Skip to content
BY 4.0 license Open Access Published by De Gruyter (O) June 3, 2020

The ternary phase Li8SbxSn3-x with 0.3 ≤ x ≤ 1.0

  • Patric Berger , Clemens Schmetterer , Herta Silvia Effenberger and Hans Flandorfer EMAIL logo


In the frame of the studies on the phase relations in the ternary system Li–Sb–Sn at 300 °C the new ternary phase Li8SbxSn3-x (0.3 ≤ x ≤ 1.0) was synthesized and characterized predominantly by single crystal and powder X-ray diffraction. The title compound crystallizes trigonally in the space group R3m (no. 166), the lattice parameters are a = 4.6962(11) Å and c = 31.536(6) Å. The crystal structure of Li8SbxSn3-x is described in the present paper. In addition, the stereochemical and topological relations to the phases with similar composition, namely Li13Sn5, Li5Sn2 as well as cubic Li3Sb, besides native Li are discussed.

1 Introduction

Currently employed rechargeable Li-ion batteries (LiBs) are widely applied in handheld electronics such as notebooks, cell phones or cameras. However, for high-power applications, like stationary storage systems or electric mobility, they are insufficient in terms of capacity and cyclability. To improve the performance of today’s LiBs, further developments are required for the three fundamental constituents, namely the anode, the cathode and the electrolyte.

For the anode, intermetallic materials seem possible alternatives to the commonly used graphite (372 mAh g−1), since e. g. Sb and Sn show theoretical specific capacities of 660 and 994 mAh g−1, respectively [1]. Unfortunately, expected volume changes during (de)lithiation lead to mechanical degradation of the electrode and restrict the lifetime of respective batteries. One tentative possibility to overcome this limitation is the use of Sb–Sn alloys instead of the pure elements only [2], [3], [4]. Lithiation of these alloys starts with the formation of the phase Li3Sb going along with the precipitation of Sn. The fine grained tin matrix can buffer mechanical stress from formation of lithiated phases. Another possibility to overcome these drawbacks is the co-precipitation of a material which is completely inactive against lithium (like the transition metals Ti or Nb), in combination with the formation of lithiated Sb- and Sn-compounds. Especially the ternary compounds SbSnTi and Ti1−yNbySnSb (0 ≤ y ≤ 1) have been studied in detail [5], [6]. The lithiation path of these ternary compounds, synthesized by ball milling of the pure elements, was investigated by operando-X-ray diffraction (XRD) and Mössbauer techniques and resulted in a simultaneous conversion to Li–Sb and Li–Sn alloys during the lithiation process in which pure transition metals are precipitated contemporaneously. Thereby, Li3Sb and Li7Sn2 represent the highest lithiated phases. However, the authors have not been aware of phase relations in Li–Sb–Sn and Ti–Sb–Sn. In order to understand, control and design the (de)lithiation processes, further and detailed knowledge about relevant phases, phase transitions, thermodynamic properties and crystal structures of the involved phases is indispensable.

The Li–Sn system [7] contains at least seven intermetallic compounds which exhibit narrow homogeneity ranges. Five of them are Li-rich and close in composition where the Li17Sn4 compound is the richest in Li. In addition to the seven well described intermetallic phases, there are hints from e.m.f. investigations for a further compound, Li8Sn3 [8], [9]. Interestingly, in the Li–Pb system, an equilibrium compound Li8Pb3 was found [10], [11], while in the Li–Sn system the existence of the corresponding analogue was never proved or described by means of thermal analysis or crystal structure investigations. So far, this compound has not yet been isolated or structurally characterized; calculations of the stability or the electronic properties are available from [12], who calculated the formation energy of several thousands of hypothetical Li–Sn compounds using various crystal structures through discrete Fourier transform (DFT). From these they selected those with the lowest formation energies as possibly stable phases in addition to the known ones, including Li8Sn3. However, none of these additional compounds have been found so far and described through experiments. Furthermore, their calculations are only valid for a temperature of 0 K.

In the Li–Sb system [13], the most Li-rich compound is Li3Sb [14], [15], the second stable phase is Li2Sb. Phase diagrams of both systems exhibit similarities in their general outline (steep Li-rich liquidus line and high melting regime), but differ in the type and number of intermetallic phases. The system Sb–Sn was recently discussed and described in detail by Schmetterer et al. [16], [17]. The well-known Sb–Sn phase which exhibits a large homogeneity range was found to be incommensurately modulated, the compound Sb2Sn3 was proved to be an artifact while a new commensurate phase Sb3Sn4 with small homogeneity range was established.

There is no clear evidence for a ternary compound in Li–Sb–Sn from literature. Rönnebro et al. [18] report a compound Li2.78Sn0.22Sb which was supposed to form by the lithiation of an Ag–Sb–Sn alloy (Ag36.4Sb15.6Sn48) by ball milling. The existence of such a phase was discussed in more detail in another paper [19] by the same group doing electrochemical lithiation of Ag36.4Sb15.6Sn48. However, the evaluation of powder-XRD (PXRD) patterns is not at all convincing. Positions and intensities of peaks allocated to Li2+xSbSn1-x cannot be distinguished from these of Li3Sb which is as well formed on lithiation. In addition, a further isotypic phase, Ag2-xLi1+xSn is present in some of the alloys investigated, whose XRD-pattern can hardly be distinguished from a hypothetic Li2+xSbSn1-x. No single crystals were measured and for our opinion there is no clear proof for the existence of such a Li–Sb–Sn compound.

During our work on the phase diagram of the ternary Li–Sb–Sn system, the existence of a new ternary Li-rich compound was indicated from powder X-ray patterns of several samples showing additional but comparable diffraction peaks. From the single-crystal structure refinement the chemical formula Li8SbxSn3-x (0.3 ≤ x ≤ 1.0) was derived.

2 Experimental

2.1 Synthesis

For sample preparation, pure Li wire (Alfa Aesar, 3.2 mm, 99.8%, with oil coating) which was stored in a glovebox under Ar atmosphere (Ar 5.0; O2 < 1 ppm; H2O < 1 ppm), tin rods (Alfa Aesar, 999985) and Sb (Alfa Aesar, ingot, 99,999%) were used. Initially, the lithium wire was cleaned outside the glovebox in an ultrasonic bath in cyclohexane, followed by vacuum evaporation of the solvent in the antechamber of the glovebox. Surface oxidation products of the lithium wire caused by short time exposure to air were mechanically scraped off with a shear before use. Sb was further purified by filtration of the liquid metal through quartz glass wool under vacuum.

The single crystal for the present structure determination was selected from a sample with input composition Li75Sb12.5Sn12.5 that was initially prepared for the determination of the phase equilibria. This sample was made through weighing the required amounts of the pure elements inside a glovebox (Ar 5.0; O2 < 1 ppm; H2O < 1 ppm) and put into a tantalum crucible. The Ta-crucible was placed on a water cooled Cu mount in an electric arc furnace and sealed by arc welding under Ar-atmosphere after evacuating and flushing the chamber of the arc furnace three times. The sealed crucible was transferred into an induction furnace, where the sample was melted at 1,000 °C for 10–20 s. To ensure homogeneous mixing of the elements, the melting process was repeated twice with turning the Ta-crucible upside down after each melting sequence. In a final step, the crucible itself was sealed in a quartz glass tube under vacuum and annealed at 300 °C in a muffle furnace for four weeks. Thereafter, the sample was quenched in cold water, put back to the glovebox again and opened with a crusher. The PXRD pattern of the above mentioned sample Li75Sb15Sn15 shows the ternary phase Li8SbxSn3-x besides the binary Li3Sb (Figure 1), which is in accordance with preliminary phase diagram data from work in progress. The broad humps in the background stem from the amorphous polycarbonate cap of the sample holder (see below).

Figure 1: Powder X-ray diffraction (PXRD) pattern of the sample obtained from an input Li75Sb12.5Sn12.5: observed, calculated and difference pattern.
Figure 1:

Powder X-ray diffraction (PXRD) pattern of the sample obtained from an input Li75Sb12.5Sn12.5: observed, calculated and difference pattern.

2.2 Powder X-ray diffraction (PXRD)

Small amounts of the obtained alloy were powdered in a glovebox using a Durit® mortar. Powder patterns were recorded on a Bruker D8 diffractometer with Bragg-Brentano pseudo-focusing geometry (θ/2θ-geometry) with Cu X-radiation (40 kV/40 mA/Ni filter). The scattered intensities were detected using a one dimensional Si-strip detector (LynxEye detector, Bruker, Germany). The powdered alloy was fixed with dried petroleum jelly on a silicon-single crystal, which served as sample holder. To prevent oxidation of the powder during the measurement, the sample was covered by a polycarbonate cap with an O-ring tightening under Ar-atmosphere. For phase analysis full profile Rietveld-refinement was applied using Topas4® software supplied by Bruker, Germany.

2.3 Single-crystal X-ray diffraction (SCXRD)

A small crystal chip of the title compound was fixed with and covered by vacuum grease on a glass capillary. Multiple trials were necessary to find a single crystal of sufficient quality for X-ray data collection. For investigations, a four-circle Nonius Kappa diffractometer was used (charge-coupled device (CCD) detector, 300 µm capillary optics collimator, conventional X-ray tube, monochromated Mo radiation). To prevent decomposition of the samples in air, the crystals were kept in a dry N2-gas stream at 290 K during the measurement. The unit-cell parameters were obtained by least-squares refinements from the 2ϑ values of the positions of the measured Bragg reflections. Corrections for Lorentz polarization and absorption effects (multi-scan method) were applied; complex scattering functions [20] were used. For data collection, structure solution and refinements served the programs “Collect” [21], [22], “SHELXS-97” and “SHELXL-97” [23], [24], [25], [26]. Reflection statistics suggested centrosymmetry. The extinction symbol and possible space-group symmetries were found from the extinction rules; R3m is evident from structure refinements. The two positions of the atoms M = (Sb and/or Sn) were obtained by direct methods in combination with Patterson summations. Successive Fourier and difference Fourier summations revealed the four remaining atom sites.

3 Results and discussion

3.1 Crystal structure determination and refinements

In the first step of the structure refinement the scattering functions of Sb and Li were applied for the two M (=Sn, Sb) and four Li sites. However, for the Li1 site a slight excess in the electron density was observed due to a minor substitution by M atoms. During the refinement full occupation was assumed for the Li1 site but the ratio of scattering function Sn:Li was allowed to vary. However, one cannot exclude a larger amount of M atoms going along with the formation of vacancies.

As the chemical elements Sb and Sn are adjacent in the periodic table, a distinction using conventional X-ray sources is not likely although there is a hint for the distribution of Sb and Sn on the M1 and M2 position, respectively. The R values change slightly for structure models based on distinct scattering functions used. Trials using the scattering functions Sb/Sb – Sb/Sn – Sn/Sb – Sn/Sn for the atomic positions M1/M2, respectively, result in the following residual values R1(all reflections)/R1(observed reflections)/wR2: 0.026/0.020/0.043 – 0.024/0.019/0.044 – 0.029/0.024/0.046 – 0.026/0.021/0.046. The refinements of these distinct structure models give a hint that Sb atoms may show a site preference for the M1 position 3(a), and Sn atoms for the M2 6(c) position. It is consistent with the expected Sb:Sn ratio of 1:2 and result ideally in an ordered atomic arrangement with respect to the two M positions. Such a preferred occupation is supported by crystal chemical considerations (see Section 3 below).

Somewhat large isotropic displacement parameters were obtained for the positions Li2 and Li3. It should be mentioned that the r.m.s. amplitudes of these two atom sites indicate an approximate isometric behavior: 0.034 and 0.031 Å2 parallel to [001] respectively and 0.027 and 0.030 Å2 in (00.1). The displacement parameters of the Li2 and Li3 atoms indicate a somewhat too high scattering power assumed in the actual structure model; minor vacancies at these two sites are feasible. Due to high correlation terms with the displacement parameters the refinement of the two site occupation factors is not significant. A partial substitution at all the Li position by M atoms going along with the formation of a significant amount of vacancies cannot be ruled out. The calculated volume of the coordination polyhedra around the four crystallographically different Li atoms compare well (see Table 2).

Details of data collection and structure refinements are compiled in Table 1, structural parameters and interatomic distance in Table 2 (including [27], [28], [29]) and 3, respectively. The display of the crystal structures is based on the software ATOMS.

Table 1:

Single-crystal X-ray data-collection and structure refinements of Li8SbxSn3-x. Cell parameter obtained from PXRD are given in brackets.

a [Å]4.6962(11) [4.6975(2)]
c [Å]31.536(6) [31.538(2)]
Space groupR3m (no. 166)
V [ų]602.3
Z3 {Li7.94Sn2.06Sb}
ρcalc [g cm−3]/µ(Mo) [mm−1]3.49/9.6
Crystal dimensions [µm]70 × 80 × 90
Range of data collection (±h ± k ± l) [°]3 < 2ϑ < 70
Number of images/rotation angle per image [°]1294/2
Scan mode (at 11 distinct ω-angles)φ-scans
Scan time [s/°]/frame size (binned mode)65/621 × 576 pixels
Detector-to-sample distance [mm]32
Measured reflections2205
Unique reflections (n)/reflections with Fo > 4σ(Fo)388/352
Rint = Σ∣Fo2-Fo2(mean)∣/ΣFo20.039
Extinction parameter k: Fc* = Fc·k[1 + 0.001·Fc2λ3/sin(2ϑ)]−1/40.0019(3)
R1 = Σ(∣∣Fo∣−∣Fc∣∣)/ΣFo (for 388/352 reflections)0.024/0.019
wR2 = [Σw(Fo2Fc2)2/ΣwFo4]1/20.044
GooF = {Σ[w(Fo2Fc2)2]/(np)}0.51.06
Max Δ/σ; number of variable parameters (p)< 0.001; 20
Final difference Fourier map [eÅ−3]−0.79 to +0.86
w = 1/{σ2(Fo2) + [0.024 × P]2 + 0.75 × P}; P = ([max(0,Fo2)] + 2 × Fc2)/3]
Table 2:

Fractional atomic coordinates and displacement parameters of Li8SbxSn3-x.

AtomScattering functionWyckoff letterSite symmetryzUequivU11U33U12V [ų]CN (H)CN (O’K)

The anisotropic displacement parameters are defined as: exp [−2π2 Σ3i = 1 Σ3j = 1 Uijai ajhihj]. x = y = 0, U22 = U11, and U23 = U13 = 0 for all atoms.

V [ų]: Volume of the coordination polyhedron [29].

CN (H): coordination number according to Hoppe [27], [29].

CN (O’K): coordination number according to O’Keeffe [28], [29].

Table 3:

Interatomic bond distance (in Å) and average coordination distances for Li8SbxSn3-x. X is any atom in the coordination sphere of the central atom.

M1–Li12.811(4), 2xLi1–Li22.790(11)Li3–Li42.839(10)
M1–Li42.870(3), 6xLi1–M12.811(4)Li3–Li12.861(3), 3x
M1–Li33.310(5), 6xLi1–Li32.861(3), 3xLi3–Li42.876(4), 3x
Li1–M22.896(15), 3xLi3–Li23.012(15)
M2–M22.8544(7)Li1–Li43.294(6), 3xLi3–M13.310(5), 3x
M2–Li42.889(8)Li1–Li23.429(8), 3xLi3–M23.328(5), 3x
M2–Li12.896(15), 3x
M2―Li22.919(4), 3xLi2―Li12.791(11)Li4–Li32.839(10)
M2―Li23.240(6), 3xLi2―Li22.798(5), 3xLi4–M12.870(3), 3x
M2―Li33.328(5), 3xLi2―M22.919(4), 3xLi4–Li32.876(4), 3x
Li2–M23.240(6), 3xLi4–Li13.294(6), 3x
Li2–Li13.429(8), 3xLi4–Li43.300(9), 3x
first coordination spheresecond coordination spheretotal [14]-fold coordination

3.2 A note on the composition of the title compound

Prior to the description of the crystal structure, it is necessary to address the topic of the composition of the compound that arises from the presence of Li and the distribution of Sb and Sn.

It is impossible to determine the accurate composition of Li-containing materials by scanning electron microscope (SEM)/energy dispersive X-ray analysis (EDX) analysis since the intensity of the X-ray emission of Li is too low. Calculations of the Li-content from residual mass are highly inaccurate because of the low atomic mass of Li, especially compared to this of Sb/Sn. In addition, preparation of samples for SEM/EDX is seriously complicated by the reaction of Li-alloys with air and moisture as well as the decomposition in the vacuum chamber during the electron irradiation. Further analytical methods are not applicable as the synthesized samples are not phase pure. Individual aggregates are small and intergrowth of distinct phases always is observed. Moreover, the individual phases cannot be identified in the optical microscope as the appearance is practically identical. Consequently, the separation of a sufficient amount of phase pure material of the compound under investigation was not possible. The stoichiometric formulation of this new ternary structure in respect to the Sb/Sn ratio is therefore given as Li8SbxSn3-x. Preliminary phase diagram data suggest a lower limit of ∼3–4 at.% Sb; see also the last paragraph of this chapter. As discussed above, the structure model with Sb and Sn atoms occupying (predominantly) the M1 and M2 sites resulted in a slightly lower R-value, but a partial substitution of Sb by Sn atoms lowering the Sb-content cannot be ruled out completely. A homogeneity range of Li8SbxSn3-x up to approx. 9 at.% Sb which corresponds to full occupation of the M1 site with Sb is compatible with our investigations of phase relations. Accordingly, the homogeneity range is given as 0.3 ≤ x ≤ 1.0.

A further important aspect is the classification of Li8SbxSn3-x either as a true ternary compound or as the solid solution of Sb in a hypothetic compound Li8Sn3 [8]. This question is coupled to the existence of the binary Li8Sn3 compound (cf. Section 1), which is suggested by the investigations of [8] and [9], who observed additional steps in the e.m.f. vs. composition curves at the relevant stoichiometry. However, data have been gained from coulometric titration which easily shows artificial potential steps caused by slow lithiation kinetics. Furthermore, no direct observation of this binary compound by XRD, thermal analysis or other methods is available so far. Accordingly, the most recent version of the phase diagram by Li et al. [7] does not include the binary Li8Sn3, as well as most of other phase diagrams hitherto published. Neither phase analysis of numerous samples in the binary system Li–Sn reported in Li et al. [7] nor our present investigations in the ternary system Li–Sb–Sn reveal any phase with a binary composition Li8Sn3.

Furthermore, our results indicate that Li8SbxSn3-x qualifies as a ternary phase due to the existence of two- and three-phase fields between Li8SbxSn3-x, Li3Sb and the neighboring Li–Sn compounds Li13Sn5 and Li7Sn2. In terms of thermal stability, the ternary compound was observed at 300 °C as well as at 400 °C.

3.3 The crystal structure of the title compound

The title compound crystallizes in the trigonal crystal system with space group R3m (no. 166) and the lattice parameters a = 4.6962(11) and c = 31.536(6) Å. It is structurally related to Li8Pb3 and Li6Cu2Sn3 [11], [30]. The atomic arrangement within one unit cell is given in Figure 2.

Figure 2: The crystal structure of Li8SbxSn3-x in a projection along [11‾$&#x203e;{11}$0].
Figure 2:

The crystal structure of Li8SbxSn3-x in a projection along [110].

Due to the fact that in Li8SbxSn3-x only the sites 3(a) and 6(c) are occupied, all atoms within the unit cell are located on the (11 2 0) plane (Figure 3). The atoms within this plane form a regular arrangement: The crystal structure is composed by crystallographically identical but offset atom chains with rod symmetry P1 running parallel to [11 11 1] (one such chain is highlighted by a gray background in Figure 3). Each chain contains all atom types M1, M2, and Li1 to Li4, the repeating sequence consists of in total 11 atoms. The chains deviate only slightly from linearity; they are built by three building blocks formed by groups of atoms with distinct length and composition. With the simplification that the Li1 position is only occupied by Li atoms, all the three building blocks start with a M atom (either Sb = M1 or Sn = M2 atom), followed by two or three Li atoms (designated 2s or 3s, respectively): 2s = M2–Li2–Li2, 3s = M2–Li1–Li3–Li4, and 3s′ = M1–Li4–Li3–Li1. Nevertheless, the Li1 position is distinguished from Li2 – Li4 in Tables and Figures.

Figure 3: The (1 1 2‾$&#x203e;{2}$ 0) plane in the crystal structure of Li8SbxSn3-x, parallel to the direction [11‾$&#x203e;{11}$ 11 1]. The 3s3s′2s atomic sequence is highlighted by a gray background. Combining this sequence to chains respectively to planes, the entire crystal structure is formed. The unit cell is indicated by dotted lines (analogue projection as in Figure 2).
Figure 3:

The (1 1 2 0) plane in the crystal structure of Li8SbxSn3-x, parallel to the direction [11 11 1]. The 3s3s′2s atomic sequence is highlighted by a gray background. Combining this sequence to chains respectively to planes, the entire crystal structure is formed. The unit cell is indicated by dotted lines (analogue projection as in Figure 2).

The 2s and the 3s blocks start with a M2 atom followed by two and three Li atoms, respectively, while the 3s′ element is built from a M1 atom followed by three Li atoms arranged as the reverse of the Li sequence of 3s, because the M1 atom represents the inversion center of the rod group; i. e. the site symmetry 3m of the compound’s space group. Thereby, the whole crystal structure can be composed from these three different building blocks by forming the atom chains out of the sequence 3s, 3s′ and 2s that are in turn assembled into the layer by shifting (see Figure 3).

Further insight into the crystal structure is gained considering also the local environment of each atom in the chains, i. e. by extending the description to neighboring chains in the same layer as well as to those above and below. As a result, the crystal structure is described based on polyhedra sequences.

In Li8SbxSn3-x, the coordination number for each of the six different atoms is eight in the first coordination sphere (distorted cube) and six in the second one (distorted octahedron). The coordination figure may be described as a hexacapped cube [31] where the coordinating atoms represent the corners. This [8+6] coordination can be derived from the bcc W-structure type, which represents also the type structure of pure Li stable at ambient conditions (Figure 4 gives the coordination for each atom).

Figure 4: Polyhedral environment for all atoms in the crystal structure of the ternary compound Li8SbxSn3-x. The observed Li centered coordination polyhedra correspond with that described by [31] (see Table 4 and text). The atoms M1 and M2 were added as the title compound can be derived from the W-type structure as characteristic for Li under ambient conditions.
Figure 4:

Polyhedral environment for all atoms in the crystal structure of the ternary compound Li8SbxSn3-x. The observed Li centered coordination polyhedra correspond with that described by [31] (see Table 4 and text). The atoms M1 and M2 were added as the title compound can be derived from the W-type structure as characteristic for Li under ambient conditions.

A transformation of the unit cell of trigonal Li8SbxSn3-x according to [2/3 1/3 0.0602/−1/3 −2/3 0.0602/1/3 −1/3 −0.0602] results in the average cell with the metric a = 3.310, b = 3.310, c = 3.310 Å, α = 89.63, β = 89.63, γ = 90.37° which resembles cubic pseudo-symmetry; the unit is significantly smaller as compared to that given by Berliner et al. [32] for bcc-Li metal: a = 3.47851(1) Å. The cell volumes of the title compound and bcc-Li are 602.3 ų (for 33 atoms per unit cell) and 42.09 ų (Z = 2) giving 18.25 respectively 21.05 ų per atom; it is worth noticing that the coordination polyhedra in bcc-Li are definitely larger.

Starting point for the following discussion is the hexacapped cube that only contains Li atoms. Note that in contrast to the ideal aristotype bcc-structure type the hexacapped cubes in the ternary compound does not exhibit a regular [14]-fold coordination (and consequently deviates from an Archimedean solid), but show a slight angular distortion besides some moderate differences in the bond lengths; moreover, the point symmetries of the coordination polyhedra are only 3m or 3m in the title compound but m3m in Li metal.

Various hexacapped cubes centered by Li atoms but coordinated by various arrangements of Li, Sn (or Pb) atoms, respectively, were discussed and classified by Frank and Müller [33] who assigned the letters “a” to “n” to the various resulting hexacapped cubes. These authors focused on Li-centered hexacapped cubes occurring in Li-plumbides and −stannides, but a systematic classification for Sn- or Pb-centered hexacapped cubes is still missing. Furthermore, they studied the occurrence of various hexacapped cubes in the crystal structures of Li-plumbides and −stannides, but did not analyze the structures for any repeating hexacapped cube sequences.

Li8SbxSn3-x exhibits two hexacapped cubes S1 and S2 centered by M1 and M2 atoms, respectively (Tables 4 and 5); the hexacapped cube centered by an M1 atom is only coordinated by Li atoms, while the one centered by an M2 atom contains another M2 atom in the first coordination sphere (see Figure 4).

Table 4:

Central atom, number of ligands and nomenclature for the various polyhedra in Li8SbxSn3-x.

Central AtomLigandsNomenclature according to [33]


[10 Li + 1 M1 + 3 M2]



[4 Li + 4 Li1 + 6 M2]



[5 Li + 3 Li1 + 3M1 + 3 M2]



[7 Li + 3 Li1 + 3 M1 + 1 M2]


Nomenclature from this work


[13 Li + 1 Li1]



[10 Li + 3 Li1 + 1 M2]


Table 5:

Correlation between atom sequence and polyhedra sequence.

DesignationAtom sequencePolyhedra sequence

The atom sequences along [11 11 1] discussed above can be extended to the respective hexacapped cubes sequence; the nomenclature applied by Frank and Müller [33] for the coordination polyhedra of the Li[8+6] atoms is given in analogy for the M-centered hexacapped cubes designed as S1 (coordinated by Li atoms only) and S2 (coordinated by an M2 atom in the first coordiantion sphere besides Li atoms), respectively (see Tables 4, 5, Figure 5). Like the building blocks defined for the atom sequences, each hexacapped cube sequence starting with S1 or S2 defines the corresponding hexacapped cube building block (Figure 5). For clarity, only every second polyhedron is expressed by sticks. The hexacapped cube sequence is constructed similar to the 3s3s′2s atom sequence, (Figure 5). Thereby, the hexacapped cubes are ordered along the 3s3s′2s-sequence using the M and Li atoms as central atoms that are coordinated by the hexacapped cube types according to the scheme from Frank and Müller [33] as listed in Table 4.

Figure 5: The hexacapped cubes around atoms along [11‾$&#x203e;{11}$ 11 1] in Li8SbxSn3-x. For clarity, the hexacapped cube of only every second central atom is drawn. The sequences of central atoms and hexacapped cubes are indicated.
Figure 5:

The hexacapped cubes around atoms along [11 11 1] in Li8SbxSn3-x. For clarity, the hexacapped cube of only every second central atom is drawn. The sequences of central atoms and hexacapped cubes are indicated.

The 3s3s′2s atom sequence and the corresponding hexacapped cube sequence reflect the composition of the title compound Li:Sb:Sn = 8:1:2 (neglecting minor atom substitutions and/or mixed positions and tentative vacancies). Due to the fact that the compositions of Li–Sn phases neighbored in the phase diagram differ by a few at % only, similar atom environments as well as analogue atomic and hexacapped cube sequences are likely.

For a topological comparison, not only the occurrence of analogue hexacapped cubes but also of comparable 1D-atomic and 1D-hexacapped cube sequences, respectively, is considered. Each sequence is defined starting with the M-centered hexacapped cube (M is either a Sn- or Sb atom), followed by only Li-centered hexacapped cubes (not taking into account any mixed occupancies).

Applying these principles to the atomic arrangement in Li13Sn5 [33] and Li5Sn2 [34] revealed that both crystal structures can be described by analogue S1/S2―b―a―b and S2―d―d sequences only (Table 6) whereby both atomic positions, M1 and M2 are occupied by Sn only. However, the ratio of the individual building blocks differs to achieve the respective Sn:Li ratios. The structural and topological relationship to the other Li-Sn phases known so far in the Li-rich as well as in the Sn-rich area, requires extended discussion; it will be conducted elsewhere.

Table 6:

Polyhedra sequence in the title compound and phases with similar ratios M:Li.

CompoundsPolyhedra sequence/atom sequence
Li8SbxSn3-x [this work]S2-b-a-b/3sS1-b-a-b/3s′S2-d-d/2s
Li13Sn5 [31]S2-b-a-b/3sS1-b-a-b/3s′S2-d-d/2sS2-b-a-b/3sS2-d-d/2s
Li5Sn2 [32]S2-b-a-b/3sS2-d-d/2s
c-Li3Sb [12], [13]S1-b-a-b/3s′

Within this group formed by the three compounds it can be observed that M1[8+6] atoms are coordinated by Li atoms only (labeled S1) but M2[8+6] atoms have one M atom besides Li atoms within their first coordination sphere (labeled S2). In the crystal structure of cubic Li3Sb [14] only the sequence S1-b-a-b occurs, i. e. the 3s′ block with Sb at the M1 position forms the entire atomic arrangement (see Table 6). This is consistent with the ternary title compound Li8SbxSn3-x: the site multiplicity of the M1 and M2 position and the structure refinements suggest an allocation of the M1 and M2 positions predominantly by Sb and Sn atoms, respectively. The occurrence of S1 in Li13Sn5 shows that this cuboctahedron can also be centered by Sn, which may suggest a mixed occupancy by Sb and Sn in S1 in Li8SbxSn3-x, which explains the difference between the ordered stoichiometry “Li8SbSn2” (∼9 at.% Sb) from SC-XRD and the ∼3–4 at.% Sb according to our preliminary phase diagram analysis. In any case, Sb only centers the S1 hexacapped cube.

Recently the crystal structure of Li6Cu2Sn3 was published by Fürtauer et al. [30]. Despite the different stoichiometry and the presents of copper, its crystal structure is isotypic to Li8SbxSn3-x; however, the atom distribution over the various atom sides differs: the two M-position of Li8SbxSn3-x are occupied by Sn in Li6Cu2Sn3; Li3 corresponds to Li1* (the “*” denotes positions from Li6Cu2Sn3), which are only occupied by lithium in both structures, the remaining Li-positions have mixed Li + Cu occupancies in Li6Cu2Sn3.

Fürtauer and co-works [30] characterized this structure by different layers composed of face-sharing octahedra, tetrahedra and trigonal prisms. Li6Cu2Sn3 can be described by the repeated sequence of three layers: A1-B-A1, where A1 is made from ocathedra and tetrahedra, while B contains trigonal prisms. While at first glance this description differs from ours, the similarity can be established through identification of A1-B-A1 with our atom sequence 3s2s 3s′ = 3s3s′2s. In complete analogy the layer sequences A1-B-A1-B-A1 given for Li13Sn5 and A1-B for Li5Sn2 in [30] correspond to 3s3s′2s3s2s and 3s2s.

4 Conclusion

It is expected that atomic environments in adjacent compounds in phase diagrams may be comparable and consequently, their crystal structures exhibit similarities to some extent. As shown in this work, the title compound is derivable from the bcc structure type of the common Li modification: parts of the Li atoms are substituted by Sb or Sn atoms, the cubic cell is trigonal rhombohedrally distorted and the unit-cell volume is increased. It is demonstrated that the crystal structures of ternary compounds may be described by building blocks related to these observed in the binary constituent phase diagrams.

The compounds Li13Sn5, Li5Sn2 and Li8SbxSn3-x are built from the hexacapped cube-types a, b, d, S1 and S2. However, the crystal structures of these compounds are characterized by distinct sequences of the larger building blocks designated 3s, 3s′ and 2s. The cubic modification of Li3Sb exhibits an S1-b-a-b sequence that builds up the entire crystal structure. Additionally, as this building block is identical to that in the ternary compound, we assume that Sb atoms predominantly occupy the M1 position in Li8SbxSn3-x. The admixture of Sn at this position cannot be ruled out since S1 building blocks with Sn at the M1 position occur in Li13Sn5. Thus, these crystal structures exhibit rather general similarities with respect to their topology.

The structure models involving the hexacapped cubes labelled a, b, d, S1 and S2, as well as the building blocks 3s, 3s′ and 2s, respectively, is applicable to the compounds mentioned above. However, the topological description of Li-rich as well as Sn-rich alloys requires additional polyhedra types [33]. A further investigation of the structural relations between the ternary compound Li8SbxSn3-x and the binary Li–Sn phases is currently in progress.

Corresponding author: Hans Flandorfer, Institute of Inorganic Chemistry - functional Materials, University of Vienna, Faculty of Chemistry, Althanstraße 14 (UZAII), 1090 Vienna, Austria, E-mail:

Funding source: DFG

Award Identifier / Grant number: FL-730/1-2


We thank Mr. P. Wibner for doing preliminary experiments.

  1. Author contribution: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission.

  2. Research funding: This work was supported by the DFG project FL-730/1-2 within the priority program SPP 1473, “WeNDeLIB”.

  3. Conflict of interest statement: The authors declare no conflicts of interest regarding this article.


1. Zhang, W. J. A review of the electrochemical performance of alloy anodes for lithium-ion batteries. J. Power Sources 2011, 196, 13; doi: in Google Scholar

2. Winter, M., Besenhard, J. O. Electrochemical lithiation of tin and tin-based intermetallics and composites. Electrochim. Acta 1999, 45, 31; doi: in Google Scholar

3. Wachtler, M., Besenhard, J. O., Winter, M. Tin and tin-based intermetallics as new anode materials for lithium-ion cells. J. Power Sources 2001, 94, 189; doi: in Google Scholar

4. Trifonova, A., Wachtler, M., Winter, M., Besenhard, J.O. Sn-Sb and Sn-Bi alloys as anode materials for lithium-ion batteries. Ionics 2002, 8, 321; doi: in Google Scholar

5. Sougrati, M. T., Fullenwarth, J., Debenedetti, A., Fraisse, B., Jumas, J. C., Monconduit, L. TiSnSb a new efficient negative electrode for Li-ion batteries: mechanism investigations by operando-XRD and Mössbauer techniques. J. Mater. Chem. 2011, 21, 10069; doi: in Google Scholar

6. Marino, C., Sougrati, M. T., Darwiche, A., Fullenwarth, J., Fraisse, B., Jumas, J. C., Monconduit, L. Study of the series Ti1-yNbySnSb with 0 ≤ y ≤ 1 as anode material for Li-ion batteries. J. Power Sources 2013, 244, 736; doi: in Google Scholar

7. Li, D., Fürtauer, S., Flandorfer, H., Cupid, D.M. Thermodynamic assessment and experimental investigation of the Li-Sn system. Calphad 2014, 47, 181; doi: in Google Scholar

8. Gasior, W., Moser, Z., Zakulski, W. Thermodynamic studies and the phase diagram of the Li-Sn system. J. Non-Cryst. Solids 1996, 205, 379; doi: in Google Scholar

9. Giel, H., Henriques, D., Markus, T. Investigations of the Li8Sn3 phase in the binary Li-Sn system using an improved coulometric-titration setup. J. Electrochem. Soc. 2017, 164, A907; doi: in Google Scholar

10. Zalkin, A., Ramsey, W., Templeton, D. Intermetallic compounds between lithium and lead. II. The crystal structure of Li8Pb3. J. Chem. Phys. 1956, 60, 1275; doi: in Google Scholar

11. Cenzual, K., Gelato, L. M., Penzo, M., Parthé, E. Overlooked trigonal symmetry in structures reported with monoclinic centred Bravais lattices; trigonal description of Li8Pb3, PtTe, Pt3Te4, Pt2Te3, LiFe6Ge4, LiFe6Ge5, CaGa6Te10 and La3.266Mn1.1S6. Z. Kristallogr. 1990, 193, 217; doi: in Google Scholar

12. Mayo, M., Morris, A. J. Structure prediction of Li-Sn and Li-Sb intermetallics for lithium-ion batteries anodes. Chem. Mater. 2017, 29, 5787; doi: in Google Scholar

13. Beutl, A., Cupid, D., Flandorfer, H. The Li-Sb phase diagram part I: new experimental results. J. Alloys Compd. 2017, 695, 1052; doi: in Google Scholar

14. Brauer, G., Zintl, E. Konstitution von Phosphiden, Arseniden, Antimoniden und Wismutiden des Lithiums, Natriums und Kaliums. Z. Phys. Chem. 1937, 37, 323; doi: in Google Scholar

15. Saubanere, M., Ben Yahia, M., Lemoigno, F., Doublet, M. L. Influence of polymorphism on the electrochemical behavior of MxSb negative electrodes in Li/Na batteries. J. Power Sources 2015, 280, 695; doi: in Google Scholar

16. Schmetterer, C., Polt, J., Flandorfer, H. The phase equilibria in the Sb-Sn system - Part I: literature review. J. Alloys Compd. 2017, 728, 497; doi: in Google Scholar

17. Schmetterer, C., Polt, J., Flandorfer, H. The phase equilibria in the Sb-Sn system - Part II: experimental results. J. Alloys Compd. 2018, 743, 523; doi: in Google Scholar

18. Rönnebro, E., Yin, J., Kitano, A., Wada, M., Sakai, T. Comparative studies of mechanical and electrochemical lithiation of intermetallic nanocomposite alloys for anode materials in Li-ion batteries. Solid State Ionics 2005, 176, 2749; doi: in Google Scholar

19. Rönnebro, E., Yin, J., Kitano, A., Wada, M., Tanase, S., Sakai, T. Reaction mechanism of a Ag36.4Sb15.6Sn48 nanocomposite electrode for advanced Li-ion batteries. J. Electrochem. Soc. 2005, 152, A152; doi: in Google Scholar

20. Wilson, A. J. C. International Tables for X-ray Crystallography; Kluver: Dordrecht, The Netherlands, 1992.Search in Google Scholar

21. Otwinowski, Z., Minor, W. Processing of X-ray diffraction data collected in oscillation mode. Method. Enzymol. 1997, 276, 307; doi: in Google Scholar

22. Hooft, R. W. W. Collect; Nonius BV: Delft, The Netherlands, 1999.Search in Google Scholar

23. Sheldrick, G. M. SHELXS-97, A Program for the Solution of Crystal Structures; University of Göttingen: Germany, 1997.Search in Google Scholar

24. Sheldrick, G. M., SHELXL-97, A Program for Crystal Structure Refinement; University of Göttingen: Germany, 1997.Search in Google Scholar

25. Sheldrick, G. M. A short history of SHELX. Acta Crystallogr. A 2008, 64, 112; doi: in Google Scholar

26. Sheldrick, G. M. SHELXT – Integrated space group and crystal-structure determination. Acta Crystallogr. A 2015, 71, 3.10.1107/S2053273314026370Search in Google Scholar PubMed PubMed Central

27. Hoppe, R. Die Koordinationszahl – ein “anorganisches Chamäleon”. Angew. Chem. 1970, 82, 7; doi: in Google Scholar

28. O’Keeffe, M. A proposed rigorous definition of coordination number. Acta Crystallogr. A 1979, 35, 772; doi: in Google Scholar

29. Nowotny, H., Zobetz, E. “Kristallchemie”. Program for Solving Geometrical Problems in Crystal Structures; University of Vienna: Austria, 1982.Search in Google Scholar

30. Fürtauer, S., Effenberger, H.S., Flandorfer, H. New intermetallic phases in the Cu–Li–Sn system: the lithium-rich phases Li3CuSn and Li6Cu2Sn3. Z. Kristallogr. 2016, 231, 79; doi: in Google Scholar

31. FIZ KarlsruheA Focus on Crystallography; FIZ Karlsruhe: Germany, 2015. (accessed Jan 29, 2020).Search in Google Scholar

32. Berliner, R., Fajen, O., Smith, H. G., Hitterman, R. L. Neutron powder-diffraction studies of lithium, sodium and potassium metal. Phys. Rev. B 1989, 40, 12086; doi: in Google Scholar

33. Frank, U., Müller, W. Darstellung und Struktur der Phase Li13Sn5 und die strukturelle Verwandtschaft der Phasen in den Systemen Li-Sn und Li-Pb. Z. Naturforsch. 1975, 30b, 316; doi: in Google Scholar

34. Frank, U., Müller, W., Schäfer, H. Die Struktur der Phase Li5Sn2. Z. Naturforsch. 1975, 30b, 1.10.1515/znb-1975-1-202Search in Google Scholar

Received: 2020-02-26
Revised: 2020-04-22
Accepted: 2020-04-22
Published Online: 2020-06-03
Published in Print: 2020-07-28

© 2020 Patric Berger et al., published by De Gruyter, Berlin/Boston

This work is licensed under the Creative Commons Attribution 4.0 International License.

Downloaded on 8.2.2023 from
Scroll Up Arrow