- Email: [email protected]
Recent progress of glass and glass-ceramics as solid electrolytes for lithium secondary batteries
Solid State Ionics 177 (2006) 2715 – 2720 www.elsevier.com/locate/ssi
Recent progress of glass and glass-ceramics as solid electrolytes for lithium secondary batteries Tsutomu Minami ⁎, Akitoshi Hayashi, Masahiro Tatsumisago Department of Applied Chemistry, Graduate School of Engineering, Osaka Prefecture University, Japan Received 4 October 2005; received in revised form 20 April 2006; accepted 7 July 2006
Abstract In recent years many new glass-based solid electrolytes with high Li+ conductivity have been developed. In the present paper, we review the preparation and characterization of Li2S-based oxysulfide glasses and sulfide glass-ceramics on the basis of two strategies of enhancing Li+ conductivity: the utilization of “mixed-anion effect” by combining sulfide and oxide anions, and the precipitation of superionic metastable crystals by careful heat-treatment of glasses. The superior Li+ conducting solid electrolytes with the highest conductivity and the lowest activation energy for conduction have been achieved in the Li2S–P2S5 glass-ceramics. The use of these glass-ceramic solid electrolytes leads to the development of a bulk-type all solid-state lithium secondary battery with excellent cycling performance. © 2006 Elsevier B.V. All rights reserved. Keywords: Glass; Glass-ceramic; Lithium ion conductivity; Solid electrolyte; All-solid-state; Battery
1. Introduction Realization of fast ion conduction in solids is indispensable to producing all-solid-state devices such as batteries and chemical sensors with high safety and reliability. Inorganic ionic conductors, namely solid electrolytes, have been extensively studied in two types of solids: crystal and glass. The ionic conductivity of glassy materials is, in general, higher than that of the corresponding crystals [1–4]. In the system AgI–Ag2O–P2O5, the ambient temperature conductivity of glasses is 10− 2 S cm− 1, which is higher by 1–2 orders of magnitude than the conductivity of crystals. These glasses are named “superionic conducting glasses”. The advantage of glass in conductivity is explained by the structure of glass, which is very similar to that of the corresponding melt with high conductivity and low activation energy for conduction. In ionic conductors, Ag+ ions are known to show higher conductivity than alkali ions in spite of the fact that Ag+ ions have a large ionic radius and atomic weight. The electronic configuration of
⁎ Corresponding author. E-mail address: [email protected] (T. Minami). 0167-2738/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ssi.2006.07.017
the ions, i.e. the 4d10 configuration in the outermost orbital, plays an important role as well as the ionic radius in achieving high ion conductivities [3–5]. The most favorable ion species is Li+ because lithium secondary batteries have superior features such as high energy density and low weight. Three important discoveries for creating high Li+ conducting electrolytes have been obtained: (1) the change of an oxide matrix to a sulfide one, (2) the utilization of so-called the “mixed-anion effect” by combining sulfide and oxide anions, and (3) the precipitation of superionic metastable crystals by careful heattreatment of glasses. First of all, the development of Li+ ion conducting glasses has been achieved by changing the glass matrix from oxides to sulfides. Li+ ions acting as a “hard acid”, which is classified from a viewpoint of the “hard and soft acids and bases theory” by Pearson [6], would be more compatible to sulfide ions acting as a “soft base” [7]. In fact, the conductivity of Li2S–SiS2 sulfide glasses is in the order of 10− 4 S cm− 1 at room temperature, which can be compared with the conductivity of 10−7 S cm− 1 for the corresponding Li2O–SiO2 glasses [5,8]. In this paper, we focus on the other two strategies for developing Li+ conductivity of glass-based solid electrolytes and review the preparation and characterization of Li2S-based
2716
T. Minami et al. / Solid State Ionics 177 (2006) 2715–2720
oxysulfide glasses and sulfide glass-ceramics. The rapid quenching method using a twin-roller apparatus and the mechanochemical method using a planetary ball mill apparatus were used for preparation of glasses. All-solid-state lithium secondary batteries with glasses and glass-ceramics as a solid electrolyte were fabricated, and their performance is reported here. 2. Experimental 2.1. Preparation of glass-based solid electrolytes Li2S-based sulfide and oxysulfide glasses were prepared by melt-quenching and mechanical milling techniques. In the former case, the mixture of staring materials such as Li2S, SiS2, and Li4SiO4 in the composition of (100 − x)(0.6Li2S · 0.4SiS2) · xLi4SiO4 (mol%) was melted in a carbon crucible at 1000– 1100 °C for 2 h in a dry N2-filled glove box. The molten samples were rapidly cooled using a twin-roller quenching apparatus to prepare flake-like glasses with a thickness of 20 μm. The glasses were also prepared by the mechanochemical method using a planetary ball mill apparatus (Fritsch Pulverisette 7). The mechanochemical treatment using an Al2O3 pot and balls was performed for the mixture of Li2S and P2S5 crystals. Fine glassy powders were obtained after mechanical milling for several hours at a constant rotation speed of 370 rpm. All the processes were carried out at room temperature in a dry Ar-filled glove box. Glass-ceramic materials were prepared by heating the mechanically milled glasses over their crystallization temperatures. 2.2. Characterization of glass-based solid electrolytes
3. Results and discussion 3.1. Li2S-based oxysulfide glassy electrolytes — the utilization of mixed-anion effect Ionic conducting glasses are commonly prepared by the melt-quenching method. A large number of studies revealed that increase in lithium ion concentration in glassy materials is a key point to achieve higher electrical conductivity and lower activation energy for Li+ transport. A rapid quenching technique using a twin-roller apparatus is useful to expand the glassforming region, and thereby oxide and sulfide glasses with large amounts of lithium ions have been prepared. One of the techniques to improve Li+ conductivity is the utilization of “mixed-anion effect” [9,10]. The “mixed-anion effect” is defined as the phenomenon that the enhancement or maximum in composition dependence of conductivity is observed when two different types of anions are mixed. For example, the ionic glasses in the system Li4SiO4–Li3BO3 exhibit the maximum conductivity at the composition with the equal mole of two anions: SiO44− and BO33−. On the basis of the oxide mixed-anion systems, the oxysulfide glasses with combining SiS44− and SiO44− anions have been prepared, and the mixing effects of the two anions on electrical properties and local structure have been investigated. Fig. 1 shows the composition dependence of electrical conductivities at 25 °C (σ25) and Tc − Tg of the oxysulfide glasses in the systems (100 − x)(0.6Li2S · 0.4SiS2) · xLixMOy (M = Si, P, Ge, B, Al, Ga and In) (mol%) prepared by the twin-roller quenching technique [8,11,12]. The solid lines are a guide to the eye. The addition of 5 mol% of LixMOy (M = Si, P, Ge, B, and
Differential thermal analysis (DTA) was carried out using a thermal analyzer (Rigaku, Thermo-plus 8110) for the glassy powders sealed in an Al pan in order to determine their glass transition temperature (Tg) and crystallization temperature (Tc). Electrical conductivities were measured for the flake-like glasses or pelletized glasses for the powders obtained by mechanical milling. AC impedance measurements were carried out in dry Ar atmosphere using a Solartron 1260 impedance analyzer in a frequency range of 100 Hz to 15 MHz. X-ray diffraction (XRD) measurements (CuKα) were performed using a diffractometer (M18XHF22-SRA, MAC Science) to identify crystals in glass-ceramics. 2.3. Fabrication of all-solid-state batteries The composite positive electrode materials were prepared by mixing LiCoO2, the glass-ceramics and acetylene-black with the weight ratio of 20:30:3. The composite powder (20 mg) acting as a positive electrode and the glass-ceramics powder (80 mg) acting as a solid electrolyte were placed in a polycarbonate tube (φ = 10 mm) and pressed together under 3.6 × 108 Pa, and then an indium foil with a thickness of 0.1 mm as a negative electrode was pressed under 2.5 × 108 Pa on the pellet. The obtained In/LiCoO2 cells were charged and discharged at room temperature in an Ar atmosphere.
Fig. 1. Composition dependence of the ambient temperature conductivity (σ25) and Tc − Tg of the oxysulfide glasses in the (100 − x)(0.6Li2S · 0.4SiS2) · xLixMOy (LixMOy = Li4SiO4, Li3PO4, Li4GeO4, Li3BO3, Li3AlO3, Li3GaO3 and Li3InO3) systems.
T. Minami et al. / Solid State Ionics 177 (2006) 2715–2720
Al) to the 60Li2S · 40SiS2 sulfide system increases Tc − Tg as a measure of glass stability towards crystallization. Further addition of LixMOy decreases Tc − Tg in all the systems. On the other hand, the addition of 5 mol% of LixMOy keeps the high conductivity of 10− 3 S cm− 1, while further addition of LixMOy decreases the conductivity. The addition of oxide to sulfide is generally understood to monotonically decrease conductivity because oxide glasses show much lower conductivity than sulfide glasses. Nevertheless it is noteworthy that the oxysulfide glasses with 5 mol% of LixMOy maintain a high conductivity of 10− 3 S cm− 1. The electrochemical stability of the oxysulfide glasses against Li electrode was also examined by cyclic voltammetry [13]. The oxysulfide glass with 5 mol% Li4SiO4 exhibited a wide electrochemical window at least more than 10 V, while the addition of 20 mol% Li4SiO4 lowered the electrochemical stability. The oxysulfide glasses with small amounts of LixMOy exhibited high conductivity and electrochemical stability. Local structure of the glasses was investigated by several spectroscopic techniques. The main structural unit estimated by solidstate NMR and XPS in the oxysulfide glasses with small amounts of LixMOy is shown in Fig. 2 [14]. In the Si2OS66− dimer unit, silicon atoms are coordinated with six nonbridging sulfur atoms and one bridging oxygen (BO) atom, which would work as a weak trap of lithium ions. The reason for a low conductivity of lithium oxide glasses is based on the presence of a strong trap of nonbridging oxygen (NBO) atoms for lithium ions. NBO is not mainly present in the oxysulfide glasses added with 5 mol% of LixMOy, while further addition of LixMOy increased the NBO content in the oxysulfide glasses. The doping of oxygen as BO must be responsible for high conductivity and wide electrochemical window of the oxysulfide glasses. The Li2S–SiS2–Li4SiO4 oxysulfide glasses were also synthesized by mechanochemical treatment using a planetary ball mill apparatus [15,16]. Mechanochemical synthesis has the following advantages as a new preparation technique of glasses: the whole process is performed at room temperature, and fine electrolyte powders, which can be directly applied to lithium secondary batteries, are obtained without an additional pulverizing procedure. Fig. 3 shows the temperature dependence of electrical conductivity of the 95(0.6Li2S · 0.4SiS2) · 5 Li4SiO4 (mol%) oxysulfide samples prepared by mechanical milling. Numbers in the figure denote the milling period and the conductivity data were obtained on compressed pellets. The conductivity of the as-mixed sample is much less than the order of 10− 9 S cm− 1, while the conductivities of mechanically milled samples dramatically increase with an increase in the milling
Fig. 2. Main structural unit expected to be present in the 95(0.6Li2S · 0.4SiS2) · 5Li4SiO4 oxysulfide glass.
2717
Fig. 3. Temperature dependence of electrical conductivities of the 95 (0.6Li2S · 0.4SiS2) · 5Li4SiO4 oxysulfide samples prepared by mechanical milling.
period. The conductivity of the sample milled for more than 5 h is higher than 10− 4 S cm− 1 at room temperature, which is comparable to the conductivity of the corresponding quenched glass of a compressed powder pellet. Activation energies for conduction decrease from 78 to 32 kJ mol− 1 with increasing the milling period. It is noteworthy that the milling for only 1 h drastically increases the conductivity of the oxysulfide samples by more than three orders of magnitude. 29Si MAS-NMR spectra [15] revealed that the SiS4 and SiOS3 tetrahedral units, which are main units in the highly conductive Li2S–SiS2– Li4SiO4 melt-quenched glasses, were formed, and the SiS4 units with two edge-sharing in the SiS2 crystal almost vanished after milling for 1 h. The NMR spectrum for the glass prepared by milling for 20 h was very similar to that of the corresponding melt-quenched glass. The similarity of conductivity between milled and quenched glasses as shown in Fig. 3 is explained from the viewpoint of the similarity of their local structures. It has been considered that the amorphization mechanism by mechanical milling is not due to a local melt-quenching, but to a solid-state interdiffusion reaction [17], suggesting that there is a great possibility of preparing new amorphous materials which could not be prepared by melt-quenching. For example, the Li2S–Al2S3 binary glasses could not be prepared by a conventional melt-quenching technique. The Li2S–P2S5 glasses can be prepared by a melt-quenching method. However, the melting reaction has to be carried out in sealed quartz tubes because of high vapor pressure of P2S5. Recently, these glasses were synthesized by the mechanochemical method at room temperature under normal pressure [18,19]. Fig. 4 shows the composition dependence of electrical conductivities at 25 °C (σ25) and activation energies for conduction (Eα) for the Li2S– MxSy (M = Al, Si and P) glasses prepared by mechanical milling. In all the systems, the σ25 values increase with an increase in the Li2S content and maximize at the composition 60 mol% Li2S. Further increase in the Li2S content decreases the σ25 values. It
2718
T. Minami et al. / Solid State Ionics 177 (2006) 2715–2720
Fig. 5. Temperature dependence of electrical conductivities of the 54Li2S · 46PS2.5 (=70Li2S · 30P2S5) glass and glass-ceramic solid electrolytes. The conductivities of the crystals obtained by conventional solid-state reaction are also shown. Fig. 4. Composition dependence of electrical conductivities at 25 °C (σ25) and activation energies for conduction (Ea) for the Li2S–MxSy (M = Al, Si and P) glasses prepared by mechanical milling.
is noteworthy that the conductivities in the Li2S-MxSy (M = Al, Si and P) systems are maximized at nearly the same Li2S content. The composition dependence of the Ea values corresponds to that of σ25 values. The 60Li2S · 40SiS2 and 60Li2S · 40PS2.5 (mol%) glasses exhibit higher conductivity of 10− 4 S cm− 1 at room temperature than the 60Li2S · 40AlS1.5 (mol%) glass. The SiS2-based glasses using several lithium sources such as Li3N [20] and Li2O [21] instead of Li2S were also prepared by the mechanochemical method, and exhibited Li+ conductivity over 10− 5 S cm− 1 at room temperature.
ceramic is the lowest value in the lithium ion conductor reported so far. On the other hand, the crystal obtained by conventional solid-state reaction of Li2S and P2S5 exhibits much lower ambient temperature conductivity of 2.6 × 10− 8 S · cm− 1 and a higher activation energy of 55 kJ · mol− 1 for conduction. The conductivity enhancement is supposed to be mainly due to the precipitation of superionic crystals from the Li2S–P2S5 glasses. In order to identify crystalline phases precipitated from the 54Li2S · 46PS2.5 glass, XRD measurements were performed for the glass-ceramic. Fig. 6 shows XRD patterns of (a) the
3.2. Li2S-based glass-ceramic electrolytes — the precipitation of metastable crystal from glass In general, crystallization of glassy materials is well known to lower conductivities [3]. However, the enhancement of conductivity by careful heat-treatment of the Li2S–P2S5 glasses has been reported [22–24]. Fig. 5 shows the temperature dependences of electrical conductivities of the 54Li2S · 46PS2.5 (mol%, the composition corresponds to 70Li2S · 30P2S5) samples prepared by mechanical milling and solid-state reaction. Open and filled circles respectively denote the mechanically milled glass and the glass-ceramic obtained by crystallization of the glass at 360 °C. Filled diamonds denote the conductivity of the corresponding sample obtained by solid-state reaction of Li2S and P2S5. The ambient temperature conductivity of the 54Li2S · 46PS2.5 glass is 5.4 × 10−5 S · cm− 1. The conductivities of the glass-ceramic are much higher in the whole temperature range than those of the pristine glass; the conductivity of the glass-ceramic is 3.2 × 10− 3 S · cm− 1 at room temperature. Even if the glass-ceramic was repeatedly heated and cooled, the temperature dependence of conductivity was almost the same as that of the glass-ceramic shown in Fig. 5 (filled circles). The activation energy of the glass is 38 kJ · mol− 1, while the one of the glass-ceramic is 12 kJ · mol− 1. The activation energy (12 kJ· mol− 1) of the glass-
Fig. 6. XRD patterns of (a) the 54Li2S · 46PS2.5 (=70Li2S · 30P2S5) glass, (b) the glass-ceramic and (c) the crystals obtained by solid-state reaction.
T. Minami et al. / Solid State Ionics 177 (2006) 2715–2720
2719
3.3. Cycling performance of all-solid-state battery with Li2S– P2S5 glass-ceramic electrolytes
Fig. 7. Charge–discharge curves at the 500th cycle of In/LiCoO2 cells with the 67Li2S · 33PS2.5 (=80Li2S · 20P2S5) glass-ceramic. The inset shows cycling performance of the rechargeable capacities and charge–discharge efficiencies for the cell under the current density of 64 μA cm− 2.
54Li2S · 46PS2.5 glass, (b) the glass-ceramic and (c) the crystal obtained by solid-state reaction. XRD peaks in (c) are attributable to both thermodynamically stable crystals Li3PS4 and Li4P2S6, and these crystals are known to show low lithium ion conductivities smaller than 10− 7 S · cm− 1 at room temperature. On the other hand, the XRD pattern of the glass-ceramic in (b) is different from that of thermodynamically stable crystals in the Li2S–P2S5 system reported so far, such as Li3PS4, Li4P2S6 and Li7PS6 (JCPDS #34-0688). Furthermore, the pattern is also different from that of a series of thio-LISICONs such as the Li2S–P2S5 [25] and Li4GeS4–Li3PS4 [26] solid solutions, whose analogues were formed in the Li2S–P2S5 glass-ceramics at the compositions with 60 mol% Li2S or more [23]. Thus, the XRD pattern shown in Fig. 6 (b) implies the existence of a new crystal which is precipitated only by crystallization of the glass and could not be prepared by the solidstate reaction. Since the Li4P2S6 and Li3PS4 crystals were formed in the glass-ceramic heated at 550 °C, the precipitated new crystal would be a metastable one. Raman spectra of the glass ceramic revealed that the metastable crystal is mainly composed of the P2S74− (pyrothiophosphate) ions [24], suggesting that the crystal structure is different from that of the socalled best lithium ion conductor, thio-LISICONs consisting of PS44− (ortho-thiophosphate) ions [26]. Diffraction studies for the clarification of the structure of the metastable crystal are now in progress. Improvement in crystallinity of the metastable crystal is a key point to further enhance the conductivity of the glassceramics. Optimizing a heat treatment process, which controls both nucleation and crystal growth, for preparing the glassceramics is required.
The all-solid-state In/LiCoO2 cells using Li2S–P2S5 glassceramics as a solid electrolyte were assembled [27,28]. The glassceramics with ambient temperature conductivity of about 10− 3 S cm− 1 were prepared by heat-treatment of the mechanically milled 67Li2S · 33PS2.5 (mol%, the compositon corresponds to 80Li2S · 20P2S5) glass [23]. Fig. 7 shows charge–discharge curves of the all-solid-state cell at the 500th cycle under constant current density of 64 μA cm− 2. The inset shows cycling performance of the cell. Although an irreversible capacity is initially observed at the first few cycles, the all-solid-state cell maintains the reversible capacity of about 100 mA h g− 1 and the charge–discharge efficiency of 100% (no irreversible capacity) for 500 cycles, suggesting that the cell works as a lithium secondary battery without the decomposition of the glassy electrolyte. The excellent cycling performance with no capacity loss is achieved by using the highly conductive Li2S–P2S5 glass-ceramic electrolytes. In order to improve the energy density of batteries, the development of positive electrode materials is important because in general the capacity of positive electrode materials is much lower than that of negative electrode materials. Elemental sulfur has been of great interest as a positive electrode material because of its large theoretical capacity of 1672 mA h g− 1, low cost, and environmental friendliness. Unfortunately, the Li/S batteries with conventional liquid electrolytes suffer from rapid capacity fading on cycling because polysulfides formed during a discharge process dissolved in liquid electrolytes [29–31]. The utilization of inorganic solid electrolytes resolves a key problem in Li/S batteries. The glass-based solid electrolytes in the systems Li2S– SiS2 and Li2S–P2S5 have been used for all-solid-state Li/S batteries [32,33]. Fig. 8 shows the cycling performance of an allsolid-state Li–In/S–Cu cell. The positive electrode materials with S particles covered with CuS were syntheized by mechanical
Fig. 8. Cycling performance of an all-solid-state cell Li–In/67Li2S · 33PS2.5 (=80Li2S·20P2S5) glass-ceramic/sulfur-based mixture (S/Cu = 3).
2720
T. Minami et al. / Solid State Ionics 177 (2006) 2715–2720
milling for 15 min from S and Cu powders with the molar ratio of S/Cu = 3 [33]. The 67Li2S · 33PS2.5 glass-ceramic was used as a solid electrolyte. The all-solid-state cell works as rechargeable lithium batteries at room temperature under a constant current density of 64 μA cm− 2. The cell retains high reversible capacity over 650 mA h g− 1 (corresponding to about 1100 mA h per weight of sulfur) for 20 cycles. The CuS crystal, one component in the composite electrode, is known to act as an active material for lithium secondary batteries [34]. The obtained capacity is almost twice as large as the theoretical capacity of CuS, suggesting that S as well as CuS in the positive electrode is reversibly utilized as active materials on the charge–discharge cycling process. The Li4/3Ti5/3O4 crystal [28] or SnO/SnS-based glassy materials [35,36] were used as a negative electrode material for all-solid-state batteries. The cell with Li4/3Ti5/3O4 showed superior cycling performance of retaining 130 mA h g− 1 for 300 cycles, suggesting that the favorable feature of negligible volume change during charge-discharge processes [37] suppresses the loss of electrical contact in composite electrode materials. The Sn-based glasses such as SnO–B2O3 and SnS– P2S5 systems exhibited the constant capacity of about 400 mA h g− 1 for 50 cycles. In particular, the use of SnS–P2S5 as an electrode provides a continuous P2S5 glassy network on solid interface between electrode and electrolytes, which would reduce the interfacial polarization during ion transfer. The Li2S– P2S5 glass-ceramic electrolytes are compatible with various electrode materials for all-solid-state cells with excellent cycling performance. 4. Conclusions The conductivity and local structure of Li2S-based oxysulfide glasses and sulfide glass-ceramics were demonstrated. The utilization of “mixed-anion effect” and the precipitation of superionic metastable crystals from glasses are effective ways to improve the Li+ conductivity of glass-based solid electrolytes. In particular, the superior Li+ conducting behavior was achieved in the Li2S–P2S5 glass-ceramics with superionic metastable crystals, which were synthesized by careful heat-treatment of the corresponding glasses. Structural analysis reveal that characteristic pyro-structural anions such as Si2OS66− and P2S74− play an important role in the fast diffusion of Li+ in the glass-based solid electrolytes studied here. The cycling performance of a bulk-type all-solid-state lithium secondary batteries is developed by use of highly conductive glass-ceramic electrolytes in the system Li2S– P2S5. The enhancement of Li+ conductivity by increasing the crystallinity of superionic metastable crystals, and the formation of nanosized solid interface where electrolyte and electrode are closely attached, are the next key issues to be addressed in developing much better properties of all-solid-state batteries for practical use. Acknowledgment This work was supported by the Grant-in-Aid for Scientific Research on Priority Areas (B) and Section (B) from the
Ministry of Education, Culture, Sports, Science and Technology of Japan. References [1] T. Minami, Y. Takuma, M. Tanaka, J. Electrochem. Soc. 124 (1977) 1659. [2] T. Minami, H. Nambu, M. Tanaka, J. Am. Ceram. Soc. 60 (1977) 467. [3] T. Minami, N. Machida, Mater. Sci. Eng., B, Solid-State Mater. Adv. Technol. 13 (1992) 203. [4] T. Minami, Bull. Inst. Chem. Res., Kyoto Univ. 72 (1994) 305. [5] T. Minami, J. Non-Cryst. Solids 73 (1985) 273. [6] R.G. Pearson, J. Chem. Educ. 45 (1968) 581. [7] N. Machida, T. Minami, J. Am. Ceram. Soc. 71 (1988) 784. [8] T. Minami, A. Hayashi, M. Tatsumisago, Solid State Ionics 136-137 (2000) 1015. [9] T. Minami, N. Machida, Mater. Chem. Phys. 23 (1989) 63. [10] M. Tatsumisago, N. Machida, T. Minami, J. Ceram. Soc. Jpn. 95 (1987) 197. [11] S. Kondo, K. Takada, Y. Yamamura, Solid State Ionics 53-56 (1992) 1183. [12] M. Tatsumisago, K. Hirai, T. Minami, K. Takada, S. Kondo, J. Ceram. Soc. Jpn. 101 (1993) 1315. [13] A. Hayashi, M. Tatsumisago, T. Minami, J. Electrochem. Soc. 146 (1999) 3472. [14] A. Hayashi, M. Tatsumisago, T. Minami, Y. Miura, Phys. Chem. Glasses 39 (1998) 145. [15] H. Morimoto, H. Yamashita, M. Tatsumisago, T. Minami, J. Ceram. Soc. Jpn. 108 (2000) 128. [16] M. Tatsumisago, H. Yamashita, A. Hayashi, H. Morimoto, T. Minami, J. Non-Cryst. Solids 274 (2000) 30. [17] R.B. Schwarz, C.C. Koch, Appl. Phys. Lett. 49 (1986) 146. [18] A. Hayashi, S. Hama, H. Morimoto, M. Tatsumisago, T. Minami, J. Am. Ceram. Soc. 84 (2001) 477. [19] A. Hayashi, T. Fukuda, S. Hama, H. Yamashita, H. Morimoto, T. Minami, M. Tatsumisago, J. Ceram. Soc. Jpn. 112 (2004) S695. [20] K. Iio, A. Hayashi, H. Morimoto, M. Tatsumisago, T. Minami, Chem. Mater. 14 (2002) 2444. [21] A. Hayashi, K. Iio, H. Morimoto, T. Minami, M. Tatsumisago, Solid State Ionics 175 (2004) 637. [22] A. Hayashi, S. Hama, H. Morimoto, M. Tatsumisago, T. Minami, Chem. Lett. (2001) 872. [23] A. Hayashi, S. Hama, T. Minami, M. Tatsumisago, Electrochem. Commun. 5 (2003) 111. [24] F. Mizuno, A. Hayashi, K. Tadanaga, M. Tatsumisago, Adv. Mater. 17 (2005) 918. [25] M. Murayama, N. Sonoyama, R. Kanno, Solid State Ionics 170 (2004) 173. [26] R. Kanno, M. Murayama, J. Electrochem. Soc. 148 (2001) 742. [27] F. Mizuno, S. Hama, A. Hayashi, K. Tadanaga, T.. Minami, M. Tatsumisago, Chem. Lett. 2002 (2002) 1244. [28] M. Tatsumisago, Solid State Ionics 175 (2004) 13. [29] H. Yamin, A. Gorenshtein, J. Penciner, Y. Sternberg, E. Peled, J. Electrochem. Soc. 135 (1988) 1045. [30] D. Marmorstein, T.H. Yu, K.A. Striebel, F.R. McLarnon, J. Hou, E.J. Cairns, J. Power Sources 89 (2000) 219. [31] J. Shim, K.A. Striebel, E.J. Cairns, J. Electrochem. Soc. 149 (2002) A1321. [32] N. Machida, K. Kobayashi, Y. Nishikawa, T. Shigematsu, Solid State Ionics 175 (2004) 247. [33] A. Hayashi, T. Ohtomo, F. Mizuno, K. Tadanaga, M. Tatsumisago, Electrochem. commun. 5 (2003) 701. [34] J.S. Chung, H.J. Sohn, J. Power Sources 108 (2002) 226. [35] A. Hayashi, M. Nakai, M. Tatsumisago, T. Minami, M. Katada, J. Electrochem. Soc. 150 (2003) A582. [36] A. Hayashi, T. Konishi, K. Tadanaga, T. Minami, M. Tatsumisago, J. Power Sources 146 (2005) 496. [37] T. Ohzuku, A. Ueda, N. Yamamoto, J. Electrochem. Soc. 142 (1995) 1431.
Recent progress of glass and glass-ceramics as solid electrolytes for lithium secondary batteries Tsutomu Minami ⁎, Akitoshi Hayashi, Masahiro Tatsumisago Department of Applied Chemistry, Graduate School of Engineering, Osaka Prefecture University, Japan Received 4 October 2005; received in revised form 20 April 2006; accepted 7 July 2006
Abstract In recent years many new glass-based solid electrolytes with high Li+ conductivity have been developed. In the present paper, we review the preparation and characterization of Li2S-based oxysulfide glasses and sulfide glass-ceramics on the basis of two strategies of enhancing Li+ conductivity: the utilization of “mixed-anion effect” by combining sulfide and oxide anions, and the precipitation of superionic metastable crystals by careful heat-treatment of glasses. The superior Li+ conducting solid electrolytes with the highest conductivity and the lowest activation energy for conduction have been achieved in the Li2S–P2S5 glass-ceramics. The use of these glass-ceramic solid electrolytes leads to the development of a bulk-type all solid-state lithium secondary battery with excellent cycling performance. © 2006 Elsevier B.V. All rights reserved. Keywords: Glass; Glass-ceramic; Lithium ion conductivity; Solid electrolyte; All-solid-state; Battery
1. Introduction Realization of fast ion conduction in solids is indispensable to producing all-solid-state devices such as batteries and chemical sensors with high safety and reliability. Inorganic ionic conductors, namely solid electrolytes, have been extensively studied in two types of solids: crystal and glass. The ionic conductivity of glassy materials is, in general, higher than that of the corresponding crystals [1–4]. In the system AgI–Ag2O–P2O5, the ambient temperature conductivity of glasses is 10− 2 S cm− 1, which is higher by 1–2 orders of magnitude than the conductivity of crystals. These glasses are named “superionic conducting glasses”. The advantage of glass in conductivity is explained by the structure of glass, which is very similar to that of the corresponding melt with high conductivity and low activation energy for conduction. In ionic conductors, Ag+ ions are known to show higher conductivity than alkali ions in spite of the fact that Ag+ ions have a large ionic radius and atomic weight. The electronic configuration of
⁎ Corresponding author. E-mail address: [email protected] (T. Minami). 0167-2738/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ssi.2006.07.017
the ions, i.e. the 4d10 configuration in the outermost orbital, plays an important role as well as the ionic radius in achieving high ion conductivities [3–5]. The most favorable ion species is Li+ because lithium secondary batteries have superior features such as high energy density and low weight. Three important discoveries for creating high Li+ conducting electrolytes have been obtained: (1) the change of an oxide matrix to a sulfide one, (2) the utilization of so-called the “mixed-anion effect” by combining sulfide and oxide anions, and (3) the precipitation of superionic metastable crystals by careful heattreatment of glasses. First of all, the development of Li+ ion conducting glasses has been achieved by changing the glass matrix from oxides to sulfides. Li+ ions acting as a “hard acid”, which is classified from a viewpoint of the “hard and soft acids and bases theory” by Pearson [6], would be more compatible to sulfide ions acting as a “soft base” [7]. In fact, the conductivity of Li2S–SiS2 sulfide glasses is in the order of 10− 4 S cm− 1 at room temperature, which can be compared with the conductivity of 10−7 S cm− 1 for the corresponding Li2O–SiO2 glasses [5,8]. In this paper, we focus on the other two strategies for developing Li+ conductivity of glass-based solid electrolytes and review the preparation and characterization of Li2S-based
2716
T. Minami et al. / Solid State Ionics 177 (2006) 2715–2720
oxysulfide glasses and sulfide glass-ceramics. The rapid quenching method using a twin-roller apparatus and the mechanochemical method using a planetary ball mill apparatus were used for preparation of glasses. All-solid-state lithium secondary batteries with glasses and glass-ceramics as a solid electrolyte were fabricated, and their performance is reported here. 2. Experimental 2.1. Preparation of glass-based solid electrolytes Li2S-based sulfide and oxysulfide glasses were prepared by melt-quenching and mechanical milling techniques. In the former case, the mixture of staring materials such as Li2S, SiS2, and Li4SiO4 in the composition of (100 − x)(0.6Li2S · 0.4SiS2) · xLi4SiO4 (mol%) was melted in a carbon crucible at 1000– 1100 °C for 2 h in a dry N2-filled glove box. The molten samples were rapidly cooled using a twin-roller quenching apparatus to prepare flake-like glasses with a thickness of 20 μm. The glasses were also prepared by the mechanochemical method using a planetary ball mill apparatus (Fritsch Pulverisette 7). The mechanochemical treatment using an Al2O3 pot and balls was performed for the mixture of Li2S and P2S5 crystals. Fine glassy powders were obtained after mechanical milling for several hours at a constant rotation speed of 370 rpm. All the processes were carried out at room temperature in a dry Ar-filled glove box. Glass-ceramic materials were prepared by heating the mechanically milled glasses over their crystallization temperatures. 2.2. Characterization of glass-based solid electrolytes
3. Results and discussion 3.1. Li2S-based oxysulfide glassy electrolytes — the utilization of mixed-anion effect Ionic conducting glasses are commonly prepared by the melt-quenching method. A large number of studies revealed that increase in lithium ion concentration in glassy materials is a key point to achieve higher electrical conductivity and lower activation energy for Li+ transport. A rapid quenching technique using a twin-roller apparatus is useful to expand the glassforming region, and thereby oxide and sulfide glasses with large amounts of lithium ions have been prepared. One of the techniques to improve Li+ conductivity is the utilization of “mixed-anion effect” [9,10]. The “mixed-anion effect” is defined as the phenomenon that the enhancement or maximum in composition dependence of conductivity is observed when two different types of anions are mixed. For example, the ionic glasses in the system Li4SiO4–Li3BO3 exhibit the maximum conductivity at the composition with the equal mole of two anions: SiO44− and BO33−. On the basis of the oxide mixed-anion systems, the oxysulfide glasses with combining SiS44− and SiO44− anions have been prepared, and the mixing effects of the two anions on electrical properties and local structure have been investigated. Fig. 1 shows the composition dependence of electrical conductivities at 25 °C (σ25) and Tc − Tg of the oxysulfide glasses in the systems (100 − x)(0.6Li2S · 0.4SiS2) · xLixMOy (M = Si, P, Ge, B, Al, Ga and In) (mol%) prepared by the twin-roller quenching technique [8,11,12]. The solid lines are a guide to the eye. The addition of 5 mol% of LixMOy (M = Si, P, Ge, B, and
Differential thermal analysis (DTA) was carried out using a thermal analyzer (Rigaku, Thermo-plus 8110) for the glassy powders sealed in an Al pan in order to determine their glass transition temperature (Tg) and crystallization temperature (Tc). Electrical conductivities were measured for the flake-like glasses or pelletized glasses for the powders obtained by mechanical milling. AC impedance measurements were carried out in dry Ar atmosphere using a Solartron 1260 impedance analyzer in a frequency range of 100 Hz to 15 MHz. X-ray diffraction (XRD) measurements (CuKα) were performed using a diffractometer (M18XHF22-SRA, MAC Science) to identify crystals in glass-ceramics. 2.3. Fabrication of all-solid-state batteries The composite positive electrode materials were prepared by mixing LiCoO2, the glass-ceramics and acetylene-black with the weight ratio of 20:30:3. The composite powder (20 mg) acting as a positive electrode and the glass-ceramics powder (80 mg) acting as a solid electrolyte were placed in a polycarbonate tube (φ = 10 mm) and pressed together under 3.6 × 108 Pa, and then an indium foil with a thickness of 0.1 mm as a negative electrode was pressed under 2.5 × 108 Pa on the pellet. The obtained In/LiCoO2 cells were charged and discharged at room temperature in an Ar atmosphere.
Fig. 1. Composition dependence of the ambient temperature conductivity (σ25) and Tc − Tg of the oxysulfide glasses in the (100 − x)(0.6Li2S · 0.4SiS2) · xLixMOy (LixMOy = Li4SiO4, Li3PO4, Li4GeO4, Li3BO3, Li3AlO3, Li3GaO3 and Li3InO3) systems.
T. Minami et al. / Solid State Ionics 177 (2006) 2715–2720
Al) to the 60Li2S · 40SiS2 sulfide system increases Tc − Tg as a measure of glass stability towards crystallization. Further addition of LixMOy decreases Tc − Tg in all the systems. On the other hand, the addition of 5 mol% of LixMOy keeps the high conductivity of 10− 3 S cm− 1, while further addition of LixMOy decreases the conductivity. The addition of oxide to sulfide is generally understood to monotonically decrease conductivity because oxide glasses show much lower conductivity than sulfide glasses. Nevertheless it is noteworthy that the oxysulfide glasses with 5 mol% of LixMOy maintain a high conductivity of 10− 3 S cm− 1. The electrochemical stability of the oxysulfide glasses against Li electrode was also examined by cyclic voltammetry [13]. The oxysulfide glass with 5 mol% Li4SiO4 exhibited a wide electrochemical window at least more than 10 V, while the addition of 20 mol% Li4SiO4 lowered the electrochemical stability. The oxysulfide glasses with small amounts of LixMOy exhibited high conductivity and electrochemical stability. Local structure of the glasses was investigated by several spectroscopic techniques. The main structural unit estimated by solidstate NMR and XPS in the oxysulfide glasses with small amounts of LixMOy is shown in Fig. 2 [14]. In the Si2OS66− dimer unit, silicon atoms are coordinated with six nonbridging sulfur atoms and one bridging oxygen (BO) atom, which would work as a weak trap of lithium ions. The reason for a low conductivity of lithium oxide glasses is based on the presence of a strong trap of nonbridging oxygen (NBO) atoms for lithium ions. NBO is not mainly present in the oxysulfide glasses added with 5 mol% of LixMOy, while further addition of LixMOy increased the NBO content in the oxysulfide glasses. The doping of oxygen as BO must be responsible for high conductivity and wide electrochemical window of the oxysulfide glasses. The Li2S–SiS2–Li4SiO4 oxysulfide glasses were also synthesized by mechanochemical treatment using a planetary ball mill apparatus [15,16]. Mechanochemical synthesis has the following advantages as a new preparation technique of glasses: the whole process is performed at room temperature, and fine electrolyte powders, which can be directly applied to lithium secondary batteries, are obtained without an additional pulverizing procedure. Fig. 3 shows the temperature dependence of electrical conductivity of the 95(0.6Li2S · 0.4SiS2) · 5 Li4SiO4 (mol%) oxysulfide samples prepared by mechanical milling. Numbers in the figure denote the milling period and the conductivity data were obtained on compressed pellets. The conductivity of the as-mixed sample is much less than the order of 10− 9 S cm− 1, while the conductivities of mechanically milled samples dramatically increase with an increase in the milling
Fig. 2. Main structural unit expected to be present in the 95(0.6Li2S · 0.4SiS2) · 5Li4SiO4 oxysulfide glass.
2717
Fig. 3. Temperature dependence of electrical conductivities of the 95 (0.6Li2S · 0.4SiS2) · 5Li4SiO4 oxysulfide samples prepared by mechanical milling.
period. The conductivity of the sample milled for more than 5 h is higher than 10− 4 S cm− 1 at room temperature, which is comparable to the conductivity of the corresponding quenched glass of a compressed powder pellet. Activation energies for conduction decrease from 78 to 32 kJ mol− 1 with increasing the milling period. It is noteworthy that the milling for only 1 h drastically increases the conductivity of the oxysulfide samples by more than three orders of magnitude. 29Si MAS-NMR spectra [15] revealed that the SiS4 and SiOS3 tetrahedral units, which are main units in the highly conductive Li2S–SiS2– Li4SiO4 melt-quenched glasses, were formed, and the SiS4 units with two edge-sharing in the SiS2 crystal almost vanished after milling for 1 h. The NMR spectrum for the glass prepared by milling for 20 h was very similar to that of the corresponding melt-quenched glass. The similarity of conductivity between milled and quenched glasses as shown in Fig. 3 is explained from the viewpoint of the similarity of their local structures. It has been considered that the amorphization mechanism by mechanical milling is not due to a local melt-quenching, but to a solid-state interdiffusion reaction [17], suggesting that there is a great possibility of preparing new amorphous materials which could not be prepared by melt-quenching. For example, the Li2S–Al2S3 binary glasses could not be prepared by a conventional melt-quenching technique. The Li2S–P2S5 glasses can be prepared by a melt-quenching method. However, the melting reaction has to be carried out in sealed quartz tubes because of high vapor pressure of P2S5. Recently, these glasses were synthesized by the mechanochemical method at room temperature under normal pressure [18,19]. Fig. 4 shows the composition dependence of electrical conductivities at 25 °C (σ25) and activation energies for conduction (Eα) for the Li2S– MxSy (M = Al, Si and P) glasses prepared by mechanical milling. In all the systems, the σ25 values increase with an increase in the Li2S content and maximize at the composition 60 mol% Li2S. Further increase in the Li2S content decreases the σ25 values. It
2718
T. Minami et al. / Solid State Ionics 177 (2006) 2715–2720
Fig. 5. Temperature dependence of electrical conductivities of the 54Li2S · 46PS2.5 (=70Li2S · 30P2S5) glass and glass-ceramic solid electrolytes. The conductivities of the crystals obtained by conventional solid-state reaction are also shown. Fig. 4. Composition dependence of electrical conductivities at 25 °C (σ25) and activation energies for conduction (Ea) for the Li2S–MxSy (M = Al, Si and P) glasses prepared by mechanical milling.
is noteworthy that the conductivities in the Li2S-MxSy (M = Al, Si and P) systems are maximized at nearly the same Li2S content. The composition dependence of the Ea values corresponds to that of σ25 values. The 60Li2S · 40SiS2 and 60Li2S · 40PS2.5 (mol%) glasses exhibit higher conductivity of 10− 4 S cm− 1 at room temperature than the 60Li2S · 40AlS1.5 (mol%) glass. The SiS2-based glasses using several lithium sources such as Li3N [20] and Li2O [21] instead of Li2S were also prepared by the mechanochemical method, and exhibited Li+ conductivity over 10− 5 S cm− 1 at room temperature.
ceramic is the lowest value in the lithium ion conductor reported so far. On the other hand, the crystal obtained by conventional solid-state reaction of Li2S and P2S5 exhibits much lower ambient temperature conductivity of 2.6 × 10− 8 S · cm− 1 and a higher activation energy of 55 kJ · mol− 1 for conduction. The conductivity enhancement is supposed to be mainly due to the precipitation of superionic crystals from the Li2S–P2S5 glasses. In order to identify crystalline phases precipitated from the 54Li2S · 46PS2.5 glass, XRD measurements were performed for the glass-ceramic. Fig. 6 shows XRD patterns of (a) the
3.2. Li2S-based glass-ceramic electrolytes — the precipitation of metastable crystal from glass In general, crystallization of glassy materials is well known to lower conductivities [3]. However, the enhancement of conductivity by careful heat-treatment of the Li2S–P2S5 glasses has been reported [22–24]. Fig. 5 shows the temperature dependences of electrical conductivities of the 54Li2S · 46PS2.5 (mol%, the composition corresponds to 70Li2S · 30P2S5) samples prepared by mechanical milling and solid-state reaction. Open and filled circles respectively denote the mechanically milled glass and the glass-ceramic obtained by crystallization of the glass at 360 °C. Filled diamonds denote the conductivity of the corresponding sample obtained by solid-state reaction of Li2S and P2S5. The ambient temperature conductivity of the 54Li2S · 46PS2.5 glass is 5.4 × 10−5 S · cm− 1. The conductivities of the glass-ceramic are much higher in the whole temperature range than those of the pristine glass; the conductivity of the glass-ceramic is 3.2 × 10− 3 S · cm− 1 at room temperature. Even if the glass-ceramic was repeatedly heated and cooled, the temperature dependence of conductivity was almost the same as that of the glass-ceramic shown in Fig. 5 (filled circles). The activation energy of the glass is 38 kJ · mol− 1, while the one of the glass-ceramic is 12 kJ · mol− 1. The activation energy (12 kJ· mol− 1) of the glass-
Fig. 6. XRD patterns of (a) the 54Li2S · 46PS2.5 (=70Li2S · 30P2S5) glass, (b) the glass-ceramic and (c) the crystals obtained by solid-state reaction.
T. Minami et al. / Solid State Ionics 177 (2006) 2715–2720
2719
3.3. Cycling performance of all-solid-state battery with Li2S– P2S5 glass-ceramic electrolytes
Fig. 7. Charge–discharge curves at the 500th cycle of In/LiCoO2 cells with the 67Li2S · 33PS2.5 (=80Li2S · 20P2S5) glass-ceramic. The inset shows cycling performance of the rechargeable capacities and charge–discharge efficiencies for the cell under the current density of 64 μA cm− 2.
54Li2S · 46PS2.5 glass, (b) the glass-ceramic and (c) the crystal obtained by solid-state reaction. XRD peaks in (c) are attributable to both thermodynamically stable crystals Li3PS4 and Li4P2S6, and these crystals are known to show low lithium ion conductivities smaller than 10− 7 S · cm− 1 at room temperature. On the other hand, the XRD pattern of the glass-ceramic in (b) is different from that of thermodynamically stable crystals in the Li2S–P2S5 system reported so far, such as Li3PS4, Li4P2S6 and Li7PS6 (JCPDS #34-0688). Furthermore, the pattern is also different from that of a series of thio-LISICONs such as the Li2S–P2S5 [25] and Li4GeS4–Li3PS4 [26] solid solutions, whose analogues were formed in the Li2S–P2S5 glass-ceramics at the compositions with 60 mol% Li2S or more [23]. Thus, the XRD pattern shown in Fig. 6 (b) implies the existence of a new crystal which is precipitated only by crystallization of the glass and could not be prepared by the solidstate reaction. Since the Li4P2S6 and Li3PS4 crystals were formed in the glass-ceramic heated at 550 °C, the precipitated new crystal would be a metastable one. Raman spectra of the glass ceramic revealed that the metastable crystal is mainly composed of the P2S74− (pyrothiophosphate) ions [24], suggesting that the crystal structure is different from that of the socalled best lithium ion conductor, thio-LISICONs consisting of PS44− (ortho-thiophosphate) ions [26]. Diffraction studies for the clarification of the structure of the metastable crystal are now in progress. Improvement in crystallinity of the metastable crystal is a key point to further enhance the conductivity of the glassceramics. Optimizing a heat treatment process, which controls both nucleation and crystal growth, for preparing the glassceramics is required.
The all-solid-state In/LiCoO2 cells using Li2S–P2S5 glassceramics as a solid electrolyte were assembled [27,28]. The glassceramics with ambient temperature conductivity of about 10− 3 S cm− 1 were prepared by heat-treatment of the mechanically milled 67Li2S · 33PS2.5 (mol%, the compositon corresponds to 80Li2S · 20P2S5) glass [23]. Fig. 7 shows charge–discharge curves of the all-solid-state cell at the 500th cycle under constant current density of 64 μA cm− 2. The inset shows cycling performance of the cell. Although an irreversible capacity is initially observed at the first few cycles, the all-solid-state cell maintains the reversible capacity of about 100 mA h g− 1 and the charge–discharge efficiency of 100% (no irreversible capacity) for 500 cycles, suggesting that the cell works as a lithium secondary battery without the decomposition of the glassy electrolyte. The excellent cycling performance with no capacity loss is achieved by using the highly conductive Li2S–P2S5 glass-ceramic electrolytes. In order to improve the energy density of batteries, the development of positive electrode materials is important because in general the capacity of positive electrode materials is much lower than that of negative electrode materials. Elemental sulfur has been of great interest as a positive electrode material because of its large theoretical capacity of 1672 mA h g− 1, low cost, and environmental friendliness. Unfortunately, the Li/S batteries with conventional liquid electrolytes suffer from rapid capacity fading on cycling because polysulfides formed during a discharge process dissolved in liquid electrolytes [29–31]. The utilization of inorganic solid electrolytes resolves a key problem in Li/S batteries. The glass-based solid electrolytes in the systems Li2S– SiS2 and Li2S–P2S5 have been used for all-solid-state Li/S batteries [32,33]. Fig. 8 shows the cycling performance of an allsolid-state Li–In/S–Cu cell. The positive electrode materials with S particles covered with CuS were syntheized by mechanical
Fig. 8. Cycling performance of an all-solid-state cell Li–In/67Li2S · 33PS2.5 (=80Li2S·20P2S5) glass-ceramic/sulfur-based mixture (S/Cu = 3).
2720
T. Minami et al. / Solid State Ionics 177 (2006) 2715–2720
milling for 15 min from S and Cu powders with the molar ratio of S/Cu = 3 [33]. The 67Li2S · 33PS2.5 glass-ceramic was used as a solid electrolyte. The all-solid-state cell works as rechargeable lithium batteries at room temperature under a constant current density of 64 μA cm− 2. The cell retains high reversible capacity over 650 mA h g− 1 (corresponding to about 1100 mA h per weight of sulfur) for 20 cycles. The CuS crystal, one component in the composite electrode, is known to act as an active material for lithium secondary batteries [34]. The obtained capacity is almost twice as large as the theoretical capacity of CuS, suggesting that S as well as CuS in the positive electrode is reversibly utilized as active materials on the charge–discharge cycling process. The Li4/3Ti5/3O4 crystal [28] or SnO/SnS-based glassy materials [35,36] were used as a negative electrode material for all-solid-state batteries. The cell with Li4/3Ti5/3O4 showed superior cycling performance of retaining 130 mA h g− 1 for 300 cycles, suggesting that the favorable feature of negligible volume change during charge-discharge processes [37] suppresses the loss of electrical contact in composite electrode materials. The Sn-based glasses such as SnO–B2O3 and SnS– P2S5 systems exhibited the constant capacity of about 400 mA h g− 1 for 50 cycles. In particular, the use of SnS–P2S5 as an electrode provides a continuous P2S5 glassy network on solid interface between electrode and electrolytes, which would reduce the interfacial polarization during ion transfer. The Li2S– P2S5 glass-ceramic electrolytes are compatible with various electrode materials for all-solid-state cells with excellent cycling performance. 4. Conclusions The conductivity and local structure of Li2S-based oxysulfide glasses and sulfide glass-ceramics were demonstrated. The utilization of “mixed-anion effect” and the precipitation of superionic metastable crystals from glasses are effective ways to improve the Li+ conductivity of glass-based solid electrolytes. In particular, the superior Li+ conducting behavior was achieved in the Li2S–P2S5 glass-ceramics with superionic metastable crystals, which were synthesized by careful heat-treatment of the corresponding glasses. Structural analysis reveal that characteristic pyro-structural anions such as Si2OS66− and P2S74− play an important role in the fast diffusion of Li+ in the glass-based solid electrolytes studied here. The cycling performance of a bulk-type all-solid-state lithium secondary batteries is developed by use of highly conductive glass-ceramic electrolytes in the system Li2S– P2S5. The enhancement of Li+ conductivity by increasing the crystallinity of superionic metastable crystals, and the formation of nanosized solid interface where electrolyte and electrode are closely attached, are the next key issues to be addressed in developing much better properties of all-solid-state batteries for practical use. Acknowledgment This work was supported by the Grant-in-Aid for Scientific Research on Priority Areas (B) and Section (B) from the
Ministry of Education, Culture, Sports, Science and Technology of Japan. References [1] T. Minami, Y. Takuma, M. Tanaka, J. Electrochem. Soc. 124 (1977) 1659. [2] T. Minami, H. Nambu, M. Tanaka, J. Am. Ceram. Soc. 60 (1977) 467. [3] T. Minami, N. Machida, Mater. Sci. Eng., B, Solid-State Mater. Adv. Technol. 13 (1992) 203. [4] T. Minami, Bull. Inst. Chem. Res., Kyoto Univ. 72 (1994) 305. [5] T. Minami, J. Non-Cryst. Solids 73 (1985) 273. [6] R.G. Pearson, J. Chem. Educ. 45 (1968) 581. [7] N. Machida, T. Minami, J. Am. Ceram. Soc. 71 (1988) 784. [8] T. Minami, A. Hayashi, M. Tatsumisago, Solid State Ionics 136-137 (2000) 1015. [9] T. Minami, N. Machida, Mater. Chem. Phys. 23 (1989) 63. [10] M. Tatsumisago, N. Machida, T. Minami, J. Ceram. Soc. Jpn. 95 (1987) 197. [11] S. Kondo, K. Takada, Y. Yamamura, Solid State Ionics 53-56 (1992) 1183. [12] M. Tatsumisago, K. Hirai, T. Minami, K. Takada, S. Kondo, J. Ceram. Soc. Jpn. 101 (1993) 1315. [13] A. Hayashi, M. Tatsumisago, T. Minami, J. Electrochem. Soc. 146 (1999) 3472. [14] A. Hayashi, M. Tatsumisago, T. Minami, Y. Miura, Phys. Chem. Glasses 39 (1998) 145. [15] H. Morimoto, H. Yamashita, M. Tatsumisago, T. Minami, J. Ceram. Soc. Jpn. 108 (2000) 128. [16] M. Tatsumisago, H. Yamashita, A. Hayashi, H. Morimoto, T. Minami, J. Non-Cryst. Solids 274 (2000) 30. [17] R.B. Schwarz, C.C. Koch, Appl. Phys. Lett. 49 (1986) 146. [18] A. Hayashi, S. Hama, H. Morimoto, M. Tatsumisago, T. Minami, J. Am. Ceram. Soc. 84 (2001) 477. [19] A. Hayashi, T. Fukuda, S. Hama, H. Yamashita, H. Morimoto, T. Minami, M. Tatsumisago, J. Ceram. Soc. Jpn. 112 (2004) S695. [20] K. Iio, A. Hayashi, H. Morimoto, M. Tatsumisago, T. Minami, Chem. Mater. 14 (2002) 2444. [21] A. Hayashi, K. Iio, H. Morimoto, T. Minami, M. Tatsumisago, Solid State Ionics 175 (2004) 637. [22] A. Hayashi, S. Hama, H. Morimoto, M. Tatsumisago, T. Minami, Chem. Lett. (2001) 872. [23] A. Hayashi, S. Hama, T. Minami, M. Tatsumisago, Electrochem. Commun. 5 (2003) 111. [24] F. Mizuno, A. Hayashi, K. Tadanaga, M. Tatsumisago, Adv. Mater. 17 (2005) 918. [25] M. Murayama, N. Sonoyama, R. Kanno, Solid State Ionics 170 (2004) 173. [26] R. Kanno, M. Murayama, J. Electrochem. Soc. 148 (2001) 742. [27] F. Mizuno, S. Hama, A. Hayashi, K. Tadanaga, T.. Minami, M. Tatsumisago, Chem. Lett. 2002 (2002) 1244. [28] M. Tatsumisago, Solid State Ionics 175 (2004) 13. [29] H. Yamin, A. Gorenshtein, J. Penciner, Y. Sternberg, E. Peled, J. Electrochem. Soc. 135 (1988) 1045. [30] D. Marmorstein, T.H. Yu, K.A. Striebel, F.R. McLarnon, J. Hou, E.J. Cairns, J. Power Sources 89 (2000) 219. [31] J. Shim, K.A. Striebel, E.J. Cairns, J. Electrochem. Soc. 149 (2002) A1321. [32] N. Machida, K. Kobayashi, Y. Nishikawa, T. Shigematsu, Solid State Ionics 175 (2004) 247. [33] A. Hayashi, T. Ohtomo, F. Mizuno, K. Tadanaga, M. Tatsumisago, Electrochem. commun. 5 (2003) 701. [34] J.S. Chung, H.J. Sohn, J. Power Sources 108 (2002) 226. [35] A. Hayashi, M. Nakai, M. Tatsumisago, T. Minami, M. Katada, J. Electrochem. Soc. 150 (2003) A582. [36] A. Hayashi, T. Konishi, K. Tadanaga, T. Minami, M. Tatsumisago, J. Power Sources 146 (2005) 496. [37] T. Ohzuku, A. Ueda, N. Yamamoto, J. Electrochem. Soc. 142 (1995) 1431.