Ionic Conductivity and Microstructure of Li₄GeO₄-Based Solid Electrolytes

Abstract

Li₄GeO₄-based solid electrolytes can be synthesized at low temperatures, and their formability is improved by adding Li₂SO₄. Glass–ceramic Li₄GeO₄ exhibits a relatively high ionic conductivity of approximately 10⁻⁶ S cm⁻¹ at room temperature, which is higher than that of glass Li₄GeO₄. Thus, Li₄GeO₄ has promising applications in all-solid-state lithium ion batteries. To understand the correlation between the ionic conductivity, formability, and microstructure of the synthesized materials, the microstructures and crystallization process of glass and glass–ceramic Li₄GeO₄ and 80Li₄GeO₄·20Li₂SO₄ (Li_3.6Ge_0.8S_0.2O₄) were observed by transmission electron microscopy (TEM). Since Li_3.6Ge_0.8S_0.2O₄ glass exhibits a halo diffraction pattern, the addition of Li₂SO₄ to Li₄GeO₄ stabilizes its amorphous phase. In addition, glass–ceramic samples were found to be characterized by an amorphous state containing nanocrystallites with a crystallinity degree of approximately 40%, which improves the ionic conductivity of the material.

1. Introduction

All-solid-state batteries containing a nonflammable solid electrolyte are next-generation batteries owing to their safety and high energy density.^1,2) To develop advanced all-solid-state batteries, it is important to explore solid electrolytes that are nonflammable, chemically stable, and exhibit high ionic conductivity. In this field, two types of solid electrolytes have attracted significant attention. The first type is sulfide-based electrolytes, such as Li₂S–P₂S₅ binary systems and Li₄SnS₄,^3–5) which have ionic conductivities >10⁻⁴ S cm⁻¹ at room temperature. The second type is oxide-based electrolytes, such as Li₇La₃Zr₂O₁₂,⁶⁾ which have ionic conductivities greater than 10⁻⁴ S cm⁻¹ at room temperature. Although sulfide electrolytes typically exhibit the highest conductivity between the two types of electrolytes, oxide electrolytes have advantages in terms of chemical stability and production costs.

On the other hand, although oxide-based electrolytes exhibit high atmospheric stability and low toxicity compared with sulfide-based electrolytes, it is difficult to use them in constructing an electrode-electrolyte interface because of their low formability. High-temperature sintering of such materials is not desirable as unfavorable side reactions occur at the electrode-electrolyte interface at high temperatures.^7–9) Therefore, there is a need to develop low temperature synthesis techniques for oxide-based electrolytes that exhibit improved formability and high ionic conductivity.

Recently, the properties of Li₄GeO₄-based solid electrolytes have been reported.¹⁰⁾ Li₄GeO₄-based glass–ceramics have been synthesized at a relatively low temperature (∼250°C) with the precipitation of a metastable hexagonal crystal phase in the P6₃/mmc space group, which has a similar structure as Li₄SnS₄.⁵⁾ Additionally, the ionic conductivity and formability of Li₄GeO₄ glass–ceramics can be improved by adding Li₂SO₄, which exhibits ionic conductivity of ∼10⁻⁶ S cm⁻¹ at room temperature. Lithium superionic conductor (LISICON) crystals with the Pnma space group are obtained in Li_4−2xGe_1−xS_xO₄, and they exhibit high ionic conductivity greater than 10⁻⁵ S cm⁻¹ at room temperature when x = 0.2 (80Li₄GeO₄·20Li₂SO₄ or Li_3.6Ge_0.8S_0.2O₄).^11,12) Although X-ray diffraction and spectroscopy measurements have revealed the crystal structures in a bulk state, the microstructural properties, such as nanocrystallite size and amorphous matrix, have not been clarified yet. It is, therefore, important to investigate the microstructure of the Li₄GeO₄-based solid electrolytes to better understand their high ionic conduction property.

To achieve high conductivity in glass electrolytes, the correlation analysis of ionic conduction property and crystallization behavior is important. So far, we have examined the microstructure and crystallization process of sulfide-based glass and glass–ceramics such as Li₃PS₄, Li₁₀GeP₂S₁₂, and Na₃PS₄ using transmission electron microscopy (TEM). The Li₃PS₄ glass shows a high conductivity by sintering at approximately 180°C. TEM observation revealed that the microstructure at 180°C is characterized by an amorphous state containing nanocrystallites with 5% crystallinity.⁴⁾ That is, the early stage of crystallization relates to high conductivity in the Li₃PS₄ glass. Na₃PS₄ exhibits ionic conductivity of more than 10⁻⁴ S cm⁻¹ at room temperature when it is annealed to crystallize a cubic structure.¹³⁾ In situ TEM observation revealed that the crystallite size significantly increases during the transition from the cubic phase to the tetragonal phase.¹⁴⁾ The composite comprising nanocrystallites and amorphous domains would be likely responsible for high conductivity in the cubic Na₃PS₄ glass–ceramics. For oxide-based glass and glass–ceramics, however, little research on the correlation between conductivity and microstructure has been conducted. To apply glass and glass–ceramics to all-solid-state cells, information such as crystallite size and crystallinity is absolutely critical. Thus, in this study, glass and glass–ceramics of Li₄GeO₄ and Li_3.6Ge_0.8S_0.2O₄ were investigated by TEM. To understand the origins of the high formability and ionic conductivity of Li₄GeO₄ and Li_3.6Ge_0.8S_0.2O₄, their crystallite size, crystallinity degree, and spatial distribution in an amorphous matrix were analyzed by dark-field (DF) TEM and high-resolution TEM (HR-TEM) imaging. Besides, the amorphous matrix structure was investigated by pair-distribution function (PDF) analysis of their electron diffraction (ED) patterns.

2. Experimental Procedure

2.1 TEM observations

As-milled samples of Li₄GeO₄ and Li_3.6Ge_0.8S_0.2O₄ were prepared via mechanochemical treatment in a planetary ball mill following the procedure reported in the literature.¹⁰⁾ Subsequent heat treatment was conducted in an electric furnace at 250°C for 2 h to obtain glass–ceramics samples. The sample subjected to heat treatment is hereafter referred to as HT 250°C.

TEM observations were conducted using a JEM-2100F field-emission-type microscope at an acceleration voltage of 200 kV. Although oxide-based electrolytes are relatively stable in air, all the samples were mounted on Cu grids in a glove box filled with dry argon gas. To observe the microstructures of the Li₄GeO₄ and Li_3.6Ge_0.8S_0.2O₄ samples, ex situ TEM was conducted using a double-tilt vacuum transfer TEM holder (Gatan model 648). To observe the crystallization behavior of Li_3.6Ge_0.8S_0.2O₄, in situ TEM observations were also conducted using a double-tilt heating holder in a temperature range of 150°C–450°C at 100°C intervals. The average crystalline size was estimated using machine learning and DF images.¹⁵⁾ ProcessDiffraction was used to integrate the two-dimensional ED pattern azimuthally and convert it to a one-dimensional scattering intensity profile I(Q), which was then used to identify the precipitated crystalline phase by comparing the intensity profile I(Q) with simulation results.¹⁶⁾ The crystallization degree of the glass–ceramic sample was quantitatively evaluated based on the intensity profile of the ED pattern.¹⁷⁾

2.2 PDF analysis

Owing to the short wavelength of high energy electrons, scattering information over a wide range of the reciprocal lattice space can be obtained utilizing electron diffraction.¹⁸⁾ Therefore, electron PDF analysis is an effective method for characterizing amorphous structures. It has been employed to study amorphous or polycrystal materials, such as Mg₂F and Fe_100−xB, using their ED patterns.^19,20) More details about electron PDF analysis have been reported by Gorelik et al.²¹⁾ In this study, ED patterns for PDF analysis were recorded using imaging plates (IPs) for an exposure time of 60 s. The intensity of the IPs was recorded by an IP reader (DITABIS-MICRON). Azimuthal integration of ED patterns from IP films was also conducted using ProcessDiffraction software, and reduced PDFs were obtained using SUePDF software.²²⁾ ED has issues, including inelastic and multiple scattering. These influence the intensity of the peaks in the PDF results and affect the determination of the atomic coordination numbers, but the positions of the peaks remain unchanged.^14,23) Therefore, in this study, only the peak positions in the PDF are discussed.

3. Results and Discussions

Figure 1 shows TEM images of the Li₄GeO₄ glass sample. Ionic conductivity of the Li₄GeO₄ glass is 1.2 × 10⁻⁶ S cm⁻¹.¹⁰⁾ The ED pattern (Fig. 1(a)) shows the coexistence of a halo-ring pattern with Debye–Scherrer rings. The ED intensity profile shows that the diffraction peaks are attributable to the hexagonal Li₄GeO₄ structure. The DF image (Fig. 1(b)) shows precipitated Li₄GeO₄ nanocrystals (bright regions). The average crystallite size is 17 nm, as shown in the nanocrystallite size distribution (Fig. 1(c)). Based on the ED pattern, the crystallinity degree (x) is approximately 22%. These data indicate that the Li₄GeO₄ glass phase comprises amorphous domains containing small nanocrystallites.

Fig. 1

Transmission electron microscopy (TEM) images of the Li₄GeO₄ glass. (a) Electron diffraction (ED) pattern and its corresponding intensity profile. The simulated X-ray diffraction (XRD) pattern of the hexagonal crystal structure with the P6₃/mmc space group is shown by blue lines. (b) DF image. (c) Crystallite size distribution.

The microstructure of the as-milled Li_3.6Ge_0.8S_0.2O₄ glass was different from that of Li₄GeO₄ glass. Ionic conductivity of the as-milled Li_3.6Ge_0.8S_0.2O₄ glass is 1.4 × 10⁻⁶ S cm⁻¹.¹⁰⁾ Figure 2 shows TEM images of the Li_3.6Ge_0.8S_0.2O₄ glass sample. The ED pattern shows a halo-ring pattern without diffraction spots. The HR-TEM image shows an amorphous pattern and no nanocrystallites. These TEM results indicate that the addition of Li₂SO₄ could stabilize the amorphous structure of the sample.

Fig. 2

High-resolution TEM (HR-TEM) images and the corresponding ED pattern of Li_3.6Ge_0.8S_0.2O₄ glass.

Figure 3 shows TEM images of Li₄GeO₄ HT 250°C. Ionic conductivity of Li₄GeO₄ HT 250°C is 2.6 × 10⁻⁶ S cm⁻¹.¹⁰⁾ The ED pattern and DF image (Fig. 3(a) and (c), respectively) indicate that crystallization proceeded during heat treatment. The intensity of diffraction peaks derived from the hexagonal structure increased. The HR-TEM image (Fig. 3(d)) was obtained taken from the dotted square region in (c). A hexagonal Li₄GeO₄ nanocrystallite exhibiting lattice fringes was observed in an amorphous matrix, as indicated by the dotted circle. Fast Fourier transform (FFT) calculations were conducted for a nanocrystallite region of approximately 10 nm × 10 nm. In the FFT pattern, 010 diffraction spots are observed. Thus, the lattice fringe is assigned to the distance of the (010) plane in the hexagonal structure. Consistent with the ED pattern, furthermore, the crystallite size increased to approximately 27 nm, as shown in Fig. 3(b). Based on the ED pattern, the crystallinity degree was 43%.

Fig. 3

TEM images of Li₄GeO₄ HT 250°C. (a) ED pattern and its corresponding intensity profile. The simulated XRD pattern of the hexagonal crystal structure with the P6₃/mmc space group is shown by blue lines. (b) Crystallite size distribution. (c) DF image. (d) HR-TEM image obtained from the dotted square region in (b).

Similar trends due to heat treatment were observed in the Li₂SO₄-added Li₄GeO₄. Figure 4 shows the TEM images of Li_3.6Ge_0.8S_0.2O₄ HT 250°C. Ionic conductivity of Li_3.6Ge_0.8S_0.2O₄ HT 250°C is 2.3 × 10⁻⁶ S cm⁻¹.¹⁰⁾ The ED pattern (Fig. 4(a)) shows Debye–Scherrer rings with a halo-ring pattern. The ED intensity profile and DF image (Fig. 4(c)) show that the Li_3.6Ge_0.8S_0.2O₄ HT 250°C sample has a hexagonal structure, which is characterized by precipitated nanocrystals with a size of approximately 47 nm. The crystallinity degree was 39%. The HR image (Fig. 4(d)) shows hexagonal Li_3.6Ge_0.8S_0.2O₄ nanocrystallites dispersed in an amorphous matrix, as indicated by the dotted circles. Consequently, the coexistence of amorphous and nanocrystallites, with a crystallinity degree of approximately 40%, played a key role in the ionic conductivity of the Li₄GeO₄ and Li_3.6Ge_0.8S_0.2O₄ HT 250°C samples.

Fig. 4

TEM images of Li_3.6Ge_0.8S_0.2O₄ HT 250°C. (a) ED pattern and its corresponding intensity profile. The simulated XRD pattern of the hexagonal crystal structure with the P6₃/mmc space group is shown by blue lines. (b) Crystallite size distribution. (c) DF image. (d) HR-TEM image.

Furthermore, PDF analysis was conducted to better understand the local structures of the glass and glass–ceramics samples. Figure 5 shows the PDF result for each sample. All samples showed prominent peaks at 1.79 and 3.12 Å. The crystal structure of hexagonal Li₄GeO₄ is shown on the right side of the figure. Oxygen atoms form the hexagonal closest packing structure, with Li and Ge randomly occupying tetrahedral sites.¹⁰⁾ Thus, the peak at 1.79 Å indicates the average distance from an atom in the center to an atom at the vertex of LiO₄ or GeO₄ tetrahedron. For Li_3.6Ge_0.8S_0.2O₄, the peak at 1.79 Å indicates that additional S atoms occupy the tetrahedral sites, and the average distance from the center atom to the vertex atom shows a little change. The second most prominent peak at 3.12 Å indicates the distance of O–O correlation in the tetrahedra. No discernible peaks greater than 3.12 Å were observed for the as-milled Li_3.6Ge_0.8S_0.2O₄ glass, indicating no long-range order in its amorphous phase. In contrast, the prominent peaks greater than 3.12 Å in the patterns of the Li₄GeO₄ and Li_3.6Ge_0.8S_0.2O₄ glass–ceramics infer long-range order in their structures after crystallization. The PDF analysis results support the crystallization process induced by heat treatment of 250°C.

Fig. 5

Pair-distribution function of Li₄GeO₄ and Li_3.6Ge_0.8S_0.2O₄ glass and glass–ceramics samples and the crystal structure of hexagonal Li₄GeO₄. The hexagonal structure is constructed from O atoms, and Li or Ge atoms randomly occupy the tetrahedral sites. Correlation ① represents the distance from the center to a vertex in a tetrahedron (Li–O, Ge–O, or S–O correlation), whereas correlation ② represents the distance between any two vertices in a tetrahedron (O–O correlation).

To further understand the crystallization behavior, in situ TEM observations were also conducted for the as-milled Li_3.6Ge_0.8S_0.2O₄ sample. Figure 6 shows the crystallization process of Li_3.6Ge_0.8S_0.2O₄ glass. The ED patterns and corresponding bright-filed (BF) and DF images in (Figs. 6(a)–(e)) were obtained at 25°C, 150°C, 250°C, 350°C, and 450°C, respectively. As shown in Fig. 6(a), Li_3.6Ge_0.8S_0.2O₄ glass showed a halo pattern before heating, indicating an amorphous state, which is consistent with Fig. 2. At 150°C, crystallization occurred, and weak diffraction spots appeared in the ED pattern. Nanocrystallites with an average size of 10 nm were observed (bright areas indicated by arrows in Fig. 6(b)). Increasing heating resulted in further crystallization, and diffraction spots increased in the ED pattern at 250°C, as shown in Fig. 6(c). The average nanocrystallite size increased to approximately 45 nm. In the temperature range of 250°C–350°C, crystallization progressed slowly, and no significant change in crystallite size was observed (Fig. 6(d)). Diffraction spots and Debye–Scherrer rings observed at 150°C–350°C are attributed to the hexagonal phase and consistent with the ED patterns in Fig. 4(a). When heated at 450°C, crystallization significantly proceeded, and high-intensity diffraction spots appeared on the low angle region of the ED pattern, as shown by the arrows in Fig. 6(e). In addition, large nanocrystallites with a size of approximately 400 nm precipitated, as indicated by the dotted circle in the DF image. Based on the analysis of the corresponding intensity profile, the diffraction spots on the low angle region are attributed to d₂₁₀ = 4.12 Å in the LISICON phase. That is, the hexagonal phase precipitated from the glass as a metastable phase,⁵⁾ and then the hexagonal → LISICON phase transition occurred. In the LISICON phase, oxygen atoms form the hexagonal closest packing structure, with Li and Ge regularly occupying tetrahedral and octahedral sites.¹⁰⁾ These in situ TEM observations show that the microstructure of Li_3.6Ge_0.8S_0.2O₄ at 250°C–350°C comprises nanocrystallite and amorphous hexagonal Li_3.6Ge_0.8S_0.2O₄, and the hexagonal → LISICON phase transition occurs at approximately 400°C along with grain coarsening.

Fig. 6

In situ TEM images of as-milled Li_3.6Ge_0.8S_0.2O₄ glass. BF and DF images and the corresponding ED patterns in (a)–(e) were taken at 25°C, 150°C, 250°C, 350°C, and 450°C, respectively.

In this study, the microstructures of Li₄GeO₄-based glass and glass–ceramic were investigated by TEM. To gain high ionic conductivity in the glass, it is important to control the crystallite size and the degree of crystallinity. In the Li₄GeO₄ and Li_3.6Ge_0.8S_0.2O₄ glasses, the amorphous state with approximately 40% crystallinity degree contributes to the high formability, resulting in the high ionic conductivity because the amorphous regions can be easily deformed, as demonstrated in sulfide-based glass electrolytes.²⁴⁾ The common feature of such a glass solid electrolyte is that conductivity tends to be enhanced in a metastable phase, which is characterized by the coexistence of amorphous and nanocrystallites.^5,14) Thus, glass–ceramics with a metastable hexagonal structure would be indispensable for the fabrication of oxide-based all-solid-state cells. In amorphous materials containing nanocrystallites, the interface between amorphous matrix and crystallites may likely contribute to ionic conduction. To further understand the ionic conduction mechanism, PDF analysis will be conducted using the scanning transmission electron microscopy (STEM). Recently, PDF analysis method from nano-beam diffraction patterns by STEM has been developed.^25,26) Since the electron beam can be converged to a few nanometers in STEM observations, PDF mapping can be conducted from nano-beam diffraction patterns. As a next step, we will apply this STEM-PDF method to glass and glass–ceramics and investigate the nanoscale interface structure.

4. Conclusion

To clarify the relation between ionic conductivity and microstructure in Li₄GeO₄-based glass electrolytes, crystallite size and amorphous matrix were investigated by TEM. As-milled Li_3.6Ge_0.8S_0.2O₄ glass showed halo-ring patterns with no diffraction spots, indicating stabilization due to the addition of Li₂SO₄ in the amorphous phase of Li₄GeO₄. Li₄GeO₄ and Li_3.6Ge_0.8S_0.2O₄ glass–ceramics heat-treated at 250°C showed hexagonal nanocrystallites with an average size of 25–50 nm and an amorphous matrix. In situ TEM observations for the crystallization process demonstrated that the crystallite size at 150°C–350°C is consistent with the results in ex situ TEM observation. When heated above 400°C, furthermore, the crystallite size significantly increased to more than 100 nm, which was attributed to the hexagonal → LISICON phase transition. Therefore, heat treatment is important in the Li₄GeO₄-based glasses because it causes the precipitation of hexagonal nanocrystallites with approximately 40% crystallinity.

Acknowledgments

This work was supported by JST ALCA-SPRING (Grant JPMJAL1301), Japan. Also, this work was in part supported by JSPS KAKENHI (Grant Numbers JP19H05814 and 21H04625) and the Joint Research Center for Environmentally Conscious Technologies in Materials Science (project No. 02107, Grant No. JPMXP0618217637) at ZAIKEN, Waseda University.

REFERENCES

• 1)   S.  Ohno,  A.  Banik,  G.F.  Dewald,  M.A.  Kraft,  T.  Krauskopf,  N.  Minafra,  P.  Till,  M.  Weiss and  W.G.  Zeier: Prog. Energy 2 (2020) 022001. doi:10.1088/2516-1083/ab73dd
• 2)   M.  Tatsumisago,  M.  Nagao and  A.  Hayashi: J. Asian Ceram. Soc. 1 (2013) 17–25. doi:10.1016/j.jascer.2013.03.005
• 3)   A.  Hayashi,  S.  Hama,  T.  Minami and  M.  Tatsumisago: Electrochem. Commun. 5 (2003) 111–114. doi:10.1016/S1388-2481(02)00555-6
• 4)   H.  Tsukasaki,  S.  Mori,  S.  Shiotani and  H.  Yamamura: Solid State Ionics 317 (2018) 122–126. doi:10.1016/j.ssi.2018.01.010
• 5)   K.  Kanazawa,  S.  Yubuchi,  C.  Hotehama,  M.  Otoyama,  S.  Shimono,  H.  Ishibashi,  Y.  Kubota,  A.  Sakuda,  A.  Hayashi and  M.  Tatsumisago: Inorg. Chem. 57 (2018) 9925–9930. doi:10.1021/acs.inorgchem.8b01049
• 6)   R.  Murugan,  V.  Thangadurai and  W.  Weppner: Angew. Chem. Int. Ed. 46 (2007) 7778–7781. doi:10.1002/anie.200701144
• 7)   Y.  Kobayashi,  T.  Takeuchi,  M.  Tabuchi,  K.  Ado and  H.  Kageyama: J. Power Sources 81–82 (1999) 853–858. doi:10.1016/S0378-7753(99)00121-4
• 8)   K.  Park,  B.-C.  Yu,  J.-W.  Jung,  Y.  Li,  W.  Zhou,  H.  Gao,  S.  Son and  J.B.  Goodenough: Chem. Mater. 28 (2016) 8051–8059. doi:10.1021/acs.chemmater.6b03870
• 9)   J.  Wakasugi,  H.  Munakata and  K.  Kanamura: J. Electrochem. Soc. 164 (2017) A1022. doi:10.1149/2.0471706jes
• 10)   Y.  Yoneda,  M.  Shigeno,  T.  Kimura,  K.  Nagao,  C.  Hotehama,  A.  Sakuda,  M.  Tatsumisago and  A.  Hayashi: Solid State Ionics 363 (2021) 115605. doi:10.1016/j.ssi.2021.115605
• 11)   S.  Taminato,  T.  Okumura,  T.  Takeuchi and  H.  Kobayashi: Solid State Ionics 326 (2018) 52–57. doi:10.1016/j.ssi.2018.09.011
• 12)   M.A.K.L.  Dissanayake,  R.P.  Gunawardane,  A.R.  West,  G.K.R.  Senadeera,  P.W.S.K.  Bandaranayake and  M.A.  Careem: Solid State Ionics 62 (1993) 217–223. doi:10.1016/0167-2738(93)90375-D
• 13)   A.  Hayashi,  K.  Noi,  A.  Sakuda and  M.  Tatsumisago: Nat. Commun. 3 (2012) 856. doi:10.1038/ncomms1843
• 14)   H.  Nakajima,  H.  Tsukasaki,  J.  Ding,  T.  Kimura,  T.  Nakano,  A.  Nasu,  A.  Hirata,  A.  Sakuda,  A.  Hayashi and  S.  Mori: J. Power Sources 511 (2021) 230444. doi:10.1016/j.jpowsour.2021.230444
• 15)   T.  Ayama,  H.  Tsukasaki,  Y.  Kawasaki,  H.  Nakajima,  M.  Tatsumisago,  A.  Sakuda,  A.  Hayashi and  S.  Mori: ACS Appl. Energy Mater. 4 (2021) 6290–6295. doi:10.1021/acsaem.1c01074
• 16)   J.L.  Lábár: Ultramicroscopy 103 (2005) 237–249. doi:10.1016/j.ultramic.2004.12.004
• 17)   H.  Tsukasaki,  S.  Mori,  S.  Shiotani,  H.  Yamamura and  H.  Iba: J. Power Sources 369 (2017) 57–64. doi:10.1016/j.jpowsour.2017.09.070
• 18)   M.  Ishimaru: Mater. Trans. 62 (2021) 1420–1423. doi:10.2320/matertrans.MT-M2021094
• 19)   X.  Mu,  S.  Neelamraju,  W.  Sigle,  C.T.  Koch,  N.  Totò,  J.C.  Schön,  A.  Bach,  D.  Fischer,  M.  Jansen and  P.A.  van Aken: J. Appl. Crystallogr. 46 (2013) 1105–1116. doi:10.1107/S0021889813011345
• 20)   A.  Hirata,  Y.  Hirotsu,  T.  Ohkubo,  T.  Hanada and  V.Z.  Bengus: Phys. Rev. B 74 (2006) 214206. doi:10.1103/PhysRevB.74.214206
• 21)   T.E.  Gorelik,  R.  Neder,  M.W.  Terban,  Z.  Lee,  X.  Mu,  C.  Jung,  T.  Jacob and  U.  Kaiser: Acta Crystallogr. Sec. B 75 (2019) 532–549. doi:10.1107/S205252061900670X
• 22)   D.T.  Tran,  G.  Svensson and  C.W.  Tai: J. Appl. Crystallogr. 50 (2017) 304–312. doi:10.1107/S160057671601863X
• 23)   M.  Ishimaru: Nucl. Instrum. Methods Phys. Res. B 250 (2006) 309–314. doi:10.1016/j.nimb.2006.04.129
• 24)   M.  Nose,  A.  Kato,  A.  Sakuda,  A.  Hayashi and  M.  Tatsumisago: J. Mater. Chem. A 3 (2015) 22061–22065. doi:10.1039/C5TA05590C
• 25)   X.  Mu,  A.  Mazilkin,  C.  Sprau,  A.  Colsmann and  C.  Kübel: Microscopy 68 (2019) 301–309. doi:10.1093/jmicro/dfz015
• 26)   X.  Mu,  D.  Wang,  T.  Feng and  C.  Kübel: Ultramicroscopy 168 (2016) 1–6. doi:10.1016/j.ultramic.2016.05.009