VOx-coated samples. TXM reveals that there is also a spatial correlation between Mn3+ and Ni4+ states. We exploited the images collected around the (I), (II), and (III) spectral features to build maps of the total Mn, relative Ni4+, and relative Mn3+ distributions. In Fig. 3a,b are reported the images representing the relative Ni4+ distribution for the charge point P04 for uncoated and VOx-coated cathodes. Ni4+ results clearly present in higher amount in the bulk of the uncoated particles, which is also consistent with the results reported by Gent and coworkers.42 For the VOx-coating instead the distribution seems more homogeneous. Fig. 3c,d reports the distribution profiles of the total Mn, relative amount of Ni4+, and relative amount of Mn3+ on representative isolated particles (300150 nm) for uncoated and VOx-coated system, respectively. The distribution profiles reflect the concentration of the Ni4+ in the particle bulk for the uncoated and a more homogeneous distribution for the VOx-coated samples, while the Mn3+ follow fairly closely the Ni4+ corresponding profiles, proving that Mn reduction is a bulk phenomenon. The correlation between Ni4+ and Mn3+ along the different samples and in the space suggests a possible causeeffect, which could be explained by the idea that the strain induced on the Mn-O network by the strong Ni size variation during its oxidation is triggering the Mn reduction. The strain induced by the VOx coating could instead hamper such structural rearrangements and cause the more even Ni4+ and Mn3+ distribution within the particles. Such stiffness is likely the same that delays Ni4+ formation as seen in Fig. 2a.


Page 10 of 30

Page 11 of 30 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


Complementary information on the Mn electronic structure and its evolution as a function of the charge state and coating have been obtained by looking at the Mn Kβ emission line (Fig. 4a). We quantify the local Mn magnetic moment (µMn) from the integrated area of absolute X-ray emission Kβ line difference (IAD) with respect to a reference.43 For this purpose we compared the reported spectra to the MnO and MnO2 references. The energy position of the Kβ1,3, directly related to the spin state reflecting the effective number of unpaired 3d electrons,44 shows a similar trend as a function of the charge state, suggesting that the detected variations correspond to actual specific magnetic changes (Fig. S7). µMn is equal in the P01 state of the uncoated and VOx-coated materials. By delithiation the µMn of the VOx-coated samples stays almost unchanged, while the one of the uncoated samples significantly drops in an irreversible way. This points to different proportions of high and low spin configurations within the produced Mn3+ states in uncoated and VOx-coated samples. Generally, Mn4+ and Mn3+ in octahedral coordination are both in high spin (HS) configuration, with the latter showing a higher spin state.45-47 The HS Mn3+ Jahn-Teller active phase, has been correlated to the capacity loss in spinel LiMn2O4, with the Jahn–Teller distortions destabilizing the structure during electrochemical cycling.48-50 The strong µMn drop upon charge in the uncoated material seems to corresponds to the formation of a Mn environment corresponding to that of layered rhombohedral r-LiMnO2, where theoretical calculations predict the Mn3+ in the not Jahn–Teller active low spin (LS) configuration.47 Interestingly, the theoretical simulation of this phase predicts several potential advantages for a good electrochemical activity and agrees with the reported results, like it explains the shift of the main Mn K-edge absorption feature towards higher energy during the charge, generally associated to oxidation, while Mn is actually partially reducing. More details are reported in the supplementary information.



In the uncoated material, while the Mn4+ → Mn3+ transformation looks continuous during the first charge (Fig. 2b), the strong µMn drop indicates that the main partially irreversible formation of the LS Mn3+ phase occurs between P01 and P03, where also the Co3+/Co4+ and Ni2+/Ni4+ redox reactions occur (dashed region in Fig. 4c). During the voltage plateau, Mn4+ is instead partially converted to a mixture of LS and HS Mn3+ since the average µMn is not changing substantially. In the following cycles it appears that the LS Mn3+ phase is partially reversibly cycling, LS Mn3+ → HS Mn4+. Instead, in the VOx-coated material, in the first step (P01 → P03), Ni and Mn seem almost inactive, and the clear formation of LS Mn3+ is suppressed, most likely because of the strain induced by the coating. From our results it is tempting to hypothesize that, while increasing the extent of the Mn3+ HS formation, the coating is however suppressing the detrimental Jahn Teller distortions that would be expected with the HS configuration and are naturally absent in the Mn3+ LS. In fact, the EXAFS Fourier transforms (FTs) of P05 are similar for the two systems, highlighting a similar averaged Mn local structure (Fig. S9). In the VOx-coated case the Mn4+ → Mn3+ transformation leads to the Mn3+ phase formation in a mixed spin configuration mainly along the voltage plateau, where also the Ni2+/Ni4+ redox reaction occurs (dashed region in Fig. 4d). At the end of the voltage plateau (P04 → P05) µMn appears to irreversibly increase, suggesting the formation of an inactive HS Mn3+ phase. Only part of the formed Mn3+ seems available for the subsequent reversible cycling, similarly to the uncoated case. Finally, the position of the absorption feature (II), at slightly higher energy (Fig. S5a), and the increase of the Mn 3d-4p hybridization in the uncoated cathodes (Fig. S5c) most likely correspond both to the presence of the Mn3+ LS phase.


Page 12 of 30

Page 13 of 30 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


The reported results are globally in agreement with the qualitative evolution of the EXAFS data (Fig. S9). The intensity of the FT Mn-O peak decreases as a function of charge, highlighting an increasing disorder in the Mn-O shell, while the occurrence of coexisting different Mn sites by charge is expected to globally increase the local structural disorder as well as the increase of local distortions. As well the partially irreversible formation of Mn3+ in the charge state is in agreement with the EXAFS data which show a no completely reversible evolution of the local structure by charging. In conclusion, the local manganese electronic and magnetic properties of uncoated and VOxcoated Li[Li0.2Ni0.16Mn0.56Co0.08]O2 materials have been studied in detail as a function of the charge state in a lithium battery. The active role of Mn in the lithiation/delithiation process is identified, with Mn oxidation state evolving constantly in opposition to the main charge compensation mechanism. This supposes a supplementary oxidation localized on the other elements, including O, by charging. In parallel, the evolution of the Mn spin state as a function of the charge state in the uncoated and VOx-coated systems is unveiled. The Mn oxidation and spin states are key parameters in controlling the reaction reversibility in lithium-rich NMC cathodes. Both these parameters are linked to the Mn size and magnetic interactions, which end determining the stability of the alternative possible phases in which this cation may be involved.3, 50 In fact, it has been shown as well that the Mn spin polarization has a significant effect on the energies and relative stability of different structures.50 Importantly, we have observed that both Mn oxidation and spin state are strongly affected by the strains produced, for instance by coating or by the oxidation/reduction of Ni. In particular, the results suggest the strain induced on the Mn-O network by the strong Ni size variation during its oxidation to trigger the partially irreversible Mn reduction and to favor a Mn3+ LS phase formation in the bulk. A HS



Mn3+ phase, which could be related to spinel, most likely forms mainly during and at the end of the voltage plateau. The VOx coating, most likely participating in the redox reaction,14 appears to slower the Ni2+/Ni4+ and the consequent Mn4+/Mn3+ redox reactions. Moreover it looks as well to favor the Mn3+ HS phase, most likely because of the induced strains. This mechanism, either if further systematic studies are necessary to address quantitatively how the strain controls the Mn oxidation and spin state, provides a fundament to elaborate design strategies that can be used to control the structure, for instance hindering the spinel formation at the benefit of the electrode cycle life.

EXPERIMENTAL METHODS Materials. The previously characterized uncoated13 and VOx-coated14 Li[Li0.2Ni0.16Mn0.56Co0.08]O2 samples were herein studied via ex-situ techniques using samples at different state of charge taken from electrochemical cells assembled as detailed earlier. The samples were extracted from cycled electrodes after dismantling the respective cells inside an Ar filled glove box, and washed with dimethylcarbonate (Sigma-Aldrich) in order to minimize interferences from salt and other possible soluble species in the electrolyte. The so extracted electrodes were loaded on the sample holder for the measurements. The samples investigated in the present study cover several charge/discharge cycles, focusing, however, over the first charge step. In particular a few key state of charge points have been identified (see also Fig. S1), which are the pristine state (P01), the begin (P03) and the end (P04) of the high voltage plateau and the fully charged state (P05) during the first delithiation step, and the following fully discharged (P08), as well as fully charged after 5 (P12) and, again, fully charged after 50 (P16) cycles.


Page 14 of 30

Page 15 of 30 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


Hard X-ray absorption and emission measurements. Mn K-edge X-ray absorption and Kβ X-ray emission measurements were performed at the CLÆSS beamline of the ALBA CELLS synchrotron (Spain).51 The synchrotron radiation emitted by a wiggler source was monochromatized using a double crystal Si(311) (Si(111)) monochromator for absorption (emission) measurements. Higher harmonics were rejected choosing proper angles and coatings of the collimating and focusing mirrors. The absorption data were collected in transmission mode using three ionization chambers, for simultaneous measurements of the sample and the Mn foil used as a reference, within an incoming energy resolution below 0.3 eV. The spectra have been normalized with respect to the atomic absorption jump established by a linear fit far away from the absorption edge. The emission data were collected in back scattering horizontal geometry (parallel to the linear polarization vector of the incoming X-ray beam) by means of the CLEAR emission spectrometer. The spectrometer is based on a diced Si(333) analyzer crystal (bending radius R = 1 m) and a 1D position-sensitive Mythen detector. The emission spectra were acquired exciting the sample well above the Mn K edge and detecting the emitted Mn Kβ emission lines with a total energy resolution around 1 eV. The measurements were performed at ambient temperature under vacuum condition. Several Xray absorption and emission scans were measured to ensure the reproducibility of the spectra and to obtain high signal-to-noise-ratio. Soft X-ray absorption measurements. Energy-resolved soft X-ray transmission microscopy (ER-TXM) was performed at the MISTRAL beamline of the ALBA synchrotron.52 Samples were scratched off from the electrode and deposited on carbon-coated Au TEM grids inside an Ar filled glove box. Dimethylcarbonate (Sigma-Aldrich) is used to wet the sample powder improving its adhesion on the carbon-coated



TEM grid used as sample holder. The grids were then stored under Ar in cryogenic vials and transferred in the MISTRAL microscope in cryogenic condition (T < 110 K) under N2 vapor to avoid atmospheric contamination. The samples were kept at cryogenic temperature and under high vacuum conditions during all the measurements. Representative regions of the samples were selected from areas of 100 µm x 100 µm recorded at different zones of the sample by the means of composite or mosaic images just at two energies, above and below the O K-edge for a first localization of oxygen compounds. On the selected areas transmission images (2 s exposure time, effective pixel size 10 nm, field of view (FOV) 10 µm x 10 µm) were collected varying the energy across the O K-edge with a variable spectral sampling (0.5 - 0.1 eV). The energy resolution was determined at the nitrogen K-edge measuring the visibility of Nitrogen gas vibrational peaks at the exit slit of the monochromator following the strategy proposed in ref. 53. The energy resolution at the O – K edge was extrapolated to be around 0.5 eV (full width half maximum). A preliminary calibration of the absolute value of the energy was carefully performed before the experiment using standard references samples at different energies along the available energy range of the beamline (CaCO3, N2, TiO2, Mn2O3, and Fe2O3). No energy shift was applied to the measured sample spectra. The objective zone plate lens (outermost zone width of 25 nm, 1500 zones) and the back illuminated CCD detector (Pixis XO by Princeton Instruments with 1024 x 1024 pixels and 13 µm pixel size) positions were automatically adjusted to maintain the sample in focus and constant magnification (= 1300). The spatial resolution of the system was estimated at 520 eV using a Siemens star pattern with 30 nm smallest features to be 23 nm half pitch.54 The necessary total acquisition time for a single energy stack of transmission images was about 1.5 hours, including the flat field acquisition at each energy step. After normalization, alignment and conversion to absorbance of the transmission energy stack,


Page 16 of 30

Page 17 of 30 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


spectra were extracted only from pixels in the field of view (FOV) with a Signal to Noise ratio ≥ 2, where the Noise is defined as the average single pixel spectra standard deviation in the preedge energy region (< 526 eV) and the Signal is defined as the difference between the absorbance average value in the post edge energy region (> 544 eV) and the average value in the pre-edge energy region. The Noise is mainly due to small instabilities in the incident beam and is around 0.05 for all the measured spectra. The reported spectrum for each sample is then calculated as the average of the extracted single pixels spectra from the full FOV (fig. 1c,d) or from particular region of interest (fig. S15a,f). Spectra have been normalized with respect to the atomic absorption jump established by a linear fit far away from the absorption edge. The experimental error on the extracted spectra from the full FOV is discussed in the supplementary information.

Acknowledgements This work has been partially financed by the Spanish Ministry of Economy and Competitiveness, through the Severo Ochoa Programme for Centres of Excellence in R&D (SEV- 2015-0496). The HIU authors kindly acknowledge the basic funding of the Helmholtz Association.

Supporting Information Available: Methods, Gaussian model fit, Mn oxidation state evaluation, Mn local magnetic properties, local structure, soft X-ray elemental maps, Soft X-rays energy stack pre-treatment, Soft X-rays spectra experimental error, O K-edge spectral differences along isolated particles: border vs bulk

References



[1]

Page 18 of 30

J. M. Tarascon, M. Armand, Issues and Challenges Facing Rechargeable Lithium

Batteries, Nature 2001, 414, 359. [2]

T.-H. Kim, J.-S. Park, S. K. Chang, S. Choi, J. H. Ryu, H.-K. Song, The Current Move of

Lithium Ion Batteries Towards the Next Phase, Adv. Energy Mater. 2012, 2, 860. [3]

P. Rozier and J. M. Tarascon, Li-Rich Layered Oxide Cathodes for Next-Generation Li-

Ion Batteries: Chances and Challenges, Journal of The Electrochemical Society 2015, 162 (14), A2490. [4]

J. Wang, X. He, E. Paillard, N. Laszczynski, J. Li and S. Passerini, Lithium ‐ and

Manganese ‐ Rich Oxide Cathode Materials for High ‐ Energy Lithium Ion Batteries, Adv. Energy Mater. 2016, 6, 1600906. [5]

C. Delmas, J. P. Peres, A. Rougier, A. Demourgues, F. Weill, A. Chadwick, M.

Broussely, F. Perton, Ph. Biensan, and P. Willman, On the Behavior of the LixNiO2 System: an Electrochemical and Structural Overview, J. Power Sources 1997, 68, 120. [6]

K. Mizushima, P. C. Jones, P. J. Wiseman and J. B. Goodenough, LixCoO2 (0< x

The Role of Manganese in Lithium- and Manganese-rich Layered

Recommend Documents