Crystal structure of dicaesium strontium hexacyanidoferrate(II), Cs2Sr[Fe(CN)6], from laboratory X-ray powder data

Cs2Sr[Fe(CN)6] was synthesized. It is a potential phase to trap simultaneously caesium and strontium radionuclides. The crystal structure is isotypic with Cs2Sr[Mn(CN)6].


Chemical context
Ferrocyanides (FCN), such as Prussian blue, were discovered almost 300 years ago. The attractive properties of these materials for batteries and decontamination processes have ensured that FCNs remain an active research topic (Haas, 1993;Paolella et al., 2017). In particular, potassium copper FCN is currently being investigated for the purification of 137 Cs-contaminated water streams through partial exchange with potassium (Haas, 1993;Mimura et al., 1997). To the best of our knowledge, however, using FCNs to extract strontium either alone or with caesium has never been considered before. In the framework of the Center for Hierarchical Waste Forms (CHWM), an Energy Frontier Research Center (EFRC) funded by the US Department of Energy, we have been working on potassium copper FCN as an efficient K-ionic exchanger to capture 137 Cs and serve as a waste containment matrix (zur Loye et al., 2018). In this context, we have synthesized a cesium strontium FCN to study its efficiency in immobilizing both 90 Sr and 137 Cs, two radionuclides that are in most cases found together in radioactive water streams. Caesium strontium FCNs appear to be poorly described in the ICDD 2020 PDF4+ powder diffraction database (Gates-Rector & Blanton, 2019). Since the synthesized phase Cs 2 Sr[Fe(CN) 6 ] did not match with existing entries, we decided to characterize the structure and we report the results herein.   (Ziegler et al., 1989). As shown in Table 1, the lattice parameters of Cs 2 Sr[Fe(CN) 6 ] are slightly greater than those of Cs 2 Na[Mn(CN) 6 ], but the cell volumes differ by less than 0.3%. The crystal structure adopts the cryolite structure type and comprises a framework of corner-sharing [Sr(CN) 6 ] (dark green in Fig. 1) and [Fe(CN) 6 ] octahedra (brown in Fig. 1). Both types of octahedra exhibit site symmetry 1, with Sr situated on Wyckoff position 2 c, and Fe on 2 a. In the voids of this framework, Cs sites (light green in Fig. 1) have a distorted square-antiprismatic environment with four C and four N atoms as ligands. The substitution of manganese by iron in Cs 2 Na[Mn(CN) 6 ] can be explained by the similar ionic radii of the two elements: r Mn(III) = 0.58 Å and r Fe(II) = 0.61 Å (Shannon, 1976). For the substitution of sodium by strontium, the ionic radii differ more substantially: r Na(I) = 1.02 Å and r Sr(II) = 1.18 Å . The two crystal structures were compared numerically using COMPSTRU (de la Flor et al., 2016). The structure similarity index Á was calculated to be 0.009 (Bergerhoff et al., 1999). However, since only a few parameters (11) were refined and many parameters kept fixed in the refinement, the similarity index is not reliable.     Hence, we focused on ferrocyanides with A I = Cs and B II = Sr, for which only three phases have been reported, however with poorly described crystal data. The Cs 2 Sr[Fe(CN) 6 ] phase studied by Kuznetsov et al. (1970) is reported to crystallize in the tetragonal crystal system with a ranging from 10.72 to 10.89 Å and c from 10.75 to 10.99 Å (PDF00-024-0293 and PDF00-24-0294). The entry for the third phase (PDF00-048-1203), the hydrated ferrocyanide CsSr[Fe(CN) 6 ]Á3H 2 O reported by Slivko et al. (1988), is comprised only of reflections without further crystal data given. None of these PDF cards matched the X-ray diffraction pattern of the studied sample. As shown in Fig. 2a,b, the cubic crystal habit revealed by SEM measurements hints at a crystal structure with cubic symmetry, but the number of reflections is not consistent with such a highly symmetrical crystal system. The whole pattern can be described by a monoclinic cell and the experimental data are well reproduced by adjusting the reflections from the Cs 2 Na[Mn(CN) 6 ] phase (Ziegler et al., 1989;PDF 04-012-3126). The Cs 2 Sr[Fe(CN) 6 ] crystal structure was refined from that of Cs 2 Na[Mn(CN) 6 ] assuming complete substitution of manganese by iron and sodium by strontium. As described above, the ionic radii of the corresponding metals are close enough for these substitutions to be possible.

Synthesis and crystallization
All solutions were prepared using Millipore water.  Padigi et al. (2015). Once prepared, K 2 BaHCF was collected by centrifugation, washed and dried. Its chemical composition (K, Fe and Ba) and water content, respectively, were determined by inductively coupled plasma (ICP) analysis and thermogravimetric analysis (TGA). The dried K 2 BaHCF particles redispersed readily in water, producing a clear, slightly yellow dispersion. Cs 2 SrHCF forms immediately as a milky white precipitate (Fig. 3a) (Gray & Beach, 1963). (c) FT-IR spectra of K 2 BaHCF (black) and Cs 2 SrHCF (red). All peaks shift to higher frequencies after ion exchange. The arrows indicate the (HOH) signal at 1611 cm À1 and (OH) signals at 3527 cm À1 and 3601 cm À1 observed for K 2 BaHCF but not for Cs 2 SrHCF. These signals indicate the presence of structural water, which is common in hexacyanidoferrate particles. (d) Enlarged view of the FT-IR spectra in the 400-650 cm À1 range. the mixed CsNO 3 /Sr(NO 3 ) 2 solution to the clear yellow K 2 BaHCF dispersion. To ensure complete substitution, 2.2 moles of CsNO 3 and 1.1 moles of Sr(NO 3 ) 2 were added for every mole of K 2 BaHCF present. After being left to mix for 1 h, the formed Cs 2 SrHCF was collected by centrifugation, washed and dried. The chemical composition (Fe and Sr) of the powder was determined by ICP analysis while the Cs content was determined by atomic absorption spectroscopy (AAS). An initial characterization of the Cs 2 SrHCF powder was carried out by TGA, UV-Vis and FT-IR spectroscopy. The UV-Vis spectrum of the Cs 2 SrHCF (Fig. 3b) confirmed that the [Fe(CN) 6 ] moiety was maintained with only slight decreases in the wavelengths of the various absorption peaks (Gray & Beach, 1963). The FT-IR spectra of Cs 2 SrHCF and K 2 BaHCF are shown in Fig. 3c. While a (HOH) signal is observed for K 2 BaHCF at 1611 cm À1 along with (OH) signals at 3527 cm À1 and 3601 cm À1 , no such signals were detected for Cs 2 SrHCF. This absence of water was confirmed by TGA, which showed no mass loss between 30 and 400 C (Fig. S1 in the supporting information. The largest change was in the (M-N) stretching mode, which shifted from 421 cm À1 (Ba-N) to 439 cm À1 (Sr-N) (Fig. 3d).

Refinement
Crystal data and details of the data collection and structure refinement methods are summarized in Table 2. The observed and calculated intensities are shown in Fig. 4 (Ziegler et al., 1989) and the given individual isotropic displacement parameters were used. All occupancies were set to unity because of the experimentally determined composition. Except for cesium, all displacement parameters were kept fixed because otherwise some became negative. The positions of the nitrogen and carbon atoms were also kept fixed. Since iron and strontium atoms are in special positions, only the lattice parameters, the position of the cesium atom and its U iso value were refined, together with three profile parameters. The residual electron density is about is 6.06 e Å À3 at a distance of 0.71 Å from Cs.

Special details
Refinement. The Platon test reports 21 Alerts level C. They could be gathered in three groups for explanation : i)17 C-Alerts (out of 21) are "missing esd on x,y,z coordinates of N and C atoms. This is normal since these positions were not refined. Hence no esd was calculated by Jana2006. ii)3 C-Alerts (out of 21) are due to a slighlty larger Fourier difference density than allowed by CheckCIF. I can not enhance the quality of the data so no reduction of this value can be done.
iii) The last C-Alert is about a "calc. and reported Sum Formula which differ. I did not find the origin of the Alert.