Structure refinement and chemical analysis of Cs3Li(DSO4)4, formerly 'Cs1.5Li1.5D(SO4)2'

Wim Klooster, Ross Piltz (ANSTO), Tetsuya Uda and Sossina Haile (CalTech)

The proton transport properties of the compound 'Cs1.5Li1.5D(SO4)2' have attracted some interest because of its high symmetry at ambient temperatures, and apparently unusual hydrogen bond network. Specifically, the compound is reported to possess hydrogen-bonded chains of SO4 groups, in which the average bond occupancy by protons is 1/6.

Any understanding of the proton transport mechanisms requires a complete and accurate assessment of the proton sites within the structure. However, this is generally not possible by X-ray diffraction. The present neutron diffraction study was carried out in order to unequivocally determine the location of protons/deuterons within the structure.

Preliminary analysis of the neutron diffraction data suggested that the composition of the Cs-Li acid sulfate differed from that reported earlier, prompting chemical analysis. Measurement of the Cs and Li concentrations in the crystals was performed by ICP-MS (inductively coupled plasma mass spectrometry). Using these methods the Cs and Li mass percentages were found to be 50% (±5%) and 0.79% (±0.04%), respectively, for the deuterated compound and 51.1% (±0.6%) and 0.895% (±0.011%) for the protonated. These values imply molar ratios of Cs/Li of 3.3 (±0.3) and 2.98 (±0.04) for the two materials.

The atomic positions from the X-ray study of 'Cs1.5Li1.5D(SO4)2' were used as the starting structure. It soon became apparent that the deuterium atom position was in error, and that the Li site occupancy was significantly lower than one. A Fourier difference map was used to identify the appropriate deuterium site, and occupancies of both this atom and the Li atom were varied.

The Li atom site occupancy refined to a chemically meaningful value of ~1/3, implying, for reasons of overall charge balance, a site occupancy at the deuterium position of also 1/3. This diffraction result motivated the chemical analysis reported above, which also indicated a much lower Li content than reported in the X-ray study. Fixing the Li occupancy at 1/3 and allowing for partial occupation of the deuterium site by hydrogen yielded a D:H ratio of 85:15. The resulting stoichiometry of Cs3Li(D0.85H0.15SO4)4 implies Cs and Li mass percentages of 50.01 and 0.87 respectively, in good agreement with the chemical analysis.

Selected portions of the overall structure are shown in projection on (111) in Figure 1, and a thermal ellipsoid representation is presented in Figure 2.

Figure 1. The structure of Cs3Li(DSO4)4 shown in projection on (1 1 1); (a) structure from ~ (-0.05, -0.05, -0.05) to (0.32, 0.32, 0.32) and (b) from ~ (0.2, 0.2, 0.2) to (0.54, 0.54, 0.54). Only complete tetrahedra are shown. Those with large spacing between shading lines are LiO4 groups and the remainder SO4 groups. Numbers on Cs atoms and tetrahedral groups correspond to Cs, Li and S atom elevations in ? relative to the central S atom in (a) at 0.0867, 0.0867, 0.0867. The apparent proximity of the Li and D atoms is accounted for by the partial occupancies, as discussed in the text.

Figure 2. Thermal ellipsoid representation of the structure of Cs3Li(DSO4)4, shown in projection on (1 1 1). Portion of the structure depicted corresponds to the center of Fig. 1(a).

The overall structural features of this compound are discussed only briefly here. Eight oxygen atoms coordinate the single crystallographically distinct cesium atom with Cs-O bond distances of less than 3.35 ?, whereas both the lithium and sulfur atoms are tetrahedrally coordinated, Figure 1.

In the X-ray structure refinement [1], the Li site was assumed to be fully occupied, and, because of the weak X-ray scattering power of Li, the model reached convergence with good refinement statistics. It is noteworthy, however, that the thermal displacement parameter obtained for Li in that work was unusually high [Ueq = 0.15?2] suggesting that that crystal had a similar partial occupancy on the Li site.

In the present work, the deuterium ion position was directly located by Fourier difference maps, and because of the large scattering length of D, there is no ambiguity in its position. It was found to reside on the contracted edge of the LiO4 tetrahedron, providing an explanation for the severe distortion of this tetrahedral group. This positioning, together with the occupancies obtained for the Li and D site, can be interpreted as follows. Of the possible LiO4 tetrahedra, only 1/3 are filled with Li atoms. When the Li ion site is filled, the possible hydrogen bonds at its two edges are unoccupied.

In the remaining 2/3 of cases, both hydrogen bonds are occupied, with deuterium ions residing in double minimum potential wells such that the deuterium ion site occupancy is lowered by a factor of 2 to 1/3. The situation is illustrated in Figure 3. The fact that the thermal displacement parameters of O(2) are not enhanced relative to the other atoms in the structure suggests that the local bonding configuration, and in particular the position of O(2), is equally well suited to both the LiO4  tetrahedron and the hydrogen bond.

Figure 3. Schematic representation of the structure in the region of the hydrogen bond and lithium atom site. The configuration on the left, in which the Li site is occupied and the hydrogen bonds on the edges of the tetrahedra are unoccupied, occurs for 1/3 of the tetrahedra, whereas the configuration on the right, in which the hydrogen bonds are occupied and the Li site is unoccupied, occurs for the remaining 2/3 of the tetrahedra.

The reanalysis of the structure of the compound formerly regarded as Cs1.5Li1.5X(SO4)2, where X = H or D, leads to a reexamination of the possible proton transport mechanism. Earlier, it was suggested that protons move along hydrogen bonded chains that extend along [1 1 1] via the reorientation of bisulfate groups and subsequent H transfer along asymmetric bonds, and that this reorientation was hindered by the presence of lithium ions.

One would expect that even for the corrected structure, the overall steps of (HSO4) group reorientation and proton transfer along hydrogen bonds apply, as they do to the majority of acid sulfate and selenate compounds [9]. In Cs₃Li(XSO₄)₄, bisulfate group reorientations are likely to occur around the [1 1 1] axes, moving protons from occupied to unoccupied hydrogen bonds. The close proximity of lithium ions to the normally unoccupied hydrogen bonds, 1.394(1) ?, would render this reorientation energetically unfavourable, and could explain the high activation energy for proton conductivity.

The compound formerly reported as Cs1.5Li1.5X(SO4)2 (X = H or D) has been shown here to have a true stoichiometry of Cs3Li(XSO4)4. Sulfate groups in this compound are linked together either by pairs of hydrogen bonds or by a single LiO4 tetrahedron. The hydrogen bonds nominally exist at two of the edges of the LiO4 group, however, when the proton sites are occupied (and the hydrogen bonds present) the Li site is unoccupied and vice versa.

The average Li site occupancy is 1/3, the average hydrogen bond occupancy is 2/3, and because of the double minimum in the hydrogen bond, the average D (or H) site occupancy is 1/3. Despite the partial occupancies, Cs₃Li(XSO₄)₄ has a fixed overall stoichiometry. Proton transport is likely to occur via the reorientation of (HSO4) groups about the [1 1 1] axes, bringing protons from occupied to otherwise unoccupied hydrogen bonds.

As the latter sit, on average, on the edges of occupied LiO4 tetrahedra, this reconfiguration must be associated with a large increase in energy, and may explain the high activation energy for proton conductivity. Thus, even the high cubic symmetry and the partial occupancy of the proton/deuterium ion sites are not sufficient to ensure 'superprotonic' conductivity in Cs₃Li(XSO₄)₄ , as is known, for example, in the high temperature phase of CsHSO₄.

Reference

W. T. Klooster, R. O. Piltz, T. Uda and S. M. Haile, Journal of Solid State Chemistry, 177, 274-280, 2004.