Formation and structure of Na2S + P2S5 amorphous materials prepared by melt-quenching and mechanical milling

Journal of Non-Crystalline Solids 358 (2012) 93–98

Contents lists available at SciVerse ScienceDirect

Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/ locate/ jnoncrysol

Formation and structure of Na2S + P2S5 amorphous materials prepared by melt-quenching and mechanical milling Seth S. Berbano 1, Inseok Seo, Christian M. Bischoff, Katherine E. Schuller, Steve W. Martin ⁎ Department of Materials Science and Engineering, Iowa State University, 2220 Hoover Hall, Ames, Iowa 50011, USA

a r t i c l e

i n f o

Article history: Received 19 July 2011 Received in revised form 30 August 2011 Available online 29 September 2011 Keywords: xNa2S + (1 – x)P2S5; Solid state electrolyte; Mechanical milling; Raman spectroscopy; Infrared spectroscopy

a b s t r a c t xNa2S + (1 − x)P2S5 amorphous and partially crystalline materials were prepared by melt-quenching and mechanical milling. These products were characterized using x-ray diffraction (XRD), Fourier-transform infrared spectroscopy (FT-IR), and Raman spectroscopy. Compared to the narrower x-ray amorphous range for this system obtained through melt-quenching as found in this study, 0.50 ≤ x ≤ 0.67, the x-ray amorphous range for this system could be extended from the low-alkali ultra-thiophosphate composition of x ~ 0.25 to slightly above the high-alkali pyro-thiophosphate composition of x ~ 0.70 using mechanical milling. Mechanically milled samples with Na2S of composition x = 0.75 yielded a partially crystalline material that had diffuse XRD peaks associated to the α-Na3PS4 phase. A similar result was obtained for the x = 0.80 composition except that, as expected, it also showed peaks for unreacted (over stoichiometric) Na2S. The melt-quenched and mechanically milled samples with the same compositions 0.50 ≤ x ≤ 0.67 showed similar FT-IR and Raman spectra, indicating very similar chemical short-range structures are present in both of these amorphous materials. It was found that the Na2S + P2S5 system exhibited similar behavior to that of the Li2S + P2S5 in that chemical reaction between Na2S and P2S5 could be induced by mechanical milling near room temperature to produce both amorphous and polycrystalline materials. © 2011 Elsevier B.V. All rights reserved.

1. Introduction The accelerating demand for lithium-ion batteries in electric vehicles and portable devices will cause increased demands for lithium. Batteries are also important in load leveling cells because of the intermittent nature of energy sources (solar, wind, and tide) and energy usage (peak vs. off-peak hours). Such cells would smooth fluctuations in the electrical grid, thereby optimizing energy collection, storage, and usage. Lithiumbased materials are usually more expensive than sodium-based materials. Therefore, sodium batteries may be better candidates than lithium batteries for large-scale energy storage systems if the lower cost of sodium-based materials is accompanied by higher performance, easier manufacture, and safer operation of solid-state Na batteries. Sodium/sulfur, Na/S, cells were heavily studied in the 1960s [1,2]. In these early Na/S cells, liquid sodium served as the anode, liquid sulfur as the cathode, and ceramic sodium beta-alumina as the solid electrolyte. At the Na/S battery's high operating temperature of ~ 300 °C, both Na and S are reactive melts and highly corrosive sodium polysulfides are present as by-products. Further, the high operating temperature forces a fraction of the energy to be used to maintain

⁎ Corresponding author. Tel.: + 1 515 294 0745; fax: + 1 515 294 5444. E-mail address: [email protected] (S.W. Martin). 1 Current address: Center for Dielectric Studies, Materials Research Institute, Materials Science and Engineering, The Pennsylvania State University, University Park, PA 16802, USA. 0022-3093/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jnoncrysol.2011.08.030

this operating temperature, resulting in a lower overall efficiency. Therefore, studies of Na cells that operate at lower temperatures, are safer in operation, and yet are still cost effective become important. To facilitate the development of new sodium batteries that operate at lower temperatures, new anodes, cathodes, and sodium-ion conducting electrolytes are being researched. In one report, Na/Ni3S2 cells with liquid electrolytes have successfully been made [3]. However, cells with liquid electrolytes may form dendritic growths that may cause short circuits, similar to lithium batteries cycled at room temperature [4]. A safer alternative is to use a solid electrolyte which would provide a barrier to dendritic growths and replace flammable solvents. One choice for such a solid electrolyte is a Na + ion-conducting glass. Conventionally, glass is made by heating components above the liquidus temperature and quickly cooling to below the glass transition temperature (Tg). In a different route to the amorphous (glassy) state below the liquidus or melting point for a given phase, crystalline materials that are excessively mechanically deformed may react to create amorphous products that are chemically distinct from the starting materials. In this process known as mechanical milling, it is thought that the energy from the collisions between the milling media, milling pot, and the powdered sample are sufficient for chemical reaction and amorphization to occur only slightly above room temperature [8]. The xLi2S + (1 − x)P2S5 glassy electrolyte system has been extensively studied for use in all-solid-state batteries because of its high Li-ion conductivity and negligible electronic conductivity [5–7].

94

S.S. Berbano et al. / Journal of Non-Crystalline Solids 358 (2012) 93–98

When mechanically milled, 0.70 Li2S + 0.30 P2S5 crystalline starting powders react to form an amorphous product [7]. The resulting amorphous product was found to have higher ionic (Li +) conductivity than a comparable composition of polycrystalline material produced by a solid-state reaction [7]. For most materials, the glassy phase is found to have a higher ionic conductivity than the corresponding crystalline phase because the typically larger free volume of the glassy amorphous phase allows for greater ion mobility in the glassy phase [9]. However, Mizuno et al. have also shown that after a heat treatment, mechanically milled samples of amorphous 0.70 Li2S + 0.30 P2S5 produce a glass-ceramic material with a precipitated superionic crystal Li7P3S11[10–11]. This superionic crystal has a room temperature ionic conductivity greater than 10 -3 S/cm, which is more than two orders of magnitude higher than the conductivity of the base amorphous phase. The polycrystalline material produced by a solidstate reaction has an ionic conductivity of ~ 10 −8 S/cm at room temperature, which is a relatively low ionic conductivity, especially for Li + ions in an all-sulfide material. Subsequently, all-solid-state cells using a high conductivity glass-ceramic Li2S + P2S5 solid electrolyte have been fabricated. Some cells have shown good cycleablity up to two thousand cycles [12–14]. In this study, a comparison was made between xNa2S + (1 − x) P2S5 prepared by the conventional method of melt-quenching and by the more recently developed method of mechanical milling. In future work, these xNa2S + (1 − x)P2S5 materials are being explored to determine densities, structures via NMR, glass transition temperatures, and if this system can be heat treated to precipitate the analogous Na7P3S11 crystal, to the Li7P3S11 phase. So far, studies of the low temperature α-Na3PS4 crystalline compound have shown its ionic conductivity to be 4.17 × 10 − 6 S/cm at 50 °C [15]. High-purity, fully-reacted amorphous materials were synthesized by melt-quenching and mechanical milling. XRD was used to examine the range over which fully amorphous products could be produced. Additionally, the materials were characterized by FT-IR and Raman spectroscopies to examine the extent of chemical reaction between Na2S and P2S5 and to examine and compare their short-range structures.

for five minutes in an agate mortar and pestle and then loaded into an 80 mL ZrO2 pot. Twenty ZrO2 balls, 10 mm in diameter, were used as the milling media. The sample to ball weight ratio was ~ 1:20. The pot was sealed under vacuum with a rubber gasket and taped shut to avoid air influx and contamination while the sample was being milled outside the glovebox. Samples were mechanically milled for 40 h using a high-energy planetary mono-mill (Model Pulverisette 6, Fritsch) rotated at 370 rpm. After mechanical milling, the color of the mixed starting materials changed from dark yellow to light yellow.

2. Experimental methods

Raman spectra of the powder samples were collected using a Renishaw Raman inVia spectrometer with the 488 nm line of an Ar + laser at 50 mW power. The instrument was calibrated to within ± 1 cm − 1 using an internal silicon reference at 520 cm − 1. Bulk samples of the melt-quenched glass were loaded in a plastic sample holder and covered with transparent, amorphous tape to prevent contamination during measurement. Milled powder samples were pressed into a hollow disk sample holder and again covered with transparent, amorphous tape. Measurements were taken from multiple locations on the samples to examine sample homogeneity on the length scale of the spot size of the laser beam (10 μm) and the area over which the different spectra were collected.

2.1. Preparation by melt-quenching Reagent-grade Na2S is not commercially available. Anhydrous Na2S was prepared by dehydrating Na2S·9H2O (Aldrich, 99.99%) under vacuum at ~ 0.1 Pa pressure at 150 °C for 1 h followed by heating at 650 °C for 24 h. P2S5 (Aldrich, 99%) was used as is without further purification. xNa2S + (1 − x)P2S5 melt-quenched products were prepared by weighing appropriate amounts of Na2S and P2S5 inside a nitrogen glovebox (b1 ppm O2 and b 1 ppm H2O). The powders were ground for five minutes in an agate mortar and pestle and then melted in a covered vitreous carbon crucible inside a hermetically sealed tube furnace attached to the outside of the glovebox. Samples were heated at 650 °C for 15 min and then splat-quenched between room temperature brass plates to produce transparent, homogeneous and yellowcolored glasses. Weight loss, assumed to be primarily due to the vaporization of P2S5, was less than 1 mol% for the melted batches of composition x = 0.50 and 0.67. The melt-quenched samples were then annealed ~ 30 °C below their Tgs for 1 h. 2.2. Preparation by mechanical milling xNa2S + (1 − x)P2S5 mechanically milled products were prepared by weighing appropriate amounts of Na2S, prepared as described above, and P2S5 inside a nitrogen glovebox. The powders were ground

2.3. X-ray diffraction To verify the structural phase of the various materials, XRD data were collected on powder samples using a Scintag XDS 2000 with CuKa radiation, λ = 1.5406 Å at 40 kV and 30 mA in the 2θ range of 30° to 70° with a step size of 0.01° and scan rate of 1.0°/min. Powder samples were placed in a plastic sample cup holder and covered with Kapton tape to prevent contamination during measurement. Kapton is essentially amorphous and has no diffraction peaks in the range of interest. 2.4. Fourier-Transform Infrared Spectroscopy Infrared (IR) absorption spectra of the powder samples were collected using a Bruker IFS-66 V/s spectrometer in the mid-IR range of 4000 to 400 cm − 1 using a KBr beamsplitter and far-IR range of 750 to 150 cm − 1 using a Ge-coated mylar beamsplitter. Background spectra were collected on 100 mg of ground and previously dried CsI pressed into a round disk sample holder. The IR spectra were recorded on a ground mixture of 100 mg dried CsI and 2 mg sample powder pressed into a disk sample holder. Samples were transferred to a small nitrogen glovebox connected to the IR spectrometer in order to load and measure the samples without exposure to air. The IR spectra were obtained using 32 scans at 4 cm − 1 resolution under vacuum. 2.5. Raman spectroscopy

3. Results 3.1. X-ray diffraction Fig. 1 shows XRD patterns of Na2S and P2S5 starting materials and their milled products. The P2S5 reagent appears to be partially crystalline. After milling P2S5, it remained partially crystalline and did not become an amorphous phase after milling. The melt-quenched products were found to be x-ray amorphous for 0.50 ≤ x ≤ 0.67. Milled products were found to be x-ray amorphous for x = 0.25, 0.33, 0.50, 0.67, and 0.70. For the higher Na2S composition of x = 0.75, mechanically milled samples showed the formation of a slightly disordered α-Na3PS4 crystal [15]. The x = 0.80 composition consisted of a mixture of the α-Na3PS4 and Na2S phases.

S.S. Berbano et al. / Journal of Non-Crystalline Solids 358 (2012) 93–98

95

xNa2S + (1-x)P2S5

xNa2S + (1-x)P2S5 α− Na3PS4 Na2S

IR x = 0.80 Mechanically Milled

x = 0.80

0.75 Mechanically Milled 0.75 0.70 Mechanically Milled

0.70 0.67 Mechanically Milled

0.50 0.33 0.25 P2S5 Mechanically Milled

Absorbance (A.U.)

Intensity (A. U.)

0.67

0.50 Mechanically Milled 0.33 Mechanically Milled 0.25 Mechanically Milled

P2S5

P2S5 Mechanically Milled

Na2S

Na2S

P=S Stretch

P-S-P Stretch

JCPDS Ref. 16 30

35

40

45

50

55

60

65

70

Na2S

2θ Fig. 1. XRD powder patterns of the starting materials and mechanically milled xNa2S + (1 − x)P2S5 products.

P2S5

150

300

450

600

750

900

1050 1200 1350 1500

Wavenumber (cm-1) Fig. 2. FT-IR spectra of the starting materials and mechanically milled xNa2S + (1− x)P2S5 products.

3.2. Infrared spectroscopy Fig. 2 shows the IR spectra of Na2S and P2S5 starting materials and their mechanically milled products. Fig. 3 shows that the IR spectral features of the melt-quenched and mechanically milled samples are nearly identical. 3.3. Raman spectroscopy Fig. 4 shows the Raman spectra of Na2S and P2S5 starting materials and their milled products. Fig. 5 shows that the melt-quenched and mechanically milled samples have nearly identical Raman spectra, a result also found for the IR spectra as described above. 4. Discussion 4.1. X-ray diffraction Sodium sulfide was found to be phase pure since the peaks at 2θ = 39°, 46°, 56°, and 62° agree well with the Joint Committee on Powder Diffraction Standards (JCPDS) card No. 47–1698 [16]. The P2S5 reagent is partially crystalline due to the gas phase preparation technique used to produce this commercial product. However, the small number of peaks that are observed in the XRD pattern of P2S5 agree well with literature [17]. P2S5 was mechanically milled to see if a more crystalline product could be produced or if it could be

rendered amorphous. However, as seen in the XRD powder pattern, mechanical milling of P2S5 did not result in a transformation of the material. Literature reports of the preparations of this binary glass system using sealed silica ampoules suggest a glass formation range of 0 ≤ x ≤ 0.66 [19]. So far, sealed ampoule preparation routes have not been used in this study, but could possibly expand the glass formation range to lower values of x. For compositions of x N 0.67, there was not enough glass former to produce a homogeneous glass product on quenching and the resulting materials were polycrystalline. The compositions x = 0.25 and 0.33 could not be melted in our experimental setup without large amounts of weight loss because of the large difference in melting points between Na2S and P2S5, ~1160 °C and ~280 °C, respectively [20–21]. Mechanical milling, however, significantly extends the X-ray amorphous range of this binary system. This technique allows P2S5rich fully reacted materials to be made without weight loss from the vaporization of P2S5 and without contamination from O2 and H2O (see Section 4.2). Therefore, mechanical milling is a favorable processing technique to produce xNa2S + (1 − x)P2S5 materials because it extends the amorphous range and produces amorphous and fully reacted materials without weight loss at essentially room temperature. These attributes make mechanical milling attractive for large scale commercial manufacturing.

96

S.S. Berbano et al. / Journal of Non-Crystalline Solids 358 (2012) 93–98

xNa2S + (1-x)P2S5

xNa2S + (1-x)P2S5

IR

Raman

x = 0.80 Mechanically Milled

Absorbance (A. U.)

x = 0.67 Mechanically Milled

0.70 Mechanically Milled

0.67 Melt-quenched

0.67 Mechanically Milled

300

450

600

750

900

0.50 Mechanically Milled

Intensity (A.U.)

0.50 Mechanically Milled

0.50 Melt-quenched

150

0.75 Mechanically Milled

0.33 Mechanically Milled 0.25 Mechanically Milled

1050 1200 1350 1500

Wavenumber (cm-1)

P2S5 Mechanically Milled

Fig. 3. FT-IR spectra comparing melt-quenched and mechanically milled xNa2S + (1− x) P2S5 products for x = 0.50 and 0.67.

P2S5 Jansen et. al produced α-Na3PS4 by a solid-state reaction in a sealed silica ampoule [15]. Using Bragg's Law, 2θ angles corresponding to α-Na3PS4 were calculated from the reported d-spacings [15] and compared in Fig. 1 to those found in this study from the x = 0.75 (nominally Na3PS4) and 0.80 compositions. To the best of our knowledge, this study is the first to report the synthesis of αNa3PS4 in a single step process without either heating or slowly cooling from the melt. The products remained as α-Na3PS4 after milling for 40 h, which demonstrates the crystalline phase's compositional and thermodynamic stability. In contrast, 0.75 Li2S + 0.25 P2S5 (Li3PS4) mills into an amorphous material after 20 h and crystalline Li3PS4 appears not to be the favored phase formed with mechanical milling [7]. A reason for this difference may be due to the variation in the volumes of the milling pots, 45 mL reported by others in the lithium system and the 80 mL used here, among other differences in the milling parameters. Since the rotation speed of 370 rpm has been used in both studies, for a given milling duration it is expected that there will be larger numbers of collisions in the smaller volume milling pot than in the larger volume milling pot. Additionally, the difference in thermodynamic stability of lithium and sodium crystalline phases may be another reason the 0.75 Li2S + 0.25 P2S5 phase appears to be more easily milled into an amorphous material. Finally, from the phase diagram, one would expect the x = 0.80 composition sample to consist of a mixture of the α-Na3PS4 compound and Na2S. This is seen in the XRD pattern as peaks assigned to α-Na3PS4 become less prominent, while the main Na2S peak increases in intensity.

4.2. Infrared spectroscopy The low frequency doublet at 200 and 250 cm − 1 in IR spectra of Na2S are attributed to Na +S = rattling modes in the antifluorite crystal structure. This is because Na + is loosely (ionically) bound in

Na2S

100

200

300

400

500

600

700

800

900

1000

Raman Shift (cm-1) Fig. 4. Raman spectra of the starting materials and mechanically milled xNa2S + (1 − x) P2S5 products.

tetrahedral sulfur cages. P2S5 is built from molecular P4S10 adamantane-like units and has two main vibrational modes. The peak at 530 cm − 1 is assigned to the P–S–P stretching mode and the higher frequency peak at 690 cm − 1 is assigned to the P=S stretching mode [18]. As shown in Fig. 2, samples can be prepared with little O or OH contamination. Absorption peaks due to the bending mode of H2O were not observed at 1500 cm − 1. Likewise, peaks due to contaminant –OH were also not observed at ~ 3000 cm − 1. Similarly, P–O peaks at 1000–1500 cm − 1 are also not observed. This is attributed to the high-quality glovebox atmosphere used, the tight seal made in the zirconia milling pot, and the IR spectrometer's high-quality vacuum system. As Na2S glass modifier is added to P2S5, non-bridging sulfur (NBS) units, Na + −S–PS(S2/2), are created which depolymerizes the phosphorus-sulfur-phosphorous network. Accordingly, the P=S stretching mode decreases in frequency while the P–S–P stretching mode increases in frequency. These frequency shifts are the result of π-bond delocalization on Q 3, PS(S3/2), species that effectively lengthen the P=S terminal sulfur bond and strengthen the bridging sulfur (BS) P–S–P linkages with increasing Na2S content. These frequency shifts are analogous to those attributed to additions of Na2O in the Na2O + P2O5 system [22]. At higher Na2S contents, additions of Na2S

S.S. Berbano et al. / Journal of Non-Crystalline Solids 358 (2012) 93–98

xNa2S + (1-x)P2S5

Raman

Intensity (A.U.)

x = 0.67 Mechanically Milled

0.67 Melt-quenched

0.50 Mechanically Milled

97

increasing Na2S and merge into a broad peak which then decreases with intensity at the x = 0.75 composition. The modes around 700 cm − 1 arise from P = S vibrations and the peak at 416 cm − 1 arises from PS3/2 impurities in the phase. Sharp bands at 210, 280, and 305 cm − 1 arise from molecular P4S10 and P4S9 species [26]. The main P = S stretching mode is observed to have shifted to lower frequency for samples with increased Na2S content. Starting at x = 0.67 (pyro-thiophosphate), the main peak shifts to higher frequency with increasing x until x = 0.80. The x = 0.70 shows Raman spectra which is a combination of the spectra between the x = 0.67 and x = 0.75 (ortho-thiophosphate) compositions. Also, the P = S stretching peak at 413 cm − 1 is much sharper for x = 0.75 than for other compositions, providing further evidence to indicate the presence of the α-Na3PS4 crystalline compound. Raman spectra have been reported for the Na3PS4 compound with peaks at 410 and 280 cm − 1 and are assigned to asymmetric P-S stretching and asymmetric P–S–P stretching of this unit, respectively [24]. In the xLi2S + (1 − x)P2S5 system, the Raman spectra of melted and milled products were also very similar [10, 14]. 5. Conclusions

0.50 Melt-quenched 100

200

300

400

500

600

700

800

900

1000

Raman Shift (cm-1) Fig. 5. Raman spectra comparing melt-quenched and mechanically milled xNa2S + (1− x) P2S5 products for x= 0.50 and 0.67.

cause the low frequency doublet peak to be resolved into two separate modes at 180 cm − 1 and 260 cm − 1. The formula units corresponding to x = 0.25, 0.33, 0.50, 0.67, 0.70, and 0.75 are NaP3S8, NaP2S3, NaPS3, Na4P2S7, Na7P3S11, and Na3PS4, respectively. Until now, only brief IR and Raman results have been reported for the Na3PS4 compound [23]. The composition 0.5Na2S + 0.5P2S5 corresponds to the NaPS3 compound and a structure composed of Q 2, NaSP(S)S2/2, units. Its IR spectrum contains three dominant peaks at 496, 551, and 667 cm − 1. The composition 0.75Na2S + 0.25P2S5 corresponds to the Na3PS4 compound and a structure composed of Q 0, (NaS)3PS, units, a PS43- tetrahedral structure with one terminating NBS and three Na +−S–P NBS units. Its spectrum contains one dominant peak at 544 cm − 1 which is attributed to asymmetric P–S stretching of the PS43− tetrahedral unit and a weaker peak at 280 cm − 1 assigned to asymmetric P–S–P stretching arising from slight non-stoichiometry in the material [24]. Finally, the small peak at ~ 1015 cm − 1 is attributed to the main peak of Na2SO4 due to slight oxygen contamination in the starting materials [25]. The relative intensities of the modes at ~900 cm − 1 and 1100 cm − 1 differ between melted and milled spectra. This may be attributed to the differences in the fraction of phosphorous oxygen impurity structures Na +−O–P and P=O, respectively present in the materials. The similarity of the x = 0.50 and 0.67 melted and milled spectra suggests that mechanical milling resulted in full reaction at around room temperature during the milling process.

4.3. Raman spectroscopy As shown in Fig. 4, Na2S has one dominant peak at 190 cm − 1. This peak is similar in position to one of the modes of P2S5, which has many vibrational modes. These peaks decrease in intensity with

High-purity xNa2S + (1 − x)P2S5 amorphous and polycrystalline materials were prepared by melt-quenching and mechanical milling and compared for the first time. To the best of our knowledge, this is the first reported study on this system using the mechanical milling method. The melt-quenched products were x-ray amorphous for x = 0.50 and 0.67, whereas mechanically milled products were x-ray amorphous for x = 0.25, 0.33, 0.50, 0.67, and 0.70. For x = 0.75, samples showed the formation of a disordered α-Na3PS4 crystal. Again, to the best of our knowledge, this is the first time α-Na3PS4 has been synthesized in a single step method using mechanical milling. FT-IR spectroscopy showed that samples could be prepared with little contamination and IR spectra of the melt-quenched and mechanically milled samples of the same compositions were nearly identical. Raman spectra showed that corresponding melt-quenched and mechanically milled samples also had nearly identical spectra. These two results suggest that nearly identical chemical structures in these materials could be produced using both melt-quenching and mechanical milling. In order to use these materials as solid electrolytes, more studies are needed to determine their thermal stability and ionic conductivity. Acknowledgements Support for this research by the National Science Foundation (NSF) Grant no. DMR 0710564 is gratefully acknowledged. The authors thank the NSF's International Materials Institute for New Functionality in Glass (NSF DMR 0844014) and its directors, Professors Himanshu Jain (Lehigh University), and Carlo Pantano (Pennsylvania State University) for sponsoring one of the authors, SB, in a Research Experience for Undergraduates. Professor Masahiro Tatsumisago of the Inorganic Chemistry Group at Osaka Prefecture University, Japan is thanked for hosting SB in this collaborative international research exchange. References [1] J.T. Kummer, N. Weber, Society of Automotive, a Sodium-Sulfur Secondary Battery, Society of Automotive Engineers, New York, 1967. [2] I.W. Jones, Electrochim. Acta 22 (1977) 681. [3] J.S. Kim, H.J. Ahn, H.S. Ryu, D.J. Kim, G.B. Cho, K.W. Kim, T.H. Nam, J.H. Ahn, J. Power Sources 178 (2008) 852. [4] M. Dolle, L. Sannier, B. Beaudoin, M. Trentin, J.M. Tarascon, Electrochem. Solid St. 5 (2002) A286. [5] M. Murayama, N. Sonoyama, A. Yamada, R. Kanno, Solid State Ionics 170 (2004) 173.

98

S.S. Berbano et al. / Journal of Non-Crystalline Solids 358 (2012) 93–98

[6] R. Mercier, J.P. Malugani, B. Fahys, G. Robert, Solid State Ionics 5 (1981) 663. [7] A. Hayashi, S. Hama, H. Morimoto, M. Tatsumisago, T. Minami, J. Am. Ceram. Soc. 84 (2001) 477. [8] C. Suryanarayana, Prog. Mater. Sci. 46 (2001) 1. [9] T. Minami, N. Machida, Mater. Sci. Eng. B-Solid. 13 (1992) 203. [10] F. Mizuno, A. Hayashi, K. Tadanaga, M. Tatsumisago, Adv. Mater. 17 (2005) 918. [11] K. Minami, A. Hayashi, M. Tatsumisago, J. Ceram. Soc. Jpn. 118 (2010) 305. [12] A. Hayashi, T. Ohtomo, F. Mizuno, K. Tadanaga, M. Tatsumisago, Electrochim. Acta 50 (2004) 893. [13] M. Nagao, H. Kitaura, A. Hayashi, M. Tatsumisago, J. Power Sources 189 (2009) 672. [14] A. Hayashi, K. Minami, M. Tatsumisago, J. Solid State Electrochem. 14 (2010) 1761. [15] M. Jansen, U. Henseler, J. Solid State Chem. 99 (1992) 110.

[16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26]

E. Zintl, A. Harder, B. Dauth, Z. Elektrochem. Angew. Phys. 40 (1934) 588. H. Yamamoto, N. Machida, T. Shigematsu, Solid State Ionics 175 (2004) 707. J.O. Jensen, D. Zeroka, J. Mol. Struct. Theochem 487 (1999) 267. M. Ribes, B. Barrau, J.L. Souquet, J. Non-Cryst. Solids 38–9 (1980) 271. J. Sangster, A. Pelton, J. Phase Equilib. 18 (1997) 89. R. Foerthmann, A. Schneider, Naturwissenschaften 52 (1965) 390. J.J. Hudgens, R.K. Brow, D.R. Tallant, S.W. Martin, J. Non-Cryst. Solids 223 (1998) 21. H. Burger, H. Falius, Z. Anorg, Allg. Chem. 363 (1–2) (1968) 24. M.J.F. Leroy, G. Kaufmann, A. Muller, H.W. Roesky, C. R. Acad. Sci. Paris Ser. C 267 (1968) 563. R.M. Atkins, K.A. Gingerich, Chem. Phys. Lett. 53 (1978) 347. B. Cherry, J.W. Zwanziger, B.G. Aitken, J. Phys. Chem. B 106 (2002) 11093.