Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Li4-xGe1-xPxO4 a potential solid-state electrolyte for all-oxide microbatteries

Li4-xGe1-xPxO4 a potential solid-state electrolyte for all-oxide microbatteries Solid-state electrolytes for Li-ion batteries are attracting growing interest as they allow building safer batteries, also using lithium metal anodes. Here we studied a compound in the lithium superionic conductor (LISICON) family, i.e. Li Ge P O (LGPO). Thin films were 4-x 1-x x 4 deposited via pulsed laser deposition and their electrical properties were compared with ceramic pellets. A detailed characterization of the microstructure shows that thin films can be deposited fully crystalline at higher temperatures but also partially amorphous at room temperature. The conductivity is not strongly influenced by the presence of grain boundaries, exposure to air or lithium deficiencies. First-principles molecular dynamics simulations were employed to calculate the lithium ion diffusion profile and the conductivity at various temperatures of the ideal LGPO crystal. Simulations gives the upper limit of conductivity for -2 -1 a defect free crystal, which is in the range of 10 S cm at 300 °C The ease of thin film -1 fabrication, the room-temperature Li-ion conductivity in the range of a few µS cm make LGPO very appealing electrolyte material for thin film all–solid-state all-oxide microbatteries. Introduction energy storage. Conventional batteries use liquid organic solvents as the electrolyte, Lithium ion batteries are nowadays among which are corrosive and flammable, the most widely used technology for 1 therefore posing important safety issues. exploiting the “mixed polyanion effect”: The risk of leaks precludes the potential the introduction of different cations causes use of these batteries for biomedical a disordered redistribution of the anionic implants. A liquid electrolyte imposes also tetrahedra, which facilitates the lithium ion 11-12 severe restrictions to the possible migration. miniaturization of the battery, which is Compounds in the LISICON family are needed for on-chip power sources for also particularly interesting for the growth portable/wearable applications. Also, Li of thin films, as they can be deposited metal dendrites can easily grow into a 13-15 easily by sputtering and pulsed laser liquid electrolyte, short circuiting the deposition (PLD). Thin films of 1-2 electrodes. LISICON materials reported in the The use of solid-state electrolytes (SSE) literature are generally amorphous, which would solve or mitigate these issues, as further enhances the lithium ion SSE are not flammable, cannot leak, and conductivity. can in principle be grown as nanometric On the contrary, the growth of thin films of thin films for battery miniaturization. other, much more conductive, oxides, such Many SSE are stable within a large as LLZO and LLTO, turns out to be 3-4 electrochemical voltage window. 17-20 extremely difficult. Different reports Finally, SSE hinder the formation of 15- can be found in the literature on silicates, dendrites, although when metallic lithium 16, 21 but nothing has been published to the anode is employed, the problem cannot be 5 best of our knowledge on germanate completely avoided. materials. Inorganic oxides SSE are less conductive Here we report the deposition via PLD of at room temperature than sulfides or thin films of lithium germanium halides, but generally more stable in phosphate, with the general formula Li 4- contact with electrodes as well as when Ge P O (LGPO), a solid solution of γ- x 1-x x 4 exposed to air. 22-23 Li PO and Li GeO . Different 3 4 4 4 Different oxides show promising deposition temperatures led to films with properties for possible use as SSE in different crystalline quality and chemical batteries, for instance Li La Zr O 7 3 2 12 composition. These features were then 6-7 (LLZO) and Li La TiO (LLTO). 3x 2/3-x 3 correlated to the electrical properties. These materials offer lithium ion First-principles molecular dynamics conductivities in the order of a few mS cm at room temperature. Also, ionic simulations were then employed to conductors in the LISICON family, with a calculate the Li-ion diffusion profile and composition of Li M Y O (M = Ge, diffusivity at different temperatures in an 4±x 1−x x 4 Si; Y = P, Al, Zn, Ge, Ga, Sb) are well ideal (defect-free) LGPO crystal. known solid-state electrolytes, with a Methods conductivity at room temperature in the range between a few and a few tens of Powder synthesis and characterization -1 8-10 µS cm . Recent works highlighted that Li P Ge O (LGPO) powders were 3.2 0.80 0.20 4 the conductivity can be further improved 2 synthesized via the solid-state synthesis temperature at the same oxygen method using lithium carbonate (Li CO ), background pressure. 2 3 germanium oxide (GeO ) and lithium Thin film characterization phosphate (Li PO ). Lithium carbonate and 3 4 The crystal structure of the films was germanium oxide were mixed in characterized by X - ray diffraction (out of stoichiometric amount and grounded using plane /2 scan) with a Seifert an agate mortar, and the resulting powders diffractometer with monochromatic were pressed uniaxially with 0.5 GPa. The Cu - Kα1 radiation. The thin films pressed powders were sintered at 800 °C deposited at low temperature were for 8 hours. Resulting Li GeO and Li PO 4 4 3 4 characterized also by 2 scan. While  was powders were then mixed and pressed fixed at 1°, 2 was scanned between 10° uniaxially with 0.5 GPa. The powders were and 65°. then placed in a tubular furnace with constant oxygen flux at 900 °C for For the in-plane (along the direction of the 12 hours. The resulting pellets were 13 mm substrate surface) electrical in diameter with a thickness of 2 mm. The characterization two strip-shaped gold same procedure was repeated for sintering electrodes, 100 nm thick, were sputtered ceramic LGPO pellets used for on the surface of the films. conductivity measurements and as target Electrical characterization was performed for thin film deposition. The crystal by impedance spectroscopy (IS) in Ar structure of the synthesized LGPO flow, between 2 MHz and 1 Hz, applying powders were characterized by X-ray an excitation of 1 V. Ag paste and Au wires diffraction (Siemens D500 diffractometer were used to connect the electrodes to the with Cu-Kα radiation). measurement cell. A Solatron 1260 gain- Thin film deposition phase analyzer and the software Zplot were Thin films of LGPO were deposited via used for the impedance measurement. The software Zview was used to fit the complex pulsed laser deposition (PLD) with a KrF impedance plane plots. Thin films excimer laser, wavelength 248 nm, laser deposited at high temperature were fluence of 2.1 J/cm , repetition rate of equilibrated for 2 hours at 430 °C. The 10 Hz. To compensate the expected Li loss temperature was then decreased in steps of (a problem often encountered during the 20 °C and the samples were equilibrated growth of Li-containing thin films) a ceramic target with 10 mol% excess Li (as for 30 minutes at each temperature step Li O)was used. The distance between before acquiring a complex impedance plot. Thin films deposited at room target and substrate during deposition was temperature were equilibrated 24 hours at 60 mm. Thin films were deposited on 25 °C and data was acquired while single crystal (100) - oriented MgO increasing the temperature up to 480 °C. substrate. Silver paste was used to provide The conductivity was also measured the thermal contact between the substrate decreasing the temperature following the and the heating stage in the vacuum chamber. The oxygen pressure was set to same procedure as described above. 0.01 mbar during the deposition. Films Rutherford Back Scattering was performed were deposited at 500 °C and at room with a 2 MeV 4He beam and a silicon PIN 3 diode detector under 168°. Data were ensemble, temperature was controlled via a analyzed by the RUMP code. Elastic Nosé-Hoover thermostat for each atomic 33-34 Recoil Detection Analysis with a 13 MeV species. 127I primary ion beam was primarily used From the NVT trajectories, the Li-ion to determine the lithium-to-oxygen ratio. positions were extracted and the Rt  Recoiling ions were detected by the i combination of a time-of-flight Li-ion probability density was nr   Li spectrometer with a gas ionization calculated as: chamber. The acquired spectral data were analyzed by a custom software. (1) rR t   Li 3  2 n r  2 e     Li  First-Principles molecular dynamics To examine Li-ion diffusivity in LGPO on The normal distribution was used to the atomic scale in a perfect crystal with a account for finite statistics, setting the composition of Li Ge P O , 3.33 0.33 0.67 4 deviation to σ = 0.3 Å, and the summation extended first-principles molecular -1 was performed on a grid of 0.1 Å . dynamics (FPMD) simulations were performed. The Car-Parrinello (CPMD) In addition, we analyzed the trajectory scheme was used, based on Kohn-Sham using the SITATOR package for an 25-26 density-functional theory (DFT) in the unsupervised analysis of the diffusive plane-wave pseudopotential formalism, pathways in the system. as implemented in the cp code of the Li The tracer diffusion matrix D was Quantum ESPRESSO distribution computed from the NVT trajectories via (http://www.quantum-espresso.org/). A the mean square displacement (MSD) of plane-waves cutoff of 50 Ry for the lithium, according to the Einstein relation: wavefunctions and 400 Ry for the electron density and ultrasoft pseudopotentials from (2) Standard Solid State Pseudopotential N Li Li l l l l D  lim R t   R tR t   R t  29 30  ij i i j j (SSSP) Efficiency library 1.0 (GBRV  2 N l Li for Li and O, PSLib for Ge and P) were  lim MSD  , ij employed, with Brillouin-zone integration  2 at the Gamma point. The exchange- where is the i-component ( ) correlation functional was PBE. Rt  i  x,, y z of the Li-ion position at time t, and l th CPMD simulations, with a time step of the angular brackets define a time 4 a.u. and an electronic mass of 500 a.u., were performed for 200 – 500 ps in the average over the starting times t, that, fixed simulation (1x1x3) supercell defined following the ergodic hypothesis, we by the experimental crystallographic data corresponding to an ensemble assume at six temperatures ranging between average. We computed this average, and its 600 K and 1400 K (600, 720, 900, 1000, statistical error, by means of a block 1100, 1400 K). To sample the NVT analysis. The tracer diffusion coefficient 4 Li Li was calculated from the matrix as The electrical conductivity was measured D D tr by impedance spectroscopy (IS) with gold Li 1/3 Tr(D ), or directly from the formula (Li - ion blocking) electrodes. Two (3) representative Nyquist plots, used to calculate the total resistance, are reported Li 2 Li l l D  lim R t  R t      in Figure 1b. With our samples and tr   6 N Li t selected experimental conditions, it was not possible to clearly distinguish the grain  lim MSD  .    6 boundary and grain interior contributions to the conductivity. Therefore, only the From the tracer diffusion coefficient at total resistance could be reliably measured. each temperature, the ionic conductivity was extracted via the Nernst-Einstein LGPO thin films equation: LGPO thin films were grown by pulsed (4) laser deposition (PLD) on 2 MgO (100) - oriented single crystal Ne Li Li  T  D () T   tr substrates. A ceramic pellet fabricated in Vk T our laboratory was used as target for the ablation process. where is the Li-ions density and NV / e Li In the case of complex multi-element is the elementary charge. We assumed an materials, containing elements with very average Born effective charge for Li of different atomic masses, the target-to-film +1. compositional transfer can be a severe is the Boltzmann constant and the 38-39 issue. Many factors, other than the absolute temperature. target composition, can influence the film elemental content. Due to the interaction of Results and Discussion the ablated species with the gaseous environment during the time of flight LGPO ceramic pellets towards the substrate, pulsed laser LGPO ceramic pellets, with a nominal deposited films are typically enriched with composition of Li Ge P O were 3.2 0.2 0.8 4 the heavier elements at the expense of the prepared by solid - state synthesis, then lighters. This is why it is very difficult to pressed and sintered. The density of the grow films with the desired Li content. To -3 pellets was 2.20 g cm , corresponding to partially compensate for the Li loss, 10 % 85 % of the theoretical density. The XRD mol excess Li was added to the target as analysis and comparison with the ICSD Li O. Films with thickness between reference (Figure 1a) indicates the 150 nm and 300 nm were deposited in an formation of the expected LGPO oxygen background pressure of 0.01 mbar. orthorhombic structure (space group The samples were prepared at room P n m a). Three minor diffraction peaks temperature and at 500 °C to probe the corresponding to lithium carbonate and effects of temperature on composition, lithium phosphate were also detected. crystallinity, and transport properties. 5 Instead, no diffraction peaks can be seen in the plot of the 2θ/ω scan of the sample grown at room temperature (Figure 2a-II). For this sample, more can be learned from the plot of the 2θ scan shown in Figure 2b. This XRD pattern can be explained assuming a nanocrystalline film, which could also be partially amorphous. However, the presence of diffraction peaks clearly ascribable to the LGPO structure indicates a certain degree of crystallinity also for the films deposited at room temperature. This is noteworthy, because in the literature LISICON thin films deposited via PLD were reported to be amorphous, on the basis of the absence of diffraction peaks in the 2θ/ω scan. Our results show that this is not always the case. An accurate analysis of the crystallinity is important, as complete film amorphization is known to improve the conductivity in LISICON materials. Figure 2c shows representative complex impedance plane plots acquired at the same Figure 1 XRD and IS characterization of temperature with different samples. The LGPO (Li Ge P O ) pellet. a) X-ray 3.2 0.2 0.8 4 electrical resistance was measured in- diffraction pattern of LGPO ceramic pellet plane. MgO is a very good electrical after sintering at 900 °C, 12 h. In green the insulator, therefore no contribution to the comparison with the ICSD reference 250066. The blue stars indicate peaks originating from resistance is expected from the substrate. secondary phases. b) Nyquist plots obtained at The Nyquist plots were fitted to the temperatures of 60 °C and 25 °C in Ar gas, from the same ceramic pellet with sputtered response of an equivalent circuit composed gold electrodes. In the inset: a scheme of the of a resistance (R) and a constant phase electrode configuration is shown. element (Q) in parallel (inset in Figure 2c). In Figure 2a-I the XRD analysis of a film In the case of in - plane resistance grown at high temperature is presented. measurements of thin films on substrates, Except minor peaks stemming from the bulk and grain boundary contributions secondary phases, all diffraction peaks can to the total resistance cannot be be ascribed to the orthorhombic structure distinguished, due to the stray capacitance of LGPO. The films appear to be induced by the substrate and experimental polycrystalline, as expected, due to the set up. Therefore, only the total absence of crystallographic matching conductivity can be deduced from the IS between films and substrates. plot. 6 the plot obtained by fitting the data to the equivalent circuit shown in the inset. According to Rutherford Back Scattering and Elastic Recoil Detection Analysis the composition of the thin films deposited at room temperature and 500 °C is Li Ge P O and 1.58±0.4 0.45 0.55 3.27±0.5 Li Ge P O , respectively. The 2.7±0.3 0.49 0.5 3.7±0.3 experimental uncertainty of these measurements is about 4 % for the Ge and P content. At both temperatures, a pronounced Li deficiency was observed, as expected. The two complementary ion beam-based compositional analyses also revealed the formation of carbonates. Lithium carbonate is a very common contaminant in Li-containing material when exposed to air. The presence of carbonates is particularly high in the film grown at room temperature, while for the film deposited at high temperature only a thin layer at the surface was detected. This observation suggests a much rougher and porous micromorphology of the LGPO films grown at room temperature, with respect to those grown at 500 °C. Roughness and porosity widen the surface area of the film, therefore allowing the formation of carbonates not only at the surface, as in the case of dense films. The formation of Figure 2 a) X-ray diffraction pattern (2θ/ω) of a thin carbonate was taken into account for the film deposited at 500 °C (I) and thin film calculation of the thin film compositions. deposited at room temperature (II). The Miller indexes of the diffraction peaks of the LGPO LGPO ionic conductivity film, gold electrodes and MgO substrate are reported. b) 2θ scan (ω kept constant at 1°) of The conductivity of the LGPO ceramic the LGPO thin film deposited at room pellets fabricated for this study is about temperature. c) Nyquist plots measured at -1 250 °C on LGPO thin films deposited at 1.2 µS cm at room temperature, with an different temperatures. The thin film deposited activation energy of about 0.53 eV. at 500 °C is 150 nm and the thin film deposited Figure 3 shows a representative at 25 °C is 300 nm thick. Solid lines represent conductivity Arrhenius plot for our 7 22 samples in comparison with available data measured for single crystals and sintered 10, 43 from the literature. pellets suggests that the secondary phases in the sample, shown by XRD, do not affect significantly the transport properties. In previous works, LGPO pellets with different Ge to P ratio (from 0.25 to 1.5) showed conductivities ranging from 1 to -1 10 µS cm at room temperature and the reported activation energies are all around 10, 43 0.53 eV. The present work confirms these findings. The correlation between chemical composition, morphology and conductivity seems to be not completely clear. It is however clear that the ionic conductivity Figure 3 Conductivity dependence on the near room temperature of LGPO with inverse temperature of of LGPO ceramic different compositions ranges between 1 pellets and single crystal. Our sample has the -1 and a few tens µS cm . nominal composition Li P Ge O . Data of 3.2 0.8 0.2 4 line 1 (Li P Ge O single crystal) are 3.34 0.66 0.34 4 Figure 4 compares the conductivity of one taken from Ref. , data of line 2 of the pellets fabricated for this work with (Li P Ge O pellet) from Ref. 3.2 0.8 0.2 4 the conductivity of LGPO thin films grown Our measurements, obtained with pellets at room temperature and at 500 °C. with a nominal composition of Li Ge P O (x = 0.8), are in very good 3.2 0.2 0.8 4 agreement with values reported in the literature for pellets with the same 10, 42 composition. A very similar conductivity was reported also for LGPO with x = 0.2, whereas the peak conductivity was found for x between 0.4 and 0.6 reaching values of about 10 µS cm 1 10 . Another study reports a room- temperature conductivity of about -1 35 µS cm for LGPO pellets with x ≈ 0.25. LGPO single crystal, with x = 0.66, showed a conductivity which is slightly lower but with the same activation -1 energy. The value of 1.8 µS cm was reached at 40 °C. The very similar value of conductivity 8 Figure 4 Conductivity dependence on the An advantage of LISICON thin films is inverse temperature of LGPO pellet that the same conductivity of the (Li Ge P O ) and thin films deposited at 25 3.2 0.2 0.8 4 stoichiometric compound can be reached and 500 °C. Composition of thin films from already for low lithium content. For RBS and ERDA analysis: Li Ge P O and example, in case of solid solutions of 1.58±0.4 0.45 0.55 3.27±0.5 Li Ge P O , respectively. The point 2.7±0.3 0.49 0.5 3.7±0.3 lithium silicate – lithium phosphate (pink sphere) at 25 °C was measured for the (Li SiO – Li PO ), the conductivity 4 4 3 4 thin film deposited at room temperature “as strongly depends on the lithium grown”. After the complete thermal cycle, the stoichiometry only up to a certain threshold conductivity did not show any significant change. The inset shows the electrode ( Li/(P+Si) ~ 1). Above this value, the configuration used for the different samples. conductivity becomes independent of the lithium content as the mobile lithium sites LGPO thin films show electrical properties are saturated. Assuming that LGPO that are very similar to those of the pellets, behaves similarly to silicates, we can independently of the deposition observe that in the film deposited at room temperature. The film grown at room temperature the measured lithium temperature shows an activation energy stoichiometry is lower compared to sample that is about 0.07 eV higher. deposited at 500 °C. However, in both This range of conductivity overlaps with cases the ratio between Li and P+Ge that reported in the literature for LiPON contents is well above 1. Therefore, it is amorphous thin films (highest value 2.3 likely that other factors, such as different -1 44-45 µS cm at room temperature). LiPON morphological features and/or the presence is, so far, the only solid-state Li0-ion of Li CO , are affecting the ionic transport. 2 3 conductor that can be used as thin film Considering the results of the XRD electrolyte for microbatteries. The analysis in Figure 2a, these measurements perovskite-type oxide LLTO and the confirm also for thin films what was garnet-type oxide LLZO, with room deduced from the comparison of temperature conductivity in the range of -1 conductivities of sintered pellets and single mS cm , require sintering temperatures crystals (Figure 3): the degree of well above 1000 °C for ceramic pellets and crystallinity (average grain size, extent of deposition temperatures well above 800 °C grain boundaries) has a relatively small for thin film growth. This promotes Li influence on the conducting properties. evaporation and the stabilization of the La Ti O and La Zr O non-conductive 2 2 7 2 2 7 First-principles molecular dynamics phases that are thermodynamically simulations for bulk defect-free LGPO favored. The conductivity of LLTO thin films reported so far in the literature is in A 100 - atom (1 × 1 × 3) supercell was the same range of LiPON and LGPO (but built starting from the single crystal unit require a much higher processing cell for which crystallographic data are temperature), while much lower available in the literature (ICSD conductivity has been reported for LLZO #250066 ). Stoichiometry of x = 0.66 was films. needed, in order to ensure a P to Ge ratio 22-23 equal to 1:2. This supercell, with a 9 formula of Li Ge P O , is reported in The absence of preferential conduction 40 4 8 48 Figure 5. pathways for the ionic transport is an important advantage, since channel- To understand Li-ion density and diffusion blocking effects can lower significantly the properties in an ideal (defect-free) LGPO diffusion and have detrimental crystal, we performed extended consequences on the charge/discharge (200 ps – 500 ps) FPMD simulations on properties, when the material is used as the above described periodic supercell in 35, 44 electrolyte in batteries. the NVT ensemble, between 300 °C and 1100 °C. The isosurfaces of the Li-ion probability density at different temperatures, displayed in Figure 6 a and 6 b, indicate that the ideal LGPO crystal offers a three-dimensional isotropic conduction pathway for Li ions. This is consistent with the small degree of anisotropy shown by the Li-ion MSD along x, y, z, at the temperature of 1400 K (Figure 6 c). It is noteworthy that LGPO does not present unidimensional channels for the lithium ion conduction. Channel-blocking effects to the ionic transport can be ruled out and the diffusion is isotropic. Figure 6 The Li - ion probability density at a) 1400 K (1126 °C) and b) 600 K (327 °C), for -3 -3 - three isovalues, 0.001 Å , 0.01 Å and 0.1 Å , as purple, salmon and yellow isosurfaces, respectively. The equilibrium positions of oxygen, phosphorus and germanium are shown as red, orange, and green spheres, respectively. c) x-, y- and z-resolved mean-squared displacement (MSD) for Li as a function of time at 1400 K (1126 °C), as red, green, and blue solid lines, respectively, showing again isotropic diffusion. In the legends, the resulting diagonal elements of the diffusion matrix (and the error originating from the individual Figure 5 Side and top view of the 100 - atom blocks, dashed lines) are reported. unit cell used for the simulations of defect-free Li-ion MSDs were used to calculate the LGPO crystal (from Ref. ). Li atoms are displayed in green, oxygen atoms in red, Ge ionic diffusion coefficients and the and P atoms are at the center of the light and conductivity dependence on the dark purple tetrahedra, respectively. temperature. 10 Figure 7 compares the computational values of Li-ion conductivity in LGPO with the experimental values previously described (high-temperature deposited thin film, x=0.5) obtained in this study. The non-quantitative agreement between FPMD-calculated (0.37±0.04 eV) and IS- measured (0.51±0.01 eV) activation barriers is not new for ionic conductivities 46-48 in solid-state electrolytes. Possible reasons are the different time and length scales accessed through simulations (atomic level) and experiments (entire sample), and the absence of defects in the simulated ideal crystal, so that values calculated by FPMD may be considered as the upper limit for the conductivity in Figure 7) Li-ion conductivity Arrhenius plot LGPO (Li Ge P O ). Besides, going 3.33 0.33 0.66 4 for the simulated ideal LGPO (x=0.67) crystal (FPMD), compared with experimental data beyond the Li-ion tracer diffusion (i.e. (high-temperature deposited thin film, x=0.5). calculating the charge diffusion coefficient The underestimation of the activation energy [ ]) would be desirable. However, our by simulations compared to the experimental attempts to calculate the Haven ratio gave values was previously ascribed to the different length scales analyzed by the techniques and rise to charge diffusion coefficients with difficulty to take 1D and 2D defects into large statistical errors, that we don’t report account. here, due to LGPO being a weak ionic conductor, so that longer trajectories would Conclusions be needed for such calculations. It is In this work we synthesized and worthwhile to recall here that some of us characterized pellets and thin films of Li 4- have shown the Haven ratio not to play a Ge P O (solid solution between γ- x 1-x x 4 significant role on the activation energy for Li PO and Li GeO ) belonging to the 3 4 4 4 the diffusion for the similar system ionic conductors in the LISICON family. LGPS. Pellets with nominal composition In turn, comparison with the experimental Li Ge P O and a relative density of 3.2 0.2 0.8 4 values, obtained in this work and reported 85% show a conductivity and activation in the literature, suggests that the transport energy in agreement with previously properties are affected by the composition published data. and presence of large-scale defects. Large- scale defects affect the conductivity Thin films were deposited via PLD at measured by IS, while their effect is not different temperatures. The compositional included in the conductivity calculated by transfer from target to substrate during FPMD. laser ablation is not stoichiometric, resulting in different ratios between Ge and 11 previously for similar systems that the P and the lithium content. Samples Haven ration does not significantly affect deposited at low temperature are deficient the activation barrier. in lithium and show contaminations in the form of lithium carbonate. In addition, thin In conclusion, LGPO exhibits a number of films deposited at low temperature are not very interesting features that are potentially completely amorphous as reported important for a technological application as previously for similar materials, but appear electrolyte in thin film batteries and to possess a nanocrystalline morphology. fundamental research. In particular, the fabrication of thin films with the desired Compared to the deposition at high characteristics is particularly easy to temperature, the deposition at low achieve (also at room temperature). temperature reduces the conductivity of the films by approximately a factor of 4 at It was previously reported that LGPO (with room temperature. This is most probably formula Li Ge P O ) is stable 3.33 0.33 0.66 4 due to the higher roughness and porosity, between -0.5 and 7 V vs Li/Li , which which favors the formation of carbonates make it suitable to be used also with high throughout the film. Lithium carbonate is voltage electrodes. present as a contaminant only at the surface All these characteristics make LGPO a of the dense crystalline films without potentially interesting electrolyte material strongly affecting the conductivity after for all-solid-state all-oxide microbatteries. exposing the sample to air. . AUTHOR INFORMATION From the characterization of pellet and thin films, it appears that the conductivity is Corresponding Author almost independent of the microstructure and lithium content, if this is above the * elisa.gilardi@psi.ch threshold of saturation of the mobile lithium sites. * giuliana.materzanini@epfl.ch Conductivities at different temperatures calculated by FPMD and experimental results are then compared. A clear, but Author Contributions expected difference exists between the short length scale conductivity calculated The manuscript was written through by FPMD and the conductivity measured contributions of all authors. All authors on the whole sample by IS. This suggests that there is a blocking effect of large-scale have given approval to the final version of defects that is not detected by the experimental characterization. Effects of the manuscript. the Li-ion collective diffusion (Haven ratio) were not taken into account here, due Funding Sources to the modest conductivity of this material, that would have required longer simulation This research was supported by the times. At the same time, it was reported NCCR MARVEL, funded by the Swiss 12 National Science Foundation and by DAIMLER AG, Stuttgart, Andreas Hintennach. Notes EG, TL and DP would like to dedicate this work to their colleague, Andreas Hintennach, who unfortunately, died prematurely this year. He will be sorely missed and his personal, scientific and technical contributions will be irreplaceable. 13 (9) Hodge, I. M.; Ingram, M. D.; West, A. R. References Ionic-Conductivity of Li SiO , Li GeO , and 4 4 4 4 their Solid-Solutions. J. Am. Ceram. Soc. (1) Oudenhoven, J. F. M.; Baggetto, L.; 1976, 59 (7-8), 360-366, DOI: Notten, P. H. L. All-Solid-State Lithium-Ion 10.1111/j.1151-2916.1976.tb10985.x. Microbatteries: A Review of Various Three- (10) Muy, S.; Bachman, J. C.; Chang, H. H.; Dimensional Concepts. Adv. Energy Mater. Giordano, L.; Maglia, F.; Lupart, S.; Lamp, P.; 2011, 1 (1), 10-33, DOI: Zeier, W. G.; Shao-Horn, Y. Lithium 10.1002/aenm.201000002. conductivity and Meyer-Neldel rule in (2) Sun, C. W.; Liu, J.; Gong, Y. D.; Wilkinson, Li PO -Li VO -Li GeO lithium superionic 3 4 3 4 4 4 D. P.; Zhang, J. J. Recent advances in all-solid- conductors. Chem. Mat. 2018, 30 (16), 5573- state rechargeable lithium batteries. Nano 5582, DOI: Energy 2017, 33, 363-386, DOI: 10.1021/acs.chemmater.8b01504. 10.1016/j.nanoen.2017.01.028. (11) Deng, Y.; Eames, C.; Fleutot, B.; David, (3) Manthiram, A.; Yu, X.; Wang, S. Lithium R.; Chotard, J. N.; Suard, E.; Masquelier, C.; battery chemistries enabled by solid-state Islam, M. S. Enhancing the Lithium Ion electrolytes. Nat. Rev. Mater. 2017, 2 Conductivity in Lithium Superionic (16103), 1-16. Conductor (LISICON) Solid Electrolytes (4) Zhu, Y. Z.; He, X. F.; Mo, Y. F. Origin of through a Mixed Polyanion Effect. ACS Appl. outstanding stability in the lithium solid Mater. Interfaces 2017, 9 (8), 7050-7058, electrolyte materials: insights from DOI: 10.1021/acsami.6b14402. thermodynamic analyses based on first- (12) Zhao, G. W.; Suzuki, K.; Yonemura, M.; principles calculations. ACS Applied Hirayama, M.; Kanno, R. Enhancing Fast Materials & Interfaces 2015, 7 (42), 23685- Lithium Ion Conduction in Li GeO -Li PO 4 4 3 4 23693, DOI: 10.1021/acsami.5b07517. Solid Electrolytes. ACS Appl. Energ. Mater. (5) Krauskopf, T.; Hartmann, H.; Zeier, W. G.; 2019, 2 (9), 6608-6615, DOI: Janek, J. Toward a Fundamental 10.1021/acsaem.9b01152. Understanding of the Lithium Metal Anode (13) Famprikis, T.; Galipaud, J.; Clernens, O.; in Solid-State Batteries-An Electrochemo- Pecquenard, B.; Le Cras, F. Composition Mechanical Study on the Garnet-Type Solid Dependence of Ionic Conductivity in Electrolyte Li6.25Al0.25La3Zr3O12. Acs LiSiPO(N) Thin-Film Electrolytes for Solid- Applied Materials & Interfaces 2019, 11 (15), State Batteries. ACS Appl. Energ. Mater. 14463-14477, DOI: 2019, 2 (7), 4782-4791, DOI: 10.1021/acsami.9b02537. 10.1021/acsaem.9b00415. (6) Inaguma, Y.; Chen, L. Q.; Itoh, M.; (14) Su, Y. R.; Falgenhauer, J.; Leichtweiss, Nakamura, T.; Uchida, T.; Ikuta, H.; T.; Geiss, M.; Lupo, C.; Polity, A.; Zhou, S. Q.; Wakihara, M. High ionic-conductivity in Obel, J.; Schlettwein, D.; Janek, J.; Meyer, B. K. lithium lanthanum titanate. Solid State Electrochemical properties and optical Commun. 1993, 86 (10), 689-693, DOI: transmission of high Li+ conducting 10.1016/0038-1098(93)90841-a. LiSiPON electrolyte films. Phys. Status Solidi (7) Wu, J. F.; Chen, E. Y.; Yu, Y.; Liu, L.; Wu, Y.; B-Basic Solid State Phys. 2017, 254 (2), 8, Pang, W. K.; Peterson, V. K.; Guo, X. Gallium- DOI: 10.1002/pssb.201600088. Doped Li La Zr O Garnet-Type 7 3 2 12 (15) Kanehori, K.; Matsumoto, K.; Miyauchi, Electrolytes with High Lithium-Ion K.; Kudo, T. Thin-Film Solid Electrolyte and Conductivity. Acs Applied Materials & its application to secondary lithium cell. Interfaces 2017, 9 (2), 1542-1552, DOI: Solid State Ion. 1983, 9-10 (DEC), 1445- 10.1021/acsami.6b13902. 1448, DOI: 10.1016/0167-2738(83)90192- (8) Rodger, A. R.; Kuwano, J.; West, A. R. Li Ion Conducting Gamma-Solid Solutions in (16) Sakurai, Y.; Sakuda, A.; Hayashi, A.; the Systems Li XO -Li YO - X=Si,Ge,Ti- 4 4 3 4 Tatsumisago, M. Preparation of amorphous Y=P,As,V-Li XO -LiZO -Z=Al,Ga,Cr and 4 4 2 Li SiO -Li PO thin films by pulsed laser 4 4 3 4 Li GeO -Li CaGeO . Solid State Ion. 1985, 15 4 4 2 4 deposition for all-solid-state lithium (3), 185-198, DOI: 10.1016/0167- secondary batteries. Solid State Ion. 2011, 2738(85)90002-5. 182 (1), 59-63, DOI: 14 10.1016/j.ssi.2010.12.001. 10.1103/PhysRev.140.A1133. (17) Kim, D. H.; Imashuku, S.; Wang, L.; Shao- (27) Payne, M. C.; Teter, M. P.; Allan, D. C.; Horn, Y.; Ross, C. A. Li loss during the growth Arias, T. A.; Joannopoulos, J. D. Iterative of (Li,La)TiO thin films by pulsed laser minimization techniques for Ab-Initio total- deposition. J. Cryst. Growth 2013, 372, 9-14, energy calculations - molecular-dynamics DOI: 10.1016/j.jcrysgro.2013.02.028. and conjugate gradients. Rev. Mod. Phys. (18) Pfenninger, R.; Struzik, M.; Garbayo, I.; 1992, 64 (4), 1045-1097, DOI: Stilp, E.; Rupp, J. L. M. A low ride on 10.1103/RevModPhys.64.1045. processing temperature for fast lithium (28) Giannozzi, P.; Baroni, S.; Bonini, N.; conduction in garnet solid-state battery Calandra, M.; Car, R.; Cavazzoni, C.; Ceresoli, films. Nature Energy 2019, 4 (6), 475-483, D.; Chiarotti, G. L.; Cococcioni, M.; Dabo, I.; DOI: 10.1038/s41560-019-0384-4. Dal Corso, A.; de Gironcoli, S.; Fabris, S.; (19) Aguesse, F.; Roddatis, V.; Roqueta, J.; Fratesi, G.; Gebauer, R.; Gerstmann, U.; Garcia, P.; Pergolesi, D.; Santiso, J.; Kilner, J. Gougoussis, C.; Kokalj, A.; Lazzeri, M.; A. Microstructure and ionic conductivity of Martin-Samos, L.; Marzari, N.; Mauri, F.; LLTO thin films: Influence of different Mazzarello, R.; Paolini, S.; Pasquarello, A.; substrates and excess lithium in the target. Paulatto, L.; Sbraccia, C.; Scandolo, S.; Solid State Ion. 2015, 272, 1-8, DOI: Sclauzero, G.; Seitsonen, A. P.; Smogunov, A.; 10.1016/j.ssi.2014.12.005. Umari, P.; Wentzcovitch, R. M. QUANTUM (20) Lobe, S.; Dellen, C.; Windmuller, A.; ESPRESSO: a modular and open-source Tsai, C. L.; Vondahlen, F.; Uhlenbruck, S.; software project for quantum simulations of Guillon, O. Challenges regarding thin film materials. J. Phys.-Condes. Matter 2009, 21 deposition of garnet electrolytes for all- (39), 19, DOI: 10.1088/0953- solid-state lithium batteries with high 8984/21/39/395502. energy density. Ionics 2018, 24 (8), 2199- (29) Prandini, G.; Marrazzo, A.; Castelli, I. E.; 2208, DOI: 10.1007/s11581-018-2594-3. Mounet, N.; Marzari, N. Precision and (21) Julien C.; A., M. Pulsed Laser Deposited efficiency in solid-state pseudopotential Films for Microbatteries Coatings 2019, 9 calculations. npj Comput. Mater. 2018, 4, 13, (6), 386. DOI: 10.1038/s41524-018-0127-2. (22) Ivanov-Shitz, A. K.; Kireev, V. V. Growth (30) Garrity, K. F.; Bennett, J. W.; Rabe, K. M.; and ionic conductivity of Li P Ge O Vanderbilt, D. Pseudopotentials for high- 3+x 1-x x 4 (x=0.34) single crystals. Crystallogr. Rep. throughput DFT calculations. Comput. 2003, 48 (1), 112-115, DOI: Mater. Sci. 2014, 81, 446-452, DOI: 10.1134/1.1541752. 10.1016/j.commatsci.2013.08.053. (23) Rabadanov, M. K.; Pietraszko, A.; (31) Dal Corso, A. Pseudopotentials periodic Kireev, V. V.; Ivanov-Schitz, A. K.; Simonov, table: From H to Pu. Comput. Mater. Sci. V. I. Atomic structure and mechanism of 2014, 95, 337-350, DOI: ionic conductivity of Li Ge P O single 10.1016/j.commatsci.2014.07.043. 3.31 0.31 0.69 4 crystals. Crystallogr. Rep. 2003, 48 (5), 744- (32) Perdew, J. P.; Burke, K.; Ernzerhof, M. 749, DOI: 10.1134/1.1612594. Generalized gradient approximation made (24) Doolittle, L. R. A Semiautomatic simple. Phys. Rev. Lett. 1996, 77 (18), 3865- Algorithm for Rutherford Backscattering 3868, DOI: 10.1103/PhysRevLett.77.3865. Analysis. Nucl. Instrum. Methods Phys. Res. (33) Martyna, G. J.; Klein, M. L.; Tuckerman, Sect. B-Beam Interact. Mater. Atoms 1986, M. Nosee-Hoover chains - the canonical 15 (1-6), 227-231, DOI: 10.1016/0168- ensemble via continuous dynamics. J. Chem. 583x(86)90291-0. Phys. 1992, 97 (4), 2635-2643, DOI: (25) Hohenberg, P.; Kohn, W. 10.1063/1.463940. Inhomogeneous electron gas. Phys. Rev. B (34) Marcolongo, A.; Marzari, N. Ionic 1964, 136 (3B), B864-+, DOI: correlations and failure of Nernst-Einstein 10.1103/PhysRev.136.B864. relation in solid-state electrolytes. Phys. Rev. (26) Kohn, W.; Sham, L. J. Self-consistent Mater. 2017, 1 (2), 4, DOI: equations including exchange and 10.1103/PhysRevMaterials.1.025402. correlation effects. Physical Review 1965, (35) Kahle, L.; Musaelian, A.; Marzari, N.; 140 (4A), 1133-&, DOI: Kozinsky, B. Unsupervised landmark 15 analysis for jump detection in molecular Lithium phosphorus oxynitride. J. dynamics simulations. Phys. Rev. Mater. Electrochem. Soc. 1997, 144 (2), 524-532, 2019, 3 (5), 18, DOI: DOI: 10.1149/1.1837443. 10.1103/PhysRevMaterials.3.055404. (46) Adams, S.; Rao, R. P. Structural (36) Frenkel, D.; Smit, B. Understanding requirements for fast lithium ion migration molecular simulation: from algorithms to in Li10GeP2S12. J. Mater. Chem. (UK) 2012, 22 applications 2001; Vol. 1. (16), 7687-7691, DOI: (37) Grasselli, F.; Baroni, S. Topological 10.1039/c2jm16688g. quantization and gauge invariance of charge (47) Mo, Y. F.; Ong, S. P.; Ceder, G. First transport in liquid insulators. Nat. Phys. Principles Study of the Li GeP S Lithium 10 2 12 2019, 15 (9), 967-972, DOI: Super Ionic Conductor Material. Chem. Mat. 10.1038/s41567-019-0562-0. 2012, 24 (1), 15-17, DOI: (38) Saccoccio, M.; Yu, J.; Lu, Z. H.; Kwok, S. 10.1021/cm203303y. C. T.; Wang, J.; Yeung, K. K.; Yuen, M. M. F.; (48) Deng, Y.; Eames, C.; Chotard, J. N.; Ciucci, F. Low temperature pulsed laser Lalere, F.; Seznec, V.; Emge, S.; Pecher, O.; deposition of garnet Li La Zr Ta O Grey, C. P.; Masquelier, C.; Islam, M. S. 6.4 3 1.4 0.6 12 films as all solid-state lithium battery Structural and mechanistic insights into fast electrolytes. J. Power Sources 2017, 365, 43- lithium-ion conduction in Li SiO -Li PO 4 4 3 4 52, DOI: 10.1016/j.jpowsour.2017.08.020. solid electrolytes. J. Am. Chem. Soc. 2015, (39) Rawlence, M.; Garbayo, I.; Buecheler, S.; 137 (28), 9136-9145, DOI: Rupp, J. L. M. On the chemical stability of 10.1021/jacs.5b04444. post-lithiated garnet Al-stabilized (49) France-Lanord, A.; Grossman, J. C. Li La Zr O solid state electrolyte thin Correlations from Ion Pairing and the 7 3 2 12 films. Nanoscale 2016, 8 (31), 14746- Nernst-Einstein Equation. Phys. Rev. Lett. 14753, DOI: 10.1039/c6nr04162k. 2019, 122 (13), 6, DOI: (40) Ojeda-G-P, A.; Dobeli, M.; Lippert, T. 10.1103/PhysRevLett.122.136001. Influence of plume properties on thin film composition in pulsed laser deposition. Adv. Mater. Interfaces 2018, 5 (18), 16, DOI: 10.1002/admi.201701062. (41) Gerstl, M.; Navickas, E.; Friedbacher, G.; Kubel, F.; Ahrens, M.; Fleig, J. The separation of grain and grain boundary impedance in thin yttria stabilized zirconia (YSZ) layers. Solid State Ion. 2011, 185 (1), 32-41, DOI: 10.1016/j.ssi.2011.01.008. (42) Rodger, A. R.; Kuwano, J.; West, A. R. Li+ ion conducting gamma-solid solutions in the systems Li XO -Li YO - X=Si,Ge,Ti-Y=P,As,V 4 4 3 4 - Li4XO -LiZO -Z=Al,Ga,Cr and Li GeO - 4 2 4 4 Li CaGeO . Solid State Ion. 1985, 15 (3), 185- 2 4 198, DOI: 10.1016/0167-2738(85)90002-5. (43) Song, S. F.; Dong, Z. C.; Deng, F.; Hu, N. Lithium superionic conductors Li MP O 10 2 12 (M = Ge, Si). Funct. Mater. Lett. 2018, 11 (2), 4, DOI: 10.1142/s179360471850039x. (44) Kamaya, N.; Homma, K.; Yamakawa, Y.; Hirayama, M.; Kanno, R.; Yonemura, M.; Kamiyama, T.; Kato, Y.; Hama, S.; Kawamoto, K.; Mitsui, A. A lithium superionic conductor. Nat. Mater. 2011, 10 (9), 682-686, DOI: 10.1038/nmat3066. (45) Yu, X. H.; Bates, J. B.; Jellison, G. E.; Hart, F. X. A stable thin-film lithium electrolyte: http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Physics arXiv (Cornell University)

Li4-xGe1-xPxO4 a potential solid-state electrolyte for all-oxide microbatteries

Loading next page...
 
/lp/arxiv-cornell-university/li4-xge1-xpxo4-a-potential-solid-state-electrolyte-for-all-oxide-Sy3Uy4JkBZ
ISSN
2574-0962
eISSN
ARCH-3341
DOI
10.1021/acsaem.0c01601
Publisher site
See Article on Publisher Site

Abstract

Solid-state electrolytes for Li-ion batteries are attracting growing interest as they allow building safer batteries, also using lithium metal anodes. Here we studied a compound in the lithium superionic conductor (LISICON) family, i.e. Li Ge P O (LGPO). Thin films were 4-x 1-x x 4 deposited via pulsed laser deposition and their electrical properties were compared with ceramic pellets. A detailed characterization of the microstructure shows that thin films can be deposited fully crystalline at higher temperatures but also partially amorphous at room temperature. The conductivity is not strongly influenced by the presence of grain boundaries, exposure to air or lithium deficiencies. First-principles molecular dynamics simulations were employed to calculate the lithium ion diffusion profile and the conductivity at various temperatures of the ideal LGPO crystal. Simulations gives the upper limit of conductivity for -2 -1 a defect free crystal, which is in the range of 10 S cm at 300 °C The ease of thin film -1 fabrication, the room-temperature Li-ion conductivity in the range of a few µS cm make LGPO very appealing electrolyte material for thin film all–solid-state all-oxide microbatteries. Introduction energy storage. Conventional batteries use liquid organic solvents as the electrolyte, Lithium ion batteries are nowadays among which are corrosive and flammable, the most widely used technology for 1 therefore posing important safety issues. exploiting the “mixed polyanion effect”: The risk of leaks precludes the potential the introduction of different cations causes use of these batteries for biomedical a disordered redistribution of the anionic implants. A liquid electrolyte imposes also tetrahedra, which facilitates the lithium ion 11-12 severe restrictions to the possible migration. miniaturization of the battery, which is Compounds in the LISICON family are needed for on-chip power sources for also particularly interesting for the growth portable/wearable applications. Also, Li of thin films, as they can be deposited metal dendrites can easily grow into a 13-15 easily by sputtering and pulsed laser liquid electrolyte, short circuiting the deposition (PLD). Thin films of 1-2 electrodes. LISICON materials reported in the The use of solid-state electrolytes (SSE) literature are generally amorphous, which would solve or mitigate these issues, as further enhances the lithium ion SSE are not flammable, cannot leak, and conductivity. can in principle be grown as nanometric On the contrary, the growth of thin films of thin films for battery miniaturization. other, much more conductive, oxides, such Many SSE are stable within a large as LLZO and LLTO, turns out to be 3-4 electrochemical voltage window. 17-20 extremely difficult. Different reports Finally, SSE hinder the formation of 15- can be found in the literature on silicates, dendrites, although when metallic lithium 16, 21 but nothing has been published to the anode is employed, the problem cannot be 5 best of our knowledge on germanate completely avoided. materials. Inorganic oxides SSE are less conductive Here we report the deposition via PLD of at room temperature than sulfides or thin films of lithium germanium halides, but generally more stable in phosphate, with the general formula Li 4- contact with electrodes as well as when Ge P O (LGPO), a solid solution of γ- x 1-x x 4 exposed to air. 22-23 Li PO and Li GeO . Different 3 4 4 4 Different oxides show promising deposition temperatures led to films with properties for possible use as SSE in different crystalline quality and chemical batteries, for instance Li La Zr O 7 3 2 12 composition. These features were then 6-7 (LLZO) and Li La TiO (LLTO). 3x 2/3-x 3 correlated to the electrical properties. These materials offer lithium ion First-principles molecular dynamics conductivities in the order of a few mS cm at room temperature. Also, ionic simulations were then employed to conductors in the LISICON family, with a calculate the Li-ion diffusion profile and composition of Li M Y O (M = Ge, diffusivity at different temperatures in an 4±x 1−x x 4 Si; Y = P, Al, Zn, Ge, Ga, Sb) are well ideal (defect-free) LGPO crystal. known solid-state electrolytes, with a Methods conductivity at room temperature in the range between a few and a few tens of Powder synthesis and characterization -1 8-10 µS cm . Recent works highlighted that Li P Ge O (LGPO) powders were 3.2 0.80 0.20 4 the conductivity can be further improved 2 synthesized via the solid-state synthesis temperature at the same oxygen method using lithium carbonate (Li CO ), background pressure. 2 3 germanium oxide (GeO ) and lithium Thin film characterization phosphate (Li PO ). Lithium carbonate and 3 4 The crystal structure of the films was germanium oxide were mixed in characterized by X - ray diffraction (out of stoichiometric amount and grounded using plane /2 scan) with a Seifert an agate mortar, and the resulting powders diffractometer with monochromatic were pressed uniaxially with 0.5 GPa. The Cu - Kα1 radiation. The thin films pressed powders were sintered at 800 °C deposited at low temperature were for 8 hours. Resulting Li GeO and Li PO 4 4 3 4 characterized also by 2 scan. While  was powders were then mixed and pressed fixed at 1°, 2 was scanned between 10° uniaxially with 0.5 GPa. The powders were and 65°. then placed in a tubular furnace with constant oxygen flux at 900 °C for For the in-plane (along the direction of the 12 hours. The resulting pellets were 13 mm substrate surface) electrical in diameter with a thickness of 2 mm. The characterization two strip-shaped gold same procedure was repeated for sintering electrodes, 100 nm thick, were sputtered ceramic LGPO pellets used for on the surface of the films. conductivity measurements and as target Electrical characterization was performed for thin film deposition. The crystal by impedance spectroscopy (IS) in Ar structure of the synthesized LGPO flow, between 2 MHz and 1 Hz, applying powders were characterized by X-ray an excitation of 1 V. Ag paste and Au wires diffraction (Siemens D500 diffractometer were used to connect the electrodes to the with Cu-Kα radiation). measurement cell. A Solatron 1260 gain- Thin film deposition phase analyzer and the software Zplot were Thin films of LGPO were deposited via used for the impedance measurement. The software Zview was used to fit the complex pulsed laser deposition (PLD) with a KrF impedance plane plots. Thin films excimer laser, wavelength 248 nm, laser deposited at high temperature were fluence of 2.1 J/cm , repetition rate of equilibrated for 2 hours at 430 °C. The 10 Hz. To compensate the expected Li loss temperature was then decreased in steps of (a problem often encountered during the 20 °C and the samples were equilibrated growth of Li-containing thin films) a ceramic target with 10 mol% excess Li (as for 30 minutes at each temperature step Li O)was used. The distance between before acquiring a complex impedance plot. Thin films deposited at room target and substrate during deposition was temperature were equilibrated 24 hours at 60 mm. Thin films were deposited on 25 °C and data was acquired while single crystal (100) - oriented MgO increasing the temperature up to 480 °C. substrate. Silver paste was used to provide The conductivity was also measured the thermal contact between the substrate decreasing the temperature following the and the heating stage in the vacuum chamber. The oxygen pressure was set to same procedure as described above. 0.01 mbar during the deposition. Films Rutherford Back Scattering was performed were deposited at 500 °C and at room with a 2 MeV 4He beam and a silicon PIN 3 diode detector under 168°. Data were ensemble, temperature was controlled via a analyzed by the RUMP code. Elastic Nosé-Hoover thermostat for each atomic 33-34 Recoil Detection Analysis with a 13 MeV species. 127I primary ion beam was primarily used From the NVT trajectories, the Li-ion to determine the lithium-to-oxygen ratio. positions were extracted and the Rt  Recoiling ions were detected by the i combination of a time-of-flight Li-ion probability density was nr   Li spectrometer with a gas ionization calculated as: chamber. The acquired spectral data were analyzed by a custom software. (1) rR t   Li 3  2 n r  2 e     Li  First-Principles molecular dynamics To examine Li-ion diffusivity in LGPO on The normal distribution was used to the atomic scale in a perfect crystal with a account for finite statistics, setting the composition of Li Ge P O , 3.33 0.33 0.67 4 deviation to σ = 0.3 Å, and the summation extended first-principles molecular -1 was performed on a grid of 0.1 Å . dynamics (FPMD) simulations were performed. The Car-Parrinello (CPMD) In addition, we analyzed the trajectory scheme was used, based on Kohn-Sham using the SITATOR package for an 25-26 density-functional theory (DFT) in the unsupervised analysis of the diffusive plane-wave pseudopotential formalism, pathways in the system. as implemented in the cp code of the Li The tracer diffusion matrix D was Quantum ESPRESSO distribution computed from the NVT trajectories via (http://www.quantum-espresso.org/). A the mean square displacement (MSD) of plane-waves cutoff of 50 Ry for the lithium, according to the Einstein relation: wavefunctions and 400 Ry for the electron density and ultrasoft pseudopotentials from (2) Standard Solid State Pseudopotential N Li Li l l l l D  lim R t   R tR t   R t  29 30  ij i i j j (SSSP) Efficiency library 1.0 (GBRV  2 N l Li for Li and O, PSLib for Ge and P) were  lim MSD  , ij employed, with Brillouin-zone integration  2 at the Gamma point. The exchange- where is the i-component ( ) correlation functional was PBE. Rt  i  x,, y z of the Li-ion position at time t, and l th CPMD simulations, with a time step of the angular brackets define a time 4 a.u. and an electronic mass of 500 a.u., were performed for 200 – 500 ps in the average over the starting times t, that, fixed simulation (1x1x3) supercell defined following the ergodic hypothesis, we by the experimental crystallographic data corresponding to an ensemble assume at six temperatures ranging between average. We computed this average, and its 600 K and 1400 K (600, 720, 900, 1000, statistical error, by means of a block 1100, 1400 K). To sample the NVT analysis. The tracer diffusion coefficient 4 Li Li was calculated from the matrix as The electrical conductivity was measured D D tr by impedance spectroscopy (IS) with gold Li 1/3 Tr(D ), or directly from the formula (Li - ion blocking) electrodes. Two (3) representative Nyquist plots, used to calculate the total resistance, are reported Li 2 Li l l D  lim R t  R t      in Figure 1b. With our samples and tr   6 N Li t selected experimental conditions, it was not possible to clearly distinguish the grain  lim MSD  .    6 boundary and grain interior contributions to the conductivity. Therefore, only the From the tracer diffusion coefficient at total resistance could be reliably measured. each temperature, the ionic conductivity was extracted via the Nernst-Einstein LGPO thin films equation: LGPO thin films were grown by pulsed (4) laser deposition (PLD) on 2 MgO (100) - oriented single crystal Ne Li Li  T  D () T   tr substrates. A ceramic pellet fabricated in Vk T our laboratory was used as target for the ablation process. where is the Li-ions density and NV / e Li In the case of complex multi-element is the elementary charge. We assumed an materials, containing elements with very average Born effective charge for Li of different atomic masses, the target-to-film +1. compositional transfer can be a severe is the Boltzmann constant and the 38-39 issue. Many factors, other than the absolute temperature. target composition, can influence the film elemental content. Due to the interaction of Results and Discussion the ablated species with the gaseous environment during the time of flight LGPO ceramic pellets towards the substrate, pulsed laser LGPO ceramic pellets, with a nominal deposited films are typically enriched with composition of Li Ge P O were 3.2 0.2 0.8 4 the heavier elements at the expense of the prepared by solid - state synthesis, then lighters. This is why it is very difficult to pressed and sintered. The density of the grow films with the desired Li content. To -3 pellets was 2.20 g cm , corresponding to partially compensate for the Li loss, 10 % 85 % of the theoretical density. The XRD mol excess Li was added to the target as analysis and comparison with the ICSD Li O. Films with thickness between reference (Figure 1a) indicates the 150 nm and 300 nm were deposited in an formation of the expected LGPO oxygen background pressure of 0.01 mbar. orthorhombic structure (space group The samples were prepared at room P n m a). Three minor diffraction peaks temperature and at 500 °C to probe the corresponding to lithium carbonate and effects of temperature on composition, lithium phosphate were also detected. crystallinity, and transport properties. 5 Instead, no diffraction peaks can be seen in the plot of the 2θ/ω scan of the sample grown at room temperature (Figure 2a-II). For this sample, more can be learned from the plot of the 2θ scan shown in Figure 2b. This XRD pattern can be explained assuming a nanocrystalline film, which could also be partially amorphous. However, the presence of diffraction peaks clearly ascribable to the LGPO structure indicates a certain degree of crystallinity also for the films deposited at room temperature. This is noteworthy, because in the literature LISICON thin films deposited via PLD were reported to be amorphous, on the basis of the absence of diffraction peaks in the 2θ/ω scan. Our results show that this is not always the case. An accurate analysis of the crystallinity is important, as complete film amorphization is known to improve the conductivity in LISICON materials. Figure 2c shows representative complex impedance plane plots acquired at the same Figure 1 XRD and IS characterization of temperature with different samples. The LGPO (Li Ge P O ) pellet. a) X-ray 3.2 0.2 0.8 4 electrical resistance was measured in- diffraction pattern of LGPO ceramic pellet plane. MgO is a very good electrical after sintering at 900 °C, 12 h. In green the insulator, therefore no contribution to the comparison with the ICSD reference 250066. The blue stars indicate peaks originating from resistance is expected from the substrate. secondary phases. b) Nyquist plots obtained at The Nyquist plots were fitted to the temperatures of 60 °C and 25 °C in Ar gas, from the same ceramic pellet with sputtered response of an equivalent circuit composed gold electrodes. In the inset: a scheme of the of a resistance (R) and a constant phase electrode configuration is shown. element (Q) in parallel (inset in Figure 2c). In Figure 2a-I the XRD analysis of a film In the case of in - plane resistance grown at high temperature is presented. measurements of thin films on substrates, Except minor peaks stemming from the bulk and grain boundary contributions secondary phases, all diffraction peaks can to the total resistance cannot be be ascribed to the orthorhombic structure distinguished, due to the stray capacitance of LGPO. The films appear to be induced by the substrate and experimental polycrystalline, as expected, due to the set up. Therefore, only the total absence of crystallographic matching conductivity can be deduced from the IS between films and substrates. plot. 6 the plot obtained by fitting the data to the equivalent circuit shown in the inset. According to Rutherford Back Scattering and Elastic Recoil Detection Analysis the composition of the thin films deposited at room temperature and 500 °C is Li Ge P O and 1.58±0.4 0.45 0.55 3.27±0.5 Li Ge P O , respectively. The 2.7±0.3 0.49 0.5 3.7±0.3 experimental uncertainty of these measurements is about 4 % for the Ge and P content. At both temperatures, a pronounced Li deficiency was observed, as expected. The two complementary ion beam-based compositional analyses also revealed the formation of carbonates. Lithium carbonate is a very common contaminant in Li-containing material when exposed to air. The presence of carbonates is particularly high in the film grown at room temperature, while for the film deposited at high temperature only a thin layer at the surface was detected. This observation suggests a much rougher and porous micromorphology of the LGPO films grown at room temperature, with respect to those grown at 500 °C. Roughness and porosity widen the surface area of the film, therefore allowing the formation of carbonates not only at the surface, as in the case of dense films. The formation of Figure 2 a) X-ray diffraction pattern (2θ/ω) of a thin carbonate was taken into account for the film deposited at 500 °C (I) and thin film calculation of the thin film compositions. deposited at room temperature (II). The Miller indexes of the diffraction peaks of the LGPO LGPO ionic conductivity film, gold electrodes and MgO substrate are reported. b) 2θ scan (ω kept constant at 1°) of The conductivity of the LGPO ceramic the LGPO thin film deposited at room pellets fabricated for this study is about temperature. c) Nyquist plots measured at -1 250 °C on LGPO thin films deposited at 1.2 µS cm at room temperature, with an different temperatures. The thin film deposited activation energy of about 0.53 eV. at 500 °C is 150 nm and the thin film deposited Figure 3 shows a representative at 25 °C is 300 nm thick. Solid lines represent conductivity Arrhenius plot for our 7 22 samples in comparison with available data measured for single crystals and sintered 10, 43 from the literature. pellets suggests that the secondary phases in the sample, shown by XRD, do not affect significantly the transport properties. In previous works, LGPO pellets with different Ge to P ratio (from 0.25 to 1.5) showed conductivities ranging from 1 to -1 10 µS cm at room temperature and the reported activation energies are all around 10, 43 0.53 eV. The present work confirms these findings. The correlation between chemical composition, morphology and conductivity seems to be not completely clear. It is however clear that the ionic conductivity Figure 3 Conductivity dependence on the near room temperature of LGPO with inverse temperature of of LGPO ceramic different compositions ranges between 1 pellets and single crystal. Our sample has the -1 and a few tens µS cm . nominal composition Li P Ge O . Data of 3.2 0.8 0.2 4 line 1 (Li P Ge O single crystal) are 3.34 0.66 0.34 4 Figure 4 compares the conductivity of one taken from Ref. , data of line 2 of the pellets fabricated for this work with (Li P Ge O pellet) from Ref. 3.2 0.8 0.2 4 the conductivity of LGPO thin films grown Our measurements, obtained with pellets at room temperature and at 500 °C. with a nominal composition of Li Ge P O (x = 0.8), are in very good 3.2 0.2 0.8 4 agreement with values reported in the literature for pellets with the same 10, 42 composition. A very similar conductivity was reported also for LGPO with x = 0.2, whereas the peak conductivity was found for x between 0.4 and 0.6 reaching values of about 10 µS cm 1 10 . Another study reports a room- temperature conductivity of about -1 35 µS cm for LGPO pellets with x ≈ 0.25. LGPO single crystal, with x = 0.66, showed a conductivity which is slightly lower but with the same activation -1 energy. The value of 1.8 µS cm was reached at 40 °C. The very similar value of conductivity 8 Figure 4 Conductivity dependence on the An advantage of LISICON thin films is inverse temperature of LGPO pellet that the same conductivity of the (Li Ge P O ) and thin films deposited at 25 3.2 0.2 0.8 4 stoichiometric compound can be reached and 500 °C. Composition of thin films from already for low lithium content. For RBS and ERDA analysis: Li Ge P O and example, in case of solid solutions of 1.58±0.4 0.45 0.55 3.27±0.5 Li Ge P O , respectively. The point 2.7±0.3 0.49 0.5 3.7±0.3 lithium silicate – lithium phosphate (pink sphere) at 25 °C was measured for the (Li SiO – Li PO ), the conductivity 4 4 3 4 thin film deposited at room temperature “as strongly depends on the lithium grown”. After the complete thermal cycle, the stoichiometry only up to a certain threshold conductivity did not show any significant change. The inset shows the electrode ( Li/(P+Si) ~ 1). Above this value, the configuration used for the different samples. conductivity becomes independent of the lithium content as the mobile lithium sites LGPO thin films show electrical properties are saturated. Assuming that LGPO that are very similar to those of the pellets, behaves similarly to silicates, we can independently of the deposition observe that in the film deposited at room temperature. The film grown at room temperature the measured lithium temperature shows an activation energy stoichiometry is lower compared to sample that is about 0.07 eV higher. deposited at 500 °C. However, in both This range of conductivity overlaps with cases the ratio between Li and P+Ge that reported in the literature for LiPON contents is well above 1. Therefore, it is amorphous thin films (highest value 2.3 likely that other factors, such as different -1 44-45 µS cm at room temperature). LiPON morphological features and/or the presence is, so far, the only solid-state Li0-ion of Li CO , are affecting the ionic transport. 2 3 conductor that can be used as thin film Considering the results of the XRD electrolyte for microbatteries. The analysis in Figure 2a, these measurements perovskite-type oxide LLTO and the confirm also for thin films what was garnet-type oxide LLZO, with room deduced from the comparison of temperature conductivity in the range of -1 conductivities of sintered pellets and single mS cm , require sintering temperatures crystals (Figure 3): the degree of well above 1000 °C for ceramic pellets and crystallinity (average grain size, extent of deposition temperatures well above 800 °C grain boundaries) has a relatively small for thin film growth. This promotes Li influence on the conducting properties. evaporation and the stabilization of the La Ti O and La Zr O non-conductive 2 2 7 2 2 7 First-principles molecular dynamics phases that are thermodynamically simulations for bulk defect-free LGPO favored. The conductivity of LLTO thin films reported so far in the literature is in A 100 - atom (1 × 1 × 3) supercell was the same range of LiPON and LGPO (but built starting from the single crystal unit require a much higher processing cell for which crystallographic data are temperature), while much lower available in the literature (ICSD conductivity has been reported for LLZO #250066 ). Stoichiometry of x = 0.66 was films. needed, in order to ensure a P to Ge ratio 22-23 equal to 1:2. This supercell, with a 9 formula of Li Ge P O , is reported in The absence of preferential conduction 40 4 8 48 Figure 5. pathways for the ionic transport is an important advantage, since channel- To understand Li-ion density and diffusion blocking effects can lower significantly the properties in an ideal (defect-free) LGPO diffusion and have detrimental crystal, we performed extended consequences on the charge/discharge (200 ps – 500 ps) FPMD simulations on properties, when the material is used as the above described periodic supercell in 35, 44 electrolyte in batteries. the NVT ensemble, between 300 °C and 1100 °C. The isosurfaces of the Li-ion probability density at different temperatures, displayed in Figure 6 a and 6 b, indicate that the ideal LGPO crystal offers a three-dimensional isotropic conduction pathway for Li ions. This is consistent with the small degree of anisotropy shown by the Li-ion MSD along x, y, z, at the temperature of 1400 K (Figure 6 c). It is noteworthy that LGPO does not present unidimensional channels for the lithium ion conduction. Channel-blocking effects to the ionic transport can be ruled out and the diffusion is isotropic. Figure 6 The Li - ion probability density at a) 1400 K (1126 °C) and b) 600 K (327 °C), for -3 -3 - three isovalues, 0.001 Å , 0.01 Å and 0.1 Å , as purple, salmon and yellow isosurfaces, respectively. The equilibrium positions of oxygen, phosphorus and germanium are shown as red, orange, and green spheres, respectively. c) x-, y- and z-resolved mean-squared displacement (MSD) for Li as a function of time at 1400 K (1126 °C), as red, green, and blue solid lines, respectively, showing again isotropic diffusion. In the legends, the resulting diagonal elements of the diffusion matrix (and the error originating from the individual Figure 5 Side and top view of the 100 - atom blocks, dashed lines) are reported. unit cell used for the simulations of defect-free Li-ion MSDs were used to calculate the LGPO crystal (from Ref. ). Li atoms are displayed in green, oxygen atoms in red, Ge ionic diffusion coefficients and the and P atoms are at the center of the light and conductivity dependence on the dark purple tetrahedra, respectively. temperature. 10 Figure 7 compares the computational values of Li-ion conductivity in LGPO with the experimental values previously described (high-temperature deposited thin film, x=0.5) obtained in this study. The non-quantitative agreement between FPMD-calculated (0.37±0.04 eV) and IS- measured (0.51±0.01 eV) activation barriers is not new for ionic conductivities 46-48 in solid-state electrolytes. Possible reasons are the different time and length scales accessed through simulations (atomic level) and experiments (entire sample), and the absence of defects in the simulated ideal crystal, so that values calculated by FPMD may be considered as the upper limit for the conductivity in Figure 7) Li-ion conductivity Arrhenius plot LGPO (Li Ge P O ). Besides, going 3.33 0.33 0.66 4 for the simulated ideal LGPO (x=0.67) crystal (FPMD), compared with experimental data beyond the Li-ion tracer diffusion (i.e. (high-temperature deposited thin film, x=0.5). calculating the charge diffusion coefficient The underestimation of the activation energy [ ]) would be desirable. However, our by simulations compared to the experimental attempts to calculate the Haven ratio gave values was previously ascribed to the different length scales analyzed by the techniques and rise to charge diffusion coefficients with difficulty to take 1D and 2D defects into large statistical errors, that we don’t report account. here, due to LGPO being a weak ionic conductor, so that longer trajectories would Conclusions be needed for such calculations. It is In this work we synthesized and worthwhile to recall here that some of us characterized pellets and thin films of Li 4- have shown the Haven ratio not to play a Ge P O (solid solution between γ- x 1-x x 4 significant role on the activation energy for Li PO and Li GeO ) belonging to the 3 4 4 4 the diffusion for the similar system ionic conductors in the LISICON family. LGPS. Pellets with nominal composition In turn, comparison with the experimental Li Ge P O and a relative density of 3.2 0.2 0.8 4 values, obtained in this work and reported 85% show a conductivity and activation in the literature, suggests that the transport energy in agreement with previously properties are affected by the composition published data. and presence of large-scale defects. Large- scale defects affect the conductivity Thin films were deposited via PLD at measured by IS, while their effect is not different temperatures. The compositional included in the conductivity calculated by transfer from target to substrate during FPMD. laser ablation is not stoichiometric, resulting in different ratios between Ge and 11 previously for similar systems that the P and the lithium content. Samples Haven ration does not significantly affect deposited at low temperature are deficient the activation barrier. in lithium and show contaminations in the form of lithium carbonate. In addition, thin In conclusion, LGPO exhibits a number of films deposited at low temperature are not very interesting features that are potentially completely amorphous as reported important for a technological application as previously for similar materials, but appear electrolyte in thin film batteries and to possess a nanocrystalline morphology. fundamental research. In particular, the fabrication of thin films with the desired Compared to the deposition at high characteristics is particularly easy to temperature, the deposition at low achieve (also at room temperature). temperature reduces the conductivity of the films by approximately a factor of 4 at It was previously reported that LGPO (with room temperature. This is most probably formula Li Ge P O ) is stable 3.33 0.33 0.66 4 due to the higher roughness and porosity, between -0.5 and 7 V vs Li/Li , which which favors the formation of carbonates make it suitable to be used also with high throughout the film. Lithium carbonate is voltage electrodes. present as a contaminant only at the surface All these characteristics make LGPO a of the dense crystalline films without potentially interesting electrolyte material strongly affecting the conductivity after for all-solid-state all-oxide microbatteries. exposing the sample to air. . AUTHOR INFORMATION From the characterization of pellet and thin films, it appears that the conductivity is Corresponding Author almost independent of the microstructure and lithium content, if this is above the * elisa.gilardi@psi.ch threshold of saturation of the mobile lithium sites. * giuliana.materzanini@epfl.ch Conductivities at different temperatures calculated by FPMD and experimental results are then compared. A clear, but Author Contributions expected difference exists between the short length scale conductivity calculated The manuscript was written through by FPMD and the conductivity measured contributions of all authors. All authors on the whole sample by IS. This suggests that there is a blocking effect of large-scale have given approval to the final version of defects that is not detected by the experimental characterization. Effects of the manuscript. the Li-ion collective diffusion (Haven ratio) were not taken into account here, due Funding Sources to the modest conductivity of this material, that would have required longer simulation This research was supported by the times. At the same time, it was reported NCCR MARVEL, funded by the Swiss 12 National Science Foundation and by DAIMLER AG, Stuttgart, Andreas Hintennach. Notes EG, TL and DP would like to dedicate this work to their colleague, Andreas Hintennach, who unfortunately, died prematurely this year. He will be sorely missed and his personal, scientific and technical contributions will be irreplaceable. 13 (9) Hodge, I. M.; Ingram, M. D.; West, A. R. References Ionic-Conductivity of Li SiO , Li GeO , and 4 4 4 4 their Solid-Solutions. J. Am. Ceram. Soc. (1) Oudenhoven, J. F. M.; Baggetto, L.; 1976, 59 (7-8), 360-366, DOI: Notten, P. H. L. All-Solid-State Lithium-Ion 10.1111/j.1151-2916.1976.tb10985.x. Microbatteries: A Review of Various Three- (10) Muy, S.; Bachman, J. C.; Chang, H. H.; Dimensional Concepts. Adv. Energy Mater. Giordano, L.; Maglia, F.; Lupart, S.; Lamp, P.; 2011, 1 (1), 10-33, DOI: Zeier, W. G.; Shao-Horn, Y. Lithium 10.1002/aenm.201000002. conductivity and Meyer-Neldel rule in (2) Sun, C. W.; Liu, J.; Gong, Y. D.; Wilkinson, Li PO -Li VO -Li GeO lithium superionic 3 4 3 4 4 4 D. P.; Zhang, J. J. Recent advances in all-solid- conductors. Chem. Mat. 2018, 30 (16), 5573- state rechargeable lithium batteries. Nano 5582, DOI: Energy 2017, 33, 363-386, DOI: 10.1021/acs.chemmater.8b01504. 10.1016/j.nanoen.2017.01.028. (11) Deng, Y.; Eames, C.; Fleutot, B.; David, (3) Manthiram, A.; Yu, X.; Wang, S. Lithium R.; Chotard, J. N.; Suard, E.; Masquelier, C.; battery chemistries enabled by solid-state Islam, M. S. Enhancing the Lithium Ion electrolytes. Nat. Rev. Mater. 2017, 2 Conductivity in Lithium Superionic (16103), 1-16. Conductor (LISICON) Solid Electrolytes (4) Zhu, Y. Z.; He, X. F.; Mo, Y. F. Origin of through a Mixed Polyanion Effect. ACS Appl. outstanding stability in the lithium solid Mater. Interfaces 2017, 9 (8), 7050-7058, electrolyte materials: insights from DOI: 10.1021/acsami.6b14402. thermodynamic analyses based on first- (12) Zhao, G. W.; Suzuki, K.; Yonemura, M.; principles calculations. ACS Applied Hirayama, M.; Kanno, R. Enhancing Fast Materials & Interfaces 2015, 7 (42), 23685- Lithium Ion Conduction in Li GeO -Li PO 4 4 3 4 23693, DOI: 10.1021/acsami.5b07517. Solid Electrolytes. ACS Appl. Energ. Mater. (5) Krauskopf, T.; Hartmann, H.; Zeier, W. G.; 2019, 2 (9), 6608-6615, DOI: Janek, J. Toward a Fundamental 10.1021/acsaem.9b01152. Understanding of the Lithium Metal Anode (13) Famprikis, T.; Galipaud, J.; Clernens, O.; in Solid-State Batteries-An Electrochemo- Pecquenard, B.; Le Cras, F. Composition Mechanical Study on the Garnet-Type Solid Dependence of Ionic Conductivity in Electrolyte Li6.25Al0.25La3Zr3O12. Acs LiSiPO(N) Thin-Film Electrolytes for Solid- Applied Materials & Interfaces 2019, 11 (15), State Batteries. ACS Appl. Energ. Mater. 14463-14477, DOI: 2019, 2 (7), 4782-4791, DOI: 10.1021/acsami.9b02537. 10.1021/acsaem.9b00415. (6) Inaguma, Y.; Chen, L. Q.; Itoh, M.; (14) Su, Y. R.; Falgenhauer, J.; Leichtweiss, Nakamura, T.; Uchida, T.; Ikuta, H.; T.; Geiss, M.; Lupo, C.; Polity, A.; Zhou, S. Q.; Wakihara, M. High ionic-conductivity in Obel, J.; Schlettwein, D.; Janek, J.; Meyer, B. K. lithium lanthanum titanate. Solid State Electrochemical properties and optical Commun. 1993, 86 (10), 689-693, DOI: transmission of high Li+ conducting 10.1016/0038-1098(93)90841-a. LiSiPON electrolyte films. Phys. Status Solidi (7) Wu, J. F.; Chen, E. Y.; Yu, Y.; Liu, L.; Wu, Y.; B-Basic Solid State Phys. 2017, 254 (2), 8, Pang, W. K.; Peterson, V. K.; Guo, X. Gallium- DOI: 10.1002/pssb.201600088. Doped Li La Zr O Garnet-Type 7 3 2 12 (15) Kanehori, K.; Matsumoto, K.; Miyauchi, Electrolytes with High Lithium-Ion K.; Kudo, T. Thin-Film Solid Electrolyte and Conductivity. Acs Applied Materials & its application to secondary lithium cell. Interfaces 2017, 9 (2), 1542-1552, DOI: Solid State Ion. 1983, 9-10 (DEC), 1445- 10.1021/acsami.6b13902. 1448, DOI: 10.1016/0167-2738(83)90192- (8) Rodger, A. R.; Kuwano, J.; West, A. R. Li Ion Conducting Gamma-Solid Solutions in (16) Sakurai, Y.; Sakuda, A.; Hayashi, A.; the Systems Li XO -Li YO - X=Si,Ge,Ti- 4 4 3 4 Tatsumisago, M. Preparation of amorphous Y=P,As,V-Li XO -LiZO -Z=Al,Ga,Cr and 4 4 2 Li SiO -Li PO thin films by pulsed laser 4 4 3 4 Li GeO -Li CaGeO . Solid State Ion. 1985, 15 4 4 2 4 deposition for all-solid-state lithium (3), 185-198, DOI: 10.1016/0167- secondary batteries. Solid State Ion. 2011, 2738(85)90002-5. 182 (1), 59-63, DOI: 14 10.1016/j.ssi.2010.12.001. 10.1103/PhysRev.140.A1133. (17) Kim, D. H.; Imashuku, S.; Wang, L.; Shao- (27) Payne, M. C.; Teter, M. P.; Allan, D. C.; Horn, Y.; Ross, C. A. Li loss during the growth Arias, T. A.; Joannopoulos, J. D. Iterative of (Li,La)TiO thin films by pulsed laser minimization techniques for Ab-Initio total- deposition. J. Cryst. Growth 2013, 372, 9-14, energy calculations - molecular-dynamics DOI: 10.1016/j.jcrysgro.2013.02.028. and conjugate gradients. Rev. Mod. Phys. (18) Pfenninger, R.; Struzik, M.; Garbayo, I.; 1992, 64 (4), 1045-1097, DOI: Stilp, E.; Rupp, J. L. M. A low ride on 10.1103/RevModPhys.64.1045. processing temperature for fast lithium (28) Giannozzi, P.; Baroni, S.; Bonini, N.; conduction in garnet solid-state battery Calandra, M.; Car, R.; Cavazzoni, C.; Ceresoli, films. Nature Energy 2019, 4 (6), 475-483, D.; Chiarotti, G. L.; Cococcioni, M.; Dabo, I.; DOI: 10.1038/s41560-019-0384-4. Dal Corso, A.; de Gironcoli, S.; Fabris, S.; (19) Aguesse, F.; Roddatis, V.; Roqueta, J.; Fratesi, G.; Gebauer, R.; Gerstmann, U.; Garcia, P.; Pergolesi, D.; Santiso, J.; Kilner, J. Gougoussis, C.; Kokalj, A.; Lazzeri, M.; A. Microstructure and ionic conductivity of Martin-Samos, L.; Marzari, N.; Mauri, F.; LLTO thin films: Influence of different Mazzarello, R.; Paolini, S.; Pasquarello, A.; substrates and excess lithium in the target. Paulatto, L.; Sbraccia, C.; Scandolo, S.; Solid State Ion. 2015, 272, 1-8, DOI: Sclauzero, G.; Seitsonen, A. P.; Smogunov, A.; 10.1016/j.ssi.2014.12.005. Umari, P.; Wentzcovitch, R. M. QUANTUM (20) Lobe, S.; Dellen, C.; Windmuller, A.; ESPRESSO: a modular and open-source Tsai, C. L.; Vondahlen, F.; Uhlenbruck, S.; software project for quantum simulations of Guillon, O. Challenges regarding thin film materials. J. Phys.-Condes. Matter 2009, 21 deposition of garnet electrolytes for all- (39), 19, DOI: 10.1088/0953- solid-state lithium batteries with high 8984/21/39/395502. energy density. Ionics 2018, 24 (8), 2199- (29) Prandini, G.; Marrazzo, A.; Castelli, I. E.; 2208, DOI: 10.1007/s11581-018-2594-3. Mounet, N.; Marzari, N. Precision and (21) Julien C.; A., M. Pulsed Laser Deposited efficiency in solid-state pseudopotential Films for Microbatteries Coatings 2019, 9 calculations. npj Comput. Mater. 2018, 4, 13, (6), 386. DOI: 10.1038/s41524-018-0127-2. (22) Ivanov-Shitz, A. K.; Kireev, V. V. Growth (30) Garrity, K. F.; Bennett, J. W.; Rabe, K. M.; and ionic conductivity of Li P Ge O Vanderbilt, D. Pseudopotentials for high- 3+x 1-x x 4 (x=0.34) single crystals. Crystallogr. Rep. throughput DFT calculations. Comput. 2003, 48 (1), 112-115, DOI: Mater. Sci. 2014, 81, 446-452, DOI: 10.1134/1.1541752. 10.1016/j.commatsci.2013.08.053. (23) Rabadanov, M. K.; Pietraszko, A.; (31) Dal Corso, A. Pseudopotentials periodic Kireev, V. V.; Ivanov-Schitz, A. K.; Simonov, table: From H to Pu. Comput. Mater. Sci. V. I. Atomic structure and mechanism of 2014, 95, 337-350, DOI: ionic conductivity of Li Ge P O single 10.1016/j.commatsci.2014.07.043. 3.31 0.31 0.69 4 crystals. Crystallogr. Rep. 2003, 48 (5), 744- (32) Perdew, J. P.; Burke, K.; Ernzerhof, M. 749, DOI: 10.1134/1.1612594. Generalized gradient approximation made (24) Doolittle, L. R. A Semiautomatic simple. Phys. Rev. Lett. 1996, 77 (18), 3865- Algorithm for Rutherford Backscattering 3868, DOI: 10.1103/PhysRevLett.77.3865. Analysis. Nucl. Instrum. Methods Phys. Res. (33) Martyna, G. J.; Klein, M. L.; Tuckerman, Sect. B-Beam Interact. Mater. Atoms 1986, M. Nosee-Hoover chains - the canonical 15 (1-6), 227-231, DOI: 10.1016/0168- ensemble via continuous dynamics. J. Chem. 583x(86)90291-0. Phys. 1992, 97 (4), 2635-2643, DOI: (25) Hohenberg, P.; Kohn, W. 10.1063/1.463940. Inhomogeneous electron gas. Phys. Rev. B (34) Marcolongo, A.; Marzari, N. Ionic 1964, 136 (3B), B864-+, DOI: correlations and failure of Nernst-Einstein 10.1103/PhysRev.136.B864. relation in solid-state electrolytes. Phys. Rev. (26) Kohn, W.; Sham, L. J. Self-consistent Mater. 2017, 1 (2), 4, DOI: equations including exchange and 10.1103/PhysRevMaterials.1.025402. correlation effects. Physical Review 1965, (35) Kahle, L.; Musaelian, A.; Marzari, N.; 140 (4A), 1133-&, DOI: Kozinsky, B. Unsupervised landmark 15 analysis for jump detection in molecular Lithium phosphorus oxynitride. J. dynamics simulations. Phys. Rev. Mater. Electrochem. Soc. 1997, 144 (2), 524-532, 2019, 3 (5), 18, DOI: DOI: 10.1149/1.1837443. 10.1103/PhysRevMaterials.3.055404. (46) Adams, S.; Rao, R. P. Structural (36) Frenkel, D.; Smit, B. Understanding requirements for fast lithium ion migration molecular simulation: from algorithms to in Li10GeP2S12. J. Mater. Chem. (UK) 2012, 22 applications 2001; Vol. 1. (16), 7687-7691, DOI: (37) Grasselli, F.; Baroni, S. Topological 10.1039/c2jm16688g. quantization and gauge invariance of charge (47) Mo, Y. F.; Ong, S. P.; Ceder, G. First transport in liquid insulators. Nat. Phys. Principles Study of the Li GeP S Lithium 10 2 12 2019, 15 (9), 967-972, DOI: Super Ionic Conductor Material. Chem. Mat. 10.1038/s41567-019-0562-0. 2012, 24 (1), 15-17, DOI: (38) Saccoccio, M.; Yu, J.; Lu, Z. H.; Kwok, S. 10.1021/cm203303y. C. T.; Wang, J.; Yeung, K. K.; Yuen, M. M. F.; (48) Deng, Y.; Eames, C.; Chotard, J. N.; Ciucci, F. Low temperature pulsed laser Lalere, F.; Seznec, V.; Emge, S.; Pecher, O.; deposition of garnet Li La Zr Ta O Grey, C. P.; Masquelier, C.; Islam, M. S. 6.4 3 1.4 0.6 12 films as all solid-state lithium battery Structural and mechanistic insights into fast electrolytes. J. Power Sources 2017, 365, 43- lithium-ion conduction in Li SiO -Li PO 4 4 3 4 52, DOI: 10.1016/j.jpowsour.2017.08.020. solid electrolytes. J. Am. Chem. Soc. 2015, (39) Rawlence, M.; Garbayo, I.; Buecheler, S.; 137 (28), 9136-9145, DOI: Rupp, J. L. M. On the chemical stability of 10.1021/jacs.5b04444. post-lithiated garnet Al-stabilized (49) France-Lanord, A.; Grossman, J. C. Li La Zr O solid state electrolyte thin Correlations from Ion Pairing and the 7 3 2 12 films. Nanoscale 2016, 8 (31), 14746- Nernst-Einstein Equation. Phys. Rev. Lett. 14753, DOI: 10.1039/c6nr04162k. 2019, 122 (13), 6, DOI: (40) Ojeda-G-P, A.; Dobeli, M.; Lippert, T. 10.1103/PhysRevLett.122.136001. Influence of plume properties on thin film composition in pulsed laser deposition. Adv. Mater. Interfaces 2018, 5 (18), 16, DOI: 10.1002/admi.201701062. (41) Gerstl, M.; Navickas, E.; Friedbacher, G.; Kubel, F.; Ahrens, M.; Fleig, J. The separation of grain and grain boundary impedance in thin yttria stabilized zirconia (YSZ) layers. Solid State Ion. 2011, 185 (1), 32-41, DOI: 10.1016/j.ssi.2011.01.008. (42) Rodger, A. R.; Kuwano, J.; West, A. R. Li+ ion conducting gamma-solid solutions in the systems Li XO -Li YO - X=Si,Ge,Ti-Y=P,As,V 4 4 3 4 - Li4XO -LiZO -Z=Al,Ga,Cr and Li GeO - 4 2 4 4 Li CaGeO . Solid State Ion. 1985, 15 (3), 185- 2 4 198, DOI: 10.1016/0167-2738(85)90002-5. (43) Song, S. F.; Dong, Z. C.; Deng, F.; Hu, N. Lithium superionic conductors Li MP O 10 2 12 (M = Ge, Si). Funct. Mater. Lett. 2018, 11 (2), 4, DOI: 10.1142/s179360471850039x. (44) Kamaya, N.; Homma, K.; Yamakawa, Y.; Hirayama, M.; Kanno, R.; Yonemura, M.; Kamiyama, T.; Kato, Y.; Hama, S.; Kawamoto, K.; Mitsui, A. A lithium superionic conductor. Nat. Mater. 2011, 10 (9), 682-686, DOI: 10.1038/nmat3066. (45) Yu, X. H.; Bates, J. B.; Jellison, G. E.; Hart, F. X. A stable thin-film lithium electrolyte:

Journal

PhysicsarXiv (Cornell University)

Published: Oct 6, 2020

There are no references for this article.