Investigation of SEI Layer Formation in Conversion Iron Fluoride

Apr 16, 2015 - solid electrolyte interphase (SEI) layer at the cathode/electrolyte interface. ... understanding of the SEI formation mechanism at the ...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/JPCC

Investigation of SEI Layer Formation in Conversion Iron Fluoride Cathodes by Combined STEM/EELS and XPS M. Sina,† R. Thorpe,‡ S. Rangan,‡ N. Pereira,†,§ R. A. Bartynski,‡ G. G. Amatucci,†,§ and F. Cosandey*,† †

Department of Materials Science & Engineering, and ‡Department of Physics & Astronomy, Rutgers University, Piscataway, New Jersey 08854, United States § Energy Storage Research Group, Department of Materials Science and Engineering, Rutgers University, North Brunswick, New Jersey 08902, United States S Supporting Information *

ABSTRACT: Li-ion cathodes based on conversion reactions such as iron fluoride (FeF2) can achieve in principle high specific capacity. However, significant capacity fading is observed upon cycling. This has been attributed in part to the formation and continuous growth of a solid electrolyte interphase (SEI) layer at the cathode/electrolyte interface. In this work, scanning transmission electron microscopy, electron energy loss spectroscopy, selected area electron diffraction, and X-ray photoelectron spectroscopy were used in combination to study both the structural changes of the FeF2/C active material and the growth and evolution of the SEI layer upon cycling. Two main sources of capacity loss have been found. An increasing amount of Fe2+ appeared trapped inside the SEI layer with increasing cycle number, thus resulting in the loss of active material. In addition, reconversion is strongly impeded with increasing cycle number, leaving untransformed LiF and Fe0 upon delithiation. This correlates with the irreversible growth of a SEI layer that limits electronic and ionic transport.

I. INTRODUCTION Recently, conversion materials based on metal fluorides have received considerable attention as candidate positive electrodes for Li-ion batteries, due to their theoretical high capacity and high redox potential.1−5 However, short cycle life, hysteresis, and slow kinetics have until now restricted their potential use for commercialization.3,5 Decomposition of the electrolyte at the surface of the cathode has been shown to lead to the growth of a solid electrolyte interphase (SEI), which in some cases can inhibit the transport of lithium ions into and out of the electrodes. The irreversible growth of a SEI layer has previously been linked to capacity losses in fluoride-based conversion batteries.6−10 Although it has been shown that on anode materials the presence of a stable thin SEI is beneficial during cycling, as it prevents further reaction of the electrolyte with the electrode material, the situation appears more complex in the case of cathode materials.11,12 In addition, unlike intercalation materials, conversion materials undergo considerable structural changes during lithiation−delithiation processes, involving both alteration of the crystal structures and coexistence of separate nanophases. The surface of the electrode is thus constantly modified and in principle offers a fresh reactive site leading to the electrolyte decomposition.8 Consequently, a detailed understanding of the SEI formation mechanism at the surface of conversion materials is needed. The identification and structure of SEI compounds have been addressed for several electrodes, solvents, and Lisalts.7−10,13−19 These studies point to the fact that the SEI © XXXX American Chemical Society

composition and the growth mode depend not only on the salts and solvents used in the electrolyte but also on specific cathode materials and cycling parameters (e.g., temperature, cycling rate, and voltages). In the absence of fluorine sources, SEI constituents have been identified as lithium carbonates (Li2CO3), lithium alkyl carbonates (R−O−COOLi) or lithium alkoxides (R−CH2−COLi) and are thought to be the products of both the electrolyte solvent and salt decomposition at the surface of the electrodes.15,19 In the presence of fluorine, especially in the form of lithium salts (LiPF6), the major decomposition product that has been observed is LiF, which has been attributed to the following chemical reaction: LiPF6 → LiF + PF5.6,7,9,14,20−22 In the presence of water contamination, PF5 is believed to immediately degrade and release HF that can dissolve compounds such as Li2CO3 and lead to further LiF formation.21,23 In addition, it has been reported that the formation of SEI layer is dynamic and that both composition and thickness can vary as a function of cycle life, in a reversible or nonreversible way. Nonreversible behaviors induce a continuous growth of the SEI, leading in most cases to limited Li+ mobility and consequently to capacity fading.7,9,11,13,14,21 Various approaches have been shown in the literature to improve successfully the cycling performance of conversion materials by, for instance, doping FeF2 with oxygen to form Received: March 2, 2015 Revised: April 15, 2015

A

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

electron diffraction (SAED) pattern, annular dark field scanning transmission electron microscopy (ADF-STEM), and electron energy loss spectroscopy (EELS) were recorded at 197 kV with a JEOL-2010F microscope equipped with a Gatan-200 imaging filter (GIF) spectrometer. The energy resolution of the EELS spectra measured from the full width at half magnitude (FWHM) of the zero-loss peak (ZLP) was 0.9 eV. In addition, a collection angle (β) of 20 mrad and a convergence angle (α) of 10 mrad were used for the acquisition of EELS spectra. The program called Process Diffraction25 was used to obtain SAED intensity profiles by first taking the rotational average followed by background removal. In addition, the JEMS software26 was used for the simulation of SAED intensity profiles of nanosized particles in the 2−5 nm range. These were obtained ab initio by calculating the pair distribution function and solving the Debye equation. 4. Helium Ion Microscopy (HIM). Helium ion microscopy (HIM) was performed using a Zeiss Orion Plus system, using 30 kV acceleration voltage and a 0.2−0.4 pA beam current. No beam damage was observed using these parameters. The samples were briefly exposed to air for less than 1 min before transfer in the HIM vacuum system. 5. X-ray Photoelectron Spectroscopy (XPS). XPS measurements were performed with a Thermo ESCALAB 250 Xi using monochromated 1486.6 eV Al Kα X-rays. The total instrument resolution was approximately 0.5 eV. During measurements, a flood gun using low-energy electrons and Ar+ ions was used to compensate for sample charging. Under controlled dosage, very little differential charging was observed with only moderate peak broadening and peak energy shifts limited to about 200 meV. The component of the C 1s photoelectron peak at a binding energy of −284.8 eV corresponding to a C−C environment was used as a binding energy reference. Most core levels were fitted with Voigt profiles after subtracting a Shirley background function. Fe 2p core levels were fitted using reference spectra taken from highpurity FeF2 and Fe0. The FeF2 electrode samples were transported in a nitrogen atmosphere and introduced into the vacuum system using a nitrogen-filled glovebag. In order to determine thickness and composition of the SEI layer as well as the oxidation state of iron in the cycled cathodes, XPS spectra were taken before and after sputtering with 200 eV Ar+ ions in ultrahigh vacuum. Sputter rates were calibrated with Ta2O5 and then adjusted to account for the high carbon content of the cycled FeF2 samples. A FeF2 powder reference sample was sputtered using a low ion beam current, and its oxidation state was not observed to change. At higher current, a small amount of reduction occurred due to the preferential removal of fluorine ions from the powder. Conversely, sputtering a prepared mixture of Fe0 and PVDF powder at low current resulted in the formation of FeF2 (as shown in Supporting Information Figure S1). Consequently, the amount of FeF2 in these samples is likely overestimated after low-current sputters and underestimated after high-current sputters.

isostructural FeOxF2−x24 or via modifications of electrolyte composition.8,17 The focus of this study is solely on the identification of failure modes which lead to the poor cycling stability of pure FeF2 nanocomposite model system. In this work, we study the simultaneous morphological and chemical evolutions of the SEI layer of the FeF2 cathode material upon cycling. Helium ion microscopy (HIM) was used to characterize the material’s surface morphology at the submicron scale. Scanning transmission electron microscopy and electron energy loss spectroscopy (STEM/EELS) were used to identify bulk phases and local structure. In addition, the surface sensitive technique of X-ray photoemission spectroscopy (XPS) was used in conjunction with controlled low-energy sputter depth profiling to probe both the surface of the SEI and deeper layers of the cathode material. As the chemistry of the SEI layer can be largely affected by exposure to atmosphere, transfer from battery cells to either the TEM or XPS systems was done in a controlled environment. Using a combination of bulk (TEM) and surface (XPS) sensitive techniques, the nature of the decomposition products of the electrolyte and the degradation of the cathode material are correlated to the cycling performance of FeF2.

II. EXPERIMENTAL SECTION 1. Material Synthesis and Electrode Fabrication. The iron fluoride nanopowder used in this study was synthesized via a solution process involving reacting iron metal with a fluorosilicic acid aqueous solution. After filtering the excess iron metal, drying the solution led to a FeSiF6·6H2O precursor. The precursor was then heat-treated under flowing argon at temperature 250 °C to form 20 nm FeF2 nanoparticles. The FeF2/C nanocomposites were then prepared by high-energy ball milling with 15 wt % activated carbon black (ASupra, Norit) for 1 h in a helium atmosphere.24 2. Electrochemistry. The electrodes were fabricated using the Bellcore-developed process consisting of mixing the active materials in acetone with poly(vinylidene fluoride-co-hexafluoropropylyene) (PVDF-HFP, Kynar 2801, Elf Atochem), carbon black (Super P,MMM), and dibutyl phthalate (Aldrich). After removal of the dibutyl phthalate plasticizer, the resulting electrodes typically consisted of 57 wt % active material, 12 wt % carbon black, and 31 wt % binder. The coin cells were assembled in a helium-filled glovebox. Lithium foil was used as a counter electrode with a glass-fiber separator (GF/D, Whatman) and a layer of polypropylene separator (Celgard) saturated with 1 M LiPF6 in ethylene carbonate:dimethyl carbonate electrolyte (EC:DMC 50:50 in vol %) (BASF). Since the transformation kinetics, capacity, and reversibility of conversion electrodes are dependent on the cycling rate and temperature, all samples were cycled in galvanostatic mode at 60 °C for optimum performances, using a constant current density of 50 mAg−1 between 1.5 and 4.5 V. Further experimental details can be found elsewhere.24 3. Transmission Electron Microscopy (TEM). The electrochemical cells were disassembled in a helium-filled glovebox, and the FeF2 electrodes were rinsed in DMC in order to remove most of the LiPF6 salt in contact with the SEI layer and residual nonvolatile EC. The electrodes were then scraped to produce a fine powder, which was then dispersed in DMC (BASF). Several drops of this solution were then placed on a TEM lacy carbon film supported on a copper grid. The TEM samples were loaded on a vacuum transfer holder and transferred to the TEM without exposure to air. Selected area

III. RESULTS 1. Electrochemical Performance of FeF2/C Cathode Material. A typical specific capacity curve for the FeF2/C positive electrode as a function of cycle number is shown in Figure 1. The FeF2/C nanocomposite yields a discharge capacity of 441 mAh/g in the first cycle, which is equivalent to 519 mAh/g of active FeF2 material (corresponding to 90% of its theoretical capacity). However, the specific capacity B

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

reduced to LiF and metallic Fe. After cycling (2−20 cycles), similar structures were observed for all lithiated electrodes (Supporting Information Figure S2). SAED patterns and ADFSTEM images of the fully delithiated samples after 1 cycle and after 20 cycles are compared in Figure 4. The diffraction pattern obtained after one cycle (Figure 4a) reveals full reconversion to the initial FeF2 rutile structure. Its corresponding ADF-STEM image, shown in Figure 4b, reveals a single homogeneous phase. However, after 20 cycles, the diffraction pattern of the cathode material, shown in Figure 4c, reveals the coexistence of both LiF and metallic Fe. This inhomogeneous structure is also evident from the corresponding ADF-STEM image of Figure 4d indicating phase separation, with high-Z element corresponding to Fe nanoparticles, appearing as bright features. The corresponding SAED intensity profiles at various cycle numbers from 1 to 20 are shown in Figure 5. It can be seen that reconversion to rutile FeF2 is not occurring after cycling and that the electrode remains in the form of Fe nanoparticles along with LiF. Delithiated samples, after 10 and 20 cycles, indicated the coexistence of unconverted Fe0 and LiF phases as well as reconverted FeF2 rutile phase (c.f. Figure 5). Reconversion from Fe0 and LiF phases to rutile FeF2 is clearly impeded after cycling, resulting in the observed capacity loss. The average Fe particle size also increased from approximately 2.3 to 2.7 nm as the number of cycles increased from 1 to 20. In addition, it has been observed that (c.f. Figures 6a and 6b) some agglomerates contain larger Fe nanoparticles up to 5 nm in size at the center, but with very small Fe nanoparticles (1 nm in diameter) at the edge. This Fe particle size reduction at the edge is due possibly to partial Fe dissolution at the electrode near-surface region leaving smaller Fe at the edge of the agglomerate. Further evidence of the formation of ionic Fe2+ is presented in the following EELS and XPS results. To obtain information on the chemical evolution of the FeF2 electrodes with increasing cycle number, EELS spectra were acquired on both lithiated and delithiated samples, after 1, 2, 10, and 20 cycles in two different energy window ranges comprising the Fe−M and Li−K edges (45−80 eV) and the O−K, F−K, and Fe−L2,3 edges (500−800 eV). The EELS spectra measured on the lithiated samples (shown in Supporting Information Figure S3) reveal the presence of metallic Fe and LiF. Upon cycling, however, a gradual increase of Li and F content is observed along with the appearance of oxygen. The EELS spectra measured on the delithiated samples as a function of cycle number are shown in Figure 7. Important changes can be seen, particularly the presence of a Li−K edge signal after the 10th and 20th cycle indicated by the two arrows

Figure 1. Discharge capacity vs cycle number of FeF2/C nanocomposite electrode cycled at 50 mA/g, at 60 °C.

decreases to 300 mAh/g after five cycles and to 80 mAh/g after 20 cycles. In order to probe the mechanisms causing this capacity loss, a set of cells have been prepared and disassembled in their lithiated and delithiated states after 1, 2, 10, and 20 cycles. 2. Surface Morphology. In the cathode material, the FeF2/C nanocomposite (15−20 nm) was in the form of large micron-size agglomerates. After cycling, morphological changes were observed on the agglomerate surfaces. As an example, helium ion microscopy (HIM) images of typical surfaces of the lithiated FeF2/C agglomerate sample stopped after the first cycle and after the 20th cycle are presented in Figure 2a and Figure 2b, respectively, with Figure 2c an enlarged view of Figure 2b. Whereas the surface of the FeF2 electrode appears smooth and exhibits large flat regions after one full conversion cycle, the surface of the 20 cycle FeF2 sample is characterized by a high degree of surface roughness. This roughness is interpreted as the physical evidence of the growth of a SEI overlayer, surrounding the FeF2-containing agglomerates. In the following chapters, further analysis by STEM/EELS and XPS will provide further details on chemistry and thickness of this SEI layer. 3. STEM/EELS and Selected Area Electron Diffraction (SAED) Characterization. Two aspects of the material’s evolution upon cycling can be studied in parallel using STEM/ EELS and SAED analysis, i.e., the structural transformation of the active part of the FeF2 cathode material and the evolution of the SEI layer. The SAED pattern and corresponding ADF STEM image of the lithiated FeF2 sample after 1 cycle, shown in Figures 3a and 3b, respectively, indicate that the electrode is electrochemically

Figure 2. Helium ion microscope image of the lithiated FeF2/C material (a) after 1 cycle and (b) after 20 cycles with a higher magnification shown in (c). C

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

Figure 3. (a) SAED pattern and (b) ADF image of the lithiated sample after 1 cycle.

Figure 4. SAED pattern and ADF images of the delithiated sample (a, b) after 1 cycle and (c, d) after 20 cycles.

cyclic molecule EC has been shown to decompose at high biases, leading to the accumulation of Li2CO3.8,15 ADF-STEM and EELS were also used to follow the structure and chemistry of surface protrusions observed by HIM on the cycled FeF2 cathodes (c.f. Figure 2c). In ADF-STEM images, regions typically darker than the cathode material (attributed to the presence of only low Z-elements) were often found at the edges of the FeF2/C agglomerates. Figure 8a shows one such

in Figure 7a. A gradual increase in O−K edge signal is observed after 10 and 20 cycles (c.f. Figure 7b) along with an increase in F−K edge signal with the apparition of a fluorine postpeak (indicated by arrows) attributed to LiF.27 These results point toward the formation of an unconverted LiF layer along with the presence of oxygen. The progressive increase in oxygen content after 10 cycles in both lithiated and delithiated states can be attributed to solvent decomposition. In particular, the D

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

model.28 Higher energy resolution EELS spectra from this region display distinctively different fine structures compared to the electrode material. These experimental EELS spectra are compared in Figure 9 with those obtained from possible reference compounds resulting from electrolyte decomposition (LiF, Li2CO3) or already present in the system (PVDF-HFP). A comparison of the Li−K, C−K, O−K, and F−K edges taken from the SEI layer (Figure 9a−d) suggests that Li is mainly in the form of LiF, and not in the form of Li2CO3. Specific features are the separation distance between peaks in the Li−K edge of 7.2 eV for LiF compared to 5.2 eV for Li2CO3 and the presence of strong prepeaks of the O−K edge of Li2CO3 which are not present in the O−K edge from the SEI layer. Furthermore, the F−K edge of the SEI shown in Figure 9d includes features associated with LiF such as the separation of 24.6 eV between peaks and the presence of a postpeak marked by an arrow.27,29 Although Li2CO3 has been commonly reported in the literature for similar (EC-DMC) systems,8,15 our EELS spectroscopy indicates that the main component of the SEI layer is in the form of LiF. This is also reflected in the composition of this layer (Li0.5C0.1O0.1F0.3) containing three times more F than O. Nevertheless, the detection of both C and O in the SEI layer still suggests the presence of some Li2CO3. It is to be noted that although low electron dose spectroscopy was used, damage caused by electron beam might preclude the detection of carbonates via detection of their EELS fine structures. In the F−K edge from the SEI layer shown in Figure 9d, there is also an additional peak marked by a circle at around 708.5 eV that does not correspond to any known F-containing compound likely to be found in the cycled electrode (e.g., PVDF). However, by subtracting the F−K edge of the SEI spectrum from a LiF standard spectrum as shown in Figure 10, two peaks appear (located at 708 and 720.3 eV) separated by 12.3 eV which correspond to a small Fe−L2,3 edge signal with Fe either in the Fe2+ or Fe0 oxidation state.30 Given the absence of high-Z contrast in the corresponding ADF-STEM image (Figure 8a), the presence of metallic Fe0 nanoparticles is unlikely. Thus, ionic Fe2+ is present in the SEI layer. Further evidence of its presence is described in the next section.

Figure 5. SAED intensity profiles at various cycle numbers for delithiated FeF2/C. The FeF2 rutile structure is reconverted up to 10 cycles. For 20 cycles, only unconverted metallic Fe nanoparticle and LiF are observed.

region, observed on a lithiated cathode after 20 cycles. These surface protrusions grew larger and more numerous with increasing cycle number, in both lithiated and delithiated samples. The measured thickness was inhomogeneous and ranged between 15 and 100 nm. These inhomogeneous protrusions observed in ADF-STEM mode are attributed to the developing SEI layer. A spliced EELS spectrum with energy range from 50 to 750 eV measured from this region is shown in Figure 8b, revealing the presence of Li−K, C−K, O−K, and F−K edges. Phosphorus is more difficult to observe by EELS but was not detected in this analysis. Additional measurements by X-ray EDS did show trace quantities of P on the order of 1−2%. From this spectrum, a semiquantitative atomic composition of the overlayer can be estimated as Li0.5C0.1O0.1F0.3 using theoretical ionization cross sections based on a Hartree−Slater

Figure 6. ADF-STEM images of the (a) lithiated and (b) delithiated samples after 20 cycles. E

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

Figure 7. EELS spectra of the delithiated sample after various cycle number (a) Li−K and Fe-M edges, (b) O−K, F−K, and Fe−L2,3 edges.

Figure 8. (a) ADF-STEM image of lithiated FeF2 after 20 cycles and (b) the corresponding spliced EELS spectrum revealing the presence of Li, C, O, and F.

sputtering of 10 nm of material, metallic Fe0 peaks became visible at binding energies −708 and −720 eV due to the selective removal of fluorine. However, the relative intensity of the Fe 2p does not increase, suggesting that the FeF2 electrode is relatively homogeneous below 1 nm; i.e., the SEI layer is 1 nm thick. For this and all other samples, the intensity of the Fe 2p peak is used to determine the thickness of the SEI layer; i.e., when the Fe 2p intensity stops increasing with sputter time, the SEI has been fully removed. The C 1s core level spectra (c.f. Figure 11b) reveal a rich chemistry, evolving with sputter depth. The surface spectrum can be decomposed into several components corresponding to different chemical bonds. The peak at −284.8 eV is attributed to C−C or C−H bonds from the PVDF-based binder, carbon black, and various hydrocarbons.16 The most intense peaks at −286.2 and −290.5 eV are attributed to C−O and CO3 bonds, respectively, and are both likely from the EC:DMC electrolyte solvent. The ratio of C−O to CO3 in EC:DMC should be 2:1 (Supporting Information Figure S5), so the remaining CO3 signal could be attributed to the formation of Li-alkyl carbonates (ROCO2Li), Li-alkoxides (R−CH2OLi), and/or Li2CO3.15,18,19,32 The small peak at −293 eV corresponds to CF3 bonds in PVDF-HFP by comparison with XPS of standard PVDF powder (Supporting Information Figure S4). After

4. XPS Characterization. A series of XPS core level spectra of a delithiated FeF2 electrode after 1 cycle is shown in Figure 11. The thickness appended to each spectrum (surface, 1 and 10 nm) corresponds to the estimated sputter depth at which the spectrum was acquired. Sputtering was performed using a 200 eV Ar+ beam, whose energy and flux were chosen to be small enough to minimize bond breaking and selective sputtering of lighter elements. The Fe 2p core levels at the surface and after sputtering to 1 and 10 nm are depicted in Figure 11a. Before sputtering (surface), a small Fe 2p signal is visible, suggesting that most of the Fe in the sample is located in the subsurface region and its signal is attenuated by the SEI layer. The most intense peaks in this broad spectrum are at binding energies of about −712 and −725 eV, corresponding to the 2p3/2 and 2p1/2 components of Fe2+, respectively.31 After sputtering 1 nm of material, the signature of FeF2 is more prominent, and the intensity in the region of the Fe 2p core levels can be in part fitted with a FeF2 reference spectrum (red), leaving some intensity around −720 eV. This intensity is attributed to a plasmon loss peak from the F 1s core level, since this intensity is located in the Fe 2p3/2 region and has no accompanying component in the 2p1/2 region. Reference spectra from PVDF support this attribution (Supporting Information Figure S4). After a high-current F

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

Figure 9. Comparison of SEI layer EELS fine structure in lithiated FeF2 after 20 cycles with reference compounds (LiF, Li2CO3, and PVDF) for (a) Li−K edge, (b) C−K edge, (c) O−K edge, and (d) F−K edge.

XPS spectra of the nonsputtered surface for all the samples studied in this work. Its peak position at −134 eV indicates that the electrolyte remaining on the surface was oxidized to LixPOyFz. The small F 1s component at −685.5 eV corresponds to FeF2.33 After sputtering to 1 nm depth, the LiPF6 and PVDF components are reduced in intensity, while the FeF2 peak grows, in agreement with what was observed for the Fe 2p spectra. Sputtering to 10 nm further decreases the signal from the electrolyte, while the FeF2 peak remains constant. The surface O 1s spectrum (c.f. Figure 11d) can be decomposed into three peaks. The small peak at −530.6 eV is characteristic of transition metal oxides and corresponds probably to a small amount of FeO that has formed at the surface of the electrode due to a reaction between Fe0 or FeF2 with the electrolyte or SEI.19 The main O 1s peak at −532 eV is attributed to C−O and CO3 bonds, corresponding to EC:DMC and Li2CO3, while the smaller peak at −533.6 eV is attributed to ROCO2Li and/or R−CH2OLi.15,18,19,32 Each of these components decreases after sputtering, again confirming that the SEI layer is localized at the outermost 1−2 nm of the electrode. The Fe 3p and Li 1s core levels are shown in Figure 11f, but their overlapping intensities preclude definitive peak assignments. Nevertheless, the surface Fe 3p/Li 1s spectrum can be described as a combination of Fe2+ features at −56 eV and Li 1s intensity at about −55 eV. After sputtering, the Fe2+ intensity increases, and a small Fe0 feature appears at −53 eV, also in agreement with the Fe 2p spectrum of Figure 11a. The XPS spectra for the delithiated electrode after 2 cycles were similar, except for a slightly thicker SEI layer of about 5 nm. A similar set of XPS core levels for a delithiated electrode after 10 cycles is shown in Figure 12 at which point the electrode has lost more than 50% of its initial capacity. The Fe 2p surface spectrum (Figure 12a) again contains only Fe2+

Figure 10. Subtracted EELS spectrum of SEI layer from LiF (SEI is from the same area in Figure 8a). The subtracted spectrum shows the presence of Fe−L2,3 edges.

sputtering, the C 1s intensity from EC:DMC is reduced, again suggesting that the SEI is mainly confined to the top 1 nm of the electrode. The carbon black component (C−C bonds at −284.8 eV) increases in intensity and remains constant after 10 nm of sputtering. The surface F 1s core level spectrum (c.f. Figure 11c) has two major peaks at −687 and −688 eV, which are commonly attributed to C−F and LiPF6, respectively.16 However, the P 2p peak (c.f. Figure 11e) represents only 3% of the LiPF6 signal from the F 1s peak, suggesting that this peak assignment is probably misattributed in most cases. The phosphorus intensity, although weak, was consistently observed in the G

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

Figure 11. XPS spectra of a delithiated FeF2 electrode after 1 cycle at different sputter depths with (a) Fe 2p, (b) C 1s, (c) F 1s, (d) O 1s, (e) P 2p, and (f) overlapped Fe 3p and Li 1s peaks. The Fe 2p peak can be fitted very well by a single FeF2 component, indicating that the active cathode material has fully reconverted during delithiation. SEI composition is discussed in the text.

Figure 12. XPS of a delithiated FeF2 electrode after 10 cycles at different sputter depths with (a) Fe 2p, (b) C 1s, (c) F 1s, (d) O 1s, (e) P 2p, and (f) overlapped Fe 3p and Li 1s peaks. The line shape of the Fe 2p peak is indicative of FeF2, Fe0, and a third component (possibly Li1−xFexF). The composition of the SEI is consistent with that of the 1 cycle sample, although this SEI is substantially thicker.

of the cathode at this stage, leading to some capacity losses. The Fe 2p intensity increases slightly after 10 nm of sputtering and remains constant thereafter, suggesting that the SEI layer is now 10 nm thick. Again, due to the higher sputter current, the iron was reduced and metallic Fe peaks appeared. The C 1s spectra in Figure 12b are similar to those of the 1 cycle sample (c.f. Figure 11b), except for the addition of a peak at −283.7 eV. This peak position is characteristic of CC bonds and is likely from carbon black.36 Otherwise, the similarity between the 1 cycle and 10 cycle spectra provide further evidence that the carbonaceous compounds in the SEI are constant from one cycle to the next. In Figure 12c, the F 1s surface spectrum contains a new component at −684.6 eV, which is attributed to LiF. This agrees with the TEM-EELS data, which suggests that LiF is not completely removed from the cathode upon delithiation and likely inhibits both electronic and ionic

components. However, three different components are necessary to fit the sputtered Fe 2p spectra: FeF2 (shown in red), Fe0 (gray), and a third component (yellow) derived from XPS spectra of lithiated FeF2 thin films.34 This additional component, comprising 10% of the Fe 2p intensity, has peaks at binding energies of −709 and −722 eV and intense satellite features at −715.5 and −729.5 eV. Since this component has been observed in similar quantities in XPS studies of lithiated FeF2 thin films without any sputtering, it is not believed to be caused by ion beam damage. Furthermore, this component was present in all Fe 2p spectra of cycled FeF2 materials except for the delithiated 1 cycle sample. This component might possibly be attributed to Fe2+ ions trapped in LiF in the form of FexLi2−2xF2, as suggested by previous PDF measurements of FeF2 electrodes.35 Regardless of the origin of this component, it is clear that FeF2 is not completely reformed upon delithiation H

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

Figure 13. XPS of a lithiated FeF2 electrode after 10 cycles at different sputter depths with (a) Fe 2p, (b) C 1s, (c) F 1s, (d) O 1s, (e) P 2p, and (f) overlapped Fe 3p and Li 1s peaks. The line shape of the Fe 2p peak is indicative of FeF2, Fe0, and a third component (possibly Li2−2xFexF).

increases as a function of cycle number. The F 1s surface spectrum does not contain any component corresponding to FeF2 compound, suggesting that FeO is the only iron compound present at the surface of this electrode. After sputtering, the FeF2 peak become visible and the FeO peak in the O 1s spectrum decreases. The P 2p spectra in Figure 13e reassert the presence of LixPOyFz confined to the SEI layer. Finally, as expected, the Li 1s spectra in Figure 13f show a much greater concentration of lithium in this lithiated sample than in the delithiated samples. An estimated average SEI layer thickness as a function of cycle number based on the intensity of the Fe 2p peak is shown in Table 1 with the percentages of the F 1s signal attributed to

transport in the electrochemical cell. LiF is present throughout the 10 nm depth of the SEI. The O 1s spectra in Figure 12d again have peaks characteristic of ROCO2Li, EC:DMC, and FeO, although these peaks are more intense than those in Figure 11d. The amount of FeO is 40% of the total iron content at the surface of the electrode after 10 cycles and 5% of the Fe content at 10 nm and below. The P 2p spectrum in Figure 12e indicates an increased presence of LixPOyFz. Lastly, the Li 1s spectra in Figure 12f show a greater Li intensity than those of the 1 cycle sample, as evidenced by the large, narrow peak at −56 eV for the surface and 1 nm spectra. This provides further evidence of the formation of a thick LiF-rich SEI layer after 10 cycles. The trend of increasing SEI thickness and greater LiF and FeO content at the surface continues in the 20 cycle samples, shown in Supporting Information Figures S6 and S7. A similar set of XPS spectra is shown in Figure 13 for a lithiated sample after 10 cycles, in order to provide a direct comparison to the delithiated samples. The Fe 2p surface spectrum depicted in Figure 13a has a much lower intensity than the same spectrum from the delithiated sample, suggesting that the SEI layer grows thicker upon lithiation and decreases upon delithiation. This observation is consistent across the whole set of lithiated/delithiated samples. The relatively small Fe0 intensity (gray) with respect to the FeF2 signal, visible after sputtering, is probably not representative of the real amount of metallic iron after lithiation, due to the sputter-induced formation of FeF2 discovered in this study (Supporting Information Figure S1). Additionally, the line shape of the Fe 2p core levels was best decomposed by adding a component (orange) similar to the one of Figure 11a. The Fe 2p core levels thus indicate at least three chemical environments for iron. Finally, the total Fe 2p intensity increases as a function of sputter time until about 50 nm of material has been removed from the surface indicating a 50 nm SEI overlayer thickness. The C 1s, F 1s, and O 1s spectra in Figures 13b, 13c, and 13d, respectively, contain the same peaks as those of the delithiated samples (c.f. Figure 12). This suggests that the chemical composition of the SEI does not change during cycling, although the thickness increases and the relative amount of LiF

Table 1. Summary of XPS Results for Lithiated and Delithiated Samples: SEI Thickness As Estimated by the Amount of Sputtering Necessary To Reach a Stable Fe 2p Intensity and Percentage of the Total F 1s Signal Attributed to LiF SEI thickness (nm)

LiF at surface (% of total F 1s signal)

no. of cycles

lithiated

delithiated

lithiated

delithiated

1 2 10 20

10 50 >50 >50

1 5 20 20

5% 3% 18% 36%

0% 0% 21% 13%

the LiF phase. These XPS results indicate that the thickness of this SEI is much larger for lithiated samples (from 10 to more than 50 nm) as compared to delithiated samples (from 1 to 20 nm). Furthermore, the amount of LiF at the surface is also largest upon lithiation and is partially reduced upon delithiation for all cycles. In addition to changes in SEI layer thickness, several phases have been identified throughout the SEI layer regardless of cycle number: (1) an increasingly growing LiF phase caused by reconversion of LiF + Fe back into FeF2 as well as the dissociation of EC:DMC solvent and LiPF6; (2) oxidized lithium salts LiOxPyFz from LiPF6 degradation; (3) minor phases of Li2CO3 or ROCOOLi due to EC:DMC solvent decomposition; (4) a FeO phase at the outermost I

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C surface due to reaction with the oxygen-containing solvent; (5) an increasing number of Fe2+ cations possibly trapped in the LiF phase.

IV. DISCUSSION The results of this combined STEM/EELS and XPS study reveal a certain number of features of SEI layer morphology, thickness, and composition that could explain the capacity loss upon cycling. First the main composition of this SEI layer is LiF which is only partially reduced upon delithiation and this for all cycles, i.e., (2LiF + Fe = FeF2 + 2Li). This has been observed via an increase in LiF content after delithiation as well as from the presence of unreacted Fe nanoparticles. Several additional surface phases have been observed also by XPS such as LiOxPyFz and Li2CO3 from a reaction with oxygen-containing solvents. The presence of LiF is typically interpreted as a result of incomplete reconversion and a byproduct of decomposition of LiPF6 and Li2CO3. However, very little phosphorus has been detected by both EELS and XPS, indicating either the removal of all P-containing materials upon rinsing in DMC or that another reaction path exists for LiF formation. The most significant finding observed by both XPS and EELS is the presence of Fe2+ in the LiF layer. This Fe loss seems irreversible since Fe2+ has been observed after cycling in both lithiated and delithiated electrodes. Such a Fe loss could be a factor in the measured capacity loss upon cycling via a loss of active material and the formation of regions with smaller Fe nanoparticles and having larger interparticle distance. The incorporation of Fe2+ into the SEI is consistent with work done on other metal-containing electrodes, indicating metal ions migration away from the surface of the active material.9,20,37,38 Our XPS results show that on the surface, Fe2+ is in the form of FeO. The remaining Fe2+ could be incorporated into the LiF lattice via substitution with 2Li+. Indeed, a recent work on the lithiation/delithiation of FeF235 indicates that LiF lattice parameters expand upon delithiation as a result of the substitution of 2Li+ ions in the LiF lattice by larger Fe2+ cations, leading to the formation of a new FexLi2−2xF2 compound. In our study, we observe an increase in trapped Fe2+ upon cycling with a resulting increase in untransformed LiF content leading to capacity loss. In addition, EELS and XPS results reveal the incorporation of oxygen at the surface of the electrode. From XPS data, some of the Fe2+ is in the form of FeO and not incorporated into the LiF lattice as described previously. The amount of FeO is estimated to be 10% of the iron content near the surface of the cathode based on the relative peak intensity. Furthermore, the decreased FeO intensity after sputtering suggests that Fe oxidation occurs in the near-surface region and not into the bulk of the electrode. The incorporation of oxygen into the top layer of BiF3 positive electrode from the decomposition of the SEI layer has been also reported by Gmitter et al. as a cause of poor cycling.9 Based on the observed microstructural changes (ADFSTEM) as well as chemical evolution (EELS and XPS) made in this study, we can draw the following schematic illustration of the effect of cycling on morphology and chemistry of SEI layer after 1 cycle and 20 cycles (c.f. Figures 14a and 14b, respectively). We propose that upon first lithiation, reversible conversion products are formed consisting of Fe nanoparticles embedded in LiF. Upon cycling, Fe2+ ions start accumulating in the LiF surface layer resulting in smaller Fe nanoparticles with larger interparticle distance. On the surface of the agglomerates

Figure 14. Schematic illustration showing the evolution of SEI layer in discharged FeF2 cathode after (a) 1 cycle and (b) 20 cycles.

in contact with the electrolyte, a fraction of Fe2+ is most likely in the form of FeO. The combination of increase in Fe interparticle distance with formation of oxidized surface SEI layer leads to increasing poor ionic and electronic conductivity. These kinetic limitations in the near-surface regions lead upon cycling to incomplete reconversion, resulting in capacity loss of the electrochemical cell.

V. SUMMARY AND CONCLUSIONS The SEI formation and evolution upon cycling of conversion FeF2/C cathode have been investigated by combined SAED, STEM-EELS, and XPS techniques. STEM/EELS and XPS results indicate that reconversion is strongly impeded with increasing cycle number, leaving mainly LiF and Fe0 after the 20th cycle. The combination of STEM/EELS and XPS provides clear evidence for the growth of a SEI layer at the surface of the cathode material upon cycling. The thickness of this SEI is much larger for lithiated as compared to delithiated samples. Two main mechanisms have been identified leading to capacity loss. First, an increasing amount of active Fe material is trapped as ionic Fe2+ in LiF with, on the surface, formation of FeO. Thus, part of the active material (Fe) cannot participate in the reconversion process. Second, the resulting larger Fe interparticle distance with the growth of an insulating LiFrich SEI layer can slow or even prevent ionic and electronic transport to and from the active material.

■

ASSOCIATED CONTENT

■

AUTHOR INFORMATION

* Supporting Information S

References with more than 10 authors and additional figures. This material is available free of charge via the Internet at http://pubs.acs.org Corresponding Author

*Tel.: 848 445 4942. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported as part of the NorthEast Center for Chemical Energy Storage (NECCES), an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences under Award DESC0012583 and DE-SC0001294. R.T. was supported by National Science Foundation Grant No. 0903661: Nanotechnology for Clean Energy IGERT. J

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

■

(18) Swiatowska-Mrowiecka, J.; Maurice, V.; Zanna, S.; Klein, L.; Marcus, P. XPS Study Of Li Ion Intercalation in V2O5 Thin Films Prepared by Thermal Oxidation of Vanadium Metal. Electrochim. Acta 2007, 52, 5644−5653. (19) Tian, B.; Swiatowska, J.; Maurice, V.; Zanna, S.; Seyeux, A.; Klein, L. H.; Marcus, P. Combined Surface and Electrochemical Study of the Lithiation/De-lithiation Mechanism of the Iron Oxide ThinFilm Anode for Lithium-Ion Batteries. J. Phys. Chem. C 2013, 117, 21651−21661. (20) Balasubramanian, M.; Lee, H. S.; Sun, X.; Yang, X. Q.; Moodenbaugh, A. R.; McBreen, J.; Fischer, D. A.; Fu, Z. Formation of SEI on Cycled Lithium-Ion Battery Cathodes. Soft X-ray Absorption Study. Electrochem. Solid-State Lett. 2002, 5, A22−A25. (21) Dedryvere, R.; Laruelle, S.; Grugeon, S.; Gireaud, L.; Tarascon, J. M.; Gonbeau, D. XPS Identification of the Organic and Inorganic Components of the Electrode/Electrolyte Interface Formed on a Metallic Cathode. J. Electrochem. Soc. 2005, 152, A689−A696. (22) Wang, F.; Graetz, J.; Moreno, M. S.; Ma, C.; Wu, L.; Volkov, V.; Zhu, Y. Chemical Distribution and Bonding of Lithium in Intercalated Graphite: Identification with Optimized Electron Energy Loss Spectroscopy. ACS Nano 2011, 5, 1190−1197. (23) Plakhotnyk, A. V.; Ernst, L.; Schmutzler, R. Hydrolysis in the System LiPF6Propylene CarbonateDimethyl CarbonateH2O. J. Fluorine Chem. 2005, 126, 27−31. (24) Pereira, N.; Badway, F.; Wartelsky, M.; Gunn, S.; Amatucci, G. G. Iron Oxyfluorides as High Capacity Cathode Materials for Lithium Batteries. J. Electrochem. Soc. 2009, 156, A407−A416. (25) Labar, J. L.; Adamic, M.; Barna, B. P.; Czigzny, Z.; Fogarassy, Z.; Horvath, Z. E.; Geszti, O.; Misjak, F.; Morgiel, J.; Radnoczi, G.; et al. Electron Diffraction Based Analysis of Phase Fractions and Texture in Nanocrystalline Thin Films, Part III: Application Examples. Microsc. Microanal. 2012, 18, 406−420. (26) Stadelmann, P. http://cimewww.epfl.ch/people/stadelmann/ jemswebsite/jems.html. (27) Sina, M.; Nam, K. W.; Su, D.; Pereira, N.; Yang, X. Q.; Amatucci, G. G.; Cosandey, F. Structural Phase Transformation and Fe Valence Evolution in FeOxF2−x/C Nanocomposite Electrodes During Lithiation and De-Lithiation Processes. J. Mater. Chem. A 2013, 1, 11629−11640. (28) Egerton, R. Electron Energy-Loss Spectroscopy in the Electron Microscope; Springer: New York, 1996. (29) Cosandey, F.; Su, D.; Sina, M.; Pereira, N.; Amatucci, G. G. Fe Valence Determination and Li Elemental Distribution in Lithiated FeO0.7F1.3/C Nanocomposite Battery Materials by Electron Energy Loss Spectroscopy (EELS). Micron 2012, 43, 22−29. (30) Cosandey, F.; Al-Sharab, J. F.; Badway, F.; Amatucci, G. G.; Stadelmann, P. EELS Spectroscopy of Iron Fluorides and FeFx/C Nanocomposite Electrodes Used in Li-Ion Batteries. Microsc. Microanal. 2007, 13, 87−95. (31) Grosvenor, A. P.; Kobe, B. A.; Biesinger, M. C.; McIntyre, N. S. Investigation Of Multiplet Splitting of Fe 2p XPS Spectra and Bonding in Iron Compounds. Surf. Interface Anal. 2004, 36, 1564−1574. (32) Swiatowska-Mrowiecka, J.; de Diesbach, S.; Maurice, V.; Zanna, S.; Klein, L.; Briand, E.; Vickridge, I.; Marcus, P. Li-Ion Intercalation in Thermal Oxide Thin Films of MoO3 As Studied by XPS, RBS, and NRA. J. Phys. Chem. C 2008, 112, 11050−11058. (33) Rangan, S.; Thorpe, R.; Bartynski, R. A.; Sina, M.; Cosandey, F.; Celik, O.; Mastrogiovanni, D. D. T. Conversion Reaction of FeF2 Thin Films upon Exposure to Atomic Lithium. J. Phys. Chem. C 2012, 116, 10498−10503. (34) Thorpe, R.; Rangan, S.; Sina, M.; Cosandey, F.; Bartynski, R. A. Conversion Reaction Of CoO Polycrystalline Thin Films Exposed to Atomic Lithium. J. Phys. Chem. C 2013, 117, 14518−14525. (35) Ko, J. K.; Wiaderek, K. M.; Pereira, N.; Kinnibrugh, T. L.; Kim, J. R.; Chupas, P. J.; Chapman, K. W.; Amatucci, G. G. Transport, Phase Reactions, and Hysteresis of Iron Fluoride and Oxyfluoride Conversion Electrode Materials for Lithium Batteries. ACS Appl. Mater. Interfaces 2014, 6, 10858−10869.

REFERENCES

(1) Amatucci, G. G.; Pereira, N.; Badway, F.; Sina, M.; Cosandey, F.; Ruotolo, M.; Cao, C. Formation of Lithium Fluoride/Metal Nanocomposites for Energy Storage through Solid State Reduction of Metal Fluorides. J. Fluorine Chem. 2011, 132, 1086−1094. (2) Badway, F.; Pereira, N.; Cosandey, F.; Amatucci, G. G. Carbon− Metal Fluoride Nanocomposites. Structure and Electrochemistry of FeF3:C. J. Electrochem. Soc. 2003, 150, A1209−A1218. (3) Cabana, J.; Monconduit, L.; Larcher, D.; Palacin, M. R. Beyond Intercalation-Based Li-Ion Batteries: The State of the Art and Challenges of Electrode Materials Reacting through Conversion Reactions. Adv. Mater. 2010, 22, E170−E192. (4) Wang, F.; Robert, R.; Chernova, N. A.; Pereira, N.; Omenya, F.; Badway, F.; Hua, X.; Ruotolo, M.; Zhang, R.-G.; Wu, L.-J.; et al. Conversion Reaction Mechanisms in Lithium Ion Batteries: Study of the Binary Metal Fluoride Electrodes. J. Am. Chem. Soc. 2011, 133, 18828−18836. (5) Amatucci, G. G.; Pereira, N. Fluoride Based Electrodes Materials for Advanced Energy Storage Devices. J. Fluorine Chem. 2007, 128, 243−262. (6) Cherkashinin, G.; Nikolowski, K.; Ehrenberg, H.; Jacke, S.; Dimesso, L.; Jaegermann, W. The Stability of the SEI Layer, Surface Composition, and the Oxidation State of Transition Metals at the Electrolyte−Cathode Interface Impacted by the Electrochemical Cycling: X-ray Photoelectron Spectroscopy Investigation. Phys. Chem. Chem. Phys. 2012, 14, 12321−12331. (7) Cuisinier, M.; Martin, J. F.; Moreau, P.; Epicier, T.; Kanno, R.; Guyomard, D.; Dupre, N. Quantitative MAS NMR Characterization of the LiMn1/2Ni1/2O2 Electrode/Electrolyte Interphase. Solid State Nucl. Magn. Reson. 2012, 42, 51−61. (8) Gmitter, A. J.; Badway, F.; Rangan, S.; Bartynski, R. A.; Halajko, A.; Pereira, N.; Amatucci, G. G. Formation, Dynamics, and Implication of Solid Electrolyte Interphase in High Voltage Reversible Conversion Fluoride Nanocomposites. J. Mater. Chem. 2010, 20, 4149−4161. (9) Gmitter, A. J.; Halajko, A.; Sideris, P. J.; Greenbaum, S. G.; Amatucci, G. G. Subsurface Diffusion of Oxide Electrolyte Decomposition Products in Metal Fluoride Nanocomposite Electrodes. Electrochim. Acta 2013, 88, 735−744. (10) Marino, C.; Darwiche, A.; Dupre, N.; Wilhelm, H. A.; Lestriez, B.; Martinez, H.; Dedryvere, R.; Zhang, W.; Ghamouss, F.; Lemordant, D.; et al. Study of the Electrode/Electrolyte Interface on Cycling of a Conversion Type Electrode Material in Li Batteries. J. Phys. Chem. C 2013, 117, 19302−19313. (11) Edstrom, K.; Gustafsson, T.; Thomas, J. O. The Cathode− Electrolyte Interface in the Li-Ion Battery. Electrochim. Acta 2004, 50, 397−403. (12) Aurbach, D. Review Of Selected Electrode−Solution Interactions Which Determine the Performance of Li and Li Ion Batteries. J. Power Sources 2000, 89, 206−218. (13) Aurbach, D.; Zaban, A.; Ein-Eli, Y.; Weissman, I.; Chusid, O.; Markovsky, B.; Levi, M.; Levi, E.; Schechter, A.; Granot, E. Recent Studies on the Correlation between Surface Chemistry, Morphology, Three-Dimensional Structures, and Performance of Li and Li−C Intercalation Anodes in Several Important Electrolyte Systems. J. Power Sources 1997, 68, 91−98. (14) Bryngelsson, H.; Stjerndahl, M.; Gustafsson, T.; Edstroem, K. How Dynamic is the SEI? J. Power Sources 2007, 174, 970−975. (15) Dedryvere, R.; Laruelle, S.; Grugeon, S.; Poizot, P.; Gonbeau, D.; Tarascon, J. M. Contribution of X-ray Photoelectron Spectroscopy to the Study of the Electrochemical Reactivity of CoO Toward Lithium. Chem. Mater. 2004, 16, 1056−1061. (16) Dedryvere, R.; Leroy, S.; Martinez, H.; Blanchard, F.; Lemordant, D.; Gonbeau, D. XPS Valence Characterization of Lithium Salts as a Tool To Study Electrode/Electrolyte Interfaces Of Li-Ion Batteries. J. Phys. Chem. B 2006, 110, 12986−12992. (17) Gmitter, A. J.; Gural, J.; Amatucci, G. G. Electrolyte Development for Improved Cycling Performance of Bismuth Fluoride Nanocomposite Positive Electrodes. J. Power Sources 2012, 217, 21− 28. K

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C (36) Zhou, X.; Wan, L.-J.; Guo, Y.-G. Synthesis of MoS2 Nanosheet− Graphene Nanosheet Hybrid Materials for Stable Lithium Storage. Chem. Commun. 2013, 49, 1838−1840. (37) Amine, K.; Chen, Z.; Zhang, Z.; Liu, J.; Lu, W.; Qin, Y.; Lu, J.; Curtis, L.; Sun, Y.-K. Mechanism of Capacity Fade Of MCMBLi1.1[Ni1/3Mn1/3Co1/3]0.9O2 Cell at Elevated Temperature and Additives To Improve Its Cycle Life. J. Mater. Chem. 2011, 21, 17754−17759. (38) Song, B.; Liu, Z.; Lai, M. O.; Lu, L. Structural Evolution and the Capacity Fade Mechanism upon Long-Term Cycling in Li-Rich Cathode Material. Phys. Chem. Chem. Phys. 2012, 14, 12875−12883.

L

DOI: 10.1021/acs.jpcc.5b02058 J. Phys. Chem. C XXXX, XXX, XXX−XXX