Quantitative Operando Visualization of Electrochemical Reactions

Advanced Research Division, Panasonic Corporation , 3-1-1 Yagumo-naka-machi, Moriguchi , Osaka 570-8501 , Japan. ‡ Department of Crystalline Materia...
0 downloads 0 Views 3MB Size
Letter Cite This: Nano Lett. XXXX, XXX, XXX−XXX

pubs.acs.org/NanoLett

Quantitative Operando Visualization of Electrochemical Reactions and Li Ions in All-Solid-State Batteries by STEM-EELS with Hyperspectral Image Analyses Yuki Nomura,*,†,‡ Kazuo Yamamoto,§ Tsukasa Hirayama,§,¶ Mayumi Ohkawa,† Emiko Igaki,† Nobuhiko Hojo,† and Koh Saitoh¶ †

Advanced Research Division, Panasonic Corporation, 3-1-1 Yagumo-naka-machi, Moriguchi, Osaka 570-8501, Japan Department of Crystalline Materials Science, Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8603, Japan § Nanostructures Research Laboratory, Japan Fine Ceramics Center, 2-4-1 Mutsuno, Atsuta-ku, Nagoya, Aichi 456-8587, Japan ¶ Advanced Measurement Technology Center, Institute of Materials and Systems for Sustainability, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8603, Japan

Downloaded via UNIV OF WINNIPEG on August 27, 2018 at 15:04:37 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.



S Supporting Information *

ABSTRACT: All-solid-state lithium-ion batteries (LIBs) are one of the promising candidates to overcome some issues of conventional LIBs with liquid electrolytes. However, high interfacial resistance of Li-ion transfer at the electrode/solid electrolyte limits their performance. Thus, it is important to clarify interfacial phenomena in a nanometer scale. Here, we present a new method to dynamically observe the Li-ion distribution and Co-ion electronic states in a LiCoO2 cathode of the all-solid-state LIB during charge and discharge reactions using operando scanning transmission electron microscopy (STEM) and electron energy-loss spectroscopy (EELS). By applying a hyperspectral image analysis of non-negative matrix factorization (NMF) to the STEM-EELS, we succeeded in clearly observing the quantitative Li-ion distribution in the operando condition. We found from the operando observation with NMF that the Li ions did not uniformly extract/insert during the charge/discharge reactions, and the activity of the electrochemical reaction depended on the Li-ion concentration in a pristine state. An electrochemically inactive region was formed about 10−20 nm near the LiCoO2/Li2O−Al2O3−TiO2−P2O5based solid electrolyte interfaces. The STEM-EELS, electron diffraction, and Raman spectroscopy experimentally showed that the inactive region was a mixture of LiCoO2 and Co3O4, leading to the higher interfacial resistance of the Li-ion transfer because Co3O4 does not have pathways of Li-ion diffusion in its crystal. KEYWORDS: In situ, scanning transmission electron microscopy, electron energy-loss spectroscopy, all-solid-state lithium-ion battery, LiCoO2, non-negative matrix factorization

A

by coating LiNbO3 on cathode materials, which lead to a lower interfacial resistance by 2 orders of magnitude.6 Two models have been proposed as the origin of the interfacial resistance. One is the space-charge layer model proposed by Haruyama et al.9 The other is the structural disorder layer model proposed by Sakuda et al.10,11 Both of the models consider that the resistance is originated by the significant changes in the chemical composition, which includes Li ion at the interfaces. These layers commonly form on a nanometer scale. Thus, an in situ transmission

ll-solid-state lithium-ion batteries (LIBs) have recently attracted great attention because of their safety, long lifetime, and high energy density. One of the key ways to improve their performance is to increase power density. The ionic conductivity in solid electrolytes and the ion transfer resistance at the electrode/solid-electrolyte interfaces mainly determine the power density in all-solid-state LIBs. Several kinds of solid electrolytes have exhibited a higher conductivity comparable to that of conventional liquid electrolytes.1−3 However, the interfacial resistance of the lithium-ion (Li-ion) transfer between cathodes and solid electrolytes remains large.4−7 For this reason, considerable research has been focused on reducing the interfacial resistance, for example, by inducing dielectric BaTiO3 nanoparticles to the interfaces8 or © XXXX American Chemical Society

Received: June 26, 2018 Revised: August 13, 2018 Published: August 21, 2018 A

DOI: 10.1021/acs.nanolett.8b02587 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 1. Configuration and property of an all-solid-state Li-ion battery. (a) Schematic of the battery. (b) First charge and discharge curves cycled at a constant current of 50 nA and constant voltage. (c) Low magnification ADF-STEM image around an interface between the LiCoO2 cathode film and the LASGTP solid electrolyte before the electrochemical reactions. The dotted rectangle represents the area for operando STEM-EELS observation.

operation. We applied hyperspectral image analyses of nonnegative matrix factorization (NMF)20,21 and SA/SB method22 to visualize the dynamic changes in a quantitative Li-ion concentration and the distribution of Co ions with different valence states clearly. The results of the present study reveal in detail where the Li ions easily moved in the LiCoO2 cathode and how the electrochemical reactions occurred at the interface between the LiCoO2 cathode and the LASGTP solid electrolyte. Figure 1a shows a schematic cross section of the all-solidstate LIB prepared for operando STEM-EELS observation. We used a 50 μm thick LASGTP sheet (Li2O−Al2O3−TiO2− P2O5-based glass ceramics, OHARA Inc.) as the solid electrolyte, which has a high ionic conductivity of 1 × 10−4 S/cm at room temperature.23−25 A 150 nm thick LiCoO2 film was deposited on one side of the LASGTP sheet as the cathode by pulsed laser deposition (PLD). The temperature of the LASGTP sheet was set to 550 °C during the deposition. A gold (Au) current collector was deposited onto the LiCoO2 cathode film, and a platinum (Pt) current collector was deposited onto the other side of the LASGTP sheet by RF-magnetron sputtering. Details of the sample preparation are described in the Supporting Information. Generally, when the voltage exceeding the voltage window of LASGTP is applied to the Au−Pt current correctors in this battery system, excess Li ions are inserted into the anode side of LASGTP crystals near the Pt current collector, and the Li-inserted LASGTP crystals irreversibly change to amorphous phases with a reductive reaction.25 As a result, the surface layer of the LASGTP sheet acts as an anode, which we indicate as “in situ formed anode” in Figure 1a. The thickness of the in situ formed anode is reported to be about 400−700 nm.18 The large irreversibility of the in situ formed anode was reported by Iriyama et al.24,25 Coulombic efficiency at the first cycle was less than 50%. The dotted line area in Figure 1a indicates the region thinned by a focused Ga-ion beam (FIB) for TEM observation. Figure 1b shows charge and discharge curves of the prepared all-solidstate LIB sample after FIB thinning illustrated in Figure 1a. In the charging process, the potential plateau is clear around 1.65

electron microscopy (TEM) analysis that can dynamically provide change in elemental compositions in local areas is highly valuable in characterizing the origin of the resistance. Yamamoto et al. observed the dynamic change in electric potential and the Li-ion concentration along the vertical direction to the interface using electron holography12−14 and spatially resolved electron energy-loss spectroscopy (SR-TEMEELS)15 around the interface between an in situ formed negative electrode and a Li1+x+yAlx(Ti, Ge)2−xSiyP3−yO12 (LASGTP) solid electrolyte.16,17 They showed how the Liion insertion/extraction reactions influenced the crystal structures, electronic structures, and local electric potential during the charge and discharge processes.18 However, the results were one-dimensional (1D), and Li-ion concentration was not quantitative. Recently, advanced TEM holders that use a piezo-controlled tip have been widely used for in situ experiments of all-solid-state LIBs. Wang et al. reported on two-dimensional (2D) Li-ion distribution in a LiCoO2 cathode that uses scanning transmission electron microscopy-EELS (STEM-EELS) with a simultaneous galvanostatic biasing.19 The STEM-EELS showed a disordered interfacial layer between a LiCoO2 cathode and lithium phosphorus oxynitride (LiPON) solid electrolyte at a charged state. However, the observed areas were not the same regions before and after the biasing, and the Li-ion concentration was not quantitative. As described above, in situ TEM techniques for all-solid-state LIBs have progressed, but not succeeded in quantitatively observing the dynamic change of electrochemical reactions and Li ions. To clarify the interfacial phenomena in more detail, we need to quantitatively reveal the continuous changes around the same region during the charge and discharge reactions directly. Here, we present a new method with hyperspectral image analyses to dynamically and quantitatively observe the interfacial phenomena in all-solid-state LIBs. We operated an all-solid-state LIB of a LiCoO2 positive electrode/LASGTP solid electrolyte/in situ formed negative electrode in a transmission electron microscope and performed the STEMEELS mapping in the same LiCoO2 region during the B

DOI: 10.1021/acs.nanolett.8b02587 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 2. Operando STEM-EELS observation of Li-ion distribution by hyperspectral image analysis of NMF. (a−d) ADF-STEM images at 0, 50, 100% charged and 33% discharged states, respectively. (e−l) Corresponding Li/Co ratio maps obtained from (e−h) original SIs and (i−l) NMFreconstructed SIs. (m−o) Histograms of Li/Co ratio maps of panels i−k. (p) Plots of a Li/Co ratio at A, B, and C regions in panels i−l. The scale bars are 100 nm.

side. We used the STEM-EELS to observe 2D Li-ion distribution and electronic states of Co ion in the LiCoO2 cathode at 0, 50, 100% charged states and a 33% discharged state indicated by arrows in Figure 1b. We could not apply conventional two-window or three-window methods for quantitative Li-ion mapping in LixCoO2 because EEL spectra of Li−K and Co−M edges overlap. Thus, we achieved the quantitative Li mapping by using the SA/SB method. The details of the SA/SB method are described in the Supporting Information (Figure S3). Figure 1c is an annular dark-field (ADF) STEM image around the interface between the LiCoO2 cathode and the LASGTP solid electrolyte before the electrochemical reaction. In the LASGTP region, the dark gray and black areas were observed. The dark gray area indicates the Li1+xAlxGeyTi2−x−yP3O12 and Li1+x+3zAlx(Ge,Ti)2−x(SizPO4)3 grains that have a high ionic conductivity, and the black area indicates the AlPO4 grains that do not have an ionic conductivity.18 The Li ions are extracted from the LiCoO2 cathode in the charging process, and they are

V, which corresponds to the electrochemical reactions that occur when Ti reduces in LASGTP and Co oxidizes in LiCoO2.24 The discharge capacity was much smaller than the charge capacity. This result shows that the charge and discharge reactions at the first cycle are irreversible because of the formation of the in situ formed anode.24,25 A comparison of the charge and discharge curves between the original and FIB processed thin film battery is shown in the Supporting Information (Figure S1). We confirmed that the electrochemical properties are almost the same before and after FIB processing. The results show that the irreversibility is intrinsic phenomena in this battery, not caused by FIB processing. The charge transfer resistance at the interfaces was measured by AC impedance. The resistance at the cathode side was about 450 Ωcm2, which is larger than the one at the anode side. Details of the AC impedance are described in the Supporting Information (Figure S2). We observed the dynamic changes in the LiCoO2/LASGTP interface during the charge and discharge reactions in order to understand the mechanism of the high interfacial resistance on the cathode C

DOI: 10.1021/acs.nanolett.8b02587 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 3. Operando STEM-EELS observation of the valence changes of Co ion. (a and b) ADF-STEM images at 0 and 100% charged states. (c and d) Corresponding Li/Co ratio maps obtained from NMF-reconstructed SIs. (e and f) Peak position maps of the Co-L3 edge at the dotted box in panels c and d. The brighter color indicates a higher energy loss in the EEL spectra. (g and h) 1D spatially resolved EEL spectra of Co-L2,3 from A to A′ in panels e and f where the spectrum signals along the direction parallel to the interface were integrated. Green, blue, and red lines show the peak energies of each Co-L2,3 edge. The scale bars are 100 nm.

applying NMF, which reduces the random-noise component, the Li-ion distributions became much clearer, and the dynamic change of Li-ion concentration during the reactions was clearly visualized. Total dose through the present operando STEMEELS observation was ∼1.7 × 105 e− Å−2, which is the same order of typical atomic resolution STEM for the LiCoO2 cathode (∼2.8 × 105 e− Å−2).26 We confirmed that the region scanned by operando STEM-EELS did not show any clear manifestation of the electron radiation damage during the observation (Figure S5). Thus, we considered that the electron beam illumination in this study does not have a big impact on the results. The series of the Li-ion maps (Figure 2i−l) show that Li elements in the LiCoO2 cathode are concentrated near the Au current collector side rather than the LASGTP solidelectrolyte side in the overall states. Figure 2m−o shows the histograms of each Li/Co ratio map (Figure 2i−k). The histogram at the 0% charged state (Figure 2m) is relatively broad from 0.2 to 1.0. On the other hand, the histograms at the 50 and 100% charged states (Figure 2n,o) show sharper profiles. These results show that the electrochemical reaction does not proceed uniformly in the LiCoO2 cathode film. Figure 2p shows the change of the Li/Co ratio in each point (A−C)

inserted back into the LiCoO2 cathode in the discharging process. ADF-STEM image and EEL spectrum image (SI) data sets were simultaneously acquired from the same LiCoO2 region surrounded by the dotted rectangle in Figure 1c with the different charged/discharged states (0, 50, 100% charged states and 33% discharged state). In order to identify any statistically significant features in the large SI data sets and to reduce random-noise components efficiently in a statistical manner, we applied the hyperspectral image analysis of NMF to SIs. Details of STEM-EELS acquisition and hyperspectral image analysis of NMF are described in the Supporting Information (Figure S4). Panels a−d, e−h, and i−l in Figure 2 show the series of ADF-STEM images, Li/Co ratio maps obtained from the original SIs, and those obtained from the NMFreconstructed SIs at the above four states during the charge and discharge reactions, respectively. The quantitative Li/Co ratio is displayed in different colors as indicated in the color gradation bar. Because the LASGTP solid electrolyte and Au current collector do not contain Co ions, the Li/Co ratio could not be defined in LASGTP and Au. Thus, the Li/Co ratio in the region of LASGTP and Au was set to zero instead. By D

DOI: 10.1021/acs.nanolett.8b02587 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters indicated in Figure 2i−l. The points A, B, and C have different Li-ion concentrations at the 0% charged state. Point A has a high Li-ion concentration near the Au current collector, point C has a low Li-ion concentration near the LASGTP solid electrolyte, and point B has a middle Li-ion concentration. In the point A region, the changes in the Li/Co ratio during the charging process (from 0.95 to 0.27) and during the discharging process (from 0.27 to 0.46) are larger than those in the point C region, where the Li/Co ratio in the point C decreases from 0.26 to 0.06 and increases from 0.06 to 0.08. This result shows that the activity of the electrochemical reaction was not uniform, and it depended on the original Liion concentration. The change of electronic states of Co ion also confirmed the nonuniformity of the electrochemical activity. Figure 3a,b is the ADF-STEM images at the 0 and 100% charged states, and Figure 3c,d shows the corresponding Li/Co ratio maps obtained from NMF-reconstructed SIs, respectively. As with Figure 2i−l, the Li-ion concentration near the side of LASGTP is lower than the one near the side of Au. Figure 3e,f is the peak position maps of Co-L3 edge in the dotted boxes in Figure 3c,d. The peak energy is displayed in different colors as indicated in the color gradation bar. It is well-known that the peak energy of the Co-L2,3 edge reflects the valence of Co ions.27,28 Note that all of the EEL spectra were calibrated within the energy of 0.05 eV accuracy by the Dual-EELS mode,29 which quickly acquires both zero-loss and Co-L2,3 edge spectra alternatively and corrects the energy drift of the spectra by referring the zero-loss peak position. Figure 3g,h shows the 1D spatially resolved EEL spectra of the Co-L2,3 edges from A (LASGTP side) to A′ (Au side) in the LiCoO2 region, where the spectrum signals along the direction parallel to the interface were integrated. Green, blue, and red lines show the peak energies of each Co-L2,3 edge. At the 0% charged state, the peak energies of the Co-L3 edge are almost uniform at about 780.8 eV, as shown in the color map in Figure 3e and the green line in Figure 3g. This result indicates that Co ions are almost trivalent at the 0% charged state in the whole region of the cathode film although the Li-ion concentration in the cathode film is not uniform. At the 100% charged state, the contrast of the peak position map of the Co-L3 edge became brighter (Figure 3f). This shows that the peak position of the Co-L3 edge shifted to the higher energy loss; that is, Co3+ oxidized to Co4+ because of the Li-ion extraction. The peak shift to the higher energy loss can be seen in the red line in Figure 3h. However, the spectrum in the region close to the LiCoO2/LASGTP interface only has the same peak energy as the original state, as indicated by the green line in Figure 3h. This means that Co ions in the vicinity of the interface between the LiCoO2 film and the LASGTP are trivalent even at the 100% charged state. That is, in the Li-poor region near the interface, the Li-ion concentration and the valence of Co ion did not change. This indicates that the LiCoO2 film near the LiCoO2/LASGTP interface is electrochemically inactive and is the reason for the high interfacial resistance to Li-ion transfer. The width of the inactive layer was 10−20 nm. We performed Raman spectroscopy and selected area electron diffraction (SAED) to characterize the inactive layer at the interface. Figure 4a shows the Raman spectra of (i) the LASGTP sheet, (ii) a 1 μm thick LiCoO2 film deposited onto the LASGTP sheet, and (iii) a 150 nm thick LiCoO2 film deposited onto the LASGTP sheet. The three samples were

Figure 4. Characterization of an inactive layer. (a) Raman spectra of (i) LASGTP sheet, (ii) LiCoO2 (1 μm)/LASGTP, and (iii) LiCoO2(150 nm)/LASGTP at a pristine state. Circles, triangles, and rectangles show Raman peaks of LASGTP, LiCoO2, and Co3O4, respectively. (b−e) Distribution of Co3+ and Co2+ in the cathode film. (b) ADF-STEM image. (c−e) Sets of the decomposed spectra and their distributions obtained by the NMF analysis. Component 1 shows Co3+ and 2 shows Co2+.

the 0% charged state. A Raman band was observed at 448 cm−1 for the LASGTP sheet, as indicated by the circles in part i. We observed the two strong Raman bands indicated by triangles at 476 and 590 cm−1 for the 1 μm thick LiCoO2 film. This indicates that the surface of the 1 μm thick LiCoO2 film has layered rock-salt LiCoO2 with the space group of R3m.30,31 However, the spectrum of part iii, obtained from the 150 nm thick LiCoO2 film deposited onto the LASGTP, has five Raman bands. As described above, LASGTP and layered rocksalt LiCoO2 caused the three Raman bands indicated by the circles and the triangles. The other two Raman bands at 511 and 671 cm−1 indicated by the squares are the signals from the LiCoO2/LASGTP interface because Raman detects several hundred nanometers in depth from the surface. They imply the presence of a spinel structure Co3O4 at the interface.31 The result of SAED also shows the existence of Co3O4. Details of SAED are described in the Supporting Information (Figure S6). Co3O4 has one Co2+ and two Co3+ in its crystal. Since the interfacial layer is the mixture of Co3O4 with Co2+/Co3+ and LiCoO2 with Co3+, the divalent component of the Co-L2,3 edge in Co3O4 could not be distinguished by the conventional EELS analysis of the Co-L2,3 edge (Figure 3e−h). To detect a small volume of divalent components of the Co-L2,3 edge in Co3O4, we applied the hyperspectral image analysis of the NMF to the Co-L2,3 edge SI. Figure 4b−e shows an ADF-STEM image, sets E

DOI: 10.1021/acs.nanolett.8b02587 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

reaction, Li ions are inserted back into Li1−xCoO2 (Figure 5c). Because some of the Li ions are trapped to form the in situ formed anode during the charge process, the Li-ion concentration does not return to the original state in the cathode region (Figure 2l). We developed a new method to observe the dynamic change in the quantitative Li-ion concentration and valence of Co ion in the LiCoO2 cathode during the charge and discharge reactions using operando STEM-EELS with the hyperspectral image analyses of the NMF and SA/SB method. The combination of the STEM-EELS and the hyperspectral image analyses greatly improved the reliability of quantitative and dynamic analysis of solid-state electrochemical reactions in a nanometer scale. We found from the operando observation that the Li ions did not uniformly extract/insert during the charge/discharge reactions, and the activity of the electrochemical reaction depended on the Li-ion concentration in a pristine state. The operando STEM-EELS, electron diffraction, and Raman spectroscopy experimentally showed that the 10− 20 nm thick inactive region near the interface was a mixture of LiCoO2 and Co3O4. The spinel Co3O4 is the origin of the interfacial resistance because of its low Li-ion conductivity. The proposed operando method allows us to observe electrochemical reactions occurring in the same area in the working devices; therefore, it enables us to truly understand the electrochemical mechanism in the charge and discharge processes. Furthermore, operando analysis is useful for elucidating phenomena of solid-state electrochemistry in nonequilibrium states. It would lead to the design and development of high-performance all-solid-state LIBs.

of decomposed EEL spectra of components 1 and 2, and their spatial distributions obtained by the NMF analysis, respectively. The component 2 shifted to a lower energy by 1.75 eV than the component 1. Thus, components 1 and 2 are the spectra of the trivalent and the divalent Co ions, respectively.27 The spatial distribution of component 2 (Co2+) has the higher intensity near the interface between the cathode film and the LASGTP solid electrolyte, which is consistent with the other experimental results in this study. From these results, we concluded that the electrochemical inactive layer observed by operando STEM-EELS was Co3O4. We presume that the Li ions diffused from LiCoO2 to LASGTP during the deposition of the cathode film at 550 °C. The lack of Li ions resulted in the Co3O4 phase. The formation of the Co3O4 layer caused the high resistance to the Li-ion transfer at the interface and led to the degradation of the LiCoO2 electrode. Figure 5 shows the Li-ion maps and description of the composition in the cathode film at 0% charged state, 100%



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.8b02587.

Figure 5. Li map and description of the composition in the cathode film at (a) 0% charged state, (b) 100% charged state, and (c) 33% discharged state. The regions surrounded by the black lines show the LiCoO2 grains. The other regions show the mixture of LiCoO2 and Co3O4, where Co3O4 is concentrated near the LASGTP side rather than the Au side.



Description of the sample preparation, comparison of the charge and discharge curves between the original and FIB processed battery, details of the AC impedance to measure the interfacial resistance, details of the SA/SB method for quantitative Li mapping, experimental details of the operando STEM-EELS and hyperspectral image analysis, and selected area electron diffraction of the cathode film (PDF)

AUTHOR INFORMATION

Corresponding Author

charged state, and 33% discharged state. At the pristine state (Figure 5a), the electrochemically inactive Co3O4 layer is partially formed near the cathode/solid electrolyte in the active LiCoO2 cathode film. The regions surrounded by the black lines show the single LiCoO2 grains. The other regions show the mixture of LiCoO2 and Co3O4, where Co3O4 is concentrated near the LASGTP side rather than the Au side. During the charge reaction, the Li ions are extracted from LiCoO2, and a part of Co3+ in LiCoO2 oxidizes to Co4+ (Figure 5b). The dynamics of the Li-ion concentration and the valence changes of Co ion were observed as the Li maps (Figure 2i−k) and the peak shift of the Co-L2,3 edge (Figure 3e−h). The change of the Li-ion concentration and the valence of Co ion are small near the interface because Co3O4 does not contain Li ions and is electrochemically inactive. During the discharge

*E-mail: [email protected]. ORCID

Yuki Nomura: 0000-0002-6091-5902 Author Contributions

Y.N., K.Y., and E.I. designed the experiments. Y.N. performed operando STEM-EELS, Raman spectroscopy, and SAED experiments. M.O. and N.H. prepared the LixCoO2 particle for the SA/SB method. Y.N., K.Y., T.H., and K.S. wrote the manuscript. All authors contributed to the manuscript preparation and have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest. F

DOI: 10.1021/acs.nanolett.8b02587 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters



(27) Stockhoff, T.; Gallasch, T.; Berkemeier, F.; Schmitz, G. Thin Solid Films 2012, 520, 3668−3674. (28) Taguchi, N.; Akita, T.; Sakaebe, H.; Tatsumi, K.; Ogumi, Z. J. Electrochem. Soc. 2013, 160, A2293−A2298. (29) Gubbens, A.; Barfels, M.; Trevor, C.; Twesten, R.; Mooney, P.; Thomas, P.; Menon, N.; Kraus, B.; Mao, C.; McGinn, B. Ultramicroscopy 2010, 110, 962−970. (30) Huang, W.; Frech, R. Solid State Ionics 1996, 86−88, 395−400. (31) Gopukumar, S.; Jeong, Y.; Kim, K. B. Solid State Ionics 2003, 159, 223−232.

ACKNOWLEDGMENTS This work was partly supported by the Japan Society for the Promotion of Science (JSPS), Grant-in-Aid for Scientific Research KAKENHI (JP 17H02792) and Nagoya University microstructural characterization platform as a program of “Nanotechnology Platform” of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. We are grateful to Professor Shunsuke Muto at Nagoya University for advising us on the statistical spectrum analysis. We are also grateful to Dr. Jun Kikkawa for discussing the quantitative Li mapping.



REFERENCES

(1) Kamaya, N.; Homma, K.; Yamakawa, Y.; Hirayama, M.; Kanno, R.; Yonemura, M.; Kamiyama, T.; Kato, Y.; Hama, S.; Kawamoto, K.; Mitsui, A. Nat. Mater. 2011, 10, 682−686. (2) Kato, Y.; Hori, S.; Saito, T.; Suzuki, K.; Hirayama, M.; Mitsui, A.; Yonemura, M.; Iba, H.; Kanno, R. Nat. Energy 2016, 1, 16030. (3) Sun, Y.; Suzuki, K.; Hori, S.; Hirayama, M.; Kanno, R. Chem. Mater. 2017, 29, 5858−5864. (4) Iriyama, Y.; Kako, T.; Yada, C.; Abe, T.; Ogumi, Z. Solid State Ionics 2005, 176, 2371−2376. (5) Iriyama, Y.; Kako, T.; Yada, C.; Abe, T.; Ogumi, Z. J. Power Sources 2005, 146, 745−748. (6) Takada, K.; Ohta, N.; Zhang, L.; Xu, X.; Hang, B. T.; Ohnishi, T.; Osada, M.; Sasaki, T. Solid State Ionics 2012, 225, 594−597. (7) Takada, K.; Ohta, N.; Tateyama, Y. J. Inorg. Organomet. Polym. Mater. 2015, 25, 205−213. (8) Yada, C.; Ohmori, A.; Ide, K.; Yamasaki, H.; Kato, T.; Saito, T.; Sagane, F.; Iriyama, Y. Adv. Energy Mater. 2014, 4, 1301416. (9) Haruyama, J.; Sodeyama, K.; Han, L.; Takada, K.; Tateyama, Y. Chem. Mater. 2014, 26, 4248−4255. (10) Sakuda, A.; Hayashi, A.; Tatsumisago, M. Chem. Mater. 2010, 22, 949−956. (11) Kim, K. H.; Iriyama, Y.; Yamamoto, K.; Kumazaki, S.; Asaka, T.; Tanabe, K.; Fisher, C. A. J.; Hirayama, T.; Murugan, R.; Ogumi, Z. J. Power Sources 2011, 196, 764−767. (12) Gabor, D. Nature 1948, 161, 777−778. (13) Tonomura, A. Electron holography, 2nd ed.; Springer-Verlag: Heidelberg, 1999. (14) Lichte, H.; Lehmann, M. Rep. Prog. Phys. 2008, 71, 016102. (15) Kimoto, K.; Sekiguchi, T.; Aoyama, T. J. Electron Microsc. 1997, 46, 369−374. (16) Yamamoto, K.; Iriyama, Y.; Asaka, T.; Hirayama, T.; Fujita, H.; Nonaka, K.; Miyahara, K.; Sugita, Y.; Ogumi, Z. Electrochem. Commun. 2012, 20, 113−116. (17) Yamamoto, K.; Iriyama, Y.; Hirayama, T. Microscopy 2016, 66, 50−61. (18) Yamamoto, K.; Yoshida, R.; Sato, T.; Matsumoto, H.; Kurobe, H.; Hamanaka, T.; Kato, T.; Iriyama, Y.; Hirayama, T. J. Power Sources 2014, 266, 414−421. (19) Wang, Z.; Santhanagopalan, D.; Zhang, W.; Wang, F.; Xin, H. L.; He, K.; Li, J.; Dudney, N.; Meng, Y. S. Nano Lett. 2016, 16, 3760− 3767. (20) Shiga, M.; Tatsumi, K.; Muto, S.; Tsuda, K.; Yamamoto, Y.; Mori, T.; Tanji, T. Ultramicroscopy 2016, 170, 43−59. (21) Muto, S.; Yoshida, T.; Tatsumi, K. Mater. Trans. 2009, 50, 964−969. (22) Kikkawa, J.; Terada, S.; Gunji, A.; Nagai, T.; Kurashima, K.; Kimoto, K. J. Phys. Chem. C 2015, 119, 15823−15830. (23) Fu, J. J. Am. Ceram. Soc. 1997, 80, 1901−1903. (24) Iriyama, Y.; Yada, C.; Abe, T.; Ogumi, Z.; Kikuchi, K. Electrochem. Commun. 2006, 8, 1287−1291. (25) Yada, C.; Iriyama, Y.; Abe, T.; Kikuchi, K.; Ogumi, Z. Electrochem. Commun. 2009, 11, 413−416. (26) Huang, R.; Ikuhara, Y. Curr. Opin. Solid State Mater. Sci. 2012, 16, 31−38. G

DOI: 10.1021/acs.nanolett.8b02587 Nano Lett. XXXX, XXX, XXX−XXX