First-Principles Characterization of the Unknown Crystal Structure and

Jun 27, 2016 - Using first-principles density functional theory calculations and ab initio molecular dynamics (AIMD) simulations, we demonstrate the c...
0 downloads 4 Views 2MB Size
Letter pubs.acs.org/JPCL

First-Principles Characterization of the Unknown Crystal Structure and Ionic Conductivity of Li7P2S8I as a Solid Electrolyte for HighVoltage Li Ion Batteries Joonhee Kang† and Byungchan Han*,‡ †

Department of Energy Systems Engineering, DGIST, Daegu 42988, Republic of Korea Department of Chemical and Biomolecular Engineering, Yonsei University, Seoul 03722, Republic of Korea



S Supporting Information *

ABSTRACT: Using first-principles density functional theory calculations and ab initio molecular dynamics (AIMD) simulations, we demonstrate the crystal structure of the Li7P2S8I (LPSI) and Li ionic conductivity at room temperature with its atomic-level mechanism. By successively applying three rigorous conceptual approaches, we identify that the LPSI has a similar symmetry class as Li10GeP2S12 (LGPS) material and estimate the Li ionic conductivity to be 0.3 mS cm−1 with an activation energy of 0.20 eV, similar to the experimental value of 0.63 mS cm−1. Iodine ions provide an additional path for Li ion diffusion, but a strong Li−I attractive interaction degrades the Li ionic transport. Calculated density of states (DOS) for LPSI indicate that electrochemical instability can be substantially improved by incorporating iodine at the Li metallic anode via forming a LiI compound. Our methods propose the computational design concept for a sulfide-based solid electrolyte with heteroatom doping for high-voltage Li ion batteries.

G

chemical doping of halides (Cl, Br, and I) was attempted. For example, crystal structures of Li6PS5 with halide and sulfurembedded glass−ceramic electrolytes with LiI showed Li ionic conductivity of several mS cm−1, which is significantly improved compared to that without the doping. 16−18 Furthermore, new type of sulfide-based material with LiI, Li7P2S8I (LPSI), was intensively studied as a solid electrolyte due to its strong electrochemical stability as high as 10 V (vs Li/Li+) and high Li ionic conductivity of 0.64 mS cm−1 at room temperature.19 Although these properties could be superior and even better than those of nanoporous β-Li3PS4 materials,20 detailed understanding of the crystal structure of the LPSI and the Li ion transport mechanism within should be clearly addressed before its further development to a commercialization level. In this Letter, we extensively utilized first-principles density functional theory (DFT) calculations and thermodynamics to clearly characterize the crystal structure of the LPSI. We validated our results with experimental X-ray diffraction (XRD) patterns available. Using ab initio molecular dynamics (AIMD) simulations, we also demonstrated the Li ionic conductivity and diffusion mechanism in LPSI to establish design principles for heteroatom-doped solid-state electrolytes in high-voltage Li ion battery applications. We applied two different kinds of first-principles computational methods in our study: DFT calculations enabled us to

lobal concerns on environmental degradation and energy sustainability have been strong driving forces to develop renewable energy source and storage systems.1−3 Among the widely deployed, a Li ion battery is accepted as a promising device due to its high energy density, energy storage efficiency, and long cycling life grounding in electrochemical reactions. Toward its further spread over our society, several scientific and technical challenges in the Li ion battery should be overcome, for instance, the safety risk of flammable liquid-phase organic electrolytes, especially at high voltage and power conditions.4,5 In the past decades, several solid oxides (LISICON, NASICON, perovskite, garnet, and LiPON)6−11 were proposed as alternatives to the volatile liquid-phase organic electrolytes. Solid electrolyte enables use of Li metal as an anode, possibly leading to substantial improvement of the energy density. These merits are, however, mostly nullified by significantly lower Li ionic conductivity than the organic counterpart. Recently, It was reported that sulfide-based solid electrolytes such as Li10GeP2S12 (12 mS cm−1), glass ceramic Li7P3S11 (17 mS cm−1 ), and Li9.54 Si 1.74P 1.44 S 11.7 Cl0.3 (25 mS cm−1) demonstrated features of Li superionic conductors.12−14 Additionally, it was claimed that Li ionic conductivity in sulfide-based solid-state electrolytes critically depends on atomistic distribution of PS4 tetrahedrons, which was speculated as playing a role in the formation of the Li diffusion channel. In addition, in 2015, Wang et al.15 insisted that a bcclike anion framework should be the key descriptor in sulfidebased solid-state electrolytes. On the other hand, to enhance the electrochemical stability of the sulfide solid electrolyte in high-voltage windows, © 2016 American Chemical Society

Received: May 16, 2016 Accepted: June 27, 2016 Published: June 27, 2016 2671

DOI: 10.1021/acs.jpclett.6b01050 J. Phys. Chem. Lett. 2016, 7, 2671−2675

Letter

The Journal of Physical Chemistry Letters

model for the LPSI because the formation energy is only slightly positive (ΔE = 0.056 eV/atom). The replacement of Li4GeS4 with LiI seems feasible, as represented by the formation energy and the crystal symmetry. The crystal structure of Pnma is transformed into a similar Pmm2. Surprisingly, the calculated XRD pattern of the solid LGPS_2 (Figure 1a) looks similar to the experimentally measured one

evaluate ground-state energies of various LPSI models, and AIMD simulated Li transport behaviors as a function of time and temperature. We used the Vienna ab initio simulation package (VASP)21 for our calculations. The generalized gradient approximation (GGA) parametrized by Perdew, Burke, and Ernzerhof (PBE)22 was used for the exchange− correlation functional, and the projected-augmented wave (PAW) method23 described interactions among core electrons. Plane wave basis functions for the Kohn−Sham equation were expanded up to an energy cutoff of 649 eV. We stopped each of our calculations when the total energy was converged within 10−4 eV. The γ-point scheme was employed with k-meshes of 3 × 3 × 2 to accurately integrate the Brillouin zone in the reciprocal space. AIMD simulations were carried out with a time step of 1 fs (femtosecond), and we performed a total of 30 000 steps. The diffusivity and conductivity of a Li ion were determined over different temperatures from 600 and 1000 K with 100 K increments. We applied three rigorous steps to the model successively: simplification, insertion, and replacement. The simplification step was treating a LPSI as a composite consisting of three substructures: two of PS43−, one Li6I5+, and Li+. Then, we could propose three possible candidates of AB2type crystal structures, such as fluorite, rutile, and CdI2, where Li6I5+ as the A, PS43− for the B, and Li+ ions were allocated into the interstitial sites. Iodine and P ions were deployed at regular lattice sites of the AB2 crystal, and four S and six Li were assigned around a P and the iodine as far as possible. Last, the rest of the Li atoms were put into the interstitial sites. The insertion set was to put LiI in the interstitial sites of the β-LPS. Additional Li were allocated to the Li sites, while iodine atoms were put in lattice points of space group Pnma. We realized that these candidate crystals suffered from substantial amounts of lattice distortion, breaking the local symmetry of LPSI. It means that the models are not thermodynamically stable as a bulk material and thus may not be observed in the XRD pattern. The third process, the replacement step, was the most important step in our study. It was to replace Li4GeS4 in LGPS with LiI. Because this process breaks the crystallinity of the LPSI, we reasoned that anion exchange with the iodine may restore the initial symmetry. To evaluate the thermodynamic feasibility of our model system, we calculated its formation energy as defined in eq 1 for the chemical reaction described in eq 2.19 ΔE = E LPSI − 2E γ‐LPS − E LiI

(1)

2(γ‐Li3PS4 ) + LiI → Li 7P2S8I

(2)

Figure 1. XRD pattern of the calculated LPSI system (a) without and (b) with anion mixing. The insets are model systems of LPSI. A solid green circle represents lithium, light purple is for iodine, yellow is for sulfur, and red represents the tetrahedron for PS4.

by Rangasamy et al.19 There are, however, slight deviations in the XRD intensities, especially at the two highest ones. We reasoned that it could originate from anion-mixed states. We computationally screened 16 possible structures for the anionmixed states with the two ions, S and iodine, and calculated XRD patterns to identify the most promising candidate. Figure 1b illustrates the XRD pattern and model system of the anionmixed LPSI that we optimized using the formation energy and XRD intensity. After the anion mixing, we found that the structure transformed into a monoclinic structure with a space group of Pm with a unique b-axis, as shown in Table S2. The reason is characteristic of the bond distance between Ge and S (about 2.2 Å) and the ionic radius of iodine ion (about 2.1 Å), meaning that the anion (iodine ion) can mostly fill up the vacant space caused by Ge or S in the Li4GeS4 with little distortion. Thus, the crystal structure of Pnma is transformed

Here, ELPSI, Eγ‑LPS, and ELiI are the ground-state energies of Li7P2S8I, γ-Li3PS4, and LiI, respectively, calculated by the DFT method. The more negative the formation energy, the more stable the structure and vice versa. Table S1 shows calculated formation energies of the LPSI models normalized by the number of atoms (ΔE), with varying space groups. As denoted, all formation energies are positive, implying that heat treatments are necessary for synthesizing the materials. The simplification and insertion processes indicate that the Rutile_1 (ΔE = 0.075 eV/atom) and Beta_A2 (ΔE = 0.074 eV/atom) models can be stable, while others may be difficult to fabricate in practice. Among the three models identified by the replacement step, the LGPS_2 seems to be the most thermodynamically plausible 2672

DOI: 10.1021/acs.jpclett.6b01050 J. Phys. Chem. Lett. 2016, 7, 2671−2675

Letter

The Journal of Physical Chemistry Letters either into a similar P42/nmc without anion mixing or into Pm with anion mixing. The formation energy of the anion-mixed LPSI model was about 0.048 eV/atom, which is more stable than the LGPS_2 model, implying that our model is at least thermodynamically plausible. As shown at Figure 1b, the highest and second highest intensity peaks are at 27.7 and 20.4°, representing (022) and (112) planes, respectively. Planar distances of (022) and (112) were calculated to be 3.13 and 4.37 Å. Interestingly, the (022) plane has the highest planar density because the facets of PS4 and PS3I tetrahedrons are placed in the same plane. It agrees well with the conventional theory that the XRD intensity in a crystal with low symmetry relies largely on its planar density.24 As shown in Figure 1b, the ratio of intensity of the significant peaks was calculated as 0.74, which is similar to the experimental measurement of about 0.67.19 It implies the structural similarity between our LPSI model and the experimental one. Using HSE06 hybrid functionals,25,26 we calculated the density of states (DOS) of electrons in the LPSI with and without anion mixing, as illustrated in Figure 2. Band gaps

anion-mixed LPSI, either nonbinding or binding with P. Nonbinding iodine ions are located with only Li nearby, while the binding one is at the PS3I tetrahedron with a bond distance of 2.51 Å. Additional states came from the iodine in PS3I due to the charge difference between sulfur (2−) and iodine (1−). Experimental results, however, showed that the electrochemical window of iodine-doped sulfide-based electrolytes is about 10 V (vs Li/Li+).17−19 Recently, Zhu et al.5 showed that the Li reduction products of lithium halides, which are thermodynamically stable against a Li metal anode, enhance the electrochemical stability. The calculated band gap of LiI (5 eV) supported the extremely high electrochemical stability of halide-doped sulfur-based solid-state electrolytes. Using random walk theory and AIMD simulations as a function of T (= 600, 700, 800, 900, and 1000 K), we calculated the Li ionic conductivity in the LPSI crystal structure, as shown in Figure 3.27 Using AIMD simulations, Li ionic diffusivity was

Figure 3. Li ionic conductivity (black squares) as a function of temperature in LPSI calculated by AIMD simulations. The filled black square corresponds to the linearly extrapolated (blue dashed line) conductivity at 298 K from high-temperature conductivity. The red square shows the room-temperature ionic conductivity (0.64 mS cm−1) from ref 18, also shown for comparison.

calculated by a linear fit of the mean-square displacement of Li atoms as a function of time for a given temperature. Then, the Li ionic conductivity was estimated using the Nernst−Einstein relation, eq 3 ρz 2F 2 D(T ) (3) R where σ(T) and D(T) are the ionic conductivity and diffusion coefficient of Li ions in a unit cell, respectively. The ρ, z, F, and R are the molar density, the mean valence electronic charge of a Li ion, the Faraday constant, and the gas constant, respectively. Our AIMD simulations were performed at elevated temperature (600 to 1000 K) to speed up diffusion of Li ions with effective computational cost. It was previously shown that the linear extrapolation of the diffusion coefficient calculated at high temperature (not room temperature) is reasonable without causing serious error.28 The Li ionic conductivity at room temperature obtained by linearly extrapolating values at high-temperature conductivities was 0.3 mS cm−1 with an σ (T )T =

Figure 2. Electronic projected DOS of LPSI (a) without and (b) with anion mixing using HSE06 hybrid functionals. The insets are (a) PS4 and (b) PS3I.

evaluated from the DOS are 2.46 and 3.29 eV for with and without anion mixing. The band gap is, in general, a descriptor for intrinsic electrochemical stability of materials. In the anionmixed model, the band gap decreased due to the additional midgap states that originated from the P bound with the iodine. Specifically, there are two different kinds of iodine ions in 2673

DOI: 10.1021/acs.jpclett.6b01050 J. Phys. Chem. Lett. 2016, 7, 2671−2675

The Journal of Physical Chemistry Letters



activation energy of 0.20 eV, which slightly underestimates experimental observation. The reason is that the volume of the LPSI was a fixed volume regardless of the temperature in our AIMD simulations, while, in practice, solid-state electrolytes may undergo volume expansion as the temperature increases. We found that the Li ion transports in a self-diffusion mode, preferring the direction of the c-axis rather than that of the aand b-axes, implying that the diffusivity along the c-axis is always higher than that along the a- and b-axes, the same as that in LGPS.28 Our results also showed that Li ions move mostly around iodine ions, as illustrated in Figure 4. It strongly indicates that iodine ions provide additional paths for Li ion diffusion, but the strong Li−I bonding suppresses Li ion mobility.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +82-2-2123-5759. Fax: +822-312-6401. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research was supported by the Global Frontier Program through the Global Frontier Hybrid Interface Materials (GFHIM) of the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (2013M3A6B1078882). This study was also supported by the KERI top-down research program of MSIP/NST (KERI_16-12-N0101-16).



REFERENCES

(1) Armand, M.; Tarascon, J. M. Building Better Batteries. Nature 2008, 451, 652−657. (2) Etacheri, V.; Marom, R.; Elazari, R.; Salitra, G.; Aurbach, D. Challenges in the Development of Advanced Li-ion Batteries: A Review. Energy Environ. Sci. 2011, 4, 3243−3262. (3) Goodenough, J. B.; Park, K.-S. The Li-Ion Rechargeable Battery: A Perspective. J. Am. Chem. Soc. 2013, 135, 1167−1176. (4) Kang, J.; Han, B. First-Principles Study on the Thermal Stability of LiNiO2 Materials Coated by Amorphous Al2O3 with Atomic Layer Thickness. ACS Appl. Mater. Interfaces 2015, 7, 11599−11603. (5) Zhu, Y.; He, X.; Mo, Y. Origin of Outstanding Stability in the Lithium Solid Electrolyte Materials: Insights from Thermodynamic Analyses Based on First-Principles Calculations. ACS Appl. Mater. Interfaces 2015, 7, 23685−23693. (6) Bruce, P. G.; West, A. R. The A-C Conductivity of Polycrystalline LISICON, Li2+2x Zn1−xGeO4, and a Model for Intergranular Constriction Resistances. J. Electrochem. Soc. 1983, 130, 662−669. (7) Aono, H.; Sugimoto, E.; Sadaoka, Y.; Imanaka, N.; Adachi, G. y. Ionic Conductivity of Solid Electrolytes Based on Lithium Titanium Phosphate. J. Electrochem. Soc. 1990, 137, 1023−1027. (8) Inaguma, Y.; Liquan, C.; Itoh, M.; Nakamura, T.; Uchida, T.; Ikuta, H.; Wakihara, M. High Ionic Conductivity in Lithium Lanthanum Titanate. Solid State Commun. 1993, 86, 689−693. (9) Murugan, R.; Thangadurai, V.; Weppner, W. Fast Lithium Ion Conduction in Garnet-Type Li7La3Zr2O12. Angew. Chem., Int. Ed. 2007, 46, 7778−7781. (10) Yu, X.; Bates, J. B.; Jellison, G. E.; Hart, F. X. A Stable Thin-Film Lithium Electrolyte: Lithium Phosphorus Oxynitride. J. Electrochem. Soc. 1997, 144, 524−532. (11) Kang, J.; Chung, H.; Doh, C.; Kang, B.; Han, B. Integrated Study of First Principles Calculations and Experimental Measurements for Li-ionic Conductivity in Al-doped Solid-State LiGe2(PO4)3 Electrolyte. J. Power Sources 2015, 293, 11−16. (12) 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, 682−686. (13) Seino, Y.; Ota, T.; Takada, K.; Hayashi, A.; Tatsumisago, M. A Sulphide Lithium Super Ion Conductor is Superior to Liquid Ion Conductors for Use in Rechargeable Batteries. Energy Environ. Sci. 2014, 7, 627−631. (14) Kato, Y.; Hori, S.; Saito, T.; Suzuki, K.; Hirayama, M.; Mitsui, A.; Yonemura, M.; Iba, H.; Kanno, R. High-Power All-Solid-State Batteries using Sulfide Superionic Conductors. Nat. Energy 2016, 1, 16030. (15) Wang, Y.; Richards, W. D.; Ong, S. P.; Miara, L. J.; Kim, J. C.; Mo, Y.; Ceder, G. Design Principles for Solid-State Lithium Superionic Conductors. Nat. Mater. 2015, 14, 1026−1031. (16) Rayavarapu, P. R.; Sharma, N.; Peterson, V. K.; Adams, S. Variation in Structure and Li+-Ion Migration in Argyrodite-Type

Figure 4. Transport trajectories of Li simulated using AIMD over 30 ps in LPSI. Panels (a,c) and (b,d) show trajectories of Li at 600 and 1000 K, respectively.

In summary, we developed a conceptual approach based on first-principles computing to identify the unknown crystal structure and transport behaviors in solid electrolytes in highvoltage Li ion battery applications. Using DFT calculations and AIMD simulations, the thermodynamic feasibility in fabricating our model system was estimated by the formation energy and the experimentally measured XRD pattern and Li ionic conductivity. We characterized that not only PS4 but also PS3I tetrahedrons play a role in the formation of the LPSI crystal structure. The intrinsic electrochemical stability, predicted by the band gap, decreased due to the PS3I, while the formation of LiI at the Li metal anode side turned out to extremely enhance the electrochemical stability. The Li ion preferred to diffuse along the c-axis over the a- or b-axis, and the Li ionic conductivity at around room temperature was 0.3 mS cm−1, agreeing extremely well with experimental observation. Our computational methods provide the design strategy of heteroatom-doped sulfide-based solid-state electrolytes for high-voltage Li ion battery applications.



Letter

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpclett.6b01050. Formation energies, lattice parameters, and structure parameters (PDF) 2674

DOI: 10.1021/acs.jpclett.6b01050 J. Phys. Chem. Lett. 2016, 7, 2671−2675

Letter

The Journal of Physical Chemistry Letters Li6PS5X (X = Cl, Br, I) Solid Electrolytes. J. Solid State Electrochem. 2012, 16, 1807−1813. (17) Ujiie, S.; Hayashi, A.; Tatsumisago, M. Structure, Ionic Conductivity and Electrochemical Stability of Li2S−P2S5−LiI Glass and Glass−Ceramic Electrolytes. Solid State Ionics 2012, 211, 42−45. (18) Ujiie, S.; Hayashi, A.; Tatsumisago, M. Preparation and Ionic Conductivity of (100−x)(0.8Li2S·0.2P2S5)·xLiI Glass−Ceramic Electrolytes. J. Solid State Electrochem. 2013, 17, 675−680. (19) Rangasamy, E.; Liu, Z.; Gobet, M.; Pilar, K.; Sahu, G.; Zhou, W.; Wu, H.; Greenbaum, S.; Liang, C. An Iodide-Based Li7P2S8I Superionic Conductor. J. Am. Chem. Soc. 2015, 137, 1384−1387. (20) Liu, Z.; Fu, W.; Payzant, E. A.; Yu, X.; Wu, Z.; Dudney, N. J.; Kiggans, J.; Hong, K.; Rondinone, A. J.; Liang, C. Anomalous High Ionic Conductivity of Nanoporous β-Li3PS4. J. Am. Chem. Soc. 2013, 135, 975−978. (21) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for Ab Initio Total-Energy Calculations using a Plane-Wave Basis Set. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186. (22) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (23) Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 50, 17953−17979. (24) Suryanarayana, C.; Norton, M. G. X-ray Diffraction: A Practical Approach; Springer Science & Business Media, 2013. (25) Heyd, J.; Scuseria, G. E.; Ernzerhof, M. Hybrid Functionals Based on a Screened Coulomb Potential. J. Chem. Phys. 2003, 118, 8207−8215. (26) Heyd, J.; Scuseria, G. E.; Ernzerhof, M. Erratum: “Hybrid Functionals Based on a Screened Coulomb Potential”. J. Chem. Phys. 2006, 124, 219906. (27) Ong, S. P.; Mo, Y.; Richards, W. D.; Miara, L.; Lee, H. S.; Ceder, G. Phase Stability, Electrochemical Stability and Ionic Conductivity of the Li10±1MP2X12 (M = Ge, Si, Sn, Al or P, and X = O, S or Se) Family of Superionic Conductors. Energy Environ. Sci. 2013, 6, 148−156. (28) Mo, Y.; Ong, S. P.; Ceder, G. First Principles Study of the Li10GeP2S12 Lithium Super Ionic Conductor Material. Chem. Mater. 2012, 24, 15−17.

2675

DOI: 10.1021/acs.jpclett.6b01050 J. Phys. Chem. Lett. 2016, 7, 2671−2675