Impact of Fluorination on Phase Stability, Crystal Chemistry, and

Subscriber access provided by UNIV OF DURHAM

Article

Impact of Fluorination on Phase Stability, Crystal Chemistry, and Capacity of LiCoMnO High Voltage Spinels 4

Anna Windmüller, Craig A. Bridges, Chih-Long Tsai, Sandra Lobe, Christian Dellen, Gabriel M. Veith, Martin Finsterbusch, Sven Uhlenbruck, and Olivier Guillon ACS Appl. Energy Mater., Just Accepted Manuscript • DOI: 10.1021/acsaem.7b00186 • Publication Date (Web): 02 Jan 2018 Downloaded from http://pubs.acs.org on January 5, 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 free 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 accessible to all readers and 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.

ACS Applied Energy Materials 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 41 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

ACS Applied Energy Materials

Impact of Fluorination on Phase Stability, Crystal Chemistry, and Capacity of LiCoMnO4 High Voltage Spinels Anna Windmüller*,†,‡, Craig A. Bridges#, Chih-Long Tsai†,‡, Sandra Lobe†,‡, Christian Dellen†,‡, Gabriel M. Veith§, Martin Finsterbusch†,‡, Sven Uhlenbruck†,‡, Olivier Guillon†,‡ †

Forschungszentrum Jülich GmbH, Institute of Energy and Climate Research, Materials

Synthesis and Processing (IEK-1), 52425 Jülich, Germany ‡

Jülich Aachen Research Alliance: JARA-Energy, 52425 Jülich, Germany

#

Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831,

United States §

Materials Sciences and Technology Division, Oak Ridge National Laboratory, Oak Ridge,

Tennessee 37831, United States *corresponding author, [email protected] KEYWORDS : lithium battery, high voltage spinel, neutron diffraction, synchrotron diffraction, Raman spectroscopy

ACS Paragon Plus Environment

1

ACS Applied Energy 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 41

ABSTRACT: Fluorinated LiCoMnO4-yFy (y = 0, 0.05, 0.1) spinel electrodes, electrochemically active at 5 to 5.3 V vs. Li/Li+, show enhanced phase purity and enhanced capacity with increasing y. We disclose the impact of fluorination on the phase purity and reversible capacity of LiCoMnO4 via joint Rietveld refinement of neutron and synchrotron powder diffraction data, combined with micro Raman spectroscopy. It is found that fluorination stabilizes the spinel phase and hinders precipitation of Li2MnO3 as secondary phase, which controls the cation distribution on tetrahedral and octahedral sites in spinel. That is to say, for higher fluorine content the cobalt occupancy at the tetrahedral site in spinel decreases and the lithium occupancy increases. Accordingly, the number of lithium-sites that are available for electrochemical extraction and insertion of lithium-ions are raised so that the capacity is increased. Further investigation of the lithium-ion diffusion by means of cyclic voltammetry at different scan rates and the application of the Randles-Sevcik equation were carried out to investigate the extent of capacity enhancement due to faster lithium-ion diffusion in the high voltage region.

Introduction In the discussion about advanced battery technologies for increased energy and power density, one material class is never missing: the spinel electrode for lithium ion batteries1–4. Spinel electrodes allow for high energy densities and high rate capabilities, because of their relatively high operational voltage and three-dimensional lithium ion diffusion paths5–8. Among the spinel electrodes, LiCoMnO4 exhibits the highest theoretical energy density, due to its electrochemical activity at 5 and 5.3 V vs. Li/Li+ and its theoretical specific capacity of 145 mAhg-1 provided by the Co3+/4+ redox couple6. As a result, its energy density is 10 % higher than that of the popular LiNi0.5Mn1.5O4 spinel and around 25 % higher than that of the conventional LiCoO2 electrode1,6.


2

Page 3 of 41 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


Considering these outstanding features, LiCoMnO4 represents a promising electrode candidate for future lithium ion9,10 and post lithium ion technologies11,12. LiCoMnO4 spinel electrodes, as all lithium manganese based spinel electrodes, are isomorphous to the mineral spinel MgAl2O413. Spinels are generally described in the space group Fd-3m13,14. If the origin of the unit cell chosen is at -3m, the 8a Wyckoff position, at 1/8, 1/8, 1/8, is a tetrahedrally coordinated cation site, while 8b is a tetrahedral vacant site. Similarly, the octahedrally coordinated 16d site, at 1/2, 1/2, 1/2, is occupied, while the 16c site, at 0, 0, 0 is vacant13. In LiCoMnO4, lithium ions occupy the 8a tetrahedral site, while the cobalt and manganese ions occupy 16d octahedral site15. For spinels, the 32e Wyckoff position, at x, x, x, is an anionic site. In the ideal case of x = 0.25, the anion sublattice is cubic closed packed. Changes in the anionic 32e position reflect dilation of the anionic sublattice and depend upon the dimensions of 8a and 16d sites13. For example, as the anion x parameter increases in value, the LiO4 tetrahedron increases in size while the (Co/Mn)O6 octahedron decreases in size. Typically, the 32e site is reported to be near x = 0.26 in lithium manganese based spinels15–20. Lithium ion diffusion, during the electrochemical process, takes place via site hopping along the 8a and empty 16c sites. 8a and 16c sites share a triangular face, spanned by three oxygen atoms. The dimension of the shared face influences the activation energy for the lithium ion hopping mechanism. Additionally, the activation energy is controlled by electrostatic repulsion of nearest neighbors on tetrahedral and octahedral sites, i.e. lithium and transition metal ions21. Hence, lithium ion diffusion can be controlled by the crystal chemistry of the spinel, such as lattice dimensions and positions of oxygen atoms, as well as valences, ionic radii and bond lengths of cationic species and their distribution and ordering on tetrahedral and octahedral sites17,20–22.


3


Page 4 of 41

Although LiCoMnO4 offers a very promising performance from a theoretical point of view, the LiCoMnO4 electrode in practice hardly exceeds 75 % of its theoretical capacity and suffers from rapid capacity degradation23–26. Additionally, Li2MnO3 is well known to emerge as a secondary phase15,23. The precipitation of Li2MnO3 takes place during the synthesis of lithium manganese based spinels at elevated temperatures, due to the loss of oxygen from the spinel lattice27. Oxygen loss from LiCoMnO4 samples is known to start at around 600 °C in air, accompanied by the reduction of Co3+ and Mn4+ to lower valence ions. Beyond 680 °C, Li2MnO3 coexists with a cobalt-rich and lithium-, manganese- and oxygen-depleted spinel15. Recently, we showed that fluorination of LiCoMnO4, with up to 2.5 mol-% fluorine at the anionic site, enhances its performance28. Fluorination helps to stabilize the spinel structure, which leads to higher phase purity. Accordingly, less Li2MnO3 was found in the fluorinated samples than in the non-fluorinated ones under the same synthesis conditions. Secondly, the cycling stability of fluorinated samples was improved, which agrees with other studies on the fluorination of e.g. LiNi0.5Mn1.5O4 and LiMn2O429–33 and is believed to be an effect of an enhanced structural stability during cycling as a result of fluorination33. Thirdly, the reversible capacity was increased significantly, by up to 18 %, for 2.5 mol-% fluorine ions at the anionic site. Although enhanced phase purity impacts the reversible capacity, it was found that the capacity increase upon fluorination cannot be explained by enhanced phase purity alone. Increased capacity upon fluorination has been reported earlier for spinels, such as LiNi0.5Mn1.5O4, and has been discussed as an effect of increased Mn3+ content as a result of aliovalent substitution of O2- by F- 32,34. Mn3+, in contrast to Mn4+, is an electrochemically active species in spinels. Hence, more capacity is gained for fluorinated LiNi0.5Mn1.5O4-yFy containing Mn3+. However, in our previous electrochemical measurements, no evidence was found for an


4

Page 5 of 41 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


increase in Mn3+ content upon fluorination of LiCoMnO4. This means that the improved performance of fluorinated LiCoMnO4 electrodes in terms of their capacity enhancement was not simply attributed to the increase of Mn3+. But understanding how fluorination helps to increase the LiCoMnO4 capacity can impact the development of spinel type electrodes. In this study, a combination of crystallographic, spectroscopic and electrochemical methods was applied in order to understand the capacity enhancement of fluorinated LiCoMnO4 electrodes. A joint synchrotron and neutron powder diffraction analysis was carried out and combined with micro Raman spectroscopic studies in order to reveal the crystallographic and local chemical properties of fluorinated and non-fluorinated LiCoMnO4. Complementary, electrochemical measurements were carried out to investigate the lithium ion diffusivity in fluorinated LiCoMnO4 electrodes. As a result, a complete picture of the crystal chemistry of fluorinated LiCoMnO4 electrodes is given and an explanation of their enhanced capacity is derived.

Methods Synthesis Powders with the nominal composition LiCoMnO4 (LCMOF00), LiCoMnO3.95F0.05 (LCMOF05) and LiCoMnO3.9F0.1 (LCMOF10), were synthesized by solid state reaction at 800 °C for 10 h from LiOH·H2O (AppliChem, 99 %), MnCO3 (Sigma-Aldrich, 99.9 %), Co3O4 (Alfa Aesar, 99.7 % ) and LiF (Alfa Aesar, 98.5 %) as starting materials. An additional step was added at 650 °C for 10 h to allow for oxygen uptake. The procedure and the samples are identical to the samples used in our previous study, where we confirmed the stoichiometry to be Li1.1CoMnO3.6, Li1.1CoMnO3.6F0.04 and Li1.1CoMnO3.6F0.09, respectively28.


5


Page 6 of 41

Powder Diffraction Time-of-flight neutron diffraction data were collected at the POWGEN beamline at the Spallation Neutron Source, Oak Ridge National Laboratory, USA. The samples were pre-dried at 100 °C for five hours and were loaded into vanadium cans in an argon filled glove box. Data were collected in a Q-range from 1 - 15 Å-1 at 300 K. For synchrotron powder diffraction measurements, the samples were loaded into Kapton tubes prior to the measurements. The measurements were performed at the beamline 11-BM of the Advanced Photon Source, Argonne National Laboratory, at a wavelength of 0.5170 Å in a Q-range from 1 - 10.3 Å-1 at 295 K. Data sets from neutron and synchrotron powder diffraction were refined jointly using the GSAS software package35 in the EXPGUI user interface36. The starting model consisted of a LiCoMnO4 phase in Fd-3m structure (origin at -3m), with lattice parameters as derived previously (8.067, 8.065 and 8.064 Å, for LCMOF00, LCMOF05, and LCMOF10, respectively)28 and a Li2MnO3 secondary phase in space group C2/m based on the structural data from Boulineau et al. (2009)37. Raman Spectroscopy Micro Raman spectroscopy was carried out with a Renishaw inVia Raman microscope using a solid state 532 nm excitation laser and a 1800 l/mm grating. To avoid laser induced phase changes or damage to the samples, the laser power was set to 0.25 mW. The Raman spectra were collected with a line focus lens, on an area of 100 µm × 60 µm and a 50× objective. The measurement was carried out with a 30 second acquisition time and a 1 µm step size, which yielded 6000 scans. The collected scans were averaged, baseline corrected, normalized to their integral in between 340 - 740 cm-1, and fitted in the range from 340 - 740 cm-1 by using Pseudo Voigt functions within the WiRE 4.4 software (Renishaw).


6

Page 7 of 41 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


Additionally, a high-resolution Raman mapping was carried out by applying the live track function (an auto focus software solution) within the WiRE 4.4 software. The spectra were collected for an area of 37 µm × 18 µm by using a 100× objective. A step size of 0.3 µm and a 3 second spectral acquisition time was chosen for maximal resolution and minimal time per spectra. The laser power was set to 0.25 mW. The high-resolution mapping was processed and analyzed within the WiRE 4.4 software, by using a statistical noise filter and a statistical component analysis algorithm. Electrochemical Measurements To investigate the lithium ion diffusivity, cyclic voltammetry (CV) measurements were carried out using liquid electrolyte based battery cells. First, cathode slurries were prepared by mixing the powder samples of LCMOF00, LCMOF05 and LCMOF10 with 10 wt.-% carbon black (BLACK PEARLS; Cabot) and 15 wt.-% polyvinylidene fluoride (PVdF) (Alfa Aesar) dissolved in N-methyl-2-pyrrolidone (Alfa Aesar). The slurries were painted on stainless steel current collectors, yielding ~2 mg·cm-² of active material. After drying at 80 °C for 18 hours, the cells were assembled using lithium foil as anode and 1M LiPF6 in ethylene carbonate - diethyl carbonate (EC/DEC, volume ratio 1:1) as liquid electrolyte. The CV scans were performed with a VMP-300 (Bio-Logic) in between 3.4 and 5.3 V vs. Li/Li+ at scan rates of 0.1, 0.13, 0.16, 0.21, 0.4 and 0.65 mVs-1.

Experimental results Rietveld Refinement The main phase in the LCMOF00, LCMOF05 and LCMOF10 sample was indexed to the Fd-3m spinel structure (Figure 1 for LCMOF00 and Figure S1 for LCMOF05 and LCMOF10).


7


Page 8 of 41

Additionally, a monoclinic Li2MnO3-like phase, which was identified from its 100, 020, 110, 111 and 021 reflections, was found. Upon fluorination, the Li2MnO3 phase content decreases (figure 1d and table 1).

Figure 1. Combined Rietveld refinement of LCMOF00 a) neutron diffraction (inset from 1 - 10.3 Å-1, actual fitting range from 1-15 Å-1), b) synchrotron diffraction, full Q-range, c) Visualization of spinel Fd-3m structure and cation distributions, d) cation distributions of spinel phase and phase content of Li2MnO3 (C2/m ) as a function of fluorine content - error bars are not displayed in Figure 1d, since they are smaller than the symbols.


8

Page 9 of 41 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


Table 1: Lattice parameters, phase content, and goodness of fit obtained by the Rietveld refinement of LCMOF00, LCMOF05, and LCMOF10.

Lattice parameter of main phase [Å]

Weight fraction of main phase [wt.-%]

Goodness of fit χ2

LCMOF00

8.06705(2)

89.080(10)

7.7

LCMOF05

8.06588(3)

94.290(5)

5.4

LCMOF10

8.06520(2)

95.687(3)

5.3

Sample

Li2MnO3 is structurally related to LiCoMnO4 spinel and most of its main Bragg peaks superimpose on the spinel Bragg peaks. In order to eliminate false intensity contributions to the spinel Bragg peaks, it is important to describe the Li2MnO3 contribution to the pattern very carefully. However, Li2MnO3 is known to show stacking faults, which especially affect the peak intensity ratios and broadness of the 020, 110, 11-1, 021, and 111 reflections in the low Q-range from 1.4 - 2 Å-1 38. To describe the peak profiles of Li2MnO3 in the neutron and synchrotron scattering data, the profile functions were fixed at instrument resolution, with the addition of simple Lorentzian broadening for the X-ray data and quadratic broadening for the neutron data. This clearly led to an insufficient profile fit of the weak, asymmetric monoclinic Bragg peaks in the low Q-range, where the peak profiles display a Warren-type line shape as a result of stacking faults39 (Figure 1b). However, this provided an adequate fit to the more symmetric low Q-range 00l reflections. As a result, Li2MnO3 Bragg peaks in the higher Q-range, where significant superimposition of spinel and monoclinic peaks occurred, can be described sufficiently well to provide an estimate of the phase fractions and refine the atomic parameters for the spinel phase. The joint refinement of neutron and synchrotron data holds the advantage to investigate the cation distribution of the LCMOF spinel phase very precisely, given the contrast of electron


9


Page 10 of 41

density for transition metal ions and lithium ions for X-rays and the contrast of atomic scattering lengths for cobalt (b = 2.49 fm) and manganese (b = -3.73 fm) for neutrons. Therefore, the site occupancy of lithium and cobalt cations on the tetrahedral 8a site, as well as manganese and cobalt cations on the octahedral 16d site were refined constrained to full occupancy (i.e. occupancy Li + Co = 1 and Mn + Co = 1, respectively). The distribution of fluorine and oxygen at the anionic 32e site cannot be refined, since the electron density and the atomic scattering length for neutrons are similar for oxygen and fluorine. Thus, their contrast is too low to be distinguished. Therefore, the fluorine occupation was fixed to 0.01 and 0.0225 according to the quantification of fluorine contents by nuclear reaction analysis in our previous study28, assuming that all fluorine is incorporated into the spinel phase. The oxygen occupancy parameter was found to be strongly correlated to the peak profile function parameters, the scale factors and atomic displacement parameters (Uiso). That is why the oxygen ion occupation parameter was first fixed to yield full occupation of the 32e site (i.e. O + F = 1), while all other parameters were refined. After the fit had converged, all parameters were fixed and the oxygen occupancy was freed to converge. In Figure 1d and Table S1, the refined crystallographic data of the main phases in the samples are summarized. For all samples, the structural parameters converged to a very similar unit cell in terms of atomic positions: 8a and 16d Wyckoff positions are special positions and remain constant; the general 32e oxygen atomic positions were refined freely and converged to relatively similar values of 0.26318 ± 1 × 10-5 for LCMOF00 and 0.26304 ± 1 × 10-5 for LCMOF05 and LCMF10. The cation distributions on octahedral and tetrahedral sites show changes with respect to the fluorine content, such that the occupation of cobalt cations on 8a and 16d sites decreases for


10

Page 11 of 41 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


higher fluorine content in the samples. To be specific, LCMOF05 has two percentage points less cobalt cations at both the 8a site and the 16d site as compared to LCMOF00. LCMOF10 has two percentage points less cobalt cations at the 8a site as compared to LCMOF05, while the cation distribution at the 16d site is similar within the errors of the refinement. In the subsequent refinement of the oxygen occupancy, the parameters shifted slightly away from their starting values to lower ones, i.e. 0.993 ± 0.001 , 0.985 ± 0.001 and 0.972 ± 0.001 for LCMO00, LCMOF05 and LCMOF10 respectively. This corresponds to a shift of ~ 0.006 ± 0.001 for all samples. Raman Spectroscopy Cubic spinel, in Fd-3m structure, possesses the Oh7 spectroscopic symmetry and is expected to show five Raman active modes. Three modes are correlated to vibrations of the transition metal octahedra: one symmetric stretching mode (A1g) and two asymmetric bending modes (F2g(1), F2g(2)). The other two modes belong to vibrations of Li-O bonds: one asymmetric bending mode (F2g(3)) and one symmetric stretching mode (Eg)40. Furthermore, the secondary phase Li2MnO3, which is highly Raman active, is expected to show at least six Raman active modes41. Figure 2 shows the experimental and fitted Raman spectra based on the averaged and normalized data collected for a sample area of 60 µm × 100 µm. In the LCMOF00 sample, five bands corresponding to the expected modes for spinel are found at 650 (A1g), 570 (F2g(1)), 535 (F2g(2)), 468 (F2g(3)) and 379 (Eg) cm-1. Additionally, five bands corresponding to Li2MnO3 are recorded at 613, 496, 438, 415, and 370 cm-1, numbered from v1 to v5.


11


Page 12 of 41

Figure 2. Results of the micro Raman spectroscopy for a sample area of 60 × 100 µm. Averaged, normalized and fitted spectra for a) LCMOF00, b) LCMOF05 and c) LCMOF10. Furthermore, the spectrum of LCMOF00 shows a contribution of two low intensity bands that cannot be assigned to bands of spinel or Li2MnO3. These bands are centered at 594 cm-1, visible as a shoulder at v1, and at 692 cm-1, visible as a shoulder at the high energy side of the A1g mode. In literature, the latter one was found for a LiCoMnO4 sample as well, where it was assigned to Co3O442.


12

Page 13 of 41 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


Upon fluorination, the positions of band centers do not show significant changes for neither the spinel nor the Li2MnO3 phase (Figure 2, Table S2). However, the intensity contributions of both phases change. In the LCMOF00 sample, Li2MnO3 bands dominate the spectrum, i.e. the maximum intensity in the spectrum is recorded for the v1 band of Li2MnO3. For the LCMOF10 sample on the other hand, the bands of the spinel phase are more dominant, i.e. the maximum intensity in the spectrum is recorded at 572 cm-1 for the F2g(2) band of spinel. Since a) the same sample areas where mapped and averaged, b) the mapped sample area is large enough to be representative for the whole sample, c) the powder properties such as particle size distribution and optical properties were similar, and d) the preparation method and measurement conditions for the Raman mapping were identical, the intensity of the bands of one phase should correlate to the measured volume fraction of this phase. The trend of decreasing Li2MnO3 band intensities (i.e. lower measured volume fractions of Li2MnO3) for higher fluorine content agrees with the observation from powder diffraction, where less Li2MnO3 is detected upon fluorination. Although Li2MnO3 has a large contribution to the average spectra, the result of the highresolution mapping clearly reveals spinel as the dominant phase throughout the samples (Figure 3 for LCMOF00 and Figure S2 for LCMOF05 and LCMOF10). A comparison of the two representative spectra for Li2MnO3 and spinel, in Figure 3a and 3b, demonstrates that the maximum intensity of the Li2MnO3 spectrum is at least four times higher than the maximum intensity in the spinel spectrum. This observation explains the dominance of the Li2MnO3 phase in the average spectra in Figure 2. Furthermore, Li2MnO3 can clearly be identified as a separate phase on the micrometer length scale.


13


Page 14 of 41

Figure 3. Component analysis results of the high resolution micro Raman mapping of LCMOF00. Representative spectra of the individual components are given in a) for spinel, b) for Li2MnO3 (note: different intensity-scale than in (a) and (c)), and c) for a Co-rich component. The recorded third component cannot be identified as an individual phase on this length scale. However, the presence of this third component can be recognized from its relatively intense band centered at 692 cm-1. This band position is a fingerprint for Co3O4 and corresponds to its high intensity A1g mode. Two further low intensity bands are expected for Co3O4 as well43, which can be found at 535 and 485 cm-1. Obviously, the Co3O4 spectrum is superimposed by the spectrum of the LCMOF spinel, which implies that the dimension of a separate Co3O4 phase is below the resolution of the Raman microscope. Figure 4 shows the band position of A1g modes of lithium manganese based spinels as a function of the average ionic radius of the octahedrally coordinated transition metals in various spinels. The band position of the A1g mode was shown to shift to higher wavenumbers, if the average


14

Page 15 of 41 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


ionic radius of the transition metals in the octahedron gets lower44. This is directly caused by a decreased bond length, leading to a higher frequency of the vibrational mode. The displayed averaged radii of the transition metals on the octahedral site belong to the compounds LiMn2O4 (Mn3+ : Mn4+ = 1 : 1)40, LiNi0.5Mn1.5O4 (Ni2+ : Mn4+ = 0.5 : 1.5)44, LiCo0.5Mn1.5O4 (Co3+ : Mn3+ : Mn4+ = 0.5 : 0.5 : 1)44 and LiCoMnO4 (Co3+ : Mn4+ = 1 : 1)15,26. The ionic radii were taken from Shannon (1976)45. For our samples, an average radius of 0.543 Å is assumed (according to a ratio of Co3+ : Mn3+ : Mn4+ = 1 : 0.1 : 0.9), as will be discussed later. All three samples share a very similar Raman band position of the A1g mode and are consistent with the trend derived from literature.

Figure 4. Correlation of the position of the band center of the A1g mode to the average ionic radius of the transition metal cations on the octahedral site in lithium manganese based spinels (space group Fd-3m) for literature data15,26,40,44 and for the samples LCMOF00, LCMOF05 and LCMOF10.


15


Page 16 of 41

Cyclic Voltammetry The CV results for all three samples at different scan rates are given in Figure 5. Three oxidation and reduction reactions are recorded, which are centered at 3.9, 4.9 and at 5.1 V vs. Li/Li+. They belong to the Mn3+/4+ (at 4 V vs. Li/Li+) and the Co3+/4+ redox reactions (at 4.9 and 5.1 V vs. Li/Li+).

Figure 5. CV results of a) LCMOF00, b) LCMOF05, and c) LCMOF10 at different scan rates: 0.1, 0.13, 0.16, 0.21, 0.4, and 0.65 mVs-1. With increasing scan rates, we observe increasing peak current densities for all the samples. This behavior is expected, since the time for charge transfer decreases for higher scan rates (according to I = Q/t). While the peak currents for peak 1, at 4 V vs. Li/Li+, and peak 2, at 4.9 V vs. Li/Li+, clearly display the expected behavior, peak 3 at 5.1 V vs. Li/Li+ behaves differently: the forward peak current rises, while the reverse peak current decreases with increasing scan rates. Peak 3 behaves differently, because the oxidation reaction of Co3+ to Co4+ is not completed yet at the cut-off voltage (5.3 V vs Li/Li+). This effect gets more and more severe for the higher scan rates, where less time for lithium ion diffusion is given. As a consequence, the peak current densities for peak 3 in the reduction path decrease for higher scan rates. The cut-off voltage was


16

Page 17 of 41 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


not be raised beyond 5.3 V vs. Li/Li+ in purpose, compromising the instability of the liquid electrolyte in the high voltage region46. In our previous study, we reported results from the direct measurement of capacities in a galvano-static charge discharge experiment, where 18 % higher discharge capacities were recorded for LCMOF10 compared to LCMOF00. A possible reason for the increased capacity could be an improved lithium ion diffusivity in the fluorinated spinels, which allows for a faster diffusion and effectively more lithium extraction during charging before the cut-off voltage is reached. To investigate the lithium ion diffusivity, the Randles-Sevcik equation was applied. According to the Randles-Sevcik equation, the peak current of a certain redox peak in a CV experiment is proportional to the square root of the diffusion coefficient (D) and the square root of the scan rate (υ): ‫ܫ‬௣ = 2.67 · 10ହ · ‫݊ܣ‬ଵ.ହ ‫ܥ‬଴ ‫ܦ‬଴.ହ ߭ ଴.ହ

(1)

Here, A is the surface area of the electrode in cm², n is the number of electrons involved in one single reaction and C0 the concentration of lithium ions at the respective reaction center in mol/cm³. A first comparison of the lithium diffusivity in the different samples is given by the discharge peak current density as a function of the square root of the scan rate. In Figure 6, the linear regression of Ip vs. υ0.5 is displayed for the peaks centered at 4 V vs. Li/Li+ (peak 1) and 4.9 V vs. Li/Li+ (peak 2). Peak 3 is not evaluated, because of the incomplete reaction at 5.1 V vs. Li/Li+, as a reason of the low cut-off voltage, as explained before.


17


Page 18 of 41

Figure 6. Linear regression of square root of CV scan rate and the peak current for peak 1 (centered at 3.9 V vs. Li/Li+) and peak 2 (centered at 4.9 V vs. Li/Li+). It is observed that the slopes of the linear regression lines for peak 2 are significantly higher than those for peak 1. Furthermore, the lines for peak 1 display very similar slopes for all samples, whereas peak 2 shows a trend of increasing slopes in the following order: LCMO00 (9.7 As0.5V0.5 -1

g ) < LCMOF05 (10.7 As0.5V-0.5g-1) < LCMOF10 (10.9 As0.5V-0.5g-1).

If A and C0 are known, lithium ion diffusion coefficients can be calculated from equation 1. In here, we approximate A from the experimental setup by 1 cm². C0 is approximated by the discharge capacity at the respective voltage, assuming that for every electron transferred a lithium ion will be inserted into the structure. Because of the assumptions regarding A and C0 the obtained diffusion coefficients can only be interpreted semi-quantitatively, in order compare the three different materials, since the loading and the preparation processes of the electrodes were similar. Very similar lithium ion diffusion coefficients are calculated for peak 1 (Mn3+/4+),


18

Page 19 of 41 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


namely 6 × 10-13 cm²/s for LCMOF00, 7 × 10-13 cm²/s for LCMOF05, and 9 × 10-13 cm²/s for LCMOF10. For peak 2 (Co3+/4+), the calculated lithium ion diffusion coefficients display slightly increasing values: LCMOF00 (1.37 × 10-11 cm²/s) < LCMOF05 (1.79 × 10-11 cm²/s) < LCMOF10 (1.86 × 10-11 cm²/s).

Discussion A solid-state reaction synthesis was chosen to synthesize fluorinated LiCoMnO4 powders. Unlike in studies about fluorinated LiNi0.5Mn1.5O4 spinels where fluorination was carried out in a postsynthesis step30,31,33,47, in this study, fluorination was carried out during the synthesis. While fluorination via post-synthesis was reported to lead to fluorine enriched particle surfaces31, the presented method was chosen for bulk fluorination, which we confirmed by nuclear reaction analysis in our previous study28. The application of solid-state reaction synthesis, however, required relatively high synthesis temperatures (800 °C) to guarantee a complete reaction between the starting materials. At these temperatures the secondary phase Li2MnO3 is formed, which can be partially transformed to spinel at 650 °C. Therefore, we annealed our samples at this temperature after the initial reaction. As in other studies where LiCoMnO4 was not synthesized only at low temperatures and where high oxygen partial pressure atmospheres were not applied, the samples still contain Li2MnO3 as secondary phase48–50. Diffraction and spectroscopic investigations in this study demonstrate that fluorination impacts the phase stability of LiCoMnO4 and helps to reduce the fraction of Li2MnO3. This is of particular importance as so far only the synthesis in high oxygen partial pressure atmospheres


19


Page 20 of 41

has been reported to produce phase pure LiCoMnO415,48. Via stabilizing the spinel lattice by fluorination, synthesis routes for LiCoMnO4 can be further improved. The reduction of the Li2MnO3 fraction has direct consequences on the reversible capacity of the samples. However, the reduction of the Li2MnO3 fraction, irrespective if Li2MnO3 is considered as inert or not51–53, cannot explain the capacity enhancement of fluorinated LiCoMnO4 completely, as we have demonstrated previously28. Therefore, further efforts were made to characterize fluorinated LiCoMnO4 by diffraction, spectroscopy, and cyclic voltammetry within this study, as will be discussed in the following sections. Crystallography and Crystal Chemistry of the Spinel Phase Joint Rietveld refinement of neutron and synchrotron diffraction confirmed that the structure of the main phase in all samples is the Fd-3m spinel structure, with lattice parameters similar to those obtained previously28 and comparable to the values in literature15. The structural formulas were

determined

to

[Li0.88Co0.12]8a[CoMn]16dO3.94Fy

[Li0.86Co0.14]8a[Co1.05Mn0.95]16dO3.97

be for

LCMOF05

and

for

LCMOF00

[Li0.9Co0.1]8a[CoMn]16dO3.89Fy

for

LCMOF10, respectively. It was not possible to determine y with the analytical methods used. Thus, y is assumed to be 0.04 (for LCMOF05) and 0.09 (for LCMOF10), as derived by nuclear reaction analysis in our previous study28. The sample stoichiometry, as given in our previous study28, is Li1.1CoMnO3.6, Li1.1CoMnO3.6F0.04 and Li1.1CoMnO3.6F0.09, respectively. The overall oxygen stoichiometry is the sum of the oxygen contents of the spinel and the Li2MnO3 secondary phase. If the overall oxygen content is calculated from the amount of the main and secondary phase and their stoichiometry, as derived from the powder diffraction results, a value of 3.63, 3.63 and 3.62 is obtained for LCMOF00,


20

Page 21 of 41 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


LCMOF05 and LCMOF10. This is consistent with the values of the oxygen content as derived from the chemical analysis. The oxygen occupancy parameter for the spinel phase in all three samples shifted only slightly away from full occupancy of the anionic site. As a result, around 0.6 % oxygen vacancies for the spinel phase in all samples is suggested. This confirms that the tolerance for oxygen vacancies in spinel is low and oxygen loss is partially compensated by Li2MnO3 formation, as is consistent with the literature15,18. Since the samples consist of both spinel and the monoclinic Li2MnO3, the changing cation stoichiometry of the spinel phase is a consequence of Li2MnO3 formation, which leads to lithium and manganese depletion and cobalt enrichment of the residual spinel during synthesis15,27. As fluorination hinders Li2MnO3 formation during synthesis, samples with higher fluorine contents (less secondary Li2MnO3 phase) exhibit spinels with a lower Co/Li and Co/Mn cation ratio. Although fluorination impacts the phase purity positively, the recorded impact of fluorination on the crystal lattice is minor. The obtained lattice parameters decrease by less than 0.02 % upon fluorination. For fluorinated LiNi0.5Mn1.5O4, an increase of the lattice parameter is observed upon fluorination, due to an increased Mn3+ content as a result of aliovalent substitution of O2- by F32,34

. In this study, the Mn3+ content for the samples does not change upon fluorination, as was

shown in our previous study28 and will be discussed later. However, from the diffraction results in this study we know that there is a simultaneous change in cation stoichiometries, i.e. the Li/Co and Co/Mn ratios. Therefore, the simultaneous effects of fluorination and changing cation ratios on the lattice paraments needs to be considered for this sample system. The impact of changing cation ratios and fluorination is further expected to be reflected in the anionic site position of the three samples, which also shows hardly any deviations. Due to the


21


Page 22 of 41

higher electronegativity of the fluorine anion, shortening of cation-anion bonds could be expected32. Additionally, a change in the cation ratio on tetrahedral and octahedral sites is usually accompanied by changes in the anionic site position and/or the lattice parameter, which is influenced by the dimensions of tetrahedra and octahedra13. Therefore, the similar obtained anionic positions and lattice parameters of the studied spinels reflect similar dimensions of the coordination polyhedral primarily. Changing cation ratios on octahedral and tetrahedral sites without changing the dimensions of the coordination polyhedra is only possible if cations with very similar radii are interchanged. This is true for Mn4+ and low spin Co3+ in octahedral coordination, with ionic radii of 0.53 and 0.545 Å, respectively; and for Li+ and Co2+ in tetrahedral coordination, with ionic radii of 0.58 and 0.59 Å, respectively45. This result indirectly identifies Co2+ in the studied spinels, which is in agreement with Reeves et al. (2016)15. According to their study, Co2+ was found to migrate to the tetrahedral site, after reduction of Co3+ due to oxygen loss and Li2MnO3 formation during synthesis15. Parallel to the reduction of Co3+ to Co2+, Mn4+ is reduced to Mn3+ during synthesis15. Our previous electrochemical measurements confirmed very similar Mn3+ contents (around 10 %) for all samples28. A very similar Mn3+ content for the studied spinels is strongly suggested by the powder diffraction results in this study as well. Considering the very large radius of the Mn3+ ion (0.645 Å in octahedral coordination), any changes in Mn3+ contents would be reflected in the lattice parameters and/or in the oxygen positions, which is not the case. Since all manganese ions in the studied spinels are located at the octahedral sites, it is concluded that Co3+, Mn3+, and Mn4+ occupy the octahedral site in a ratio of 1.05 : 0.1 : 0.85 for LCMOF00 and 1 : 0.1 : 0.9 for LCMOF05 and LCMOF10.


22

Page 23 of 41 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


Local Structure of the Spinel Phase In agreement with the powder diffraction results for the lattice properties and atomic positions, the Raman spectra of the spinel phase do not suggest any significant changes upon fluorination regarding the bond lengths within the tetrahedra and octahedra. The insertion of the fluorine anion into the spinel lattice is expected to impact the bond strength distribution of the cations to the anions, which causes a lowering of the local coordination symmetry and an alteration of the average bond length. This would cause broadening of bands and shifts in the band positions. Since none of these effects are observed upon fluorination, we conclude that the amount of fluorine anions might to be too low in LCMOF05 and LCMOF10 to cause significant and observable shifts of band positions or broadening of bands in comparison to LCMOF00. Nevertheless, the increased phase purity of the spinel with increasing fluorination of the samples is clearly visible in the averaged Raman spectra for representative sample areas of 60 × 100 µm each, where lower band intensities (corresponding to lower volume fractions of Li2MnO3) where measured for fluorinated samples, in agreement with the powder diffraction results. Additionally, the high resolution mapping implies the occurrence of Co3O4 as an additional impurity phase. Co3O4 could be identified from its A1g mode in the Raman spectra of LCMO samples in other studies as well26,42. However, several observations let us question whether Co3O4 segregates as an individual phase or if rather the LCMOF phase exhibits a Co3O4-type local environment. The first observation is that there could no spectrum be found in the Raman mapping that showed Co3O4 without any bands of LCMOF and in every spectrum of LCMOF we always find the low intensity band at 692 cm-1 matching the A1g mode of Co3O4. From these observations it is already very likely that the presence of the “Co3O4 band” suggests a Co3O4-like local order in the spinel, with Co2+ being


23


Page 24 of 41

tetrahedrally coordinated by oxygen and second nearest neighbors being octahedrally coordinated Co3+-ions. Co2+ has been identified at the tetrahedral sites via diffraction. Interestingly, the amount of tetrahedral Co2+ decreases with increasing fluorine content, as does the intensity of the A1g mode in the average Raman spectra of the LCMOF samples. This observation further confirms a correlation of the Co2+ contents and certain extend of Co3O4-type local ordering in LCMOF spinel. Deviations in local order cannot be fitted within a Rietveld refinement, whereas considering a separate Co3O4 phase should improve the fit quality, if this phase is present. Since the lattice parameter of Co3O4 is 8.072 Å, which is slightly bigger than the lattice parameters of the LCMOF spinel phases (8.065 - 8.067 Å), any intensity contribution from Co3O4 would be observed as shoulders at the LCMOF Bragg peaks in the higher Q-range and would result in asymmetric peak shapes of the LCMOF peaks in the lower Q-range (assuming stoichiometric Co3O4). Missing intensities due to asymmetries and shoulders were not observed in the difference plot. Still, we tried to include Co3O4 in the refinements but the fit quality did not improve and the Co3O4 scale factor remained close to zero. These observations are further indications that Co3O4 is not a separate phase and that the detected Raman bands strongly suggest a Co3O4-like local order of Co2+ and Co3+ in LCMOF spinel. As discussed in the previous section, the studied spinels show a mixed cation occupation at octahedral and tetrahedral sites. Especially the mixed cation occupation at the octahedral site is mirrored in the Raman spectra: For all samples, the A1g mode of LCMOF spinel was described best by the superimposition of a relatively sharp and a relatively broad pseudo Voigt function. The A1g mode is the symmetric stretching vibration of the transition metal octahedron. In our samples, Mn3+, Mn4+ and Co3+ share the same octahedral environment. Mn3+ causes a local


24

Page 25 of 41 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


distortion of the octahedral symmetry, due to the Jahn-Teller effect54. Mn4+ and Co3+ maintain the symmetric octahedral coordination, but exhibit rather different bond strengths. This leads to local expansion and contraction of the transition metal octahedra. Hence, the broadness of the A1g mode is directly caused by different bond lengths and local distortions, due to the occupation of different ions in different valence states at the octahedral site. In stoichiometric spinel (LiCoMnO4), with Mn4+ and Co3+ in a 1:1 ratio with ionic radii of 0.53 and 0.545 Å, the average radius of the cations in the octahedron is 0.538 Å. For the spinels in this study, 10 % Mn3+ (with a radius of is 0.645 Å) raises the average radius to 0.543 Å. This leads to an increased bond length, which weakens the bond strength and causes a shift to lower frequencies of the symmetric stretching vibration of the transition metal octahedron. Thus, the higher Mn3+ content in the studied spinels is directly reflected in the position of the A1g mode in the recorded Raman spectra. Structural Features vs. Electrochemical Performance The theoretical capacity of an electrode can be calculated from the possible lithium ion insertion in a given voltage range. In the voltage range starting from 5.3 to 3.4 V vs. Li/Li+, lithium ions are inserted on tetrahedral 8a sites in spinel8. Based on the structural data derived from Rietveld refinement, we expect a capacity enhancement for the fluorinated samples because the availability of 8a sites for lithium ion insertion gets higher for the fluorinated samples, where less cobalt ions reside at 8a sites. In fact, calculating the theoretical capacity from the structural formulas as obtained from Rietveld analysis yields ~120 mAhg-1 for [Li0.86Co0.14]8a[Co1.05Mn0.95]16dO3.97, ~124 mAhg-1 for [Li0.88Co0.12]8a[CoMn]16dO3.94F0.04 and ~127 mAhg-1 for [Li0.9Co0.1]8a[CoMn]16dO3.89F0.09. Hence,


25


Page 26 of 41

we expect an increase in capacity by ~ 6 % as an effect of 5 % less cobalt ions on the tetrahedral site. Increasing capacities upon fluorination were measured directly in our previous study, where we reported discharge capacities of 85.1 mAhg-1 for LCMOF00, 97.0 mAhg-1 for LCMOF05, and 104.0 mAhg-1 for LCMOF1028. Taking inert Li2MnO3 into account, these values need to be normalized to the amount of the spinel phase. Considering 89 wt.-% spinel in LCMOF00, 94 wt.% spinel in LCMOF05, and 96 wt.-% spinel in LCMOF10, as found by the Rietveld refinement in this study, yields ~96 mAhg-1spinel for LCMOF00, ~103 mAhg-1spinel for LCMOF05 and ~108 mAhg-1spinel for LCMOF10. Accordingly, the capacity increases by around 8 % from LCMOF00 to LCMOF05 and by around 5 % from LCMOF05 to LCMOF10. As an additional reason for the enhanced reversible capacity, improved lithium diffusion properties in the high voltage range were suggested by our previous study28. In the current study, the lithium ion diffusivity was investigated by approximation of the lithium ion diffusion coefficient from the Randles-Sevcik equation. For LiMn2O4 spinel powders, the sensitivity of this method was shown to be rather high for the comparison of differently doped materials55–57. For the presented samples, systematic errors are expected due to cell degradation caused by the instability of the used liquid electrolyte, which decomposes upon charging above 4.7 V vs. Li/Li+46. Up to now, there is no liquid electrolyte system yet available, which is stable up to 5.5 V vs. Li/Li+ and would allow for full utilization of the capacity of the LiCoMnO4 spinel and diminish capacity degradation. Furthermore, systematic errors can be expected because of assumptions regarding the electrode area and the lithium ion concentration at the investigated CV peaks. In particular the cell degradation leads to a non-linear decrease in the Ip vs. υ0.5 trend, which might explain the deviation of the derived values for the diffusion coefficient (in the order


26

Page 27 of 41 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


10-11) from the theoretically calculated diffusion coefficient for spinels (in the order 10-9)21. Nevertheless, the observed results can be interpreted qualitatively, which offers valuable insight about how fluorination affects the lithium ion diffusivity. First of all, significantly higher lithium ion diffusion coefficients are observed at 4.9 V vs. Li/Li+ than at 3.9 V vs. Li/Li+. It is a well-known phenomenon for spinel electrodes that lithium ion diffusion is facilitated when more lithium ion positions in the structure are vacant. At 4 V vs. Li/Li+, at least 90 % of the lithium ion positions are occupied (as calculated from the concentration of lithium ions, which is described in the results section), causing a low lithium ion diffusivity. At 4.8 V vs. Li/Li+, only around 50 % of the lithium ion positions are occupied. Accordingly, lithium ion diffusivity is enhanced in the high voltage region, due to a higher amount of lithium vacancies. Upon fluorination, we observed a slightly enhanced lithium ion diffusion at 4.8 V vs. Li/Li+. Just as coulombic repulsion raises the activation energy for lithium ions hopping near neighboring lithium ions21, the occupation of the 8a site by cobalt ions is expected to cause an increase in the activation energy as well. Considering the slightly lower values of the cobalt ion occupancy at the 8a site in LCMOF10 (10 %) and LCMOF05 (12 %) in comparison with LCMOF00 (14 %), a slight increase in lithium ion diffusivity is plausible from a structural point of view and agrees to the observed trend of rising lithium ion diffusion coefficients upon fluorination. This explains the observation of higher capacities upon fluorination, due to faster lithium ion extraction during charging in the high voltage region before the cut-off voltage is reached.


27


Page 28 of 41

Conclusion The investigation of fluorinated LiCoMnO4 cathodes in this study revealed the origin of their enhanced capacity to be an effect of the lithium and cobalt distribution on the tetrahedral 8a sites in the Fd-3m spinel structure. The cobalt distribution in spinel is controlled by Li2MnO3 precipitation. Li2MnO3 precipitation leads to cobalt enrichment of the spinel phase, causing Co2+ ions to migrate to the 8a tetrahedral site. Due to the cobalt ion occupation of the 8a site, two effects can be identified which impact the performance of the lithium cobalt manganese spinel electrodes. Firstly, there are more tetrahedral sites available for lithium insertion, if less cobalt resides at 8a sites. As a consequence, the theoretical capacity is increased directly, i.e. lowering the cobalt ion occupancy on tetrahedral sites by 5 % increases the theoretical capacity by 6 %. Secondly, the impedance for lithium ion conduction is lower, if the cobalt ion occupancy at the 8a site is lower. Therefore, lithium ion diffusivity is enhanced, leading to faster lithium ion diffusion, which results in more lithium ions being inserted and extracted at a given current before the cut-off voltage in the high voltage region is reached, thus raising the value of the capacity indirectly.

ASSOCIATED CONTENT Supporting Information. The following files are available free of charge in a PDF file: Refined structural data of LCMOF00, LCMOF05 and LCMOF00 as table. Figures for the refinement of LCMOF05 and


28

Page 29 of 41 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


LCMOF10. Raman spectroscopy: Table of band centers and analyzed high resolution mappings for LCMOF00, LCMOF05 and LCMOF10. AUTHOR INFORMATION Corresponding Author Anna Windmüller, [email protected] Author Contributions The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

ACKNOWLEDGMENT A. Windmüller, S. Lobe, and S. Uhlenbruck gratefully acknowledge financial support from the German Federal Ministry of Education and Research, as part of the DESIREE project (project no. 03SF0477A) and A. Windmüller and M. Finsterbusch as part of the High-Temperature and Energy Materials project (project no. 03EK3032). C.-L. Tsai, C. Dellen, M. Finsterbusch, S. Uhlenbruck und O. Guillon gratefully acknowledge financial support from the HelmholtzGemeinschaft Deutscher Forschungszentren e.V. under the grant “Speicher und vernetzte Infrastrukturen” under support code PoF3-SCI. C. A. Bridges and G. M. Veith gratefully acknowledge financial support from the Materials Sciences and Engineering Division, Office of Basic Energy Sciences, U.S. Department of Energy under contract with UT-Battelle, LLC. The authors would like to thank Dr. Ashfia Huq, Spallation Neutron Source, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, United States, for the neutron diffraction


29


Page 30 of 41

measurements and Dr. Saul Lapidus, Advanced Photon Source, Argonne National Laboratory, Argonne, IL 60439, United States, for synchrotron diffraction measurements.

ABBREVIATIONS CV, Cyclic voltammetry; PVdF, polyvinylidene fluoride; EC/DEC, ethylene carbonate - diethyl carbonate. REFERENCES (1) Ohzuku, T.; Brodd, R. J. An overview of positive-electrode materials for advanced lithiumion batteries. J. Power Sources 2007, 174, 449–456. (2) Berg, E. J.; Villevieille, C.; Streich, D.; Trabesinger, S.; Novák, P. Rechargeable Batteries: Grasping for the Limits of Chemistry. J. Electrochem. Soc. 2015, 162, A2468-A2475. (3) Etacheri, V.; Marom, R.; Elazari, R.; Salitra, G.; Aurbach, D. Challenges in the development of advanced Li-ion batteries: a review. Energy Environ. Sci. 2011, 4, 3243–3262. (4) Cairns, E. J.; Albertus, P. Batteries for Electric and Hybrid-Electric Vehicles. Annu. Rev. Chem. Biomol. Eng. 2010, 1, 299–320. (5) Thackeray, M. M. Spinel Electrodes for Lithium Batteries. J. Am. Chem. Soc. 1999, 82, 3347–3354. (6) Kawai, H.; Nagata, M.; Tukamoto, H.; West, A. R. High-voltage lithium cathode materials. J. Power Sources 1999, 81-82, 67–72. (7) Kraytsberg, A.; Ein-Eli, Y. Higher, Stronger, Better… A Review of 5 Volt Cathode Materials for Advanced Lithium-Ion Batteries. Adv. Energy Mater. 2012, 2, 922–939. (8) Julien, C. M.; Mauger, A. Review of 5-V electrodes for Li-ion batteries: Status and trends. Ionics 2013, 19, 951–988. (9) Kim, J.-H.; Pieczonka, Nicholas P. W.; Yang, L. Challenges and Approaches for HighVoltage Spinel Lithium-Ion Batteries. ChemPhysChem 2014, 15, 1940–1954. (10) Tan, S.; Ji, Y. J.; Zhang, Z. R.; Yang, Y. Recent Progress in Research on High-Voltage Electrolytes for Lithium-Ion Batteries. ChemPhysChem 2014, 15, 1956–1969. (11) Uhlenbruck, S.; Dornseiffer, J.; Lobe, S.; Dellen, C.; Tsai, C.-L.; Gotzen, B.; Sebold, D.; Finsterbusch, M.; Guillon, O. Cathode-electrolyte material interactions during manufacturing of inorganic solid-state lithium batteries. J. Electroceram. 2016, 1–10. (12) Li, J.; Ma, C.; Chi, M.; Liang, C.; Dudney, N. J. Solid Electrolyte: The Key for HighVoltage Lithium Batteries. Adv. Energy Mater. 2015, 5, n/a-n/a. (13) Sickafus, K. E.; Wills, J. M.; Grimes, N. W. Structure of Spinel. J. Am. Chem. Soc. 1999, 82, 3279–3292. (14) Hahn, T.; Fuess, H.; Wondratschek, H.; Müller, U.; Shmueli, U.; Prince, E.; Authier, A.; Kopský, V.; Litvin, D. B.; Rossmann, M. G.; Arnold, E.; Hall, S.; McMahon, B. Int. Tables Crystallogr., 5th rev. ed A; International Union of Crystallography: Chester, England, 2006. (15) Reeves-McLaren, N.; Sharp, J.; Beltran-Mir, H.; Rainforth, W. M.; West, A. R. Spinel-rock salt transformation in LiCoMnO4-d. Proc. R. Soc. A 2016, 472, 20140991.


30

Page 31 of 41 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


(16) Kanno, R.; Kondo, A.; Yonemura, M.; Gover, R.; Kawamoto, Y.; Tabuchi, M.; Kamiyama, T.; Izumi, F.; Masquelier, C.; Rousse, G. The relationships between phases and structures of lithium manganese spinels. J. Power Sources 1999, 81-82, 542–546. (17) Samarasingha, P. B.; Andersen, N. H.; Sørby, M. H.; Kumar, S.; Nilsen, O.; Fjellvåg, H. Neutron diffraction and Raman analysis of LiMn1.5Ni0.5O4 spinel type oxides for use as lithium ion battery cathode and their capacity enhancements. Solid State Ionics 2016, 284, 28– 36. (18) Strobel, P.; Le Cras, F.; Seguin, L.; Anne, M.; Tarascon, J. M. Oxygen Nonstoichiometry in Li–Mn–O Spinel Oxides: A Powder Neutron Diffraction Study. J. Solid State Chem. 1998, 135, 132–139. (19) Bellitto, C.; Bauer, E. M.; Righini, G.; Green, M. A.; Branford, W. R.; Antonini, A.; Pasquali, M. The effect of doping LiMn2O4 spinel on its use as a cathode in Li-ion batteries: Neutron diffraction and electrochemical studies. J. Phys. Chem. Solids 2004, 65, 29–37. (20) Cabana, J.; Casas-Cabanas, M.; Omenya, F. O.; Chernova, N. A.; Zeng, D.; Whittingham, M. S.; Grey, C. P. Composition-Structure Relationships in the Li-Ion Battery Electrode Material LiNi0.5Mn1.5O4. Chem. Mater. 2012, 24, 2952–2964. (21) Ma, X.; Kang, B.; Ceder, G. High Rate Micron-Sized Ordered LiNi0.5Mn1.5O4, 2010. J. Electrochem. Soc., 157 (8). (22) Shin, D. W.; Bridges, C. A.; Huq, A.; Paranthaman, M. P.; Manthiram, A. Role of Cation Ordering and Surface Segregation in High-Voltage Spinel LiMn1.5Ni0.5–xMxO4 (M = Cr, Fe, and Ga) Cathodes for Lithium-Ion Batteries. Chem. Mater. 2012, 24, 3720–3731. (23) Hu, M.; Tian, Y.; Su, L.; Wei, J.; Zhou, Z. Preparation and Ni-Doping Effect of Nanosized Truncated Octahedral LiCoMnO4 As Cathode Materials for 5 V Li-Ion Batteries. ACS Appl. Mater. Interfaces 2013, 5, 12185–12189. (24) Alcantara, R.; Jaraba, M.; Lavela, P.; Tirado, J. L. Electrochemical, 6Li MAS NMR, and X-ray and Neutron Diffraction Study of LiCoxFeyMn2-(x+y)O4 Spinel Oxides for High-Voltage Cathode Materials. Chem. Mater. 2003, 15, 1210–1216. (25) Kuwata, N.; Kudo, S.; Matsuda, Y.; Kawamura, J. Fabrication of thin-film lithium batteries with 5-V-class LiCoMnO4 cathodes. Solid State Ion. 2014, 262, 165–169. (26) Li, H.; West, W. C.; Motoyama, M.; Iriyama, Y. Deep-discharged LiCoMnO4 Lithium-Ion cathodes with high rate capability and long cycle life. Thin Solid Films 2016, 615, 210–214. (27) Thackeray, M. M.; Mansuetto, M. F.; Dees, D. W.; Vissers, D. R. The thermal stability of lithium-manganese-oxide spinel phases. Mat. Res. Bull. 1996, 31, 133–140. (28) Windmüller, A.; Tsai, C.-L.; Möller, S.; Balski, M.; Sohn, Y. J.; Uhlenbruck, S.; Guillon, O. Enhancing the performance of high-voltage LiCoMnO4 spinel electrodes by fluorination. J. Power Sources 2017, 341, 122–129. (29) Hagh, N. M.; Amatucci, G. G. Effect of cation and anion doping on microstructure and electrochemical properties of the LiMn1.5Ni0.5O4−δ spinel. J. Power Sources 2014, 256, 457– 469. (30) Choi, W.; Manthiram, A. Superior Capacity Retention Spinel Oxyfluoride Cathodes for Lithium-Ion Batteries. Electrochem. Solid State Lett. 2006, 9, A245-A248. (31) Höweling, A.; Stenzel, D.; Gesswein, H.; Kaus, M.; Indris, S.; Bergfeldt, T.; Binder, J. R. Variations in structure and electrochemistry of iron- and titanium-doped lithium nickel manganese oxyfluoride spinels. J. Power Sources 2016, 315, 269–276. (32) Luo, Q.; Muraliganth, T.; Manthiram, A. On the incorporation of fluorine into the manganese spinel cathode lattice. Solid State Ion. 2009, 180, 703–707.


31


Page 32 of 41

(33) Okumura, T.; Fukutsuka, T.; Matsumoto, K.; Orikasa, Y.; Arai, H.; Ogumi, Z.; Uchimoto, Y. Role of Local and Electronic Structural Changes with Partially Anion substitution Lithium Manganese Spinel Oxides on Their Electrochemical Properties: X-ray Absorption Spectroscopy Study. Dalton Trans 2011, 40, 9752–9764. (34) Gutierrez, A.; Manthiram, A. Understanding the Effects of Cationic and Anionic Substitutions in Spinel Cathodes of Lithium-Ion Batteries. J. Electrochem. Soc. 2013, 160, A901-A905. (35) A.C. Larson and R.B. Von Dreele. General Structure Analysis System (GSAS); Los Alamos National Laboratory Report LAUR 86-748, 1994. (36) B. H. Toby. EXPGUI, a graphical user interface for GSAS; J. Appl. Cryst. (2001). 34, 210213. (37) Boulineau, A.; Croguennec, L.; Delmas, C.; Weill, F. Reinvestigation of Li 2 MnO 3 Structure: Electron Diffraction and High Resolution TEM. Chem. Mater. 2009, 21, 4216–4222. (38) Bréger, J.; Jiang, M.; Dupré, N.; Meng, Y. S.; Shao-Horn, Y.; Ceder, G.; Grey, C. P. Highresolution X-ray diffraction, DIFFaX, NMR and first principles study of disorder in the Li2MnO3–Li[Ni1/2Mn1/2]O2 solid solution. J. Solid State Chem. 2005, 178, 2575–2585. (39) Warren, B. E.; Averbach, B. L. The Separation of Stacking Fault Broadening in Cold‐ Worked Metals. J. Appl. Phys. 1952, 23, 1059. (40) Julien, C. M.; Massot, M. Lattice vibrations of materials for lithium rechargeable batteries I. Lithium manganese oxide spinel. J. Mater. Sci. Eng. B 2003, 97, 217–230. (41) Julien, C. M.; Massot, M. Lattice vibrations of materials for lithium rechargeable batteries III. Lithium manganese oxides. J. Mater. Sci. Eng. B 2003, 100, 69–78. (42) Dokko, K.; Anzue, N.; Mohamedi, M.; Itoh, T.; Uchida, I. Raman spectro-electrochemistry of LiCoxMn2−xO4 thin film electrodes for 5 V lithium batteries. Electrochem. Commun. 2004, 6, 384–388. (43) V G Hadjiev and M N Iliev and I V Vergilov. The Raman spectra of Co 3 O 4. J. Phys. C: Solid State Phys. 1988, 21, L199. (44) Julien, C. M.; Gendron, F.; Amdouni, A.; Massot, M. Lattice vibrations of materials for lithium rechargeable batteries. VI: Ordered spinels. J. Mater. Sci. Eng. B 2006, 130, 41–48. (45) Shannon, R. D. Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides. Acta Crystallogr., Sect. A 1976, 32, 751–767. (46) Xu, W.; Chen, X.; Ding, F.; Xiao, J.; Wang, D.; Pan, A.; Zheng, J.; Li, X. S.; Padmaperuma, A. B.; Zhang, J.-G. Reinvestigation on the state-of-the-art nonaqueous carbonate electrolytes for 5 V Li-ion battery applications. J. Power Sources 2012, 213, 304–316. (47) Matsumoto, K.; Fukutsuka, T.; Okumura, T.; Uchimoto, Y.; Amezawa, K.; Inaba, M.; Tasaka, A. Electronic structures of partially fluorinated lithium manganese spinel oxides and their electrochemical properties. Selected Papers presented at the 14th INTERNATIONAL MEETING ON LITHIUM BATTERIES (IMLB-2008) 2009, 189, 599–601. (48) Pasero, D.; Souza, S. de; Reeves, N.; West, A. R. Oxygen content and electrochemical activity of LiCoMnO4-[small delta]. J. Mater. Chem. 2005, 15, 4435–4440. (49) Zhecheva, E.; Stoyanova, R.; Alcántara, R.; Lavela, P.; Tirado, J. L. EPR studies of Li deintercalation from LiCoMnO4 spinel-type electrode active material. J. Power Sources 2006, 159, 1389–1394. (50) Hu, M.; Tian, Y.; Wei, J.; Wang, D.; Zhou, Z. Porous hollow LiCoMnO4 microspheres as cathode materials for 5 V lithium ion batteries. J. Power Sources 2014, 247, 794–798.


32

Page 33 of 41 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


(51) Rana, J.; Stan, M.; Kloepsch, R.; Li, J.; Schumacher, G.; Welter, E.; Zizak, I.; Banhart, J.; Winter, M. Structural Changes in Li 2 MnO 3 Cathode Material for Li-Ion Batteries. Adv. Energy Mater. 2014, 4, 1300998. (52) Wang, Y.; Yang, Z.; Qian, Y.; Gu, L.; Zhou, H. New Insights into Improving Rate Performance of Lithium-Rich Cathode Material. Adv. Mater. 2015, 27, 3915–3920. (53) Xiao, L.; Xiao, J.; Yu, X.; Yan, P.; Zheng, J.; Engelhard, M.; Bhattacharya, P.; Wang, C.; Yang, X.-Q.; Zhang, J.-G. Effects of structural defects on the electrochemical activation of Li2MnO3. Nano Energy 2015, 16, 143–151. (54) Yamada, A.; Tanaka, M.; Tanaka, K.; Sekai, K. Jahn–Teller instability in spinel Li–Mn–O. J. Power Sources 1999, 81–82, 73–78. (55) Xi, L. J.; Wang, H.-E.; Lu, Z. G.; Yang, S. L.; Ma, R. G.; Deng, J. Q.; Chung, C. Y. Facile synthesis of porous LiMn2O4 spheres as positive electrode for high-power lithium ion batteries. Journal of Power Sources 2012, 198, 251–257. (56) Xiong, L.; Xu, Y.; Lei, P.; Tao, T.; Dong, X.; Song, J. The electrochemical performance of sodium-ion-modified spinel LiMn2O4 used for lithium-ion batteries. Journal of Solid State Electrochemistry 2014, 18, 713–719. (57) Xiong, L.; Xu, Y.; Xiao, X.; Wang, J.; Li, Y. The effect of K-Ion on the electrochemical performance of spinel LiMn2O4. Electronic Materials Letters 2015, 11, 138–142.


33


Page 34 of 41

TOC


34

Page 35 of 41 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


Figure 1. Combined Rietveld refinement of LCMOF00 a) neutron diffraction (inset from 1 - 10.3 Å-1, actual fitting range from 1-15 Å-1), b) synchrotron diffraction, full Q-range, c) Visualization of spinel Fd-3m structure and cation distributions, d) cation distributions of spinel phase and phase content of Li2MnO3 (C2/m ) as a function of fluorine content - error bars are not displayed in Figure 1d, since they are smaller than the symbols. 177x99mm (300 x 300 DPI)



Figure 2. Results of the micro Raman spectroscopy for a sample area of 60 × 100 µm. Averaged, normalized and fitted spectra for a) LCMOF00, b) LCMOF05 and c) LCMOF10. 84x134mm (300 x 300 DPI)


Page 36 of 41

Page 37 of 41 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


Figure 3. Component analysis results of the high resolution micro Raman mapping of LCMOF00. Representative spectra of the individual components are given in a) for spinel, b) for Li2MnO3 (note: different intensity-scale than in (a) and (c)), and c) for a Co-rich component. 84x76mm (300 x 300 DPI)



Figure 4. Correlation of the position of the band center of the A1g mode to the average ionic radius of the transition metal cations on the octahedral site in lithium manganese based spinels (space group Fd-3m) for literature data15,26,40,44 and for the samples LCMOF00, LCMOF05 and LCMOF10. 106x89mm (300 x 300 DPI)


Page 38 of 41

Page 39 of 41 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


Figure 5. CV results of a) LCMOF00, b) LCMOF05, and c) LCMOF10 at different scan rates: 0.1, 0.13, 0.16, 0.21, 0.4, and 0.65 mVs-1. 353x121mm (300 x 300 DPI)



Figure 6. Linear regression of square root of CV scan rate and the peak current for peak 1 (centered at 3.9 V vs. Li/Li+) and peak 2 (centered at 4.9 V vs. Li/Li+). 179x150mm (300 x 300 DPI)


Page 40 of 41

Page 41 of 41 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


(graphical abstract) 84x45mm (300 x 300 DPI)


Impact of Fluorination on Phase Stability, Crystal Chemistry, and

Recommend Documents