Solid State Ionics 278 (2015) 228–232 Contents lists available at ScienceDirect Solid State Ionics journal homepage: www.elsevier.com/locate/ssi Characterization of the Li-ionic conductivity of La(2/3 − x)Li3xTiO3 ceramics used for all-solid-state batteries Le Dinh Trong a, Tran Thi Thao b, Nguyen Nang Dinh b,⁎ a b Hanoi Pedagogical University 2, Phuc Yen, Vinh Phuc, Viet Nam University of Engineering and Technology, Vietnam National University in Hanoi, 144 Xuan Thuy, Cau Giay, 1000 Hanoi, Viet Nam a r t i c l e i n f o Article history: Received 16 December 2014 Accepted 30 May 2015 Available online xxxx Keywords: Double mechanical alloying (dMA) LLTO ceramics Quenching Lithium ionic conductivity Impedance All-solid-state battery a b s t r a c t With the aim to improve the ionic conductivity of perovskite materials used for all-solid-state batteries, La(2/3) − xLi3xTiO3 with x = 0.11 (LLTO11) ceramics was prepared by a double mechanical alloying method The influence of thermal treatments (furnace-cooling, SC and quenching, QC) on the crystalline structure and Li-ion conductive properties of the LLTO ceramics has been studied by X-ray powder diffraction (XRD), Raman scattering and impedance spectroscopy XRD patterns of SC-samples exhibited a doubled perovskite with a tetragonal structure, whereas those of quenched samples indicated a simple cubic perovskite The increase in the ionic conductivity of the LLTO11 ceramics was attributed to the disordered morphology that has promoted 3D-conductive mechanism At room temperature, the grain and grain-boundary conductivities of the quenched LLTO11 ceramics reached values as large as 1.8 × 10−3 S·cm−1 and 7.2 × 10−5 S·cm−1, respectively All-solidstate batteries made from the LLTO11 solid-state electrolyte combining with LiMn2O4, and SnO2 thin films as cathode and anode, respectively, possessed a charge-discharge efficiency of ~ 61% and a charging capacity of ~3.0 μAh/(cm2 · μm) at a voltage of 1.6 V © 2015 Elsevier B.V All rights reserved Introduction Solid-state electrolytes or fast ionic conductors have increasingly been studied due to their potential applications in energy storage [1], all-soldstate ionic batteries [2], environmental monitoring electrochemical sensors, and other electrochemical devices [3] These materials can be served as a non-toxic solid electrolyte that exhibits easy preservation and comfortable use Recently, many research groups have shown that a new family of the perovskite structure of La(2/3) − xLi3xTiO3 (0.03 ≤ × ≤ 0.167) (abbreviated as LLTO) materials possess a large grain Li-ionic conductivity at room temperature For instance, Inaguma et al indicated that lanthanum lithium titanate substituted with mol% Sr showed higher ionic conductivity of the bulk part (σ = 1.5 × 10−3 S cm−1 at 300 K) than the pure lanthanum lithium titanate [4,5] However, for LLTO ceramics which consists of numerous grains, and the grain-boundary conductivity is smaller than 10−3 S cm−1 [6] The low grain-boundary conductivity results in the decrease of the total conductivity of LLTO ceramic materials To enhance the ionic conductivity, great efforts have been devoted to analyzing structural and other intrinsic attributes of the LLTO, such as the concentrations of charge carriers and vacancies, structural distortions [7], lithium coordination [8], vacant site distribution [9,10], charge ⁎ Corresponding author E-mail address: dinhnn@vnu.edu.vn (N.N Dinh) http://dx.doi.org/10.1016/j.ssi.2015.05.027 0167-2738/© 2015 Elsevier B.V All rights reserved carrier mobility [11], etc Recently, by a double mechanical alloying (dMA) method, we have successfully prepared LLTO ceramics with x = 11 (called LLTO11) and obtained a significantly improved Li-ionic conductivity (i.e 1.5 × 10−3 S.cm−1 at room temperature [12]) Earlier authors have shown that the electrical performance of LLTO ceramics is strongly affected by the structural factors such as the chemical homogeneity, particle size, morphology and disordered arrangement in the crystalline structure of the samples [13–15] These important factors depend very much on both the samples fabrication methods and the post-treatments There are different polymorphs of LLTO that have been found by researching groups For instance, Harada et al indicated that a simple cubic perovskite La0.67 − xLi3xTiO3 (x = 0.06–0.15) with disordered arrangement of the A-site ions was prepared by quenching from 1350 °C into liquid N2 Whereas a tetragonal, low-temperature form of LLTO has superstructure cell, doubled along the c-axis, due to the alternate arrangement of La-rich (Li+ and vacancy-poor) layers and La-poor (Li+ and vacancy-rich) layers along the c-axis [3,7,9,13] In Li-rich LLTO samples with disordering of cations and vacancies A-site, the ionic conductivity displays a three-dimensional 3D character (shortly called 3D-conductivity) In bulk samples like LLTO ceramics the ionic conductivity can be increased by improving the 3D-conductivity factor With the aim to enhance ionic conductivity of LLTO ceramics by, LLTO11 ceramics were prepared using dMA method, followed by two ways of thermal treatments: quenching (quickly cooled, QC) and furnace-cooling (slowly cooled, SC) The crystalline structure and ionic L.D Trong et al / Solid State Ionics 278 (2015) 228–232 229 conductivity of the QC-and SC-samples were characterized The chargecharging performance of the batteries made from the LLTO ceramics served as solid electrolyte was also presented Experimental The La(2/3) − xLi3xTiO3 with x = 0.11 (LLTO11) crystalline powders was prepared from stoichiometric mixtures of dehydrated La2O3 (4 N purity), Li2CO3 (5 N) and TiO2 (4 N) powders which were purchased from Aldrich Ltd The powder of LLTO11 solid solutions was synthesized by the dMA method, according to the procedure reported in our previous work [12] To prepare ceramic specimens, this LLTO11 ceramics was isostatically pressed under 450 MPa into pellets with 12.5 mm in diameter and 1.5 mm in thickness Afterward, the pellets were sintered at temperature of 1200 °C for h with a heating rate of °C/min Then from 1200 °C, a part of these samples was quenched in liquid nitrogen (QC-samples) and the other part was furnace-cooled with cooling rate of °C/min to room temperature (SC-samples) Phase purity and crystalline structure of the synthesized materials were determined by X-ray powder diffraction (XRD) on a Siemen D5000 diffractometer, using CuKα radiation in the 2θ range from 10o to 70o with a step of 0.02o To characterize the molecular structure of samples, the Raman spectra was done on “LABRAM-1B” Micro-Raman spectrometer (backscattering configuration regime) using an excitation radiation source of He–Ne laser with a wavelength of 632.8 nm The ion conductive properties of all samples were characterized by impedance analysis on AutoLab Potentiostat-PGS30 using FRA-2 impedance The gold thin film electrodes were prepared by vacuum deposition technique on two parallel surfaces of the samples The impedance spectroscopy (IS) measurements were recorded under normal atmosphere, in temperature range from room temperature (RT) to 200 °C and in the frequency range of 0.1 to 106 Hz with an amplitude of 20 mV The resistances of grains and grain-boundaries for the QC-samples and SC-samples at different temperatures were obtained by fitting experimental data with the theoretical curves using appropriate equivalent schema The detailed technique for the fitting can be seen elsewhere [16] Results and discussion 3.1 Crystalline structure Fig shows the influence of thermal treatments on XRD patterns of LLTO11 samples For the furnace-cooled (SC) sample, the XRD patterns, including additional superstructural peaks (indicated by “stars” symbols in Fig 1a), were indexed well in a tetragonal, doubled cell (ap × ap × 2ap; Fig XRD patterns of slowly (a) and quickly (b) cooled LLTO11 samples Fig Raman scattering spectra of slowly (a) and quickly (b) cooled samples LLTO11 space group P4/mmm) with alternate arrangement of La-rich (Livacancy-poor) layers and La-poor (Li-vacancy-rich) layers along the caxis The superstructure lines completely disappeared when the samples were quenched from 1200 °C into liquid nitrogen (QC-samples) The XRD patterns obtained from quenched samples (Fig 1.b) were entirely consistent with the results reported in [13] Thus, by quenching, the obtained samples were kept at structural phase which was formed during the high-temperature synthesizing process All the XRD peaks for QCsamples were indexed in a simple cubic perovskite cell (a = ap, space group Pm3m) with disordered arrangement of the A-site cations and vacancies Fig shows Raman spectra of both the QC- and SC-samples recorded at room temperature The most remarkable influences of quenching treatments on Raman scattering spectra were the general broadening of all scattering modes and the change in their intensity For example the intensity of the mode at ~ 320 cm−1 decreased, while for the mode at ~ 450 cm− it increased Despite these differences, the spectrum of quenched sample is seen as quite similar to those of slowly cooled sample These spectra can be explained by assuming a tetragonal structure [17] Small changes in Raman spectra of quenched sample are mainly ascribed to the disorder of A-site cations and vacancies, which increases with quenching treatments Fig The CI diagrams of LLTO11 were measured at room temperature: (■) slowly and (●) quickly cooled or quenched samples 230 L.D Trong et al / Solid State Ionics 278 (2015) 228–232 Fig Equivalent schema for fitting experimental CIS curves given from the impedance measurements on QC- and SC-LLTO11 samples in which the grain size is larger than μm 3.2 Ionic conductivity Fig shows the complex impedance spectra (CIS) SC- and QCsamples measured at room temperature These samples have the same sizes (thickness d = 1.2 mm; gold electrode area A = 0.5 cm2) From Fig one can see that all the CIS consist of two parts: the first part at high frequencies is the semicircle relating to grain-boundaries conductivity and the second one is the line obtained at low frequencies that relates to the diffusion process in the Helmholtz layer The fact that the quenching resulted in the diminishing of the semicircles proves that the grain-boundaries conductance markedly increased and ending at much higher frequencies than those obtained from the furnace-cooled sample For the accurate determination of the ionic conductivity, the fitting method between experimental curves and the theoretical curves obtained from equivalent schema was used The experimental CIS diagrams recorded on the structure of Au|LLTO|Au for QC- and SCsamples with large grains were well fitted by the equivalent schema which is shown in Fig Herein Rb and C b, respectively describe the real resistance and capacitance characterizing ionic conductance in grains; Rgb and Cgb – characterizing ionic conductance in grain boundaries; Rct – chargetransfer resistance; W and Cdl – Warburg impedance characterizing charge diffusion and the capacitance of the Helmholtz layer at the electrode/LTTO interface; Cin – the capacitance of the passive layer of the contact surface of the electrode/LTTO electrolyte; Rc and Qc, respectively are the resistance and constant phase element (CPE) characterizing the porosity of the electrode From the fitting data of Rb and Rgb one can determine the ionic conductivity (grain and grain-boundary) by using the following formula: σ¼ d RÂA ð1Þ Fig Schematic drawing of an all-solid-state battery made from LLTO11 ceramics as solid electrolyte, LiMn2O4 and SnO2 films as cathode and anode, respectively where d is the sample thickness, A — area of the gold contact and R — resistance corresponding to grain or grain-boundary conductivity The calculated results of the ionic conductivity of both the QC- and SC-samples showed that the grain conductivity (σg) of the quenched samples increased slightly compared with which of slowly cooled samples, whereas, the grain-boundary conductivity increased markedly Indeed, at RT the grain and grain boundary conductivities of the quenched samples are 1.8 × 10−3 S·cm−1 and 7.2 × 10−5 S·cm−1, respectively, while which of furnace-cooled samples are 1.5 × 10−3 S·cm−1 and 3.8 × 10−5 S·cm−1, respectively Fig shows the temperature dependence of the ionic grain conductivity and grain-boundary in the logarithmic function of σ vs T for both SC-samples and QC-samples The plots of ln(σT) against 1000/T were found to follow the Arrhenius law expressed as follows: σ¼ σ0 Ea exp − T kT ð2Þ where σ o is the ionic conductivity at room temperature, E a — the activation energy Deng at al [18] indicated that the transition point in Arrhenius plots of the grain and grain-boundaries conductivities can be determined for different specimens, which relate to tilt or rotate the TiO6 octahedra, leading to “open or close the bottle neck” in the perovskite structure, through which lithium ions move into nearby A-site vacancies The Fig Arrhenius plots of grain ionic conductivity (a) and grain-boundary conductivity (b) for SC-sample (■) and QC-sample (●) L.D Trong et al / Solid State Ionics 278 (2015) 228–232 231 Fig Charging (a) and discharging curves (b) for a SSIB device with a structure of Al/LiMn2O4/LLTO/SnO2/Al conductivity of the QC-sample increase due to the disordered arrangement of the A-site ions and vacancies of perovskite, improving 3D ion conductive process In addition, the increase of grain boundary conductivity is explained by the changes of grain boundary composition when the sample is quenched From Arrhenius plots in Fig it was found that the activation energy for grain-conductance of the QC-sample (Ea = 0.23 eV) is slightly smaller than that of the SC-sample (Ea = 0.27 eV), while Ea for grainboundary conductance exhibited the same value as ~ 0.32 eV The fact that conductivity of the SC-sample is lower than that of the QC-one clearly demonstrated a close relationship between the ionic conductance and activation energy The lower conductivity was attributed to the increase in activation energy that is associated with the decrease in the lattice parameter of the primitive cell [13] 3.3 Device performance To characterize the electrochemical properties of the LLTO electrolyte we prepared an all-solid-state battery using the QC-samples (namely pellets with 12.5 mm in diameter and 1.5 mm in thickness) For improvement of the performance efficiency and the contacts of the electrolyte/anode and electrolyte/cathode, the pellets were carefully polished by synthesized diamond powder of 1.0 μm in size Then a 400 nm-thick LiMn2O4 and a 500 nm-thick SnO2 films, respectively served as cathode and anode were deposited onto two sides of the LLTO pellet Finally, by using vacuum evaporation, 200 nm-thick Alelectrodes served as collectors were deposited on the tops to get a laminar structure of the all-solid-state battery, Al/LiMn2O4/LLTO/SnO2/ Al (abbreviated as SSIB), as shown in Fig Fig shows the charge and discharge curves of the SSIB at room temperature Curve “1” in Fig 7a exhibits the first cycle of charging with a current density of 20 μA·cm−2, while curve “2” is the second and more cycle of charging with the same current density (i.e 20 μA·cm−2) From the first to the second cycle, the voltage increased from 2.6 V to 3.6 V Curves “1”, “2” and “3”, respectively correspond to the first, second and fifth cycles of discharging with a current density of 2.0 μA·cm−2 During the discharge process, the voltage decreased from 2.5 V to 1.8 V The last was maintained for a long time We also determined both the discharging capacity and charge-discharge efficiency, i.e the ratio of the charge from discharging per the charge from charging in a charge-discharge cycle It has been seen that for the first cycle the efficiency was not large (21%), but it considerably increased from the second and more cycle of the charge-discharge process (up to 61%) At a voltage of 1.6 V the capacity was reached a value of ~3.0 μAh/(cm2 · μm) The limitation in the discharge capacity of the SSIB battery (Fig 7b) is explained due to the reaction between Li+ and SnO2 to form amorphous Li2O and metallic Sn [19, 20] Comparing with lithium metal electrode in terms of electrochemical properties, SnO2 film is not so good However, tin oxide is not only able to endure high fabrication temperatures, but can also preserve its capacity for a large number of cycles after the initial discharge process [21] Conclusion As a solid-state electrolyte, perovskite solid electrolyte La(2/3) − (x = 0.11) has been prepared by dMA method The cubic morphology, disordered form with disordered arrangement of the Asite cations and vacancies, was crystallized by quenching from 1200 °C into liquid nitrogen, while the tetragonal morphology ordered form with alternate arrangement of La-rich (Li-vacancy-poor) layers and Lapoor (Li-vacancy-rich) layers along the c-axis, was formed by slowly cooling from sintering temperatures The increase in the ionic conductivity of the LLTO ceramics, especially grain-boundary conductivity, can be attributed to the disordered morphology that has promoted 3D-conductive mechanism and the quenching has changed the composition and structure of grain boundaries At room temperature, the grain and grain-boundary conductivities of the quenched LLTO11-ceramics reached values as large as 1.8 × 10− S·cm− and 7.2 × 10− S·cm− 1, respectively Combining with LiMn2O4 and SnO2 thin films served as cathode and anode, the LLTO11 ceramics as solid electrolyte were used for fabrication of allsolid-state batteries with a structure of Al/LiMn2O4/LLTO/SnO2/Al The last possess a reasonable charge-discharge efficiency of 61% and a charging capacity of ~ 3.0 μAh/(cm2 · μm) at a voltage of 1.6 V xLi3xTiO3 Acknowledgment This research was funded by the Vietnam National Foundation for Science and Technology (NAFOSTED) under grant number 103.02-2013.39 References [1] P.G Bruce, S.A Freunberger, L.J Hardwick, J.M Tarascon, Nat Mater 11 (2012) 19–29 [2] D Linden, T.B Reddy, Handbook of batteries, Mc Graw-Hill Companies Inc., 2002 [3] C.M Lampert, C.G Granqvist, Large-area Chromogenics: Materials and Devices for Transmittance Control, vol 154, SPIE Optical Engineering Press, Bellingham, 1990 [4] Y Inaguma, C Liquan, M Itoh, T Nakamura, T Uchida, H Ikuta, M Wakihara, Solid State Commun 86 (10) (1993) 689–693 [5] Y Inaguma, L Chen, M Itoh, T Nakamura, Solid State Ionics 70/71 (1994) 196–202 [6] K Dokko, N Akutagawa, Y Isshiki, K Hoshina, K Kanamura, Solid State Ionics 176 (2005) 2345–2348 [7] J Ibarra, A Varez, C Leon, J Santamarıa, T.L.M Martinez, J Sanz, Solid State Ionics 134 (2000) 219–228 [8] T Okumura, T Ina, Y Orikasa, H Arai, Y Uchimoto, Z Ogumi, J Mater Chem 21 (2011) 10195–10205 [9] J.A Alonso, J Sanz, J Santamaría, C Léon, A Várez, M.T Fernánez-Díaz, Angew Chem 39/3 (2000) 619–621 232 L.D Trong et al / Solid State Ionics 278 (2015) 228–232 [10] X Gao, C.A.J Fisher, T Kimura, Y.H Ikuhara, H Moriwake, A Kuwabara, H Oki, T Tojigamori, R Huang, Y Ikuhara, Chem Mater 25 (2013) 1607–1614 [11] A Rivera, C Ln, J Santamaría, A Várez, M.A Paris, J Sanz, J Non-Cryst Solids 307–310 (2002) 992–998 [12] L.D Trong, N.N Dinh, P.D Long, J Mater Sci Eng B4/4 (2014) 86–94 [13] Y Harada, T Ishigaki, H Kawai, J Kuwano, Solid State Ionics 108 (1998) 407–413 [14] O Bohnke, H Duroy, J.L Forquet, S Ronchetti, D Mazza, Solid State Ionics 149 (2002) 217–226 [15] A Várez, J Ibarra, A Rivera, C Léon, J Santamariza, M.A Laguna, M.L Sanjuán, J Sanz, Chem Mater 15 (2003) 225–231 [16] N.N Dinh, P.D Long, L.D Trong, Commun Phys 14/2 (2004) 90–94 [17] M.L Sanjuán, M.A Laguna, Phys Rev B64 (2001) 174305 (5 p) [18] Y Deng, S.J Shang, A Mei, Y.H Lin, L.Y Liu, C.W Nan, J Alloys Compd 472/1–2 (2009) 456–460 [19] N Kuwata, J Kawamura, K Toribami, T Hattori, N Sata, Electrochem Commun (2004) 417–421 [20] N Kuwata, R Kumar, K Toribami, T Suzuki, T Hattori, J Kawamura, Solid State Ionics 177 (2006) 2827–2832 [21] Y.J Park, K.S Park, J.G Kim, M.K Kim, H.G Kim, H.T Chung, J Power Sources 88 (2000) 250–254 ... characterizing charge diffusion and the capacitance of the Helmholtz layer at the electrode/LTTO interface; Cin – the capacitance of the passive layer of the contact surface of the electrode/LTTO electrolyte;... conductivity and the second one is the line obtained at low frequencies that relates to the diffusion process in the Helmholtz layer The fact that the quenching resulted in the diminishing of the semicircles... from the second and more cycle of the charge-discharge process (up to 61%) At a voltage of 1.6 V the capacity was reached a value of ~3.0 μAh/(cm2 · μm) The limitation in the discharge capacity of