Communication pubs.acs.org/cm
Relocation of Cobalt Ions in Electrochemically Delithiated LiCoPO4 Cathode Materials Quang Duc Truong,*,† Murukanahally Kempaiah Devaraju,† Yoshikazu Sasaki,‡ Hiroshi Hyodo,† Takaaki Tomai,† and Itaru Honma*,† †
Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980-8577, Japan DATUM Solution Business Operations, JEOL Ltd., Tokyo 196-0022, Japan
‡
S Supporting Information *
T
he point defects in crystal lattices including antisite cation exchange, dopants, and atomic vacancies have been the topic of extraordinary research interest in solid state physics and chemistry.1−3 The optical properties, electrical conductivity, ionic diffusion and resulting chemical properties, and mass and charge transport behavior of the materials are heavily affected by the concentration and distribution of these defects in the crystal lattices, which has triggered considerable efforts in introducing the intentional defects to provide optimal performance in devices.4 In lithium intercalation compounds, the lithium diffusion along the channel is highly anisotropic, which is strongly dependent on the cation ordering within the crystal lattices.5,6 The cation exchange disorder,7,8 namely, the occupation of Li sites by transition metals, inevitably blocks the lithium ion diffusion pathway and, thus, directly affects the cathode performance in lithium-ion batteries. The presence of antisite defects in lithium-intercalated transition-metal oxides and phosphates has been clearly confirmed by theoretical calculations, neutron diffraction, and direct observation by advanced electron microscopy.7−10 We now turn to the research on the local variations, transition metal relocation, structural change, and their correlation to the voltage/capacity fading during the charge/discharge cycling.11−22 The investigation by advanced electron microscopy on the structural change of the layered nickel manganese oxides during the synthesis or electrochemical cycling reveals that the migration and segregation of transition metal ions, i.e., Ni, may initiate the phase transformation and inhibit the battery charge/discharge rate.16−20 However, the relocation of metal ions in ordered olivine lithium metal phosphates during the intercalation− deintercalation reactions remains unknown, although this migration undoubtedly influences the mass and the charge transport behavior of the olivine materials. Herein, we observed the local variation of cobalt ions from M2 sites to vacancy M1 sites in olivine lithium cobalt phosphates upon the electrochemical delithiation process using aberration-corrected scanning transmission electron microscopy. The finding provides the insight into the capacity fading mechanism of the LiCoPO4 cathode materials. Figure 1a illustrates the olivine-type lithium metal phosphate LiCoPO4 structure. The distorted CoO6 octahedrals are crosslinked with PO4 tetrahedral oxo-anions either by corner sharing or edge sharing, forming ordered olivine frameworks in the space group Pnma. Li locates at edge-sharing octahedral M1 sites, and Co is found in the corner sharing octahedral M2 sites. © 2014 American Chemical Society
Figure 1. (a) Schematic illustration for the LiCoPO4 olivine structure with cation ordering Li at M1 sites and Co at M2 sites. (b) ABF and (c) HAADF-STEM images of LiCoPO4 particle viewed along the [010] direction. The superimposed atomic arrays on the inset indicate the locations of atom columns.
The cation exchange antisite defects between two octahedral M sites are the most favorable intrinsic point defects in the olivine structure.7,8,10 Figure 1b,c shows annular bright-field (ABF) and high-angle annular dark-field (HAADF) images viewed along [010] direction of a LiCoPO4 particle synthesized at 750 °C (see Supporting Information for detailed synthesis). The twodimensional atomic arrangement of a unit cell is overlaid on the image for the comparison. The periodically arrayed brightest contrasts in Figure 1c exhibit the Co cobalt columns. The P columns are observed locating nearby each Co site. Because the ADF image contrast roughly correlates with atomic number according to a Z1.7 relationship,11,21 O atoms (Z = 8) and Li atoms (Z = 3) are invisible even at high resolution of the HAADF mode.11 However, the O columns can be visualized in Received: February 7, 2014 Published: April 22, 2014 2770
dx.doi.org/10.1021/cm501452p | Chem. Mater. 2014, 26, 2770−2773
Chemistry of Materials
Communication
the ABF image (Figure 1b). Because of poor scattering ability, Li columns should have no contrast in the ADF imaging. Therefore, the contrast visualized in the Li sites in Figure 1c indicates the presence of Co atoms at the Li sites as antisite defects. Comparing to the HAADF image of sample synthesized at 500 °C,23 the bright contrast in the present case is much weaker, indicating the lower degree of antisite defects. It was suggested that annealing at higher temperature can reduce the concentration of the cation disorder. The delithiated Li1−xCoPO4 was obtained by electrochemically charging the original sample to 5.1 V at current rate 0.1 C. The charge profile is shown in Figure S1 in the Supporting Information. The charge capacity was 121 mA h g−1, corresponding to the extraction of 0.72 Li per formula unit. The sample on the Pt current collector was disassembled in the glovebox and washed with ethanol. The X-ray diffraction (XRD) pattern shows the formation of the Li-extracted intermediated Li1−xCoPO4 phase with diffraction peaks shifting to larger angle (Supporting Information Figure S2 and Rietveld refinement results in Supporting Information Table S1 and Figures S3 and S4). The ordered olivine structure was maintained after delithiation as indicated in high-resolution TEM images and selected-area electron diffraction (SAED) patterns (Supporting Information Figure S5). Figure 2 shows
Figure 3. Schematic illustrations for the migration of metal ion during the electrochemically delithiated process.
Figure 4 displays the HAADF and ABF images at lower magnification (compared to that of Figure 2) of the delithiated
Figure 4. (a) HAADF and (b) ABF-STEM images of delithiated Li1−xCoPO4 particle viewed along the [010] direction. The superimposed atomic arrays on the inset indicate the locations of atom columns. The red arrows indicate the Co column with faint contrast.
Li1−xCoPO4 sample. We observed some Co columns with deficiently dark contrast in the ABF image as indicated by red arrows in Figure 4b and Supporting Information Figure S6a. The weak contrast of these Co columns is presumably due to the relocation of Co ions from M2 to M1 sites. The deficiently dark contrast can be found only in the typical region while the strong contrast remained in the large observed area (Supporting Information Figure S6b). The result indicates the local segregation of vacancies at the Co sites, VCo. Such local segregation is thermodynamically favorable since the extraction of Co would lower the energy barrier for migration of the other metal ions in the same channel. Furthermore, the fairly low intensities were observed at O columns as the result of diffused O columns and Co−P column distortion upon the delithiation (see Supporting Information Figure S7 for direct comparison).21 It should be noted that the intensity of contrasts in STEM images is dependent on the atomic number Z. Thus, the migration of a small number of heavy metals (for example, 6 atoms)7 could lead to visible contrast of the Li columns. According to the olivine structure, the number of Co and Li columns in the [010] projection is not equal but follows the ratio of 4n2/(2n2 + 2n +1) with n being the number of unit
Figure 2. HAADF-STEM images of LiCoPO4 particle viewed along the [010] direction: (a) original and (b) delithiated samples. The scale bars = 5 Å. (c, d) The corresponding line profiles showing the image intensity as a function of position in the HAADF images.
HAADF images of the electrochemically delithiated Li1−xCoPO4 viewed from [010] orientation in comparison with that of the original LiCoPO4 sample. Notably, much stronger contrast was observed at the Li sites in the delithiated sample, indicating the remarkably high degree of cation disorder. The contrast comparison analyzed by line profile for original and delithiated samples shown in Figure 2c,d further confirmed the increase in the contrast at Li sites. The increased degree of defect suggests that the cobalt metal ions have diffused and occupied vacancies created by Li extraction during the delithiation process (Figure 3). The ions migration is presumably driven by the vacancy stabilization as high Li vacancy concentration promoted the increase in the CoLi disorder.24 2771
dx.doi.org/10.1021/cm501452p | Chem. Mater. 2014, 26, 2770−2773
Chemistry of Materials
Communication
In summary, we have observed, at atomic resolution by stateof-the-art ABF/HAADF STEM, the local atomic rearrangement of the olivine LiCoPO4 structure upon the electrochemical delithiation process. The extraction of Li ions from the lattice promotes the relocation of Co ion from M2 sites to the vacancy M1 sites. The vacancies created by Co ions migration appear to be locally segregated in the channel. The relocation resulted in the large capacity loss (low coulomb efficiency) in the initial discharge process. The structural analysis in the present work provides the intrinsic evidence for the capacity fading of the LiCoPO4 cathode materials.
cells. Thus, the relocation of Co atoms in a small number of columns may induce strong contrast at many Li columns. It was suggested that the M2 sites are highly unstable if left unfilled due to a local accumulation of negative charge.25 The distance between neighboring cations in edge-sharing M1 octahedral is also shorter than that in corner-sharing M2 octahedral, leading to stronger electrostatic repulsion between the cations.7 Thus, the relocation of Li from M1 sites to the vacancies VCo is also favorable, and the lithium/transition metal ions exchange may also occur during the delithiation process.26 The electrochemical performance of the synthesized LiCoPO4 has been measured to elucidate effect of the cobalt relocation, and the result is shown in Figure 5. The charge
■
ASSOCIATED CONTENT
S Supporting Information *
Experimental section, XRD patterns, result of Rietveld refinement, HRTEM images, SAED patterns, ABF-STEM images, electrochemical performance of the original and delithiated samples, and more discussion. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Authors
*E-mail:
[email protected] (Q.D.T.). *E-mail:
[email protected] (I.H.). Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This research work was financially supported by Japan Society for Promotion of Science (JSPS, Grant No. P13070), Funding Program for World-Leading Innovative R&D on Science and Technology (FIRST) and Core Technology Consortium for Advanced Energy Devices, Tohoku University, Japan.
Figure 5. Charge/discharge profile of the synthesized LiCoPO4 in Liion batteries tested in the potential range of 3.0−5.1 V at 0.1 C rate.
curves possesses two voltage plateaus, indicating two-step lithium delithiation.27 The LiCoPO4 exhibited a wide and flat voltage plateau at around 4.75 V versus Li/Li+ with a initial charge capacity of 121.2 mA h g−1 (0.72 Li) at C/10 rate. At the first cycle, only 0.48 Li was reinserted and the coulomb efficiency is 66.3% (Supporting Information Figure S8). From the sequential cycles, the coulomb efficiency gradually increase to 84.8% and becomes stable. The sample showed lower charge/discharge capacities compared to that in our previous work and other studies.23 The low charge capacity in the initial cycle is believed to be due to the presence of antisite defects.23 The partial occupation of Li sites by Co atoms inevitably blocks the lithium ion diffusion pathway; thus, only partial Li ions were extracted from the frameworks. The great capacity loss observed in the first cycle is presumably due to relocation of Co to the M1 site which created more blocking channels or increase the degree of disorder. It also appears that the increase in the degree of cation disorder after delithiation prevents the reinsertion of Li ion into the vacancies VLi which have been occupied by Co ions. The ability of reinsertion of Li ions into the vacancy VCo is not easy to determine and needs further experimental analyses as well as modeling calculation. The stable coulomb efficiency achieved after the first cycle indicates that the degree of site exchange defect becomes stable after initial delithiation. Notably, the charge capacity of the sequential cycle is approximate to the discharge capacity of the previous cycle (Supporting Information Figure S6). This result suggests that the unextracted Li ions in the previous charging reaction remain inactive in the sequential charging process.
■
REFERENCES
(1) Henderson, C. M. B.; Knight, K. S.; Redfern, S. A. T.; Wood, B. J. Science 1996, 271, 1713−1715. (2) Norris, D. J.; Efros, A. L.; Erwin, S. C. Science 2008, 319, 1776− 1779. (3) Esch, F.; Fabris, S.; Zhou, L.; Montini, T.; Africh, C.; Fornasiero, P.; Comelli, G.; Rosei, R. Science 2005, 309, 752−755. (4) Sato, Y.; Buban, J. P.; Mizoguchi, T.; Shibata, N.; Yodogawa, M.; Yamamoto, T.; Ikuhara, Y. Phys. Rev. Lett. 2006, 97, 106802. (5) Shao-Horn, Y.; Levasseur, S.; Weill, F.; Delmas, C. J. Electrochem. Soc. 2003, 150, A366−A373. (6) Kang, K.; Meng, Y. S.; Breger, J.; Grey, C. P.; Ceder, G. Science 2006, 311, 977−980. (7) Chung, S. Y.; Choi, S. Y.; Yamamoto, T.; Ikuhara, Y. Phys. Rev. Lett. 2008, 100, 125502. (8) Chung, S. Y.; Choi, S. Y.; Yamamoto, T.; Ikuhara, Y. Angew. Chem., Int. Ed. 2009, 48, 543−546. (9) Koyama, Y.; Arai, H.; Tanaka, I.; Uchimoto, Y.; Ogumi, Z. J. Power Sources 2013, 244, 592−596. (10) Gardiner, G. R.; Islam, M. S. Chem. Mater. 2010, 22, 1242− 1248. (11) Lu, X.; Sun, Y.; Jian, Z.; He, X.; Gu, L.; Hu, Y. S.; Li, H.; Wang, Z.; Chen, W.; Duan, X.; Li, C.; Maier, J.; Tsukimoo, S.; Ikuhara, Y. Nano Lett. 2012, 12, 6192−6197. (12) Jarvis, K. A.; Deng, Z.; Allard, L. F.; Manthiram, A.; Ferreira, P. J. Chem. Mater. 2011, 23, 3614−3621. (13) Boulineau, A.; Simonin, L.; Colin, J. F.; Bourbon, C.; Patoux, S. Nano Lett. 2013, 13, 3857−3863. (14) Yu, H.; Ishikawa, R.; So, Y. G.; Shibata, N.; Kudo, T.; Zhou, H.; Ikuhara, Y. Angew. Chem., Int. Ed. 2013, 52, 5969−5973. 2772
dx.doi.org/10.1021/cm501452p | Chem. Mater. 2014, 26, 2770−2773
Chemistry of Materials
Communication
(15) Wang, R.; He, X.; He, L.; Wang, F.; Xiao, R.; Gu, L.; Li, H.; Chen, L. Adv. Energy Mater. 2013, 10, 1358−1367. (16) Xu, B.; Fell, C. R.; Chi, M.; Meng, Y. S. Energy Environ. Sci. 2011, 4, 2223−2233. (17) Fell, C. R.; Qian, D.; Carroll, K. J.; Chi, M.; Jones, J. L.; Meng, Y. S. Chem. Mater. 2013, 25, 1621−1629. (18) Gu, M.; Belharouak, I.; Zheng, J.; Wu, H.; Xiao, J.; Genc, A.; Amine, K.; Thevuthasan, S.; Baer, D. R.; Zhang, J.-G.; Browning, N. D.; Liu, J.; Wang, C. ACS Nano 2012, 7, 760−767. (19) Gu, M.; Belharouak, I.; Genc, A.; Wang, Z.; Wang, D.; Amine, K.; Gao, F.; Zhou, G.; Thevuthasan, S.; Baer, D. R.; Zhang, J.-G.; Browning, N. D.; Liu, J.; Wang, C. Nano Lett. 2012, 12, 5186−5191. (20) Gu, M.; Genc, A.; Belharouak, I.; Wang, D.; Amine, K.; Thevuthasan, S.; Baer, D. R.; Zhang, J.-G.; Browning, N. D.; Liu, J.; Wang, C. Chem. Mater. 2013, 25, 2319−2326. (21) Gu, L.; Zhu, C.; Li, H.; Yu, Y.; Li, C.; Tsukimoo, S.; Maier, J.; Ikuhara, Y. J. Am. Chem. Soc. 2011, 133, 4661−4663. (22) Zhu, C.; Gu, L.; Suo, L.; Popovic, J.; Li, H.; Ikuhara, Y.; Maier, J. Adv. Funct. Mater. 2014, 24, 312−318. (23) Truong, Q. D.; Devaraju, M. K.; Tomai, T.; Honma, I. ACS Appl. Mater. Interfaces 2013, 5, 9926−9932. (24) Badi, S. P.; Wagemaker, M.; Ellis, B. L.; Singh, D. P.; Borghols, W. J. H.; Kan, W. H.; Ryan, D. H.; Mulder, F. M.; Nazar, L. F. J. Mater. Chem. 2011, 21, 10085−10093. (25) Kuss, C.; Liang, G.; Schougaard, S. B. Presented at the 223rd ECS Meeting, 2013, Abstract #401. (26) Yu, H.; Qian, Y.; Ohtani, M.; Tang, D.; Guo, S.; Zhu, Y.; Zhou, H. Energy Environ. Sci. 2014, 7, 1068−1078. (27) Bramnik, N. N.; Nikolowski, K.; Baehtz, C.; Bramnik, K. G.; Ehrenberg, H. Chem. Mater. 2007, 19, 908−915.
2773
dx.doi.org/10.1021/cm501452p | Chem. Mater. 2014, 26, 2770−2773