Elucidating the Driving Force of Relaxation of Reaction Distribution in

Feb 17, 2016 - Hajime Tanida†, Hisao Yamashige†, Yuki Orikasa‡, Yuma Gogyo‡, Hajime Arai†, Yoshiharu Uchimoto‡, and Zempachi Ogumi†. †...
57 downloads 13 Views 2MB Size
Article pubs.acs.org/JPCC

Elucidating the Driving Force of Relaxation of Reaction Distribution in LiCoO2 and LiFePO4 Electrodes Using X‑ray Absorption Spectroscopy Hajime Tanida,*,† Hisao Yamashige,† Yuki Orikasa,‡ Yuma Gogyo,‡ Hajime Arai,† Yoshiharu Uchimoto,‡ and Zempachi Ogumi† †

Office of Society-Academia Collaboration for Innovation, Kyoto University, Kyoto, 611-0011, Japan Graduate School of Human and Environmental Studies, Kyoto University, Kyoto, 606-8501, Japan



S Supporting Information *

ABSTRACT: The reaction distribution in the composite electrodes used in lithium-ion batteries greatly affects battery performances, including rate capability and safety. In this study, the generation of the reaction distribution and its relaxation in cross sections of LiCoO2 and LiFePO4 composite electrodes were analyzed using microbeam X-ray absorption spectroscopy. The reaction distribution immediately after delithiation could be observed clearly with different oxidation states of the transition metal (i.e., different concentrations of lithium ions). The distribution in the Li1−xCoO2 electrodes disappeared, whereas that in the Li1−xFePO4 electrodes remained unchanged even after 15 h of relaxation. After comparing the potential profile of both types of electrodes, it is suggested that the potential difference between the more delithiated area and the less delithiated area in the composite electrode is the primary driving force for the relaxation.



INTRODUCTION Large-scale lithium-ion secondary batteries (LIBs) are currently in demand for use in electric vehicles and energy storage systems for the electric grid. To improve the performance of large-scale LIBs, it is important to understand the various resistive components in the composite electrodes used in LIBs. These electrodes are typically composed of an active material, a conducting carbon, and a binder. The main resistive components in the composite electrodes are the electronic resistivity (Rel), the ionic resistivities of the solid (Rion(s)) and liquid (Rion(l)) phases, and the ion transfer resistivity at the electrode/electrolyte interface (Rint),1−6 as shown in Figure 1.

The resistivity is greatly influenced by the distribution of the components in the electrode and electrolytes (internal factors) as well as the operating conditions such as the charge/discharge rate and temperature (external factors).6 An inhomogeneous distribution of the resistivity leads to reaction inhomogeneity in the composite electrodes during charge/discharge. This can affect the battery performance; for example, the specific area of an electrode with a low resistivity is utilized to a greater degree than the other areas during the charge/discharge process. This can lead to an imbalance in the degree of utilization of the electrode material and a decrease in its rate capability (owing to the formation of a thicker surface film) as well as safety issues (owing to differences in the oxidation state) . The reaction distribution in composite electrodes has been analyzed previously using modeling simulations.1−7 Direct observations of the reaction distribution in composite electrodes with a micrometer-level spatial resolution have been reported only recently. Liu et al. observed the inhomogeneous reaction in a LiFePO4 composite electrode using microbeam Xray diffraction (XRD).8 Murayama et al. proposed operando spectroscopic XRD for analyzing the reaction distribution in LiNi1/3Co1/3Mn1/3O2 composite electrodes.9 Zhang et al. performed operando neutron depth profiling in LiFePO4 composite electrodes,10 while energy-dispersive XRD analysis has been employed to directly observe the reaction distribution in LiFePO411 and Ag2VP2O8.12 In addition, two-dimensional

Figure 1. Schematic presentation of the resistivity components in a composite electrode in a battery: electronic resistivity (Rel), ionic resistivity of the solid (Rion(s)), ionic resistivity of the liquid phase (Rion(l)), and ion transfer resistivity at the interface (Rint). © 2016 American Chemical Society

Received: October 19, 2015 Revised: February 16, 2016 Published: February 17, 2016 4739

DOI: 10.1021/acs.jpcc.5b10210 J. Phys. Chem. C 2016, 120, 4739−4743

Article

The Journal of Physical Chemistry C (2D) X-ray absorption spectroscopy (XAS),13−15 full-field transmission X-ray microscopy,16 and microbeam XAS (μXAS)17 have also been employed to directly observe the reaction distribution. Among these, the μ-XAS method, which is sensitive to the chemical states of transition-metal atoms, is one of the more promising techniques for analyzing the reaction distribution in composite electrodes, owing to the fact that XAS combined with the microbeam technique has a spatial resolution of less than 1 μm. Further, one can use a variety of detectors with higher sensitivities and dynamic ranges. The disappearance of the reaction distribution during a relaxation period after the charge/discharge process is also an important phenomenon, since an unrelaxed reaction distribution can accelerate the inhomogeneous use of the electrode and cause severe degradation during repeated charging/discharging. Nevertheless, the relaxation of the reaction distribution has been studied mainly through mathematical models,18 and no experimental investigation has yet been performed. It is thus important to observe the generation and relaxation of the reaction distribution and to elucidate the origins of these phenomena for optimizing the battery components and operating conditions. In this study, the relaxation behavior of the reaction distribution in composite electrodes was observed using μXAS. The oxidation states of the transition metals in the active materials were analyzed at different depths of the cross section, and the redox changes occurring during the relaxation period were examined. In particular, we elucidated the effect of the potential difference on the relaxation of the reaction distribution. Two types of electrodes having different charge/ discharge profiles, namely, a Li1−xCoO2 electrode that forms a solid solution (0.2 < x < 1.0) with a gradient potential profile19,20 and a biphased Li1−xFePO4 electrode with a flat potential profile,21 were tested. It was experimentally determined that the reaction distribution in Li1−xCoO2 is relaxed faster than that in Li1−xFePO4, owing to the difference in the slopes of the potential curves of the two materials.

relaxation for 15 and 30 h for LiCoO2 and LiFePO4, respectively. The μ-XAS measurements were performed at the undulator beamline BL37XU of SPring-8 (Hyogo, Japan).22 The experimental setup is shown in Figure S1. The synchrotron radiation X-ray from the storage ring was monochromated using a Si 111 double-crystal monochromator. The monochromatized X-rays were focused on the sample using a Kirkpatrick-Baez mirror with beam dimensions of 1.3 μm (horizontal) × 0.8 μm (vertical). The μ-X-ray fluorescence images at the cobalt and iron K-edges were obtained using a silicon drift detector, while the μ-X-ray absorption spectra were measured in the transmission mode using two ionization chambers, which were located before and after the sample. In contrast to the surface-sensitive fluorescence measurements, the transmission measurements allowed us to analyze the bulks of the electrodes. To correlate the X-ray absorption spectra profile to the state of charge (SOC) of the electrode material, the cobalt and iron K-edge spectra of several Li1−xCoO2 and Li1−xFePO4 samples after full relaxation were measured by a conventional transmission method with a nonfocused beam to obtain the averaged profile of the sample.



RESULTS AND DISCUSSION Scanning electron microscopy (SEM) and X-ray fluorescence images of the cross sections of the Li1−xCoO2 and Li1−xFePO4 electrodes taken immediately after charging are shown in Figure 2 using rainbow-scale color mapping; the colors red and blue



EXPERIMENTAL METHODS The LiCoO2 composite electrodes were fabricated using LiCoO2 powder (particle size of approximately 10 μm) mixed with 5 wt % acetylene black (AB) and 5 wt % polyvinylidene fluoride (PVDF), while the LiFePO4 composite electrodes were fabricated using LiFePO4 powder (particles submicron sized in diameter) mixed with 15 wt % AB and 15 wt % PVDF. Slurries of the mixture in N-methyl pyrrolidone were spread on aluminum current collectors and dried in a vacuum oven at 80 °C. The prepared composite electrodes, which had a thickness of approximately 100 μm, and lithium metal and an electrolytesoaked separator (Celgard #2500) were assembled into a stainless steel flat cell (HS, HOSEN). 1 M LiPF6 in ethylene carbonate/ethyl methyl carbonate (3:7) as the solvent was used as the electrolyte. The cell was assembled in an Ar-atmosphere glovebox. The cells were charged at a constant current of 1 C rate, in order to adjust the value of x in Li1−xCoO2 and Li1−xFePO4 to approximately 0.5. The cells were disassembled immediately after the charging process or after being relaxed for 15 h in the open-circuit state. The electrodes were then washed twice with dimethyl carbonate, dried, and cut with a microtome in slices with a thickness of 50 μm, and the cross sections were observed using an SEM system (VE-8800, Keyence Co.). The open-circuit potential of the electrodes was measured. A series of current pulses were applied at 0.05 C; this was followed by

Figure 2. SEM (upper) and X-ray fluorescence (lower) images of the cross section of the LiCoO2 (left) and LiFePO4 (right) electrodes displayed using the same monotone and rainbow-scale color maps. The intensity is normalized for each image. The left sides of the images correspond to the area near the current collector. The X-ray absorption spectra were measured at A, which was near the collector, and at B, which was at the center of the electrode.

represent high and low intensities, respectively. The voltage profiles for the sample preparation are shown in Figure S2. In both images, the left and right sides correspond to the current collector and the separator layer with electrolyte, respectively. The incident X-ray energies were higher than the absorption edge energies of cobalt and iron. In each image, the intensity was normalized; the positions with high count rates indicate high concentrations of cobalt and iron. Therefore, the images show course and fine particles of LiCoO2 and LiFePO4. Once the presence of cobalt and iron had been confirmed from the 4740

DOI: 10.1021/acs.jpcc.5b10210 J. Phys. Chem. C 2016, 120, 4739−4743

Article

The Journal of Physical Chemistry C

position A for both electrode materials. The absorption edge energies in the spectra are correlated to the oxidation states of cobalt and iron in the electrodes and can be translated into SOC of the electrodes. We measured the X-ray absorption spectra in several different positions in these electrodes, and using the known relationship between the spectra profiles and the SOC values of Li1−xCoO223 and Li1−xFePO424 electrodes, which are shown in Figure S3, the spectra profiles were converted into the position-labeled SOC. The results are shown in Figure S4. The SOC values (x in Li1−xCoO2 and Li1−xFePO4) varied depending on the position, and those immediately after charging ranged from 0.3 to 0.5 for Li1−xCoO2 and 0.45 to 0.6 for Li1−xFePO4. The results indicate that the cobalt and iron in the center of the electrode were oxidized largely than those near the current collector, suggesting that the reaction distribution was generated by the charge dynamics. This result is consistent with a previously reported reaction distribution determined by using microbeam XRD, which showed that the active materials LiFePO4 near the current collector is utilized to a lower degree at high rate charging.8 Further, this simulation study suggested that a slow supply of lithium ions around the current collector is responsible for the observed inhomogeneity.18 After the relaxation period of 15 h, the distribution in the Li1−xCoO2 electrode disappeared, as shown in Figure 3c, whereas the distribution in the Li1−xFePO4 electrode remained nearly unchanged, as shown in Figure 3d. This suggested that the distribution of the oxidation state of cobalt in the Li1−xCoO2 electrode relaxed within 15 h, whereas that of the iron in the Li1−xFePO4 electrode did not. This difference in the relaxation behaviors can be explained by the difference in the potential profiles of Li1−xCoO2 and Li1−xFePO4, as shown in Figure 4 (upper). The SOC values for positions (A) (red square) and (B) (blue circle) in Figure 3 for both immediately after charging (solid lines) and after relaxation (broken lines) are marked in each profile. There was a significant potential difference between the two positions in the case of the Li1−xCoO2 electrode immediately after

images, two representative positions were assigned based on the optimal X-ray transmission intensity, namely, A on the current collector side and B, which was the center of the electrode, as shown in Figure 2. Then, the cobalt and iron K-edge X-ray absorption spectra were measured at positions A and B. Each measured area was approximately 1 μm2. The fluorescence images and the corresponding transmission X-ray absorption spectra were also obtained after the electrodes had been disassembled after 15 h of relaxation. Figure 3 shows the cobalt and iron K-edge X-ray absorption spectra of the LiCoO2 and LiFePO4 electrodes (normalized in

Figure 3. Cobalt and iron K-edge X-ray absorption spectra measured immediately after the charging of (a) LiCoO2 and (b) LiFePO4 and 15 h after the charging of (c) LiCoO2 and (d) LiFePO4 at positions A (red square) and B (blue circle) of the images in Figure 3.

the edge jump = 1). The spectra in Figure 3a,b, which were obtained immediately after the charging process, indicate that the absorption edge energy at position B was higher than that at

Figure 4. (Upper) Open-circuit potential versus x for (a) Li1−xCoO2 and (b) Li1−xFePO4 measured immediately after charging (solid line) and 15 h after charging (broken line) at positions A (red square) and B (blue circle) in the images in Figure 1. (Lower) Schematic presentations of the potential versus the distance from the current collector. The darker the color, the lower the oxidation state. The electrons in LiCoO2 moved 15 h after charging. 4741

DOI: 10.1021/acs.jpcc.5b10210 J. Phys. Chem. C 2016, 120, 4739−4743

Article

The Journal of Physical Chemistry C charging, and the oxidation states of cobalt at positions (A) and (B) voluntarily relaxed to the balanced value, as indicated by the sloped electrochemical potential profile in this region. In contrast, the Li1−xFePO4 electrode exhibited an extremely flat charge/discharge potential profile over the range 0.05 < x < 0.9, and this is expected to lead to the reaction distribution remaining even after 15 h of relaxation. Schematic diagrams of the insides of the electrodes are shown in Figure 4 (lower). After the cell is charged at 1 C rate, a reaction distribution is generated along the cross section of the electrode. The center of the composite electrode is more oxidized than the current collector side; this means that transition-metal ions are higher oxidation states at the center of the electrode. The inhomogeneity with more and less oxidized parts is energetically unfavorable, and this could have relaxed homogeneous states to increase the entropy. While the relaxation can quickly occur in single particles owing to facile electron and lithium-ion transportation inside the particles, the relaxation among separated particles (in a sub-millimeter scale) will require the exchange of electrons through the conductive material and lithium ions through the electrolyte to compensate the charge. It is thus suggested from the result shown above that the potential difference is a key to promote the kinetically sluggish exchange reaction and to relax the inhomogeneous states, leading to inhomogeneity relaxation in the Li1−xCoO2 electrode and no change in the inhomogeneous Li1−xFePO4 electrode. Since the lithium-ion transportation via the electrolyte, being slower than the electron transportation, is the ratedetermining step in such relaxation, the relaxation behavior itself is affected by the distance between the particles as well as by the constitution of the composite electrode. In this sense, fine Li1−xFePO4 particles should be advantageous in the lithium-ion transportation among the active material particles. Nevertheless, there was no relaxation in the Li1−xFePO4 electrode, suggesting that the potential gradient primarily determines the relaxation behavior.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +81-791-58-0803, ext. 3880. Fax: +81-791-58-1819. Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank Yasuko Terada of the Japan Synchrotron Radiation Research Institute (JASRI) for her support with the μ-XAS measurements, and Misaki Katayama of Ritsumeikan University for his help with the SEM imaging. This work was supported by the “Research and Development Initiative for Science Innovation of New Generation Battery (RISING Project)” of the New Energy and Industrial Technology Development Organization (NEDO), Japan. The synchrotron radiation experiments were performed with the approval of the Japan Synchrotron Radiation Research Institute (JASRI) (Proposal Nos. 2009B1034, 2010A1020, and 2010B1030).



REFERENCES

(1) Srinivasan, V.; Newman, J. Discharge Model for the Lithium IronPhosphate Electrode. J. Electrochem. Soc. 2004, 151, A1517−A1529. (2) Prosini, P. P. Modeling the Voltage Profile for LiFePO4. J. Electrochem. Soc. 2005, 152, A1925−A1929. (3) Yu, D. Y. W.; Donoue, K.; Inoue, T.; Fujimoto, M.; Fujitani, S. Effect of Electrode Parameters on LiFePO4 Cathodes. J. Electrochem. Soc. 2006, 153, A835−A839. (4) Roscher, M. A.; Vetter, J.; Sauer, D. U. Characterisation of Charge and Discharge Behaviour of Lithium Ion Batteries with Olivine Based Cathode Active Material. J. Power Sources 2009, 191, 582−590. (5) Chen, Y.-S.; Hu, C.-C.; Li, Y.-Y. Effects of Cathode Impedance on the Performances of Power-Oriented Lithium Ion Batteries. J. Appl. Electrochem. 2010, 40, 277−284. (6) Fongy, C.; Jouanneau, S.; Guyomard, D.; Badot, J. C.; Lestriez, B. Electronic and Ionic Wirings Versus the Insertion Reaction Contributions to the Polarization in LiFePO4 Composite Electrodes. J. Electrochem. Soc. 2010, 157, A1347−A1353. (7) Ferguson, T. R.; Bazant, M. Z. Nonequilibrium Thermodynamics of Porous Electrodes. J. Electrochem. Soc. 2012, 159, A1967−A1985. (8) Liu, J.; Kunz, M.; Chen, K.; Tamura, N.; Richardson, T. J. Visualization of Charge Distribution in a Lithium Battery Electrode. J. Phys. Chem. Lett. 2010, 1, 2120−2123. (9) Murayama, H.; Kitada, K.; Fukuda, K.; Mitsui, A.; Ohara, K.; Arai, H.; Uchimoto, Y.; Ogumi, Z.; Matsubara, E. Spectroscopic X-ray Diffraction for Microfocus Inspection of Li-Ion Batteries. J. Phys. Chem. C 2014, 118, 20750−20755. (10) Zhang, X.; Verhallen, T. W.; Labohm, F.; Wagemaker, M. Direct Observation of Li-Ion Transport in Electrodes under Nonequilibrium Conditions Using Neutron Depth Profiling. Adv. Energy Mater. 2015, 5, 1500498. (11) Strobridge, F. C.; Orvananos, B.; Croft, M.; Yu, H.-C.; Robert, R.; Liu, H.; Zhong, Z.; Connolley, T.; Drakopoulos, M.; Thornton, K.; et al. Mapping the Inhomogeneous Electrochemical Reaction Through Porous LiFePO4-Electrodes in a Standard Coin Cell Battery. Chem. Mater. 2015, 27, 2374−2386.



CONCLUSIONS In conclusion, the μ-XAS method was used to observe the generation and relaxation of the reaction distribution in the composite electrodes used in LIBs. This enabled us to observe the local electronic structure of the electrode materials and elucidate the reaction distribution on the micrometer scale. Comparing LiCoO2 with LiFePO4 suggests that the driving force for the relaxation of the reaction distribution is chiefly the difference in the electrochemical potentials and that this phenomenon is distinct from the effect of the lithium-ion concentration. The unrelaxed inhomogeneity in the electrode in a sub-millimeter scale could result in the uneven use of the active material, resulting in possibly deterioration of the electrode. Thus, the results obtained in this study demonstrate the importance of observing the relaxation behavior in designing batteries that are more efficient.



absorption spectra of Li1−xCoO2 and Li1−xFePO4, and estimated x using normalized absorbance versus distance from current collector of Li1−xCoO2 and Li1−xFePO4 (PDF)

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.5b10210. Figures showing the experimental setup for the micro-Xray absorption spectroscopy measurements, voltage profiles for sample preparation of LiCoO2 and LiFePO4 composite electrodes, Co and Fe K-edge X-ray 4742

DOI: 10.1021/acs.jpcc.5b10210 J. Phys. Chem. C 2016, 120, 4739−4743

Article

The Journal of Physical Chemistry C (12) Kirshenbaum, K.; Bock, D. C.; Lee, C.-Y.; Zhong, Z.; Takeuchi, K. J.; Marschilok, A. C.; Takeuchi, E. S. In Situ Visualization of Li/ Ag2VP2O8 Batteries Revealing Rate-Dependent Discharge Mechanism. Science 2015, 347, 149−154. (13) Tanida, H.; Yamashige, H.; Orikasa, Y.; Fujimoto, T.; Oishi, M.; Murayama, H.; Arai, H.; Katayama, M.; Inada, Y.; Ohta, T.; et al. In Situ Two-Dimensional Micro-Imaging XAFS with CCD Detector. J. Phys.: Conf. Ser. 2013, 430, 012021. (14) Katayama, M.; Sumiwaka, K.; Miyahara, R.; Yamashige, H.; Arai, H.; Uchimoto, Y.; Ohta, T.; Inada, Y.; Ogumi, Z. X-ray Absorption Fine Structure Imaging of Inhomogeneous Electrode Reaction in LiFePO4 Lithium-Ion Battery Cathode. J. Power Sources 2014, 269, 994−999. (15) Ouvrard, G.; Zerrouki, M.; Soudan, P.; Lestriez, B.; Masquelier, C.; Morcrette, M.; Hamelet, S.; Belin, S.; Flank, A. M.; Baudelet, F. Heterogeneous Behaviour of the Lithium Battery Composite Electrode LiFePO4. J. Power Sources 2013, 229, 16−21. (16) Wang, J. J.; Chen-Wiegart, Y.-c. K.; Wang, J. In Operando Tracking Phase Transformation Evolution of Lithium Iron Phosphate with Hard X-ray Microscopy. Nat. Commun. 2014, 5, 4570. (17) Nonaka, T.; Okuda, C.; Seno, Y.; Nakano, H.; Koumoto, K.; Ukyo, Y. In situ XAFS and Micro-XAFS Studies on LiNi0.8Co0.15Al0.05O2 Cathode Material for Lithium-ion Batteries. J. Power Sources 2006, 162, 1329−1335. (18) Fuller, T. F.; Doyle, M.; Newman, J. Relaxation Phenomena in Lithium-Ion-Insertion Cells. J. Electrochem. Soc. 1994, 141, 982−990. (19) Mizushima, K.; Jones, P. C.; Wiseman, P. J.; Goodenough, J. B. LixCoO2 (0 < x